Imaging noradrenergic influence on amyloid pathology in mouse models of Alzheimer’s disease
- First Online:
- Cite this article as:
- Winkeler, A., Waerzeggers, Y., Klose, A. et al. Eur J Nucl Med Mol Imaging (2008) 35: 107. doi:10.1007/s00259-007-0710-0
Molecular imaging aims towards the non-invasive characterization of disease-specific molecular alterations in the living organism in vivo. In that, molecular imaging opens a new dimension in our understanding of disease pathogenesis, as it allows the non-invasive determination of the dynamics of changes on the molecular level.
Imaging of AD characteristic changes by μPET
The imaging technology being employed includes magnetic resonance imaging (MRI) and nuclear imaging as well as optical-based imaging technologies. These imaging modalities are employed together or alone for disease phenotyping, development of imaging-guided therapeutic strategies and in basic and translational research.
In this study, we review recent investigations employing positron emission tomography and MRI for phenotyping mouse models of Alzheimers’ disease by imaging. We demonstrate that imaging has an important role in the characterization of mouse models of neurodegenerative diseases.
KeywordsMolecular imagingμPETMouse modelsAD
Alzheimer’s disease (AD) is the most common neurodegenerative disorder and can only be diagnosed with certainty by post-mortem examination of brain tissue. The disease is characterised by extracellular deposition of senile (amyloid) plaques, intracellular neurofibrillary tangles (NFT), synaptic loss and neurodegeneration. The amyloid plaques contain insoluble fibrils of the amyloid-beta (Aβ) fragment of a larger precursor protein (amyloid precursor protein [APP]), whereas intracellular NFT consist of hyperphosphorylated tau protein [1, 2]. Mutations in presenilin 1 (PS1), presenilin 2 (PS2) and APP genes have been associated with familial forms of AD and have been shown to alter normal processing of APP by proteases (secretases) causing the extracellular accumulation of amyloid plaques [3–7]. Cross-sectional studies on individuals with AD or elderly controls have contributed to the current understanding of the disease but are limited by the fact that analyses can only be performed in post-mortem brain. Therefore, the advantages of mouse models are as follows: (1) analysing the disease-specific pathophysiological mechanisms, (2) understanding the genetic alterations or interactions, as well as (3) testing therapeutic interventions at definite timepoints. Nowadays, these mouse models of AD are typically transgenic (tg) mice carrying mutations in the APP, PS1 or PS2 gene. Several tg mouse models have been generated, which are reviewed by Higgins and Jacobsen . Some of them, e.g., the tg mouse models Tg2576 or APP23 overexpress the human APP (hAPP) in combination with distinct mutations like the ‘Swedish’ mutation (APPK670N, M671L [9, 10]) or in case of the PDAPP mouse model the ‘London’ mutation (APPV717F ). Others like the PS1 M146L or M146V model express mutant PS1  or, by crossing, the tg animals create double-tg mice (PS1/APP), carrying two of these mutations in the mouse genome . Except for the PS1 model, these murine models of AD typically express sufficiently high levels of hAPP and Aβ to insure amyloid deposition. Besides being useful tools in the analysis, understanding and possible treatment of the disease based on findings in histology, biochemistry, molecular biology and behavioural testing, these mouse models have been of help in characterisation of amyloid-imaging agents and have been used for non-invasive phenotyping by multi-tracer positron emission tomography (PET).
PET-based imaging in AD
PET allows non-invasive assessment of physiological, metabolic and molecular processes in humans and animals in vivo. With the achievements in detector technology, spatial resolution of PET has been considerably improved (1–2 mm), enabling for the first time investigations in small experimental animals such as mice. With the developments in radiochemistry and tracer technology, a variety of endogenously expressed and exogenously introduced genes can be analysed by PET. This opens up the exciting and rapidly evolving field of molecular imaging, aiming towards the noninvasive localisation of a biological process of interest in normal and diseased cells, both in animal models and humans in vivo. The main and most intriguing advantage of molecular imaging is the kinetic analysis of a given molecular event in the same experimental subject over time. This will allow non-invasive characterisation and “phenotyping” of animal models of human disease at various disease stages, under certain pathophysiological stimuli or after therapeutic intervention, respectively.
