Cholinergic imaging in dementia spectrum disorders

The multifaceted nature of the pathology of dementia spectrum disorders has complicated their management and the development of effective treatments. This is despite the fact that they are far from uncommon, with Alzheimer’s disease (AD) alone affecting 35 million people worldwide. The cholinergic system has been found to be crucially involved in cognitive function, with cholinergic dysfunction playing a pivotal role in the pathophysiology of dementia. The use of molecular imaging such as SPECT and PET for tagging targets within the cholinergic system has shown promise for elucidating key aspects of underlying pathology in dementia spectrum disorders, including AD or parkinsonian dementias. SPECT and PET studies using selective radioligands for cholinergic markers, such as [11C]MP4A and [11C]PMP PET for acetylcholinesterase (AChE), [123I]5IA SPECT for the α4β2 nicotinic acetylcholine receptor and [123I]IBVM SPECT for the vesicular acetylcholine transporter, have been developed in an attempt to clarify those aspects of the diseases that remain unclear. This has led to a variety of findings, such as cortical AChE being significantly reduced in Parkinson’s disease (PD), PD with dementia (PDD) and AD, as well as correlating with certain aspects of cognitive function such as attention and working memory. Thalamic AChE is significantly reduced in progressive supranuclear palsy (PSP) and multiple system atrophy, whilst it is not affected in PD. Some of these findings have brought about suggestions for the improvement of clinical practice, such as the use of a thalamic/cortical AChE ratio to differentiate between PD and PSP, two diseases that could overlap in terms of initial clinical presentation. Here, we review the findings from molecular imaging studies that have investigated the role of the cholinergic system in dementia spectrum disorders.


Introduction
The population of the developed world is ageing and this has led to an increasing prevalence of dementia. It has recently been estimated that 47.5 million people worldwide suffer from dementia [1]. Dementia is a challenge to society, both in terms of economic cost and social burden, as most patients require long-term care in their home or nursing home [2]. There is clearly a real incentive to find more effective methods of diagnosing and treating this set of diseases, which are not yet as fully understood as many of their other neurodegenerative diseases.
The cholinergic system plays a key role in functional and structural remodelling of cortical circuits underlying cognitive processing [3]. The three major cholinergic projection systems of the central nervous system include (1) the nucleus basalis of Meynert, which supplies cholinergic projections throughout the cerebral cortex and hippocampus [4], (2) the pedunculopontine nucleus pars compacta, which projects to the forebrain as well as various subcortical structures such as the thalamus [5], and (3) cholinergic neurons intrinsic to the striatum [6]. Post-mortem studies have found that muscarinic acetylcholine receptors (mAChR) are highly expressed in the caudate nucleus and nucleus accumbens in normal monkey and human brain [7][8][9]. In the somatosensory, primary motor and temporal cortices, mAChR concentrations are about 60 % of the levels in the caudate whilst the cerebellar cortex expresses the lowest levels [8]. Nicotinic acetylcholine receptors (nAChR) are abundantly expressed in the entorhinal, temporal and primary motor cortices, and the hippocampus and thalamus of post-mortem normal human brain tissue [10][11][12].
Dysfunction of the ascending cholinergic systems from the basal forebrain and brainstem and the associated loss of cholinergic neurotransmission in the cerebral cortex has been suggested as an underlying substrate of cognitive decline (cholinergic hypothesis of dementia), supporting the use of acetylcholinesterase (AChE) inhibitors in dementia [13]. Many post-mortem studies investigating pathophysiological mechanisms of dementia have focused on alterations in functional components of the cholinergic system, such as AChE, the vesicular acetylcholine transporter (VAChT), nAChR, and mAChR [14,15]. Molecular imaging techniques such as SPECT and PET with selective radioligands for targets within the cholinergic system have led to significant advances in the understanding of the neurobiology and pathophysiology of dementia ( Fig. 1).
