Introduction

It is well recognized that increased life expectancy results in an increased frequency of neurodegenerative diseases. In the last two to three decades, the development of new diagnostic and therapeutic methods has been intensified in neurodegenerative diseases. Molecular radiopharmaceutical-based neuroimaging is a growing field and provides several new diagnostic methods to investigate and characterize neurodegenerative diseases during life. This review summarizes the facts for the most frequent neurodegenerative diseases from a nuclear medicine point of view. Furthermore, the review provides information on approved radiotracers and ongoing research activities in the development of new radiotracers for imaging neurodegenerative diseases as well as a short passage about the need for novel radiotracers.

Neurodegenerative diseases – important histopathological and clinical facts

Alzheimer’s disease (AD)

Alzheimer’s disease (AD) is the most common neurodegenerative disease causing dementia in the elderly. Histopathological characteristics of this disease are a progressive accumulation of β-amyloid (Aβ) plaques and hyperphosphorylated neurofibrillary tau protein (tau). However, only approximately 50 % of the patients have solely Alzheimer’s pathology. Many patients show additional pathologic changes related to other neurodegenerative diseases in autopsy studies (Alzheimer’s Association report 2018 - https://www.sciencedirect.com/science/article/pii/S1552526018300414). Along with the typical clinical features like memory impairment, especially in the semantic and episodic domain and the executive dysfunction (McKhann et al. 2011), atypical variants of AD exist. As such, posterior cortical atrophy (PCA) and logopenic variant primary progressive aphasia (lvPPA) are labeled as atypical AD, since the histopathological changes (i.e. Aβ and tau accumulation) in these neurodegenerative diseases determine the “typical” AD features, although the distributional pattern of the pathologic changes seems to be different (Crutch et al. 2012; Harris and Jones 2014). However, in both PCA and lvPPA a significant minority of cases showed other underlying pathologies than AD, e.g. Lewy bodies, transactive response DNA binding protein of about 43 kDa (TAR DNA-binding protein 43, TDP-43) proteinopathies, “pure” tauopathy or cerebrovascular disease (Crutch et al. 2012; Harris and Jones 2014) (Table 1). The core clinical features of patients with PCA are caused by a decline in visual processing and other posterior cognitive functions, e.g. space and/or object perception deficits, simultanagnosia or constructional dyspraxia (Crutch et al. 2017). The core clinical features of lvPPA are impaired single-word retrieval in spontaneous speech and naming as well as impaired repetition of sentences and phrases (Harris and Jones 2014).

Table 1 Summary of histopathological findings of the different neurodegenerative syndromes
Full size table

Frontotemporal lobar degeneration (FTLD)

FTLD is a potpourri of clinically, histopathologically and genetically different disorders that become united due to predominant pathological involvement of the frontal and temporal brain regions. Three distinct clinical phenotypes of FTLD are recognized including a behavior/dysexecutive syndrome - the behavioral variant of frontotemporal dementia (bvFTD); language disorders - the primary progressive aphasia (PPA): semantic variant (svPPA) and non-fluent/agrammatic variant (nfvPPA); and motor disorders (amyotrophic lateral sclerosis, corticobasal and progressive supranuclear palsy syndromes).

In general, an individual FTLD disorder can be ascribed to different histopathologies such as tauopathy (FTLD-tau), TDP-43 proteinopathy (FTLD-TDP-43) and the FET protein family that consists of Fused in sarcoma, Ewing sarcoma and TATA-binding protein associated factor 15 proteinopathy (FTLD-FET) (Table 1). Here, some histopathology is more often seen in one than in another FTLD disorder (Harris and Jones 2014; Bang et al. 2015; Mackenzie and Neumann 2016). Thus, more than 70% of the patients with svPPA have TDP-43 proteinopathy. Over 50% of patients with nfvPPA have FTLD-tau, approximately 20% show TDP-43 proteinopathy (Harris and Jones 2014). Almost half of the patients with bvFTD have FTLD-tau (Pressman and Miller 2014) and more than 25% TDP-43 pathology (Bang et al. 2015). Of interest, tauopathies can be differentiated in at least 5 subtypes according to their molecular subtype (Mackenzie and Neumann 2016).

