Abstract

Non-invasive imaging of tau pathology in the living brain would be useful for accurately diagnosing Alzheimer’s disease, tracking disease progression, and evaluating the treatment efficacy of disease-specific therapeutics. In this study, we evaluated the clinical usefulness of a novel tau-imaging positron emission tomography tracer 18F-THK5105 in 16 human subjects including eight patients with Alzheimer’s disease (three male and five females, 66–82 years) and eight healthy elderly controls (three male and five females, 63–76 years). All participants underwent neuropsychological examination and 3D magnetic resonance imaging, as well as both 18F-THK5105 and 11C-Pittsburgh compound B positron emission tomography scans. Standard uptake value ratios at 90–100 min and 40–70 min post-injection were calculated for 18F-THK5105 and 11C-Pittsburgh compound B, respectively, using the cerebellar cortex as the reference region. As a result, significantly higher 18F-THK5105 retention was observed in the temporal, parietal, posterior cingulate, frontal and mesial temporal cortices of patients with Alzheimer’s disease compared with healthy control subjects. In patients with Alzheimer’s disease, the inferior temporal cortex, which is an area known to contain high densities of neurofibrillary tangles in the Alzheimer’s disease brain, showed prominent 18F-THK5105 retention. Compared with high frequency (100%) of 18F-THK5105 retention in the temporal cortex of patients with Alzheimer’s disease, frontal 18F-THK5105 retention was less frequent (37.5%) and was only observed in cases with moderate-to-severe Alzheimer’s disease. In contrast, 11C-Pittsburgh compound B retention was highest in the posterior cingulate cortex, followed by the ventrolateral prefrontal, anterior cingulate, and superior temporal cortices, and did not correlate with 18F-THK5105 retention in the neocortex. In healthy control subjects, 18F-THK5105 retention was ∼10% higher in the mesial temporal cortex than in the neocortex. Notably, unlike 11C-Pittsburgh compound B, 18F-THK5105 retention was significantly correlated with cognitive parameters, hippocampal and whole brain grey matter volumes, which was consistent with findings from previous post-mortem studies showing significant correlations of neurofibrillary tangle density with dementia severity or neuronal loss. From these results, 18F-THK5105 positron emission tomography is considered to be useful for the non-invasive assessment of tau pathology in the living brain. This technique would be applicable to the longitudinal evaluation of tau deposition and allow a better understanding of the pathophysiology of Alzheimer’s disease.

Introduction

Senile plaques and neurofibrillary tangles are considered the major pathological hallmarks of Alzheimer’s disease (Braak and Braak, 1991). Senile plaques consist of extracellular aggregates of amyloid-β peptide cleaved from a longer amyloid precursor protein (Masters et al., 2006). The neocortical deposition of senile plaques is considered one of the earliest pathological alterations in Alzheimer’s disease and is observed even in the presymptomatic stages (Mintun et al., 2006; Rowe et al., 2007; Price et al., 2009). Recently proposed research diagnostic criteria for preclinical Alzheimer’s disease include cognitively intact elderly with abnormal amyloid-β deposition in the brain (Sperling et al., 2011). Preclinical Alzheimer’s disease is associated with future cognitive decline and mortality (Vos et al., 2013); however, several neuropathological studies have shown no significant association between density of amyloid-β plaques and the severity of dementia or neuronal loss (Arriagada et al., 1992; Bierer et al., 1995; Gomez-Isla et al., 1997), suggesting the involvement of other key factors in Alzheimer’s disease-related neurodegeneration.

Neurofibrillary tangles are comprised of paired helical filaments that result from the abnormal aggregation of tau protein (Grundke-Iqbal et al., 1986a, b; Lee et al., 1991). Initial neurofibrillary tangle lesions occur in the trans-entorhinal cortex, followed by entorhinal cortex and hippocampus involvement, progressing to temporal neocortex and finally to the other neocortical areas (Arnold et al., 1991; Braak and Braak, 1991). In contrast with senile plaques, neurofibrillary tangle formation correlates well with cognitive impairment severity (Arriagada et al., 1992; Berg et al., 1993; Bierer et al., 1995), an association that is considered to continue throughout the disease course (Abner et al., 2011). Furthermore, the inhibition of abnormal tau hyperphosphorylation and its aggregation appear to be promising therapeutic strategies in Alzheimer’s disease. Thus, non-invasive imaging of tau pathology would be useful to assist in the early and differential diagnosis of dementia, track the progression of disease-related pathology, and monitor the efficacy of anti-tau treatments.

18F-FDDNP has been reported to detect neurofibrillary tangle deposition (Shoghi-Jadid et al., 2002) and successfully differentiate subjects with Alzheimer’s disease and mild cognitive impairment from those with no cognitive impairment (Small et al., 2006). However, this tracer detects the combined signals of senile plaques and neurofibrillary tangles (Shoghi-Jadid et al., 2002). Several radiotracers have been developed for the selective visualization of neurofibrillary tangles in the living brain (Chien et al., 2013, 2014; Maruyama et al., 2013). Early clinical PET studies successfully differentiated patients with Alzheimer’s disease from cognitively normal elderly. However, the selective binding ability of these radiotracers to tau has not been fully validated in vivo.

For the development of a selective tau radiotracer, we screened β-sheet-binding small molecules and identified novel quinoline derivatives with high binding selectivity to tau deposits in Alzheimer’s disease brain samples (Okamura et al., 2005; Fodero-Tavoletti et al., 2011; Harada et al., 2013). Through a compound optimization process, we developed a novel 18F-labelled 2-arylquinoline derivative, 18F-THK5105 (Fig. 1), which showed high binding affinity and selectivity to tau protein deposits in Alzheimer’s disease brain sections (Okamura et al., 2013). This 18F-labelled radiotracer also exhibited high blood–brain barrier permeability and no defluorination in mice (Okamura et al., 2013). The present clinical study evaluated whether 18F-THK5105 PET could selectively bind to tau pathology in living patients with Alzheimer’s disease.