This review discusses the use of cholinergic radioligands in molecular imaging (Tables 1 and 2), with the aim of improving the understanding, and thereby the diagnosis, monitoring and treatment, of dementia spectrum disorders.
Cholinergic system in Alzheimer's disease AD is the leading cause of dementia in the western world, accounting for more than 60 % of cases [44,45]. AD is characterized pathologically by accumulation of amyloid β peptide (Aβ) in extracellular plaques, intracellular deposits of tau protein, and neuronal loss [46][47][48]. Evidence for the involvement of the cholinergic system in the pathogenesis of AD was provided as early as the mid-1970s in post-mortem studies showing loss of choline acetyltransferase (ChAT) and AChE in the cortex, hippocampus and amygdala of AD brain samples [14,[49][50]. Moreover, reduced ChAT activity has been found to be correlated with increased Aβ plaque load and with cognitive decline [50,52]. Degeneration of the cholinergic system affects not only cortical regions but also the nucleus basalis of Meynert, where cholinergic neurons are severely decreased in post-mortem brain tissue of AD patients, highlighting the role of subcortical cholinergic dysfunction in the pathogenesis of AD [53].

Presynaptic cholinergic dysfunction in Alzheimer's disease
PET with N-[ 11 C]methyl-piperidin-4-yl propionate ([ 11 C]PMP), a selective substrate for AChE [54], has shown a reduction in AChE activity in AD patients [17,19,20]. The degree of AChE activity reduction ranged between 9 % and 33 %, depending on the severity of cognitive impairment in the cohort of AD patients examined. Kuhl [17]. The disagreement in cholinergic and metabolic imaging patterns of degeneration suggests a different mechanism underlying cholinergic and metabolic decline in AD pathophysiology. Patients with early AD (mean MMSE score 22) showed more modest reductions in cortical AChE activity (9 -11 %) with greater decreases in the lateral temporal cortex (15 %) [19,20]. Overall, in vivo AChE losses tend to be less pronounced than those seen in post-mortem studies, in which 55 % reductions in cholinergic markers have been observed [55].
The modest degree of cholinergic denervation despite the severity of cognitive decline raises questions about the specific role of the cholinergic system in episodic memory processes. In this regard, Bohnen et al. [20] found that decreases in cortical AChE activity were negatively correlated with performance in the WAIS-III digit span test, whereas they were not  scores. These results suggest that cholinergic dysfunction is linked to attention and working memory rather than episodic memory. Although cortical cholinergic denervation from the nucleus basalis of Meynert is a feature of AD, the pontine cholinergic projection system to the thalamus has been found to be spared in the disease [28]. PET with N-[ 11 C]methyl-4-piperidyl acetate ([ 11 C]MP4A), another selective AChE radioligand, has shown that AD patients with apolipoprotein E allele ɛ4 (ApoE4) have significantly less pronounced reductions in cortical AChE activity than those negative for the ApoE4 allele [22]. This suggests that the ApoE4 allele has a protective role against the widespread loss of AChE activity in AD, although the underlying mechanism remains unknown. [ 11 C]MP4A has shown higher specificity for AChE relative to butyryl cholinesterase [56,57], but it has a higher hydrolysis rate and thus radioligand uptake in regions of high AChE activity such as basal ganglia is strongly dependent on the rate of transport into the brain [54]. In contrast, [ 11 C]PMP exhibits a rate of hydrolysis three to four times slower than that of [ 11 C]MP4A, allowing more precise estimates of the AChE activity in regions of moderate to high AChE concentration [58].