Cellular hexokinase by 2-[18F]fluoro-2-deoxy-d-glucose ([18F]FDG);
Acetylcholine esterase (AchE) by [11C]-N-methyl-4-piperidinylacetate ([11C]MP4A);
Benzodiazepine receptors by [11C]flumazenil ([11C]FMZ).
These tracers are clinically applied for early detection of AD ([18F]FDG, [11C]MP4A) and in the assessment of neuronal integrity after stroke ([11C]FMZ, [18F]FDG) [14, 15]. The overall attractive goal in neuroscience research is that, with the advent of μPET, these tracers can be applied in animal models for AD for non-invasive phenotyping of these models as well as in the development of experimental therapeutics, such as growth factor administration, gene therapy and cell transplantation. Animal PET will serve as a platform to test these therapies in a relatively efficient manner, allowing a more rapid progression from pre-clinical to clinical studies. Most importantly, the same in vivo imaging parameters can be used pre-clinically and clinically, enabling a direct comparison when translating experimental molecular imaging markers into clinical application.
To date, the cerebral glucose metabolism as measured by [18F]FDG-PET is a sensitive non-invasive surrogate marker for the diagnosis of AD [16, 17]. Typically, [18F]FDG-PET shows reduced glucose metabolism in posterior cingulate, temporoparietal and prefrontal association cortex with preservation of primary sensorimotor cortex and basal ganglia. The largest reduction is found in the posterior cingulate cortex , and this metabolic alteration becomes even apparent before the onset of cognitive impairment in persons at genetic risk for AD . Even very mild probable AD patients (defined by Mini Mental Score greater than 24) can be diagnosed with a 84% sensitivity and 93% specificity . Moreover, degeneration of cholinergic neurons located in basal ganglia, particular in the nucleus basalis of Meynert and subsequent alteration of cortical acetylcholine esterase activity is one of the central features in AD and Lewy body dementia, which can be assessed by [11C]MP4A-PET. After passing the blood–brain barrier, [11C]MP4A is hydrolysed by AChE and trapped within the brain proportional to regional AChE activity . Currently, the value of [11C]MP4A-PET is being investigated in the differential diagnosis of AD from other forms of dementia [21–24]. [11C]FMZ binds to central benzodiazepine receptors, and, as measured by PET, it is an early indicator of preserved cortical neuronal integrity . [18F]FDG together with [11C]FMZ was shown to be also helpful in the diagnosis of other forms of dementia including variants of Creutzfeld–Jakob disease .
Although [18F]FDG-, [11C]MP4A- and [11C]FMZ-PET reveal direct molecular information on glucose metabolism, AChE activity and binding to benzodiazepine receptors, they can only serve as surrogate markers for cellular density or neuronal integrity and thus as indirect surrogate measures of presence and progression of AD. Therefore, a further imaging marker that permits direct detection of the disease process, such as neurofibrillary tangles or Aβ plaques in vivo, would be very helpful. Direct imaging of these neuropathological changes would improve early diagnostic certainty of the disease, rather than waiting until behavioural manifestations can be detected. Most importantly, repetitive imaging of senile plaques in combination with glucose and acetylcholine metabolism as surrogate markers would provide a readout of the efficacy of experimental therapeutics aimed at removing these neuropathologic lesions. So far, invasive direct imaging of Aβ plaques has been developed by using fluorescein-labelled Aβ, anti-Aβ antibodies and fluorescent derivatives of thioflavin T and Congo red as specific targeting agents and multiphoton microscopy in live tg mice [26–28]. Although multiphoton microscopy requires craniotomy, these studies indicated that exogenous fluorescent targeting agents bind to existing plaques in vivo if delivery to the brain is sufficient. Intriguingly, this imaging method has been successfully used to follow plaque-directed immunotherapy in the same experimental animal over time .