Materials and methods

Participants

Sixteen subjects, including eight patients with probable Alzheimer’s disease (three male and five females, age range 66–82 years) and eight age-matched healthy control subjects (three male and five females, age range 63–76 years), underwent both 18F-THK5105 and 11C-labelled Pittsburgh compound B (11C-PiB) PET scans (Table 1). Written informed consent was obtained from all participants. Study approval was obtained from the Austin Health Human Research Ethics Committee. Elderly healthy controls were recruited by advertisements in the community, and patients with Alzheimer’s disease were recruited from tertiary Memory Disorders Clinics or physicians who sub-specialize in dementia care. All participants were reviewed and classified on the basis of their clinical and neuropsychological performance by the consensus of a neurologist and a neuropsychologist who were blind to their PET results. The diagnosis of Alzheimer’s disease was made according to the National Institute of Neurological and Communicative Disorders and Stroke/Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria.

Table 1

Demographic characteristics of healthy control and Alzheimer’s disease subjects

 Healthy controls (n = 8) Alzheimer’s disease (n = 8) 
Age 70.5 ± 4.4 74.1 ± 6.9 
Gender (M/F) 3/5 3/5 
Years of education 15.4 ± 2.4 11.3 ± 3.2* 
CDR 0.0 0.9 ± 0.5* 
CDR-SOB 0.0 6.1 ± 4.9* 
MMSE score 28.8 ± 1.5 17.3 ± 6.6* 
Episodic memory scores −0.1 ± 0.8 −3.8 ± 0.3* 
Non-memory scores −0.1 ± 0.5 −3.0 ± 1.9* 
Grey matter volume (cm3302.7 ± 12.9 272.9 ± 22.6* 
Hippocampal volume (cm34.8 ± 0.5 4.0 ± 0.6* 
 Healthy controls (n = 8) Alzheimer’s disease (n = 8) 
Age 70.5 ± 4.4 74.1 ± 6.9 
Gender (M/F) 3/5 3/5 
Years of education 15.4 ± 2.4 11.3 ± 3.2* 
CDR 0.0 0.9 ± 0.5* 
CDR-SOB 0.0 6.1 ± 4.9* 
MMSE score 28.8 ± 1.5 17.3 ± 6.6* 
Episodic memory scores −0.1 ± 0.8 −3.8 ± 0.3* 
Non-memory scores −0.1 ± 0.5 −3.0 ± 1.9* 
Grey matter volume (cm3302.7 ± 12.9 272.9 ± 22.6* 
Hippocampal volume (cm34.8 ± 0.5 4.0 ± 0.6* 

*P<0.05 by the Mann–Whitney U test.

Neuropsychological evaluation

Cognitive impairment and dementia severity were evaluated with the Mini-Mental State Examination (MMSE), the Clinical Dementia Rating (CDR) and the CDR scale sum of boxes (CDR-SOB). In addition, composite episodic memory and non-memory scores were generated as previously described (Villemagne et al., 2011). Briefly, a composite episodic memory score was calculated by taking the average of the z-scores for the Rey Complex Figure Test, the long delay California Verbal Learning Test, Second Edition, and the Logical Memory II subscale of the Wechsler Memory Scale. A composite non-memory score was calculated by taking the average of the z-scores for the Boston Naming Test, letter fluency, category fluency, digit span forwards and backwards, digit symbol-coding, and Rey Complex Figure Test copy.

Image acquisition

MRI scanning was performed on a 3 T Siemens TRIO magnetic resonance system (Siemens Healthcare) using the ADNI 3D MPRAGE sequence with 1 × 1 mm in-plane resolution and 1.2 mm slice thickness, repetition time/echo time/inversion time = 2300/2.98/900, flip angle 9°, field of view 240 × 256, and 160 slices. T2 fast spin echo and FLAIR sequences were also obtained.

Two radiotracers, 18F-THK5105 and 11C-PiB, were prepared in the Centre for PET at Austin Hospital. 18F-THK5105 was synthesized by nucleophilic substitution of the tosylate precursor as described previously (Okamura et al., 2013). The decay-corrected average radiochemical yield of the production of 18F-THK5105 was 45%, with a radiochemical purity >95% and a specific activity of 229.6 GBq/µmol (6.2 ± 3.3 Ci/μmol). 11C-PiB was synthesized using the one-step 11C-methyl triflate approach as previously described (Rowe et al., 2007). The decay-corrected average radiochemical yield for 11C-PiB was 30%, with a radiochemical purity >98% and a specific activity of 30 ± 7.5 GBq/µmol.

A list-mode emission acquisition on an Allegro™ PET camera (Philips Medical Systems) was performed in 3D mode from 0–50 min and between 90–120 min after injection of 200 MBq 18F-THK5105. List-mode raw data for the initial 50 min of the acquisition were sorted off-line into 6 × 30-s, 7 × 1-min, 4 × 2.5-min, 2 × 5-min, and 6 × 10-min frames. The final 30 min were acquired as 6 × 5-min frames. The sorted sinograms were reconstructed using a 3D RAMLA algorithm. A 30-min acquisition (6 × 5-min frames) on an Allegro™ PET camera began 40 min after intravenous injection of 300 MBq 11C-PiB.

Figure 1

Chemical structures of 18F-THK523 and 18F-THK5105.

Figure 1

Chemical structures of 18F-THK523 and 18F-THK5105.

Image analysis

Hippocampal and cortical grey matter volumes were obtained using an automated volumetric measurement program (NeuroQuant: CorTechs Labs Inc) applied to the 3D MP RAGE MRI images. The primary MRI outcome measures were the grey cortical matter and hippocampal volumes normalized to total intracranial volume.