Parkinsonian syndromes

In Parkinson’s disease (PD), a degeneration of the nigrostriatal system occurs causing a reduction of the neurotransmitter dopamine. Degeneration of dopaminergic neurons in the substantia nigra pars compacta is an inherent neuropathological sign of PD. Histologically, most patients suffering from PD exhibit intracellular accumulation of protein inclusions mainly constituted of α-synuclein (Lewy bodies). However, some patients with specific genetic forms of PD do not have Lewy body pathology and it remains unclear how Lewy bodies and neuronal loss are connected to each other. The typical clinical criteria of PD consist of symptoms such as bradykinesia, rigidity or resting tremor as cardinal motor manifestations (Postuma et al. 2015). Important supportive criteria include response to dopamine replacement, unilateral onset, olfactory dysfunction and REM sleep behavior disorder (Berg et al. 2013; Postuma et al. 2015). Corticobasal degeneration (CBD), progressive supranuclear palsy (PSP) and multisystem atrophy (MSA) are classified as atypical parkinsonisms. Histopathologically, CBD and PSP belong to primary tauopathies while MSA, DLB and PD are characterized by a pathologic accumulation of α-synuclein protein (α-synucleinopathies). The clinical presentation of CBD can be subdivided into 4 phenotypes: corticobasal syndrome, behavior spatial syndrome, nfvPPA and PSP (Armstrong et al. 2013) demonstrating that the accuracy to diagnose CBD ante-mortem is still limited.

Clinical core features of PSP are akinetic-rigid syndrome, postural instability or falls and supranuclear ophthalmoplegia (Bensimon et al. 2009). The most recent version of PSP criteria published by the Movement Disorder Society includes eleven clinical phenotypes of PSP (Ali and Josephs 2018). Overall, the clinical diagnoses of post-mortem-validated PSP cases were correctly established only in 19% of cases at the first clinical visit, and in 71% of cases over the course of the disease (Respondek et al. 2013). Furthermore, an average of 24% of post-mortem histopathologically diagnosed cases of MSA, PD and CBD were, ante-mortem, falsely diagnosed as PSP (Respondek et al. 2013). Also, the occurrence of concomitant AD (in approximately 36%) or PD (in approximately 20%) pathologies in PSP patients is remarkable (Dugger et al. 2014).

Histopathologic characteristics of MSA are inclusions of misfolded α-synuclein in oligodendrocytes (Jellinger 2014). According to the clinical presentation, MSA is usually subdivided into a parkinsonian subtype (MSA-P) and a cerebellar subtype (MSA-C). For the clinical phenotype, autonomic failure is a core symptom which must be present to establish the diagnosis of MSA (Gilman et al. 2008).

Lewy body dementia pathologies

Dementia with Lewy-bodies (DLB) is an α-synucleinopathy, characterized by widespread accumulation of Lewy bodies and Lewy neurites in the brain-stem, limbic system and cortical areas (Braak and Braak 2000) and a higher percentage of DLB compared to PD patients with dementia (PDD) show Aβ deposits in histopathological examinations. Consequently, approximately 80% of DLB patients have a positive Aβ PET scan (Drzezga 2010) (Table 1) indicating a significant overlap between AD and DLB, which is reflected in the clinical presentation of the patients. Only 50% of patients with DLB pathology show the typical symptoms of DLB (McKeith et al. 2016). Core clinical features are fluctuating cognition, recurrent visual hallucinations, rapid eye movement (REM) sleep behaviour disorder and parkinsonism (McKeith et al. 2017). However, “the likelihood that the observed neuropathology explains the DLB clinical syndrome is directly related to the severity of Lewy-related pathology, and inversely related to the severity of concurrent AD-type pathology” (McKeith et al. 2005).

More than 75% of PDD patients develop in the long term clinical course of the disease (Aarsland et al. 2003). PDD is characterized by an intracellular accumulation of α-synuclein (Lewy bodies). Approximately 15% of PDD patients show cerebral β-amyloid plaques (Lucero et al. 2015; Edison et al. 2008). Due to an overlap of clinical and morphological features there is a continuous debate (Friedman 2018) of whether DLB and PDD are the same disease with different phenotypic representation of Lewy body disease spectrum.