Direct AD plaque imaging by PET
Based on findings with the amyloid-binding compound thioflavin, neutral benzothiazoles were further modified and investigated as potential PET agents. Klunk et al.  described the first human study using the PET radiotracer N-methyl-[11C]2-(4′-methylaminophenyl)-6-hydroxybenzothiazole ([11C]6-OH-BTA-1) also named “Pittsburgh Compound B” ([11C]PIB) and could demonstrate two- to threefold higher retention of the radio-labelled compound in areas of association cortex in patients with AD as compared to healthy controls. An Aβ-binding compound, which can be radio-labelled by 18F, has also been developed. 2-(1-(6-[(2-[18F]fluoroethyl)(methyl)amino]-2-naphthyl)ethylidene)malononitrile labelled senile plaques in tissue sections from AD brain; however, its in vivo specificity is not clarified yet due to its low clearance from white matter [30, 31].
Direct imaging of Aβ plaques in tg mouse models has been performed using the PIB compound. Toyama et al.  used the Tg2576 mouse model of AD to evaluate the feasibility of μPET imaging using the radio-labelled compound [11C]6-OH-BTA-1 to image and quantify Aβ plaques in vivo. Interestingly, they found excellent brain uptake of radioactivity in tg and wild-type mice but no specific binding of the labelled compound to Aβ plaques in tg animals compared with the wild type, although histochemical staining with thioflavin S showed numerous Aβ plaque deposits in the tg animals but not in the wild type. Similar findings have been reported by Klunk et al. [33, 34] in PIB μPET studies using the PS1/APP mouse model. To further consolidate these findings and to compare the differences in binding of PIB in human versus tg mouse brain, they performed in vivo, ex vivo and post-mortem in vitro studies with brain tissue from PS1/APP tg mice . Similar to the Tg2576 murine model, PS1/APP mice showed extensive plaque staining [34, 35], but no significant difference in the retention of [11C]PIB. Ex vivo studies (with increased sensitivity as compared to in vivo μPET) analysing the binding capacity of [11C]PIB in brains of 15-month-old PS1/APP mice did not show any significant difference either. Comparison of in vitro [3H]PIB binding to human AD brain as well as PS1/APP brain demonstrated a high affinity [3H]PIB-binding site in AD brains but not in PS1/APP mouse brain. The Bmax for binding of [3H]PIB to PS1/APP brain was more than 1,000-fold lower than the Bmax for binding of [3H]PIB to AD brain. A possible explanation for this suggests that there is low efficiency for forming high affinity PIB-binding sites during plaque deposition in tg mice, which may be due to differences in the time course of Aβ deposition or the presence of tissue-specific factors during deposit formation.
Imaging of AD characteristic changes by μPET
Our results demonstrate the usefulness of radiotracers originally applied to show a characteristic pattern of altered brain glucose metabolism, acetylcholine esterase activity or neuronal integrity in patients with AD, also in multi-modal μPET imaging for characterisation and non-invasive phenotyping of mouse models of AD. However, it should be pointed out that μPET imaging in mice still has major limitations with regards to spatial resolution of currently available μPET scanners, attenuation correction and correct quantification, as well as anesthesia-induced changes of radiotracer uptake. These technical considerations have to be taken appropriately into account and further elaborated to enable μPET-based phenotyping of mouse models of brain disease as a tool to investigate the dynamics of disease-specific molecular alterations in vivo.
Our work is supported in part by the Deutsche Forschungsgemeinschaft (DFG-Ja98/1-2), Center for Molecular Medicine Cologne (CMMC-C5), 6th FW EU grants EMIL (LSHC-CT-2004-503569) and DiMI (LSHB-CT-2005-512146).
Conflict of interest statement
The authors declare that they have no relevant financial or any other interests in this manuscript.