PET images were processed using a semi-automatic region of interest method. Firstly, standardized uptake value (SUV) images of 18F-THK5105 and 11C-PiB were obtained by normalizing tissue radioactivity concentration by injected dose and body weight. Subsequently, individual MRI T1 images were anatomically co-registered into individual PET images using Statistical Parametric Mapping software (SPM8: Wellcome Trust Centre for Neuroimaging, London, UK). Co-registered MRI and PET images were then spatially normalized to an MRI T1 template in Talairach space using SPM8. After spatial normalization, a region of interest template was placed on individual axial images in the cerebellar hemisphere, ventrolateral frontal cortex [Brodmann areas (BA) 10, 44, 45 and 46], lateral and medial orbitofrontal cortex (BA 11 and 12), superior temporal cortex (BA 22), inferior temporal cortex (BA 20 and 37), parietal cortex (BA 39 and 40), lateral occipital cortex (BA 18 and 19), anterior cingulate cortex, posterior cingulate cortex, mesial temporal cortex (BA 27, 28, 34 and 35), putamen, pons, and subcortical white matter. Regional SUVs were sampled using PMOD software (PMOD Technologies, Ltd).The ratio of regional SUV to cerebellar cortex SUV ratio (SUVR) was used as an index of tracer retention. Neocortical tau and amyloid-β burden were expressed as the average SUVR for the following cortical regions of interest: frontal, parietal, lateral temporal, and posterior cingulate for THK5105 and PiB, respectively. As in previous studies, a PiB SUVR threshold of 1.5 was used to categorize high and low amyloid-β burden.

Statistical analysis

Mann-Whitney’s U-tests were applied for comparison of the Alzheimer’s disease and healthy control groups. For comparison of regional radiotracer uptake, one-way repeated measures analysis of variance (ANOVA) followed by Bonferroni’s tests were performed. To examine the regional difference of tracer retention between neocortex and mesial temporal cortex, Wilcoxon matched-pairs signed rank tests were performed. Effect size coefficients (Cohen’s d) were calculated for the evaluation of group differences in PET measurements. Statistical significance for each analysis was defined as P < 0.05. Data are presented as mean ± standard deviation (SD).

Results

Healthy control and Alzheimer’s disease subject demographics are shown in Table 1. There were no significant differences between healthy control and Alzheimer’s disease groups with regard to age or gender; however, the Alzheimer’s disease group was significantly less educated than the healthy control group. As expected, significant differences between the two groups were observed for CDR and CDR-SOB scores, cognitive performance (MMSE, episodic memory, and non-memory scores), and brain volumetrics (grey matter and hippocampal volumes).

No toxic event was observed in the current clinical PET study. After intravenous administration of 18F-THK5105, all subjects showed rapid entry of the tracer into grey matter areas. The SUV time activity curves of 18F-THK5105 PET are shown in Fig. 2. The peak uptake and clearance rates of 18F-THK5105 in the cerebellar cortex were similar between healthy control (Fig. 2A) and Alzheimer’s disease (Fig. 2B) groups. In patients with Alzheimer’s disease, the inferior temporal cortex, which is an area known to contain high densities of neurofibrillary tangles in Alzheimer’s disease (Bouras et al., 1994), showed 18F-THK5105 retention compared to the cerebellum, especially at the later time points. In contrast, time activity curves in the inferior temporal cortex of healthy control subjects were nearly identical to those in the cerebellum. The subcortical white matter region showed relatively lower entry and slower clearance than grey matter areas, but no significant differences were observed for time activity curves between healthy control and Alzheimer’s disease groups (data not shown). The ratio of inferior temporal cortex to cerebellar SUVR became constant in all participants ∼90 min after injection of 18F-THK5105 (Fig. 2C). Therefore, we selected SUVR values from 90–100 min post-injection for the following analysis.

Figure 2

(A and B) 18F-THK5105 SUV TACs in the cerebellum (open circles) and inferior temporal cortex (filled circles) of eight healthy control subjects (A) and eight patients with Alzheimer’s disease (B). (C) SUVR time activity curves of 18F-THK5105 PET in eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD) (filled circles). Each point represents the mean ± SD.

Figure 2

(A and B) 18F-THK5105 SUV TACs in the cerebellum (open circles) and inferior temporal cortex (filled circles) of eight healthy control subjects (A) and eight patients with Alzheimer’s disease (B). (C) SUVR time activity curves of 18F-THK5105 PET in eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD) (filled circles). Each point represents the mean ± SD.

Summed SUVR images from 90–100 min post-injection for healthy control and Alzheimer’s disease subjects are shown in Fig. 3. Contrasted with a lack of remarkable 18F-THK5105 retention in the grey matter of healthy control subjects, patients with Alzheimer’s disease showed high grey matter 18F-THK5105 retention in the lateral and mesial temporal regions. 18F-THK5105 retention was additionally observed in the brain stem; however, similar retention in these areas was detected in healthy control subjects. When comparing the 90–100 min regional SUVR in Alzheimer’s disease and healthy control subjects, 18F-THK5105 SUVRs for the ventrolateral prefrontal, medial orbitofrontal, superior, and inferior temporal, parietal, posterior cingulate, and mesial temporal cortices were significantly higher in patients with Alzheimer’s disease (Table 2 and Fig. 4). Notably, SUVR in the inferior temporal cortex showed no overlap between the Alzheimer’s disease and healthy control groups (Fig. 4). 18F-THK5105 retention in other neocortical areas was relatively lower than in the inferior temporal area. The SUVR in the parietal cortex was elevated in 62.5% (5/8) of patients with Alzheimer’s disease; however, 18F-THK5105 retention in the ventrolateral prefrontal cortex was only elevated in 37.5% (3/8) of patients with Alzheimer’s disease. Mesial temporal 18F-THK5105 retention was significantly higher in patients with Alzheimer’s disease than in healthy control subjects. However, a substantial overlap of SUVR was observed between both groups. The SUVR values in the pons and subcortical white matter were nearly identical in both groups, but higher than other neocortical regions. The effect size value between Alzheimer’s disease and healthy control subjects was highest in the inferior temporal cortex, followed by the superior temporal, posterior cingulate, parietal, and medial orbitofrontal cortices and was lowest in the other regions examined (Table 2). Regional difference of 18F-THK5105 retention was additionally examined in healthy control subjects. As a result, mesial temporal 18F-THK5105 retention (mean SUVR = 1.17) was significantly higher than neocortical 18F-THK5105 retention (mean SUVR = 1.05) in healthy control subjects.

Figure 3

18F-THK5105 PET images from 60–80 min post-injection in a healthy control subject (72-years-old, CDR 0, MMSE 29) and a patient with Alzheimer’s disease (68-years-old, CDR 1.0, MMSE 20).

Figure 3

18F-THK5105 PET images from 60–80 min post-injection in a healthy control subject (72-years-old, CDR 0, MMSE 29) and a patient with Alzheimer’s disease (68-years-old, CDR 1.0, MMSE 20).

