Abstract

The development of radiotracers for use in vivo to image β-amyloid (Aβ) plaques in cases of Alzheimer's disease (AD) is an important, active area of research. The presence of Aβ aggregates in the brain is generally accepted as a hallmark of AD. Since the only definitive diagnosis of AD is by postmortem staining of affected brain tissue, the development of techniques which enable one to image Aβ plaques in vivo has been strongly desired. Furthermore, the quantitative evaluation of Aβ plaques in the brain could facilitate evaluation of the efficacy of antiamyloid therapies currently under development. This paper reviews the current situation in the development of agents for SPECT-based imaging of Aβ plaques in Alzheimer's brains.

1. Introduction

Alzheimer's disease (AD) is an age-related, irreversible form of dementia characterized by memory loss, a progressive decline in intellectual ability, language impairment, and personality and behavioral changes that eventually interfere with daily life. The accumulation of β-amyloid (Aβ) aggregates (major protein aggregates of senile plaques) in the brain is considered one of the hallmarks of AD [1, 2]. Today, the clinical diagnosis of AD is primarily based on history and memory testing, which is often difficult and not accurate, as the early cognitive and behavioral symptoms of AD are difficult to distinguish from normal signs of aging. To facilitate the early diagnosis of this disease, there is an urgent need for the sensitive noninvasive detection of biomarkers for the pathophysiology. Toward achieving this goal, nuclear imaging techniques such as positron emission computed tomography (PET) and single photon emission computed tomography (SPECT) have been employed. Radionuclide-labeled agents targeting the Aβ plaques in the brain may greatly facilitate the diagnosis of AD and new antiamyloid therapies [37]. The differential diagnosis for AD includes a large number of other diseases such as vascular dementia, frontal temporal lobe dementia (FTLD) complex, and dementia with Lewy bodies (DLB) as well as rarer neurodegenerative diseases such as Creutzfeld-Jacob disease (CJD). Importantly, AD subjects will always have Aβ plaques, whereas Aβ is seen not at all or only sporadically in most of these other diseases. In each case, appropriate prognosis and treatment require accurate diagnostic assessment.

Developing Aβ imaging agents is currently an emerging field of research. The basic requirements for suitable Aβ imaging agents include (i) good penetration of the blood-brain barrier, (ii) selective binding to Aβ plaques, and (iii) clear and contrasting signals between plaques and nonplaques (Figure 1). Based on these requirements, several promising agents with the backbone structure of DDNP, thioflavin-T and Congo Red have been synthesized and evaluated for use in vivo as probes to image Aβ plaques in AD brain. Clinical trials in AD patients have been conducted with several PET imaging agents including [11C]PIB [810], [11C]SB-13 [6, 11], [11C]BF-227 [12], [11C]AZD2184 [13], [18F]FDDNP [1416], [18F]BAY94-9172 [7, 17, 18], [18F] AV-45 [1921], and [18F]GE-067 [22] (Figure 2), indicating the imaging of Aβ plaques in living brain tissue to be useful for the diagnosis of AD. The 11C-labeled agents limit their use to on-site cyclotrons and sophisticated radiochemistry laboratories due to the short half-life (20 min) of 11C. PET agents with the longer half-life (110 min) radioisotope 18F have recently been developed and could increase the availability of Aβ imaging to all PET facilities, but still represents a minority of modern hospitals, as only a small fraction of hospitals have a PET scanner. Since SPECT is more valuable than PET in terms of routine diagnostic use, the development of more useful Aβ imaging agents for SPECT has been a critical issue. However, progress in developing imaging agents targeting Aβ plaques is less advanced for SPECT than PET. In this review, we summarize the current situation in the development of probes for SPECT-based imaging of Aβ plaques in Alzheimer’s brains.

2. Radioiodinated Probes for Imaging of Aβ Plaques

Many radioiodinated imaging agents derived from Congo Red or thioflavin-T have been developed. Compounds 1 [29], 2 [29], 3 [40], 4 [23], 5 [24], and 6 [25] (Figure 3) are thought to be derived from Congo Red. Although 1, 2, and 3 showed unfavorable pharmacokinetics in vivo such as low uptake into the brain and a slow washout, the radioactivity pharmacokinetics of 5 and 6 was much improved. Because thioflavin-T has a lower molecular weight than Congo Red, implying greater blood-brain penetration, a number of groups have worked to develop probes for SPECT derived from thioflavin-T including 7 (IMPY) [2628], 8 (TZDM) [29], 9 (IBOX) [30], 10 (benzofuran derivatives) [31], and 11 (phenylindole derivatives) (Figure 4) [32].