SPECT with [ 123 I]IBVM, a selective ligand for presynaptic VAChT, has been used to assess presynaptic cholinergic terminal density [59]. Interestingly, in mildly demented AD patients, the [ 123 I]IBVM binding pattern differs according to age at onset [16]. In patients with early-onset AD (age at onset <65 years) loss of cholinergic terminals were observed in the neocortex and hippocampus, whereas in patients with lateonset AD significant VAChT reductions were limited to the temporal cortex and hippocampus. These findings are consistent with those of post-mortem studies, which have demonstrated more widespread cholinergic degeneration in earlyonset AD [60]. A new PET radioligand with good sensitivity a n d s p e c i f i c i t y f o r V A C h T , ( − ) 5 -1 8 Ffluoroethoxybenzovesamicol ([ 18 F]FEOBV), has recently been developed and tested in humans [61]. Both reference tissue modelling and late static scanning approaches correlated well with the full kinetic modelling with arterial sampling and plasma metabolite analysis [61]. In comparison with [ 123 I]IBVM SPECT, [ 18 F]FEOBV PET showed low binding in the lateral cerebellar cortex and high binding in the mesopontine junction and medulla, providing a robust index of VAChT binding [61]. [ 18 F]FEOBV PET is a significant advance over currently available presynaptic cholinergic imaging agents and may be a valuable tool to assess dysregulation of the cholinergic system in AD and parkinsonian patients.

Postsynaptic cholinergic dysfunction in Alzheimer's disease
Previous PET studies using [ 11 C]nicotine have shown significant reductions in nAChR binding in the frontal cortex, temporal cortex and hippocampus of patients with moderate AD [62,63]. Similar to AChE activity, decreases in cortical nAChR expression as measured by [ 11 C]nicotine are associated with attention deficit but not episodic memory impairment [32]. These [ 11 C]nicotine studies, however, were hindered by high levels of nonspecific binding, rapid metabolism, and washout from the brain, as well as a strong dependence on cerebral blood flow [64,65]. More recently, new PET and SPECT radioligands have been developed to target α 4 β 2 nAChR. Post-mortem autoradiography studies have shown that α 4 β 2 nAChR is the most affected receptor subtype in AD, with reductions of up to 50 % in the neocortex, entorhinal cortex and hippocampus [12,66,67]. [ 123 I]5-iodo-3-[2(S)-2azetidinylmethoxy] pyridine ([ 123 I]5IA) SPECT has shown significant reductions in α 4 β 2 nAChR in the frontal cortex, striatum, right medial temporal lobe and pons of AD patients [34]. [ 123 I]5IA binding is also significantly reduced in the medial temporal cortex of patients with amnestic mild cognitive impairment (MCI) and α 4 β 2 nAChR decreases correlated  AD Alzheimer's disease, DLB dementia with Lewy bodies, HC healthy control, mAChR muscarinic acetylcholine receptor, MCI mild cognitive impairment, nAChR nicotinic acetylcholine receptor, PD Parkinson's disease, PDD Parkinson's disease dementia, PSP progressive supranuclear palsy with cognitive decline [42]. However, other studies have yielded contrasting results, finding no changes in α 4 β 2 nAChR binding in cortical and thalamic regions of patients with early AD or MCI compared to healthy controls [36,41]. This discrepancy may be due to the different severities of cognitive impairment of the subjects studied and different methodologies in assessing α 4 β 2 nAChR expression. PET with 2-[ 18 F]F-A-85380 ([ 18 F]2FA), a radioligand selective for α 4 β 2 nAChR, has shown significant reductions in nAChR availability in the hippocampus, and frontal, temporal and parietal cortices of AD patients, which furthermore correlated with the level of cognitive impairment [38]. Another [ 18 F]2FA study showed significantly reduced α 4 β 2 nAChR expression in the medial frontal cortex and nucleus basalis of Meynert in AD patients, which were correlated with Frontal Assessment Battery scores, suggesting that in vivo α 4 β 2 nAChR plays role in those specific functions that may be different from episodic memory [43]. Furthermore, a negative correlation between α 4 β 2 nAChR availability and Aβ load (measured by [ 11 C]Pittsburgh compound B) was found in the same brain regions, suggesting that Aβ deposition may induce degeneration of cholinergic neurons [43]. However, [ 18 F]2FA exhibits slow brain distribution kinetics and relatively low binding potentials (≤0.6 -0.8) in extrathalamic regions including the cortex which are of high importance in studying neurodegenerative diseases [68]. In recent years, major effort has focused on the development of new α 4 β 2 nAChR PET radioligands with faster regional brain kinetics than [ 18 F]2FA that will enable further imaging studies in dementia spectrum disorders [69].