Figure 4

Regional 18F-THK5105 SUVR values from 60–80 min post-injection in healthy control and Alzheimer’s disease (AD) subjects. *P<0.05 by the Mann–Whitney U test.

Figure 4

Regional 18F-THK5105 SUVR values from 60–80 min post-injection in healthy control and Alzheimer’s disease (AD) subjects. *P<0.05 by the Mann–Whitney U test.

Table 2

Regional 18F-THK5105 SUVR values in healthy control and Alzheimer’s disease subjects

Region Healthy controls Alzheimer’s disease Cohen’s d 
Ventrolateral prefrontal 1.08 ± 0.08 1.23 ± 0.14* 1.33 
Lateral orbitofrontal 1.01 ± 0.08 1.15 ± 0.13 1.32 
Medial orbitofrontal 1.17 ± 0.06 1.29 ± 0.09* 1.55 
Superior temporal 1.04 ± 0.06 1.22 ± 0.07* 2.75 
Inferior temporal 1.09 ± 0.04 1.32 ± 0.08* 3.58 
Parietal 0.99 ± 0.08 1.16 ± 0.13* 1.59 
Lateral occipital 1.07 ± 0.06 1.18 ± 0.15 1.01 
Anterior cingulate 1.07 ± 0.11 1.12 ± 0.13 0.35 
Posterior cingulate 1.04 ± 0.08 1.20 ± 0.12* 1.61 
Mesial temporal 1.17 ± 0.05 1.26 ± 0.10* 1.17 
Putamen 1.41 ± 0.10 1.52 ± 0.17 0.83 
Pons 1.88 ± 0.14 1.89 ± 0.23 0.03 
Subcortical white matter 1.22 ± 0.15 1.22 ± 0.15 0.01 
Neocortex 1.05 ± 0.05 1.23 ± 0.08* 2.68 
Region Healthy controls Alzheimer’s disease Cohen’s d 
Ventrolateral prefrontal 1.08 ± 0.08 1.23 ± 0.14* 1.33 
Lateral orbitofrontal 1.01 ± 0.08 1.15 ± 0.13 1.32 
Medial orbitofrontal 1.17 ± 0.06 1.29 ± 0.09* 1.55 
Superior temporal 1.04 ± 0.06 1.22 ± 0.07* 2.75 
Inferior temporal 1.09 ± 0.04 1.32 ± 0.08* 3.58 
Parietal 0.99 ± 0.08 1.16 ± 0.13* 1.59 
Lateral occipital 1.07 ± 0.06 1.18 ± 0.15 1.01 
Anterior cingulate 1.07 ± 0.11 1.12 ± 0.13 0.35 
Posterior cingulate 1.04 ± 0.08 1.20 ± 0.12* 1.61 
Mesial temporal 1.17 ± 0.05 1.26 ± 0.10* 1.17 
Putamen 1.41 ± 0.10 1.52 ± 0.17 0.83 
Pons 1.88 ± 0.14 1.89 ± 0.23 0.03 
Subcortical white matter 1.22 ± 0.15 1.22 ± 0.15 0.01 
Neocortex 1.05 ± 0.05 1.23 ± 0.08* 2.68 

*P<0.05 by the Mann–Whitney U test.

As reported in previous PET studies (Klunk et al., 2004; Mintun et al., 2006; Rowe et al., 2007), 11C-PiB SUVR values were significantly greater in the neocortical regions of patients with Alzheimer’s disease compared to healthy control subjects (Table 3). All patients with Alzheimer’s disease showed marked and extensive PiB retention in neocortical areas. On the other hand, neocortical PiB retention in healthy control subjects was not significant, except for one healthy control case that only showed high 11C-PiB retention in the frontal cortex. In contrast with the highest neocortical 18F-THK5105 retention in the inferior temporal cortex of patients with Alzheimer’s disease, 11C-PiB retention in the same group was highest in the posterior cingulate cortex, followed by the ventrolateral prefrontal, anterior cingulate, and superior temporal cortices. The PiB effect size value between Alzheimer’s disease and healthy control subjects was also highest in the parietal cortex, followed by the posterior cingulate and superior temporal cortices. As shown in Fig. 5, the pattern of cortical 18F-THK5105 retention was completely different from that of 11C-PiB retention in patients with Alzheimer’s disease, which was prominent in the frontal cortex and precuneus, but not evident in the mesial temporal cortex. In contrast, 18F-THK5105 retention was evident in both lateral and mesial temporal areas but not so remarkable in other neocortical areas. There was no correlation between neocortical 18F-THK5105 and 11C-PiB SUVR values in patients with Alzheimer’s disease (r = 0.17, P = 0.703). In addition, one healthy control case showing elevated PiB retention in the frontal cortex did not show any significant retention of 18F-THK5105.

Figure 5

18F-THK5105 PET images from 60–80 min post-injection and 11C-PiB PET images from 40–70 min post-injection in an Alzheimer’s disease patient (68-years-old, CDR 1.0, MMSE 20). Co-registered magnetic resonance images are shown on the right.

Figure 5

18F-THK5105 PET images from 60–80 min post-injection and 11C-PiB PET images from 40–70 min post-injection in an Alzheimer’s disease patient (68-years-old, CDR 1.0, MMSE 20). Co-registered magnetic resonance images are shown on the right.