Initially, Zhuang and coworkers prepared iodo-strylbenzene derivatives based on the chemical structure of Congo Red, [125I]IMSB (1) and [125I]ISB (2). These ligands exhibited low brain uptake likely due to two ionizable carboxyl groups [29]. Thus, a small and neutral thioflavin-T analog, [125I]TZDM (8), was prepared [29]. In vitro binding studies of these ligands, [125I]ISB, [125I]IMSB and [125I]TZDM, showed excellent binding affinities with values of 0.08, 0.13 and 0.06 nM for aggregates of Aβ(1–40) and 0.15, 0.73 and 0.14 nM for aggregates of Aβ(1–42), respectively. Interestingly, in a competitive-binding assay, different binding sites on Aβ(1–40) and Aβ(1–42) aggregates, which are mutually exclusive, were observed for Congo Red and thioflavin-T derivatives. Biodistribution experiments in normal mice after an i.v. injection showed that [125I]TZDM exhibited good uptake and retention in the brain, much higher than [125I]ISB and [125I]IMSB. Preliminary experiments on the biodistribution of [125I]TZDM in transgenic mice, engineered to produce excess Aβ plaques in the brain as an AD model, suggested labeling of Aβ aggregates in vivo. However, [125I]TZDM is not ideal as an imaging agent in vivo, due to its labeling of white matter, which significantly increases the background activity. To improve the pharmacokinetics of uptake and retention, Kung et al. and others have prepared several compounds derived from thioflavin-T and studied features such as affinity for Aβ aggregates in vitro and biodistribution in vivo. Interestingly, the compounds with good binding to Aβ aggregates share a common structural feature: either an N-methylamino- or N,N-dimethylaminophenyl group at one end of the molecule. The structural feature required for binding to Aβ aggregates appears to be simple and unique.

[123I]IMPY has been characterized as a potential agent for SPECT-based imaging of Aβ plaques. IMPY displayed selective labeling of Aβ plaques ex vivo in autoradiographic experiments using double-transgenic mice (PSAPP) as a model of AD [33]. Preliminary clinical data on [125I]IMPY in normal and AD patients showed a distinct distribution pattern similar to that of [11C]PIB [34, 35]. However, the signal-to-noise ratio for plaque labeling is not as high as that of [11C]PIB. The low contrast may be due to the fast clearance from brain and plasma observed in AD and normal subjects. But the rapid metabolism and instability of [123I]IMPY in vivo may have led to less than optimal signal-to-noise ratios for targeting Aβ plaques in the brain. Additional candidates are being explored for SPECT imaging of Aβ plaques in the brain.

Recently, the effects of polyhydroxyflavones on the formation, extension, and destabilization of Aβ aggregates have been studied in vitro [36]. These flavones dose-dependently inhibited the formation of Aβ aggregates, as well as destabilized preformed Aβ aggregates, indicating that they could interact directly with the aggregates. The findings in that report prompted us to use flavones as a core structure in the development of Aβ imaging agents. Furthermore, some recent studies have shown that electron-donating groups such as methylamino, dimethylamino, methoxy, and hydroxy groups play a critical role in the binding to Aβ aggregates. With these considerations in mind, we designed four radioiodinated flavones with a radioiodine at the 6-position and an electron-donating group at the 4′-position (Figure 5). Then we synthesized a series of flavone derivatives and evaluated their usefulness in vivo as SPECT Aβ imaging agents [37].

Experiments on the binding of [125I]12 to aggregates of Aβ(1–40) and Aβ(1–42) were carried out. Transformation of the saturation binding of [125I]12 to Scatchard plots gave linear plots, suggesting one binding site (Figure 6). [125I]12 showed excellent affinity for both Aβ(1–40) (  nM) and Aβ(1–42) (  nM) aggregates. The binding of nonradioactive flavone derivatives (compounds 12, 13, 14, and 15) was evaluated in experiments inhibiting [125I]12 from binding Aβ(1–40) and Aβ(1–42) aggregates. As shown in Table 1, all flavone derivatives competed well with [125I]12 ( = 13–77 nM). More interestingly, when thioflavin-T, and Congo Red gave high values (>1000 nM) (Table 1), indicating little competition. This finding suggests that these flavones may have a binding site on Aβ aggregates different from that of thioflavin T and Congo Red, although additional studies regarding the selectivity of binding affinity for Aβ aggregates are required.