Huntington’s disease (HD)

HD is an autosomal-dominant neurodegenerative disease caused by a single gene mutation (Kim and Fung 2014) i.e. a CAG repeat expansion in exon 1 of the huntingtin gene (MacDonald et al. 1993). This repeated CAG expression results in an abnormal toxic protein which is named Huntingtin. This protein is expressed in all brain cells and disturbs protein degradation as well as several other cellular processes, e.g. mitochondrial function, axonal trafficking or peripheral immune regulation (Kim and Fung 2014). The cerebral accumulation of hyperphosphorylated tau aggregates seems to be a further histopathological characteristic besides the accumulation of the mutated Huntingtin protein (Vuono et al. 2015). From the clinical perspective, HD is characterized by progressive motor symptoms, cognitive decline and neuropsychiatric disturbances (Kim and Fung 2014).

Role of PET imaging in neurodegenerative diseases

Currently, the classification of neurodegenerative diseases is in permanent change and progress. Predominantly, phenotypical definitions are increasingly substituted – at least in research settings- by classifications which include biomarkers for the underlying pathophysiological process and thus lead to a more biological definition of neurodegenerative diseases (Jack et al. 2018).

This development is of great significance for nuclear medicine, as molecular imaging using PET tracers can provide biomarker information, e.g. [18F]FDG as a biomarker of neuronal injury or Aβ PET as a biomarker of AD pathology (Barthel et al. 2015). Table 2 summarizes all radiotracers mentioned in the review with abbreviation and chemical definition.

Table 2 Chemical structures of all radiotracers discussed
Full size table

Molecular imaging of neurodegeneration

[18F]FDG PET

Glucose is the energy supplier of the brain. In all neurodegenerative diseases, impairment of neuronal function and therefore reduced energy metabolism occur. [18F]FDG (Fig. 1) as a marker for neuronal injury can be used to detect this impairment, and it is well known that different neurodegenerative diseases show distinct patterns of reduced [18F]FDG uptake (Hellwig et al. 2012; Barthel et al. 2015) (Table 3). However, the hypometabolic patterns of some neurodegenerative disorders overlap. Due to its broad availability and sufficient diagnostic accuracy, [18F]FDG is currently the widely used radiotracer in imaging of neurodegenerative diseases in clinical routine.

Fig. 1
figure 1

Chemical structure of [18F]Fluordesoxyglucose ([18F]FDG)

Full size image
Table 3 Summary of typical hypometabolism patterns in [18F] FDG PET in different neurodegenerative diseases
Full size table

Synaptic density PET

In neurodegenerative diseases as well as in a variety of other neurological and psychiatric diseases, a reduction of synaptic density occurs during the course of disease (Feng et al. 1996; Kuhl et al. 1999; Roy et al. 2016) (Fig. 5d). Furthermore, [18F]FEOBV, a novel, very promising PET radioligand targeting VChAT has been developed and successfully applied in patients with AD and PD (Aghourian et al. 2017; Bohnen et al. 2019). Thus, using abovementioned radioligands, reduced activities of VChAT and AChE were demonstrated in various neurodegenerative diseases like AD, PD and DLB (Kuhl et al. 1996; Kuhl et al. 1999; Roy et al. 2016; Aghourian et al. 2017; Bohnen et al. 2019).

Monoamine system imaging

The vesicular monoamine transporter 2 (VMAT2) is a membrane protein that transports monoamines (e.g. dopamine or serotonin) into the presynaptic vesicles. [18F]AV-133 (Fig. 6) is a PET radiotracer targeting VMAT2. In patients with PD, a reduced [18F]AV-133 uptake was found in the basal ganglia, more pronounced in the putamen and contralateral to the predominantly affected side at onset (Gao et al. 2016). An accuracy in differentiating PD patients from HCs similar to that of DAT SPECT has been reported. Furthermore, [18F]AV-133 PET data might better correlate to clinical characteristics than PET/SPECT imaging data of DAT (Hsiao et al. 2014).