Cholinergic imaging in assessing Alzheimer's disease treatment
Cholinergic PET imaging techniques have also been employed to assess the efficacy of medications. PET with [ 11 C]PMP has shown 19 -39 % inhibition of cortical AChE activity following donepezil treatment in AD patients [70,71] whereas 28 -39 % AChE inhibition has been observed after donepezil and rivastigmine in the frontal temporal and parietal regions of AD patients using [ 11 C]MP4A PET [72,73]. Kadir and colleagues [74] investigated the effect of galantamine, a cholinesterase inhibitor, on AChE activity and nicotine binding to nAChR in AD patients. They found significant reductions in cortical [ 11 C]PMP following galantamine treatment at 3 weeks, and 3 and 12 months compared with baseline [74]. These findings suggest that galantamine treatment is able to increase ACh concentrations in the synaptic cleft resulting in increased cholinergic neurotransmission. No significant changes in [ 11 C]nicotine binding were found following galantamine treatment [74]. Galantamine acts on nAChR as an allosterically potentiating ligand [75], thereby sensitizing the receptors, which in turn causes nAChR upregulation to maintain a baseline level of signalling.

Cholinergic system in parkinsonian dementias
Parkinson's disease (PD) is a chronic neurodegenerative disorder characterized pathologically by degeneration of dopaminergic neurons in the substantia nigra pars compacta and formation of α-synuclein proteinaceous intraneuronal inclusions referred to as Lewy bodies and Lewy neurites [76][77][78]. In addition to classical motor symptoms, non-motor features such as cognitive decline, are a very important aspect of the disease, because they add significantly to the burden on patients and caregivers [79,80]. PD patients carry a sixfold increased risk for dementia compared to the general population, with approximately 80 % of patients developing PD dementia (PDD) over the course of the disease [81,82]. MCI in PD also appears to be common, occurring in 20 % to 50 % of PD patients, even at the time of PD diagnosis and prior to initiation of dopaminergic therapy [83,84]. Dementia with Lewy bodies (DLB) is characterized by fluctuating cognitive and attentional deficits, recurrent visual hallucinations and parkinsonism, and is the second most common cause of degenerative dementia, accounting for 15 -25 % of cases [85]. Evidence for the involvement of the cholinergic system in the development of PDD and DLB has been provided by postmortem studies, which have shown decreased AChE activity and nAChR density in cortical and subcortical brain tissue of PDD and DLB patients [86][87][88][89][90][91][92].

Presynaptic cholinergic dysfunction in parkinsonian dementias
PET studies using [ 11 C]MP4A and [ 11 C]PMP have demonstrated mild to moderate decreases in cortical AChE activity in PD patients without dementia, and severe decreases in PDD and DLB patients [21,23,26]. Using SPECT with [ 123 I]IBVM, Kuhl et al. [16] demonstrated that whilst in PD without dementia, VAChT levels are reduced only in the parietal and occipital cortices, in PDD, major losses are seen throughout the entire cortex, suggesting a greater and more widespread presynaptic cholinergic dysfunction in PDD patients. Loss of AChE activity occurs in early stages of the disease, with de novo PD patients showing significant (12 %) AChE losses in the medial occipital cortex [24]. Differences in AChE activity between early and advanced PD are nonsignificant, however, suggesting that cholinergic dysfunction occurs early, but does not progress with the disease [24]. Moreover, loss of cortical AChE activity as measured by [ 11 C]PMP is correlated with worse performance in working memory and attention tests but not with the severity of motor symptoms [23]. A relationship between striatal [ 18 F]FDOPA uptake and cortical [ 11 C]MPA reduction has been found in PDD patients, suggesting that cognitive decline in PD occurs when the disease spreads from nigral neurons to the cortex, leading to a cholinergic dysfunction in this region [21].