Table 3

Regional 11C-PiB SUVR values in healthy control and Alzheimer’s disease subjects

Region Healthy controls Alzheimer’s disease Cohen’s d 
Ventrolateral prefrontal 1.32 ± 0.39 2.92 ± 0.82* 2.49 
Lateral orbitofrontal 1.08 ± 0.25 1.66 ± 0.50* 1.46 
Medial orbitofrontal 1.36 ± 0.23 2.38 ± 0.63* 2.15 
Superior temporal 1.11 ± 0.13 2.67 ± 0.67* 3.22 
Inferior temporal 1.08 ± 0.08 2.42 ± 0.66* 2.88 
Parietal 1.22 ± 0.17 2.56 ± 0.51* 3.54 
Lateral occipital 1.25 ± 0.11 2.07 ± 0.61* 1.86 
Anterior cingulate 1.39 ± 0.28 2.79 ± 0.70* 2.64 
Posterior cingulate 1.31 ± 0.24 3.15 ± 0.78* 3.22 
Mesial temporal 1.29 ± 0.15 1.65 ± 0.32* 1.42 
Putamen 1.48 ± 0.20 2.64 ± 0.62* 2.50 
Pons 2.04 ± 0.33 2.04 ± 0.40 0.00 
Subcortical white matter 2.02 ± 0.38 2.06 ± 0.66 0.07 
Neocortex 1.21 ± 0.14 2.75 ± 0.66* 3.21 
Region Healthy controls Alzheimer’s disease Cohen’s d 
Ventrolateral prefrontal 1.32 ± 0.39 2.92 ± 0.82* 2.49 
Lateral orbitofrontal 1.08 ± 0.25 1.66 ± 0.50* 1.46 
Medial orbitofrontal 1.36 ± 0.23 2.38 ± 0.63* 2.15 
Superior temporal 1.11 ± 0.13 2.67 ± 0.67* 3.22 
Inferior temporal 1.08 ± 0.08 2.42 ± 0.66* 2.88 
Parietal 1.22 ± 0.17 2.56 ± 0.51* 3.54 
Lateral occipital 1.25 ± 0.11 2.07 ± 0.61* 1.86 
Anterior cingulate 1.39 ± 0.28 2.79 ± 0.70* 2.64 
Posterior cingulate 1.31 ± 0.24 3.15 ± 0.78* 3.22 
Mesial temporal 1.29 ± 0.15 1.65 ± 0.32* 1.42 
Putamen 1.48 ± 0.20 2.64 ± 0.62* 2.50 
Pons 2.04 ± 0.33 2.04 ± 0.40 0.00 
Subcortical white matter 2.02 ± 0.38 2.06 ± 0.66 0.07 
Neocortex 1.21 ± 0.14 2.75 ± 0.66* 3.21 

*P<0.05 by the Mann–Whitney U test.

Finally, we explored the relationship of neocortical or mesial temporal radiotracer retention to cognitive parameters in patients with Alzheimer’s disease. Neocortical 18F-THK5105 retention was significantly correlated with MMSE scores (r = −0.781, P = 0.022), CDR (r = 0.730, P = 0.050) and CDR-SOB (r = 0.779, P = 0.030), but not correlated with composite episodic memory and non-memory scores. In contrast, neocortical 11C-PiB retention showed no significant correlation with any cognitive parameters (Fig. 6). There was no correlation between radiotracer retention in mesial temporal cortex and cognitive parameters. Furthermore, we explored the relationship of radiotracer retention with brain volumetrics. When all subjects were included in the analysis, a significant correlation was observed between mesial temporal 18F-THK5105 retention and hippocampal volumes (r = −0.565, P = 0.023) and between neocortical 18F-THK5105 retention and whole brain grey matter volumes (r = −0.649, P = 0.007). When only patients with Alzheimer’s disease were included in this analysis, hippocampal volumes were significantly correlated with neocortical 18F-THK5105 retention (r = −0.765, P = 0.027), but not correlated with mesial temporal 18F-THK5105 retention. However, there were no correlations between mesial temporal 11C-PiB retention and hippocampal volumes and between neocortical 11C-PiB retention and whole brain grey matter volumes (Fig. 7).

Figure 6

Correlation of neocortical 18F-THK5105 and 11C-PiB SUVR with MMSE scores (upper) and CDR-SOB scores (lower). Data from eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD, filled circles) are shown.

Figure 6

Correlation of neocortical 18F-THK5105 and 11C-PiB SUVR with MMSE scores (upper) and CDR-SOB scores (lower). Data from eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD, filled circles) are shown.

Figure 7

Correlation of 18F-THK5105 and 11C-PiB SUVR with adjusted hippocampal volumes (upper) or grey matter volumes (lower). Data from eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD, filled circles) are shown.

Figure 7

Correlation of 18F-THK5105 and 11C-PiB SUVR with adjusted hippocampal volumes (upper) or grey matter volumes (lower). Data from eight healthy control subjects (open circles) and eight patients with Alzheimer’s disease (AD, filled circles) are shown.

Discussion

In this study, the novel radiotracer 18F-THK5105 successfully differentiated patients with Alzheimer’s disease from healthy control subjects. The pattern of 18F-THK5105 distribution in patients with Alzheimer’s disease appears similar to the reported neurofibrillary tangle distribution in the post-mortem Alzheimer’s disease brain. 18F-THK5105 retention in the inferior temporal cortex, where neurofibrillary tangle accumulation is highest in Alzheimer’s disease, was observed in most patients with Alzheimer’s disease. In contrast to the high frequency of 18F-THK5105 retention in the temporal cortices of Alzheimer’s disease cases, ventrolateral prefrontal 18F-THK5105 retention was less frequent (3/8) and was only observed in cases with moderate-to-severe Alzheimer’s disease (MMSE range 10–17). This finding is consistent with neurofibrillary tangle distribution in post-mortem Alzheimer’s disease brain, where there is a higher frequency of neurofibrillary tangles in the temporal cortex than the frontal cortex (Arnold et al., 1991; Bouras et al., 1994; Haroutunian et al., 1999). It is also in agreement with recent PET results using other novel radiotracers (18F-T807 and 18F-T808) that demonstrated higher radiotracer retention in the lateral temporal lobe compared to the frontal lobe and selective binding ability to paired helical filament tau (Chien et al., 2013, 2014). These findings suggest a spreading of tau pathology (Soto, 2012; Mohamed et al., 2013) from temporal areas to the other association cortices. A longitudinal assessment of tau pathology will help elucidate the spatial patterns of tau pathology progression in the living brain. In addition, as observed in 18F-T807 and 18F-T808 PET studies (Chien et al., 2013, 2014), 18F-THK5105 retention in the mesial temporal area was relatively lower than in the lateral temporal area in patients with Alzheimer’s disease, which conflicts with microscopic observations showing higher neurofibrillary tangle density in the entorhinal cortex and hippocampus of Alzheimer’s disease brain compared to the neocortex (Arnold et al., 1991). One possible explanation for this phenomenon is the partial volume effect of radiotracer signals (Muller-Gartner et al., 1992). 18F-THK5105 retention in the mesial temporal cortex might be underestimated in patients with severe hippocampal atrophy.