Fig. 6
figure 6

Chemical structure of (2R,3R,11bR)-9-(3-[18F]fluoranylpropoxy)-10-methoxy-3-(2-methylpropyl)-2,3,4,6,7,11b-hexahydro-1H-benzo [a]quinolizin-2-ol ([18F]AV-133)

Full size image

Imaging of misfolded proteins

β-Amyloid (Aβ) PET imaging

Aβ plaques are the histopathological hallmark of AD. Moreover, their appearance in the brain is an early, if not the causal event in AD. The most widely used Aβ-targeting PET tracer is [11C]Pittsburgh Compound B (PiB) (Fig. 7). However, the short half-life hampers the use of this tracer for clinical routine applications. Thus, three 18F-labeled radiotracers (i.e. florbetapir, florbetaben, flutemetamol) Fig. 7) have been developed and approved for clinical usage. The phase 3 data of all 3 radiotracers demonstrated high sensitivity ([18F]florbetapir: 96%, [18F]florbetaben: 98%, [18F]flutemetamol: 88%) and specificity ([18F]florbetapir: 100%, [18F]florbetaben: 89%, [18F]flutemetamol: > 80%) in detecting Aβ plaques in-vivo compared to the postmortem data (Clark et al. 2012; Sabri et al. 2015a, 2015b; Curtis et al. 2015). Other Aβ PET tracers, such as [11C]AZD2184, [18F]FIBT, and [18F]NAV4694 (Fig. 7), are under clinical examination (Ito et al. 2014; Grimmer et al. 2018). Results of the first in humans studies revealed a fast kinetics of [11C]AZD2184 and [18F] FIBT, and a time-window of 40–60 min p.i. was determined as reliable to calculate standard uptake value ratios (SUVRs) (Ito et al. 2014; Grimmer et al. 2018). The kinetics is therefore comparable to that of [18F]florbetapir and [18F]florbetaben where an acquisition start 30 min p.i. ([18F]florbetapir) and 45 min p.i. for the USA/90 min p.i. for Europe ([18F]florbetaben) is recommended (https://eanm.org/publications/guidelines/Amyloid-Guideline-J_Nucl_Med-2016-Minoshima-1316-22.pdf). Compared to the three approved 18F-labeled radiotracers, [11C]AZD2184 and [18F]NAV4694 seem to show lower white matter binding (Ito et al. 2014; Rowe et al. 2013) which principally might translate to a higher sensitivity in detecting subtle amyloid pathology.

Fig. 7
figure 7

Chemical structures of the β-amyloid targeting PET radioligands

Full size image

Over the last few years, the clinical and research diagnostic criteria especially for AD (McKhann et al. 2011; Albert et al. 2011; Dubois et al. 2014; McKeith et al. 2017; Jack et al. 2018) but also for other neurodegenerative diseases have been revised (McKeith et al. 2017; Berg et al. 2013), resulting in an implementation of biomarkers such as Aβ PET or [18F]FDG PET. However, the validation process of these biomarkers is still incomplete (Frisoni et al. 2017). Especially, clinical outcome and cost-effectiveness studies are still missing (Frisoni et al. 2017). As such studies are the decisive prerequisite for reimbursement within many healthcare systems, these PET imaging biomarkers are so far not regularly applied in clinical routine.

Tau PET imaging

Physiological tau is a phosphoprotein which stabilizes the microtubules. In the brain, six isoforms of tau exist with either three repeats (3R) or four repeats (4R) of the microtubules-binding domain (Buée et al. 2000). Aggregated tau proteins consist of post-translationally modified tau isoforms, whereby specific phenotypes/neurodegenerative diseases are associated with specific tau deposits that differ in microscopic appearance and ultrastructure (Buée et al. 2000; Villemagne et al. 2015). Importantly, the same clinical tauopathy phenotype can be caused by different misfolded tau proteins and vice-versa (Villemagne et al. 2015). In general, aggregated tau proteins are mainly located intracellularly and therefore a complex target for PET imaging. Current tau radiotracers share β-sheet binding properties. Since other misfolded proteins have similar structures, high selectivity for aggregated tau proteins is necessary (Lois et al. 2018). This is of particular interest, as tau aggregates can be co-localized to Aβ plaques with much higher concentrations of Aβ plaques compared to tau deposits (Villemagne et al. 2015). First-generation tau PET radiotracers – [18F]AV-1451, [11C]PBB3, [18F]THK5351 (Fig. 8) – showed favourable kinetics and high affinity to the 3R/4R tau isoform combination which is typical in AD (Villemagne et al. 2015; Lois et al. 2018; Villemagne 2018). However, the limitation of the first-generation tau PET radiotracers are a relevant off-target binding as well as ante-mortem vs. post-mortem inconsistencies (Villemagne et al. 2015; Harada et al. 2018; Lois et al. 2018; Villemagne 2018). Second-generation selective tau PET radiotracers, such as [18F]RO6958948, [18F]GTP1, [18F]PI-2620, [18F]MK-6240 (Fig. 8), have been preclinically evaluated and demonstrated high affinity, selectivity and specificity (Lois et al. 2018). Preliminary clinical data, partially available as conference abstracts, are promising (Mueller et al. 2017; Barret et al. 2017; Bohorquez et al. 2016; Wong et al. 2018; Bohorquez et al. 2017; Betthauser et al. 2018). However, [18F]GTP1 showed off-target binding in the basal ganglia (Bohorquez et al. 2017), while [18F]RO6958948 did so in the substantia nigra (Wong et al. 2018). So far, for [18F]MK-6240 and [18F]PI-2620 off-target binding was not observed (Betthauser et al. 2018; Barret et al. 2017). Noteworthy, preliminary data also suggest that [18F]PI-2620 might not only be able to visualize the 3R/4R tau isoform combination in AD, but also the 4R isoform in 4R-tauopathies such as PSP/CBD (https://www.alzforum.org/news/conference-coverage/next-generation-tau-pet-tracers-strut-their-stuff). Although the available data on the second-generation tau PET radiotracers are encouraging, the usefulness of these radiotracers for research and clinical approaches remains to be demonstrated in larger clinical trials.