Cholinergic dysfunction is higher in PDD patients than in AD patients with a similar degree of cognitive impairment (MMSE score 22.8 and 22.2, respectively) suggesting that the mechanisms underlying cognitive decline in PD and AD have different pathogeneses [19]. In contrast to the cholinergic system in AD, which was spared (0.7 % reduction), thalamic AChE activity was found to be decreased by 19.8 % in PDD patients [28].
Cortical AChE activity differs between patients with PD and those with progressive supranuclear palsy (PSP) [18,25]. PD patients show significant decreases in cortical [ 11 C]MP4A uptake, whereas PSP patients show significant reductions only in the thalamus [18]. Therefore, a thalamic/ cortical [ 11 C]MP4A binding ratio may be useful in distinguishing between PD and PSP. A SPECT study using an [ 123 I]IBVM ligand found that VAChT levels are significantly lower in the thalamus and anterior cingulate cortex of PSP patients compared to healthy controls, and that thalamic cholinergic dysfunction is inversely correlated with disease duration [29]. This is in line with the findings of similar PET studies reported previously, that show a preferential cholinergic denervation in the thalamus of PSP patients [83]. VAChT levels, however, are unaffected in the striatum of PSP patients [29]. This differs from PD pathology, in which cholinergic neurons in the striatum are affected, suggesting a possible role of cholinergic imaging markers in aiding the differential diagnosis of these two clinically similar diseases [29].
In other atypical parkinsonisms such as multiple system atrophy, cortical AChE activity decreases have been found to be similar to those seen in PD patients (−14.6 % and −15.3 %, respectively) [25]. Hirano et al. [27] investigated differences in AChE activity in patients with PSP, corticobasal degeneration (CBD) and frontotemporal dementia (FTD) using both voxel-based and volume of interest analysis. Statistical parametric mapping analysis showed significant AChE decreases in the paracentral region, frontal, parietal and occipital cortices in CBD patients and in the paracentral region and thalamus in PSP patients [27]. FTD patients showed no significant differences in AChE activity compared to the control group. Volume of interest analysis showed significant decreases in thalamic AChE activity only in PSP patients. Thus, cholinergic dysfunction occurs in CBD and PSP, although involving different brain regions, but it was not observed in FTD patients, which may explain why AChE inhibitors are ineffective for this condition [93].

Postsynaptic cholinergic dysfunction in parkinsonian dementias
Post-mortem studies have shown that cortical mAChR density is altered in PDD and DLB samples [94,95]. An in vivo PET st udy with N -[ 1 1 C]methyl -4-piperi dyl benzilate ([ 11 C]NMPB), a marker for mAChR, demonstrated increased mAChR levels in the frontal cortex of PD patients, probably due to denervation hypersensitivity caused by loss of the ascending cholinergic system in frontal areas [30]. Contrastingly, in PSP patients, cortical and thalamic [ 11 C]NMPB binding was preserved [30]. A SPECT study using [ 123 I]iodo-quinuclidinyl benzilate ([ 123 I]QNB), another marker for mAChR, showed significant increases in mAChR in the occipital lobe in both PDD and DLB, which may be the substrate of visual disturbances in these diseases [31]. Furthermore, there was significantly lower [ 123 I]QNB binding in the frontal and temporal lobes of PDD patients than in DLB patients. [ 11 C]NMPB and [ 123 I]QNB are high-affinity mAChR antagonists with similar chemical structures and regional brain distributions [96][97][98]. Both radioligands penetrate the blood-brain barrier efficiently but nonspecifically in relation to mAChR subtype [99].