Since the in vitro binding assays demonstrated the high affinity of the flavone derivatives for Aβ(1–40) and Aβ(1–42) aggregates, compounds 12, 13, 14, and 15 were investigated for their neuropathologic staining of Aβ plaques and NFTs in human AD brain sections (Figure 7). The compounds intensely stained Aβ plaques (Figures 7(a), 7(e), 7(i), and 7(m)), neuritic plaques (Figures 7(b), 7(f), 7(j), and 7(n)), and cerebrovascular amyloids (Figures 7(c), 7(g), 7(k), and 7(o)) with nearly the same pattern. However, as seen in Figures 7(a), 7(e), 7(i), and 7(m), these flavone compounds did not intensely stain the core region in so-called classic Aβ plaques, unlike the thioflavin-T and Congo Red derivatives previously reported as Aβ imaging probes, indicating that flavone derivatives may have somewhat distinct binding characteristics for amyloid fibrils. These flavone derivatives appear to stain not only neuritic Aβ plaques but also diffuse amyloid plaque deposits, which are known to be mainly composed of Aβ(1−42) [38] and to be the initial pathologic change in AD [39]. Thus flavone derivatives with high affinity for Aβ(1–42)-positive diffuse plaques may be more useful for presymptomatic detection of AD. Furthermore, 12, 13, 14, and 15 also showed high affinity for NFTs in AD brain sections (Figures 7(d), 7(h), 7(l), and 7(p)). These findings suggest that these flavone derivatives can bind amyloid fibrils and NFTs without the backbone structure of thioflavin-T or Congo Red and that quantitative evaluation of their cerebral localization may provide useful information on Aβ and tau pathology.

Four radioiodinated flavone ligands ([125I]12, [125I]13, [125I]14, and [125I]15) were evaluated for their biodistribution in vivo in normal mice. Previous studies suggest that the optimal lipophilicity range for brain entry is observed for compounds with log P-values between 1 and 3 [5]. All four ligands displayed optimal lipophilicity as reflected by log P-values of 1.94, 2.69, 2.41, and 1.92, respectively. As expected, these ligands exhibited high uptake ranging from 3.2% to 4.1% ID/g brain at 2 min postinjection, a level sufficient for imaging in the brain (Table 2). In addition, they displayed good clearance from the normal brain: 1.2, 1.0, 0.17, and 0.08% ID/g at 60 min postinjection for [125I]12, [125I]13, [125I]14, and [125I]15, respectively. Radioiodinated amyloid imaging agents such as [125I]m-I-stilbene (3) [40], [125I]TZDM (8) [29], [125I]IBOX (9) [30], and [125I]benzofuran (10) [31], and [125I]phenylindole (11) [32] reported previously showed good uptake, but a relatively slow washout from the normal brain. A low washout rate leads to high background activity and prevents the visualization of Aβ plaques in the AD brain. Appropriate properties in vivo (higher uptake and faster washout from the normal brain) make radioiodinated flavones useful candidates for SPECT tracers for Aβ imaging.

On the basis of this success in the development of SPECT imaging agents, to search for more useful candidates for Aβ imaging probes, we have designed a chemical modification of the flavone structure, and selected the chalcone and aurone structure as a novel core for Aβ imaging probes (Figure 5) [41, 42]. Chalcone and aurone are categorized as flavonoids containing a flavone. We newly designed and synthesized novel chalcone and aurone derivatives, and evaluated the effect of their structure–activity relationships on binding to Aβ aggregates and biodistribution in vivo using a compound with high affinity [4245]. Currently, SPECT imaging agents based on chalcone and aurone are optimized.

Most of the Aβ imaging probes reported previously have two aromatic rings. Among them, 1,4-diphenyltriazole and 2,5-diphenylthiophene derivatives have triazole and thiophene between two benzene rings, respectively, and it has been shown that they have high-binding affinity for binding to Aβ aggregates despite the kinds of substituted groups [46, 47]. In an attempt to further develop novel ligands for the imaging of Aβ plaques in AD, we designed a series of 3,5-diphenyl-1,2,4-oxadiazole (18) [48, 49] and 2,4-diphenyl-1,3,5-oxadiazole (19) [49] derivatives (Figure 8). Although the diphenyloxadiazole pharmacophore with high-binding affinity for Aβ aggregates may be useful as a backbone structure to develop novel Aβ imaging agents, additional modifications are necessary to improve the uptake and rapid clearance of nonspecifically bound radiotracers.