18F-THK5105 retention in patients with Alzheimer’s disease was closely associated with dementia severity. This finding is consistent with results from previous post-mortem studies showing significant correlations of neurofibrillary tangle density with dementia severity (Arriagada et al., 1992; Bierer et al., 1995; Berg et al., 1998). Our results further demonstrate that hippocampal atrophy is significantly correlated with 18F-THK5105 retention but not with 11C-PiB retention in the same area. In addition, the neocortical grey matter volume was negatively correlated with global 18F-THK5105 retention in the neocortex. These findings correspond with the neuropathological observation that the density of neurofibrillary tangles, but not senile plaques, is closely associated with neuronal loss (Gomez-Isla et al., 1996, 1997). Intriguingly, 18F-THK5105 retention in healthy control subjects was significantly higher in the mesial temporal cortex (SUVR = 1.17) than in the neocortex (SUVR = 1.05). This finding is likely to reflect age-related tau accumulation in this area. In future studies, the association of mesial temporal 18F-THK5105 retention with ageing should be evaluated in a large population. It is also still unclear whether or not tau accumulation precedes neuronal loss. To answer this question, mesial temporal cortex tau density should be evaluated in the mild cognitive impairment population, as well as cognitively normal individuals with amyloid-β deposition, and these results should be compared with fluorodeoxyglucose and brain atrophy in a longitudinal analysis.

The amount of neocortical 18F-THK5105 retention (SUVR = 1.23) was considerably lower than that of 11C-PiB (SUVR = 2.75) in patients with Alzheimer’s disease. This is thought to result from higher concentrations of amyloid-β fibrils than tau fibrils in the Alzheimer’s disease brain (Villemagne et al., 2012). Therefore, a tau-specific radiotracer must be highly sensitive and selective to tau protein fibrils. Our previous study demonstrated that the binding affinity of 18F-THK5105 for tau fibrils (Kd = 1.45 nM) was 25-times higher than to amyloid-β fibrils (Kd = 35.9 nM) (Okamura et al., 2013). Autoradiography studies further confirmed the preferential binding ability of 18F-THK5105 to tau protein deposits in Alzheimer’s disease brain sections. In this PET study, all Alzheimer’s disease cases were PiB-positive and showed remarkable PiB retention in broad neocortical areas. As reported by many researchers, these patients with Alzheimer’s disease showed prominent PiB retention in the ventrolateral prefrontal cortex (SUVR > 2.0), reflecting high amyloid-β pathology in this area. In contrast, 18F-THK5105 retention in the frontal cortex was not elevated in more than half of the Alzheimer’s disease cases (Fig. 4). In addition, one healthy control case showing PiB retention in the frontal cortex showed no retention of 18F-THK5105 in this area. These results support the low sensitivity of 18F-THK5105 to amyloid-β plaques.

Compared with previous 18F-THK523 PET data, specific 18F-THK5105 retention in grey matter was considerably higher whereas white matter retention was considerably lower than those observed for 18F-THK523. This observation is consistent with previous in vitro autoradiograms showing a higher signal-to-background ratio of 18F-THK5105 than 18F-THK523 in Alzheimer’s disease brain sections (Okamura et al., 2013). This is probably due to the higher binding affinity of 18F-THK5105 to tau protein fibrils. The peak brain entry of 18F-THK5105 (cerebellar SUV = 4.5), which was observed before 6 min post-injection, was higher than for 18F-THK523 and other reported radiotracers (18F-T807, 18F-T808 and 11C-PBB3) and almost identical to the reported SUV value of 11C-PiB (Klunk et al., 2004). In addition, 18F-THK5105 did not accumulate in the skull, which is often the result of defluorination and interferes with visual PET image inspection. These pharmacokinetic properties are unique advantages of 18F-THK5105 over the other reported radiotracers. Conversely, one of the drawbacks of 18F-THK5105 is the existence of non-specific tracer retention in the brainstem, thalamus, and subcortical white matter, which reflects the binding of tracer to β-sheet structures present in myelin basic protein, similar to that observed for 11C-PiB (Stankoff et al., 2011). Nevertheless, the 18F-THK5105 signal in the subcortical white matter was not visually noticeable as compared with 18F-THK523 and 11C-PiB. The relatively slower kinetics of 18F-THK5105 cause the high background signals in grey matter, which may make the white matter signals less noticeable. Actually, the clearance of 18F-THK5105 from normal grey matter tissue was slower than that of PiB because of higher lipophilicity for 18F-THK5105 (LogP = 3.03) than PiB (LogP = 1.20). As another tau-imaging radiotracer candidate, we have developed a 2-arylquinoline derivative named 18F-THK5117 (Okamura et al., 2013), which is more hydrophilic and shows faster pharmacokinetics in mice than 18F-THK5105. It is expected to show faster clearance from normal brain tissue in humans and higher signal-to-background ratio than 18F-THK5105. A clinical trial testing 18F-THK5117 is currently underway.

Tau deposits are also present in non-Alzheimer’s disease tauopathies including frontotemporal lobe degeneration, corticobasal degeneration, progressive supranuclear palsy and chronic traumatic encephalopathy. Although THK523 failed to detect tau deposits in these non-Alzheimer’s disease tauopathies, we recently observed in vitro binding ability of THK5105 and THK5117 to glial tau pathology in corticobasal degeneration and progressive supranuclear palsy. Therefore, clinical PET study in non-Alzheimer’s disease tauopathies will be necessary to decide whether 18F-THK5105 is applicable to the study of a broad range of tauopathies.

The results of the current study indicate that 18F-THK5105 has adequate safety to be used as a clinical PET tracer. 18F-THK5105 PET demonstrated high tracer retention in sites susceptible to paired helical filament-tau deposition in patients with Alzheimer’s disease and distinctly differentiated patients with Alzheimer’s disease from age-matched healthy controls. Collectively, these findings suggest that 18F-THK5105 is useful for the non-invasive evaluation of tau pathology in humans and could be employed to study longitudinal tau deposition in healthy and pathological ageing.