Fig. 8
figure 8

Chemical structures of the Tau targeting PET radioligands

Full size image

Imaging of other misfolded proteins

Following the recent success with bringing amyloid and tau PET tracers into humans, the desire for radioligands targeting other misfolded proteins like α-synuclein or TDP-43 is evident. However, develo** PET radiotracers that target misfolded proteins beyond Aβ is challenging as these proteins (i) are mainly intracellularly localized, (ii) appear in a much lower concentration than Aβ plaques (Villemagne et al. 2015; Lois et al. 2018; Harada et al. 2018; Verdurand et al. 2018), and (iii) have β-sheet binding motives which are rather similar to those of amyloid aggregates. Despite intensive efforts to develop α-synuclein- and TDP-43-targeting PET radiotracers, until now no suitable substance has been described (Mathis et al. 2017).

The FET protein family consists of fused in sarcoma (FUS), Ewing sarcoma (EWS) and TATA-binding protein associated factor 15 (TAF15) and was first discovered as components of fusion oncogenes causing specific malignancies (Mackenzie and Neumann 2016). As DNA/RNA binding proteins, they are predominantly located in the cell nucleus and are involved in DNA/RNA metabolism as well as in the maintenance of genomic stability (Mackenzie and Neumann 2016; Svetoni et al. 2016). In approximately 5–10% of all FTLD cases, the intracellular inclusions are FTLD-Tau- and TDP-43-negative in immunohistochemical examination. But they can be labeled using FUS/EWS/TAF15 antibodies and are therefore classified as FTLD-FUS or FTLD-FET group (Mackenzie and Neumann 2016). Similar to α-synuclein and TDP-43, the existing literature does not reveal any reports regarding FTLD-FET targeting radiotracers. Considering the low prevalence of these diseases (FTLD-TDP is rare, FTLD-FET is even rarer), the search for suitable radiotracers targeting these proteins is so far less active.

Summary and conclusion

In the last two to three decades, a large number of novel radiotracers for direct and indirect imaging neurodegenerative processes and their underlying pathology have been developed. Several of them have been approved and are used in clinical routine for early and differential diagnosis as well as for evaluation of disease progression. Others are appreciated as valuable research tracers. However, the more pathological components of the different neurodegenerative diseases are discovered, the more new and interesting issues occur. Such issues are (i) the classification of neurodegenerative disorders in clinical routine, (ii) the identification of targets for possible new radiotracers, (iii) the identification of novel radiotracer targets, (iv) the accurate monitoring strategy of such therapy trials. Some of them could be answered by PET studies with new radiotracers. But as important as the development of new radiotracers seems to be, at the moment it is equally important to sum up our gathered pieces of knowledge, combine them and try to get a more comprehensive understanding of the entire spectrum of neurodegenerative disorders (Fig. 9).

Fig. 9
figure 9

Overview of available imaging biomarkers detecting characteristics of neurodegenerative diseases ante-mortem/in-vivo and histopathological hallmarks which are/were, so far, only detectable post-mortem. Furthermore, the shift from a phenotypical to a biological definition of neurodegenerative diseases as it currently emerges in research settings is addressed

Full size image