Many factors such as molecular size, ionic charge, and lipophilicity affect the brain uptake of compounds. Since lipophilicity of the compounds generally increases by introduction of iodine, the large higher lipophilicity of the radioiodinated compounds may constitute one reason for the low brain uptake. In the future, introduction of hydrophilic substituted groups into the amyloid-binding scaffolds will be required to develop more promising radioiodinated tracers with in favorable in vivo pharmacokinetics.

3. 99mTc Complexes for Imaging of Aβ Plaques

99mTc (  h, 141 keV) has become the most commonly used radionuclide in diagnostics for SPECT, because it is readily produced by an  99Mo/99mTc generator, the medium gamma-ray energy it emits is suitable for detection, and its physical half-life is compatible with the biological localization and residence time required for imaging. Its ready availability, essentially 24 h a day, and easiness of use make it the radionuclide of choice. New 99mTc-labeled imaging agents will provide simple, convenient, and widespread SPECT-based methods for detecting and eventually quantifying Aβ plaques in living brain tissue.

Han and co-workers described a positively charged 99Tc-complex of Congo red (20) which binds to Aβ aggregates in vitro [50]. The basic structure of this complex is the Congo red backbone in which the biphenyl moiety is replaced by a bipyridyl moiety capable of complexing Tc in the presence of tert-butylisonitrile as a coligand. Although these Tc complexes showed high affinity for Aβ aggregates in vitro, they have not been tested in vivo. Dezutter and co-workers reported a 99mTc-labeled conjugate of Congo Red with a monoamide-monoaminedithiol (MAMA) chelating ligand [51]. However, brain uptake of this 99mTc-labeled Congo Red derivative (21) was minimal, probably because of its large size and ionized character at physiological pH. Serdons and co-workers reported the synthesis of a neutral 99mTc-labeled derivative of thioflavin-T (22), namely a benzothiazole derivative conjugated with a bisamine-bisthiol (BAT) ligand, and its biological characterization [52]. It was demonstrated that the 99mTc-labeled thioflavin-T derivative binds in vitro to Aβ plaques. Despite its high lipophilicity and neutral character, the 99mTc complex did not cross the blood-brain barrier to a sufficient degree and thus is not useful for the detection of AD in vivo. Recently, Chen et al. reported that the 99mTc-labeled thioflavin T using MAMA as a chelation ligand (23) demonstrated the binding to Aβ aggregates in sections of brain tissue from transgenic mice and AD patients [53]. In addition, 23 can penetrate the blood-brain barrier with high initial brain uptake and moderate washout. These results are encouraging for further exploration of their derivatives as imaging agents for Aβ plaques in the brain.

As described above, several 99mTc-labeled imaging probes have been developed (Figure 9) [5055], but no clinical study of them has been reported. While these 99mTc complexes showed high affinity for Aβ aggregates or Aβ plaques in vitro, they suffered the same unfavorable in vivo pharmacokinetics in normal mice, that is, a slow washout. Therefore, to make them promising probes for imaging Aβ plaques in the brain, additional molecular modifications to improve their pharmacokinetics in vivo are required.

Recently, we have developed several 99mTc complexes based on flavone, chalcone, aurone, and benzofuran derivatives with monoamine-monoamide dithiol (MAMA) and bis-amino-bis-thiol (BAT) as chelation ligands (Figures 10 and 11). MAMA and BAT were selected taking into consideration the permeability of the blood-brain barrier, because they form an electrically neutral complex with 99mTc [56]. We then evaluated their biological potential as probes by testing their affinity for Aβ aggregates and Aβ plaques in sections of brain tissue from Tg2576 mice and their uptake in and clearance from the brain in biodistribution experiments using normal mice.