Acknowledgements

We thank Prof. Michael Woodward, Dr John Merory, Dr Gordon Chan, Dr Kenneth Young, Dr David Darby, Ms Fiona Lamb, and the Brain Research Institute for their assistance with this study.

Funding

This study was supported by the research fund from GE Healthcare, the SEI (Sumitomo Electric Industries, Ltd) Group CSR Foundation, NHMRC grant 1044361 and NIRG grant from the Alzheimer's Association, the Industrial Technology Research Grant Program of the New Energy and Industrial Technology Development Organization (NEDO) in Japan (09E51025a), Health and Labor Sciences Research Grants from the Ministry of Health, Labor, and Welfare of Japan, Grant-in-Aid for Scientific Research (B) (23390297) and “Japan Advanced Molecular Imaging Program (J-AMP)” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Abbreviations

    Abbreviations
  • CDR

    clinical dementia rating

  • MMSE

    Mini-Mental State Examination

  • PiB

    Pittsburgh compound B

  • SOB

    sum of boxes

  • SUV

    standardized uptake value

  • SUVR

    ratio of regional SUV to cerebellar cortex SUV ratio

References

Abner
EL
Kryscio
RJ
Schmitt
FA
Santacruz
KS
Jicha
GA
Lin
Y
, et al.  . 
“End-stage” neurofibrillary tangle pathology in preclinical Alzheimer's disease: fact or fiction?
J Alzheimers Dis
 , 
2011
, vol. 
25
 (pg. 
445
-
53
)
Arnold
SE
Hyman
BT
Flory
J
Damasio
AR
Van Hoesen
GW
The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer's disease
Cereb Cortex
 , 
1991
, vol. 
1
 (pg. 
103
-
16
)
Arriagada
PV
Growdon
JH
Hedley-Whyte
ET
Hyman
BT
Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer's disease
Neurology
 , 
1992
, vol. 
42
 (pg. 
631
-
9
)
Berg
L
McKeel
DW
Jr
Miller
JP
Baty
J
Morris
JC
Neuropathological indexes of Alzheimer's disease in demented and nondemented persons aged 80 years and older
Arch Neurol
 , 
1993
, vol. 
50
 (pg. 
349
-
58
)
Berg
L
McKeel
DW
Jr
Miller
JP
Storandt
M
Rubin
EH
Morris
JC
, et al.  . 
Clinicopathologic studies in cognitively healthy aging and Alzheimer's disease: relation of histologic markers to dementia severity, age, sex, and apolipoprotein E genotype
Arch Neurol
 , 
1998
, vol. 
55
 (pg. 
326
-
35
)
Bierer
LM
Hof
PR
Purohit
DP
Carlin
L
Schmeidler
J
Davis
KL
, et al.  . 
Neocortical neurofibrillary tangles correlate with dementia severity in Alzheimer's disease
Arch Neurol
 , 
1995
, vol. 
52
 (pg. 
81
-
8
)
Bouras
C
Hof
PR
Giannakopoulos
P
Michel
JP
Morrison
JH
Regional distribution of neurofibrillary tangles and senile plaques in the cerebral cortex of elderly patients: a quantitative evaluation of a one-year autopsy population from a geriatric hospital
Cereb Cortex
 , 
1994
, vol. 
4
 (pg. 
138
-
50
)
Braak
H
Braak
E
Neuropathological stageing of Alzheimer-related changes
Acta Neuropathol
 , 
1991
, vol. 
82
 (pg. 
239
-
59
)
Checn
DT
Bahri
S
Szardenings
AK
Walsh
JC
Mu
F
Su
MY
, et al.  . 
Early clinical PET imaging results with the novel PHF-tau radioligand [F-18]-T807
J Alzheimers Dis
 , 
2013
, vol. 
34
 (pg. 
457
-
68
)
Chien
DT
Szardenings
AK
Bahri
S
Walsh
JC
Mu
FR
Xia
CF
, et al.  . 
Early Clinical PET Imaging Results with the Novel PHF-Tau Radioligand [F18]-T808
J Alzheimers Dis
 , 
2014
, vol. 
38
 (pg. 
171
-
84
)
Fodero-Tavoletti
MT
Okamura
N
Furumoto
S
Mulligan
RS
Connor
AR
McLean
CA
, et al.  . 
18F-THK523: a novel in vivo tau imaging ligand for Alzheimer's disease
Brain
 , 
2011
, vol. 
134
 (pg. 
1089
-
100
)
Gomez-Isla
T
Hollister
R
West
H
Mui
S
Growdon
JH
Petersen
RC
, et al.  . 
Neuronal loss correlates with but exceeds neurofibrillary tangles in Alzheimer's disease
Ann Neurol
 , 
1997
, vol. 
41
 (pg. 
17
-
24
)
Gomez-Isla
T
Price
JL
McKeel
DW
Jr
Morris
JC
Growdon
JH
Hyman
BT
Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer's disease
J Neurosci
 , 
1996
, vol. 
16
 (pg. 
4491
-
500
)
Grundke-Iqbal
I
Iqbal
K
Quinlan
M
Tung
YC
Zaidi
MS
Wisniewski
HM
Microtubule-associated protein tau. A component of Alzheimer paired helical filaments
J Biol Chem
 , 
1986a
, vol. 
261
 (pg. 
6084
-
9
)
Grundke-Iqbal
I
Iqbal
K
Tung
YC
Quinlan
M
Wisniewski
HM
Binder
LI
Abnormal phosphorylation of the microtubule-associated protein tau (tau) in Alzheimer cytoskeletal pathology
Proc Natl Acad Sci USA
 , 
1986b
, vol. 
83
 (pg. 
4913
-
7
)
Harada
R
Okamura
N
Furumoto
S
Tago
T
Maruyama
M
Higuchi
M
, et al.  . 
Comparison of the binding characteristics of [18F]THK-523 and other amyloid imaging tracers to Alzheimer's disease pathology