[ 123 I]5IA SPECT studies have shown 10 -25 % reductions in nAChR levels in cortical and subcortical regions of PD patients [33,95], with the largest decreases observed in the thalamus, and parietal and temporal cortices [33]. A [ 18 F]2FA PET study showed significant reductions in α 4 β 2 nAChR availability in various regions in PD patients, including the frontoparietal and anterior cingulate cortices, midbrain, pons and cerebellum, with the highest reduction in the left parietal cortex [40]. Furthermore, α 4 β 2 nAChR loss correlated with depression and cognitive decline as measured by the MMSE, the DemTect Scale, the Clock Drawing test, delayed recall of figures of the CERAD (Consortium to Establish a Registry for Alzheimer's Disease) battery, and the Trail Making test [40]. In nondemented PD patients, cortical and subcortical [ 123 I]5IA decreases correlated with worse performance In the Boston Naming test and Word List Intrusions test, two specific tests for detection of word-finding difficulties, learning capacity and memory for language information [98]. Decreases in [ 18 F]2FA uptake were also found in the striatum (−10 %) and substantia nigra (−14.9 %) of PD patients, but did not correlate with dopaminergic function as measured by [ 18 F]F-DOPA, nor with clinical severity [39].
[ 123 I]5IA SPECT has shown significant reductions in α 4 β 2 nAChR binding in the frontal, temporal and cingulate cortex, as well as in the striatum of DLB patients [37]. In the occipital cortex, however, α 4 β 2 nAChR levels were increased and correlated with visual hallucinations experienced by DLB patients, suggesting cholinergic dysfunction in the occipital lobe as a substrate of visual hallucinations. Because α 4 β 2 nAChRs are located in some regions presynaptically on both cholinergic and noncholinergic terminals [101], some difficulties may arise in the interpretation of nAChR imaging studies. In PD, significantly decreased α 4 β 2 nAChR levels may be a result of nAChR loss on degenerated presynaptic nigrostriatal dopaminergic neurons. A post-mortem study has shown that loss of striatal nAChR binding closely parallels the loss of nigrostriatal dopaminergic markers in PD brain tissue [102].
Postsynaptic nAChR receptors may show a compensatory increase, no change, or a decrease due to degeneration of noncholinergic systems (e.g. noradrenergic, serotoninergic, glutamatergic) to which the nAChRs are coupled [37,102]. Further studies combining cholinergic and noncholinergic radioligands may aid in elucidating the nAChR alteration patterns in parkinsonian disorders.

Conclusion
Cholinergic dysfunction has a pivotal role in the pathophysiology of cognitive decline. Given the wide variety of available markers for the cholinergic system, molecular imaging techniques provide a valuable tool to investigate pathophysiological mechanisms, and monitor progression and response to treatments in AD and parkinsonian dementia. The largest drawback of this technology, its cost, will hopefully be improved by its widening use as it becomes increasingly clear that PET and SPECT imaging holds much potential for gleaning important information about the pathology of dementia. Much remains to be learned about cholinergic dysfunction in AD, PD and related disorders. PET studies combining presynaptic and postsynaptic radioligands may be useful in unravelling alterations of cholinergic neurotransmission. Moreover, it has been suggested that the α7 nAChR subtype may have a neuroprotective role by modulating the neurotrophic system crucial for the maintenance of cholinergic neuron integrity, and also by stimulating signal transduction pathways that support neuron survival [103]. In AD, α7 nAChR may modulate β-amyloid-induced pathology [104], and deletion of the α7 nAChR gene has shown to improve cognitive impairment in animal models of AD [105]. Further PET studies using radioligands specific to the α7 nAChR such as [ 18 F]ASEM [106,107] are needed to determine the relationship between α7 nAChR and AD pathology.

Compliance with ethical standards
Conflicts of interest None.
Ethical approval This article does not describe any studies with human participants performed by any of the authors.
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.