Initially, four 99mTc-labeled chalcone derivatives and their corresponding rhenium analogues were tested as potential probes for imaging Aβ plaques (Figure 10) [57]. The chalcones showed higher affinity for Aβ(1–42) aggregates than did 99mTc complexes and, in sections of brain tissue from an animal model of AD, the four Re-chalcones intensely stained Aβ plaques. In biodistribution experiments using normal mice, 99mTc-BAT-chalcone (24) displayed high uptake in the brain (1.48%ID/g) at 2 min after injection. The radioactivity washed out from the brain rapidly (0.17%ID/g at 60 min), a highly desirable feature for an imaging agent. Although potential existence of cis- and anti-isomers was expected, one single isomer was isolated in the preparation of 24, 25, and 26. The chemical identities of 24, 25, and 26 were confirmed by NMR and MS, but their absolute configurations have not yet been determined by X-ray crystallography.

As for 99mTc complexes based on benzofuran, we evaluated binding affinity using Re-BAT-BF and Re-MAMA-BF, analogs of 99mTc-BAT-BF (27) and 99mTc-MAMA-BF (28), respectively. Both ligands inhibited the binding of [125I]IMPY to Aβ(1–42) aggregates in a dose-dependent manner, indicating an affinity for Aβ aggregates (Figure 12). Their values were 11.5 and 24.4 nM, respectively, suggesting that Re-BAT-BF displayed higher affinity than Re-MAMA-BF. Next, the affinity of 99mTc-BAT-BF (27) for Aβ plaques was investigated in vitro using sections of Tg2576 mouse brain (Figure 13). Furthermore, the radioactivity of 99mTc-BAT-BF (27) corresponded with the areas of staining with thioflavin S, a dye commonly used for Aβ plaques. In contrast, normal mouse brain displayed no detectable accumulation of  99mTc-BAT-BF (27). The results suggest that 99mTc-BAT-BF (27) binds to Aβ plaques in the mouse brain in addition to synthetic Aβ aggregates.

The biodistribution of 99mTc-BAT-BF (27) and 99mTc-MAMA-BF (28) was examined in normal mice (Table 3). 99mTc-BAT-BF (27) showed greater uptake (1.34%ID/g) than 99mTc-MAMA-BF (28) (0.74%ID/g) at 2 min after injection. The uptake of  99mTc-BAT-BF (27) peaked at 10 min after injection, reaching 1.37%ID/g, sufficient uptake for Aβ imaging, and 60% of the radioactivity had been washed out from the brain by 60 min after injection. The uptake of 99mTc-MAMA-BF (28) peaked 30 min after the injection at 1.23%ID/g, and the washout from the brain was slower than that of  99mTc-BAT-BF (27) throughout the time course, which is unsuitable for imaging in vivo. The combination of good affinity for Aβ plaques, uptake, and clearance makes  99mTc-BAT-BF (27) a promising probe for the detection of Aβ plaques in the brain. The results of the present study should provide useful information for the development of 99mTc-labeled probes for the imaging of Aβ plaques in the brain, although there are some difficulties associated with the large size of  99mTc complex in the molecular design of  99mTc-labeled Aβ imaging probes to enhance the penetration of blood-brain barrier.

4. Conclusion

Many PET probes targeting Aβ plaques in the brain have been tested clinically and demonstrated potential utility. Unfortunately, the short half-life (11C; 20 min, 18F; 110 min) of 11C- or 18F-labeled probes except 18F-FDG limits their use at major academic PET facilities with on-site cyclotrons and sophisticated radiochemistry laboratories. On the other hand, many more hospitals have the capacity to perform SPECT. Aβ imaging probes labeled with SPECT isotopes especially the inexpensive and readily available 99mTc will have more widespread clinical applicability especially in developing countries that cannot afford expensive cyclotron and PET scanners. The development of novel 123I- or 99mTc-labeled Aβ imaging probes may lead to simple and convenient SPECT imaging methods for detecting and eventually quantifying Aβ plaques in living brain tissue.

Acknowledgments

The results in this paper were partly obtained at Nagasaki University. They would like to express our sincere thanks to Professor Morio Nakayama for his helpful, continuous advice and support. The research reviewed in this paper was possible only through the dedication, enthusiasm, and creativity of all my coworkers, whose names are acknowledged in the publications from our laboratory cited here. These studies were supported by a Grant-in-aid for Young Scientists (A) and (B) from the Ministry of Education, Culture, Sports, Science and Technology, the Industrial Technology Research Grant Program from New Energy and Industrial Technology Development Organization (NEDO), the Program for Promotion of Fundamental Biomedical Innovation (NIBIO), and a Health Labour Sciences Research grant.