Eur J Nucl Med Mol Imaging
 , 
2013
, vol. 
40
 (pg. 
125
-
32
)
Harada
R
Okamura
N
Furumoto
S
Yoshikawa
T
Arai
H
Yanai
K
, et al.  . 
Use of a benzimidazole derivative BF-188 in fluorescence multispectral imaging for selective visualization of tau protein fibrils in the Alzheimer's disease brain
Mol Imaging Biol
 , 
2014
, vol. 
16
 (pg. 
19
-
27
)
Haroutunian
V
Purohit
DP
Perl
DP
Marin
D
Khan
K
Lantz
M
, et al.  . 
Neurofibrillary tangles in nondemented elderly subjects and mild Alzheimer disease
Arch Neurol
 , 
1999
, vol. 
56
 (pg. 
713
-
8
)
Klunk
WE
Engler
H
Nordberg
A
Wang
Y
Blomqvist
G
Holt
DP
, et al.  . 
Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B
Ann Neurol
 , 
2004
, vol. 
55
 (pg. 
306
-
19
)
Lee
VM
Balin
BJ
Otvos
L
Jr
Trojanowski
JQ
A68: a major subunit of paired helical filaments and derivatized forms of normal Tau
Science
 , 
1991
, vol. 
251
 (pg. 
675
-
8
)
Maruyama
M
Shimada
H
Suhara
T
Shinotoh
H
Ji
B
Maeda
J
, et al.  . 
Imaging of tau pathology in a tauopathy mouse model and in Alzheimer patients compared to normal controls
Neuron
 , 
2013
, vol. 
79
 (pg. 
1094
-
108
)
Masters
CL
Cappai
R
Barnham
KJ
Villemagne
VL
Molecular mechanisms for Alzheimer's disease: implications for neuroimaging and therapeutics
J Neurochem
 , 
2006
, vol. 
97
 (pg. 
1700
-
25
)
Mintun
MA
Larossa
GN
Sheline
YI
Dence
CS
Lee
SY
Mach
RH
, et al.  . 
[11C]PIB in a nondemented population: potential antecedent marker of Alzheimer disease
Neurology
 , 
2006
, vol. 
67
 (pg. 
446
-
52
)
Mohamed
NV
Herrou
T
Plouffe
V
Piperno
N
Leclerc
N
Spreading of tau pathology in Alzheimer's disease by cell-to-cell transmission
Eur J Neurosci
 , 
2013
, vol. 
37
 (pg. 
1939
-
48
)
Muller-Gartner
HW
Links
JM
Prince
JL
Bryan
RN
McVeigh
E
Leal
JP
, et al.  . 
Measurement of radiotracer concentration in brain gray matter using positron emission tomography: MRI-based correction for partial volume effects
J Cereb Blood Flow Metab
 , 
1992
, vol. 
12
 (pg. 
571
-
83
)
Okamura
N
Furumoto
S
Harada
R
Tago
T
Yoshikawa
T
Fodero-Tavoletti
M
, et al.  . 
Novel 18F-labeled arylquinoline derivatives for noninvasive imaging of tau pathology in Alzheimer disease
J Nucleic Med
 , 
2013
, vol. 
54
 (pg. 
1420
-
7
)
Okamura
N
Suemoto
T
Furumoto
S
Suzuki
M
Shimadzu
H
Akatsu
H
, et al.  . 
Quinoline and benzimidazole derivatives: candidate probes for in vivo imaging of tau pathology in Alzheimer's disease
J Neurosci
 , 
2005
, vol. 
25
 (pg. 
10857
-
62
)
Price
JL
McKeel
DW
Jr
Buckles
VD
Roe
CM
Xiong
C
Grundman
M
, et al.  . 
Neuropathology of nondemented aging: presumptive evidence for preclinical Alzheimer disease
Neurobiol Aging
 , 
2009
, vol. 
30
 (pg. 
1026
-
36
)
Rowe
CC
Ng
S
Ackermann
U
Gong
SJ
Pike
K
Savage
G
, et al.  . 
Imaging beta-amyloid burden in aging and dementia
Neurology
 , 
2007
, vol. 
68
 (pg. 
1718
-
25
)
Shoghi-Jadid
K
Small
GW
Agdeppa
ED
Kepe
V
Ercoli
LM
Siddarth
P
, et al.  . 
Localization of neurofibrillary tangles and beta-amyloid plaques in the brains of living patients with Alzheimer disease
Am J Geriatr Psychiatry
 , 
2002
, vol. 
10
 (pg. 
24
-
35
)
Small
GW
Kepe
V
Ercoli
LM
Siddarth
P
Bookheimer
SY
Miller
KJ
, et al.  . 
PET of brain amyloid and tau in mild cognitive impairment
N Engl J Med
 , 
2006
, vol. 
355
 (pg. 
2652
-
63
)
Soto
C
In vivo spreading of tau pathology
Neuron
 , 
2012
, vol. 
73
 (pg. 
621
-
3
)
Sperling
RA
Aisen
PS
Beckett
LA
Bennett
DA
Craft
S
Fagan
AM
, et al.  . 
Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease
Alzheimers Dement
 , 
2011
, vol. 
7
 (pg. 
280
-
92
)
Stankoff
B
Freeman
L
Aigrot
MS
Chardain
A
Dolle
F
Williams
A
, et al.  . 
Imaging central nervous system myelin by positron emission tomography in multiple sclerosis using [methyl-(1)(1)C]-2-(4'-methylaminophenyl)- 6-hydroxybenzothiazole
Ann Neurol
 , 
2011
, vol. 
69
 (pg. 
673
-
80
)
Villemagne
VL
Furumoto
S
Fodero-Tavoletti
MT
Harada
R
Mulligan
RS
Kudo
Y
, et al.  . 
The challenges of tau imaging
Future Neurol
 , 
2012
, vol. 
7
 (pg. 
409
-
21
)
Villemagne
VL
Pike
KE
Chetelat
G
Ellis
KA
Mulligan
RS
Bourgeat
P
, et al.  . 
Longitudinal assessment of Abeta and cognition in aging and Alzheimer disease
Ann Neurol
 , 
2011
, vol. 
69
 (pg. 
181
-
92
)
Vos
SJ
Xiong
C
Visser
PJ
Jasielec
MS
Hassenstab
J
Grant
EA
, et al.  . 
Preclinical Alzheimer's disease and its outcome: a longitudinal cohort study
Lancet Neurol
 , 
2013
, vol. 
12
 (pg. 
957
-
65
)