European Journal of Nuclear Medicine and Molecular Imaging

, Volume 38, Issue 2, pp 343–351

In vivo changes in microglial activation and amyloid deposits in brain regions with hypometabolism in Alzheimer’s disease

Authors

  • Masamichi Yokokura
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Norio Mori
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Shunsuke Yagi
    • Laboratory of Human Imaging Research, Molecular Imaging Frontier Research CenterHamamatsu University School of Medicine
  • Etsuji Yoshikawa
    • Central Research LaboratoryHamamatsu Photonics K.K.
  • Mitsuru Kikuchi
    • Department of Psychiatry and Neurobiology, Graduate School of Medical ScienceKanazawa University
  • Yujiro Yoshihara
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Tomoyasu Wakuda
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Genichi Sugihara
    • Research Center for Child Mental DevelopmentHamamatsu University School of Medicine
  • Kiyokazu Takebayashi
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Shiro Suda
    • Research Center for Child Mental DevelopmentHamamatsu University School of Medicine
  • Yasuhide Iwata
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
  • Takatoshi Ueki
    • Department of AnatomyHamamatsu University School of Medicine
  • Kenji J. Tsuchiya
    • Research Center for Child Mental DevelopmentHamamatsu University School of Medicine
  • Katsuaki Suzuki
    • Research Center for Child Mental DevelopmentHamamatsu University School of Medicine
  • Kazuhiko Nakamura
    • Department of Psychiatry and NeurologyHamamatsu University School of Medicine
    • Laboratory of Human Imaging Research, Molecular Imaging Frontier Research CenterHamamatsu University School of Medicine
Original Article

DOI: 10.1007/s00259-010-1612-0

Cite this article as:
Yokokura, M., Mori, N., Yagi, S. et al. Eur J Nucl Med Mol Imaging (2011) 38: 343. doi:10.1007/s00259-010-1612-0

Abstract

Purpose

Amyloid β protein (Aβ) is known as a pathological substance in Alzheimer’s disease (AD) and is assumed to coexist with a degree of activated microglia in the brain. However, it remains unclear whether these two events occur in parallel with characteristic hypometabolism in AD in vivo. The purpose of the present study was to clarify the in vivo relationship between Aβ accumulation and neuroinflammation in those specific brain regions in early AD.

Methods

Eleven nootropic drug-naïve AD patients underwent a series of positron emission tomography (PET) measurements with [11C](R)PK11195, [11C]PIB and [18F]FDG and a battery of cognitive tests within the same day. The binding potentials (BPs) of [11C](R)PK11195 were directly compared with those of [11C]PIB in the brain regions with reduced glucose metabolism.

Results

BPs of [11C](R)PK11195 and [11C]PIB were significantly higher in the parietotemporal regions of AD patients than in ten healthy controls. In AD patients, there was a negative correlation between dementia score and [11C](R)PK11195 BPs, but not [11C]PIB, in the limbic, precuneus and prefrontal regions. Direct comparisons showed a significant negative correlation between [11C](R)PK11195 and [11C]PIB BPs in the posterior cingulate cortex (PCC) (p < 0.05, corrected) that manifested the most severe reduction in [18F]FDG uptake.

Conclusion

A lack of coupling between microglial activation and amyloid deposits may indicate that Aβ accumulation shown by [11C]PIB is not always the primary cause of microglial activation, but rather the negative correlation present in the PCC suggests that microglia can show higher activation during the production of Aβ in early AD.

Keywords

Alzheimer’s diseaseMicrogliaAmyloid β proteinPositron emission tomography

Introduction

Amyloid β protein (Aβ) accumulation is important in the pathology of Alzheimer’s disease (AD). The amyloid hypothesis here posits that Aβ accumulation is responsible for triggering tau phosphorylation and neurofibrillary tangle formation, which lead to neurodegeneration in AD [1]. It has been reported that Aβ accumulation can activate microglia [24], and the activated microglia in turn accelerate Aβ accumulation by releasing proinflammatory neurotoxic substances [57]. In the AD brain, the region of increased microglial activity likely accompanies cerebral glucose hypometabolism and a higher rate of brain atrophy, suggesting that the degree of microglial activation predicts disease progression in AD [8].

Aβ accumulation would reportedly precede a decline in glucose metabolism and cognitive function and reach a plateau at an early clinical stage of AD [9, 10]. However, the Aβ accumulation seems unrelated to the reductions in glucose metabolism or Mini-Mental State Examination (MMSE) scores in AD patients [1014]. In contrast to the fibril form of Aβ, soluble Aβ oligomer that is produced earlier than formation of Aβ fibrils is considered to be a more deleterious substance which can activate microglia and stimulate secretion of cytokines [24]. Although animal experiments hinted that microglial activity could increase before Aβ accumulation [15, 16], little is known about the relationship between activated microglia and Aβ oligomer or fibril accumulation in the AD brain in vivo.

To clarify the in vivo relationships among these events, i.e. microglia activation, Aβ accumulation and glucose metabolism or neuronal function, we for the first time measured the levels of binding of radiotracers that could reflect those severities in early AD within the same day by PET.

Materials and methods

Participants

A total of 15 patients were initially enrolled in this study, but 4 patients were unable to proceed to the third PET measurement (glucose metabolism) because they unintentionally had a meal before the third scan. Accordingly, we collected data from 11 patients with AD (6 men and 5 women; mean age 70.6 ± 6.4 years) who were all able to undergo three kinds of PET scans in this study. The diagnosis of AD was based on the criteria of the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association (NINCDS/ADRDA) [17] and the Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV). The exclusion criteria were: (1) the presence of significant white matter microvascular changes on magnetic resonance imaging (MRI) over and above a few scattered lacunes compatible with normal aging, (2) smoking and drinking much alcohol regularly, (3) taking donepezil, antipsychotics and hypnotics, (4) taking anxiolytics and nonsteroidal anti-inflammatory drugs (NSAIDs) regularly and (5) having untreated hypertension or diabetes. The above exclusion criteria were also applied when selecting age-matched healthy control subjects (n = 10 for [11C](R)PK11195 measurement, n = 11 for [11C]PIB and [18F]FDG measurements). Neuropsychological assessment for all AD patients comprised a MMSE, word generation test and cubic copy as shown in Table 1. Healthy controls had no neurological problems and no abnormalities on MRI. The present study was approved by the Ethics Committee of the Hamamatsu Medical Center, and written informed consent was obtained from all participants.
Table 1

Characteristics of AD patients

Patient No.

Age

Sex

Handedness

Disease durationa

MMSE

Word generationb

Cubic copyc

(n = 11)

(years)

(years)

Correct

Incorrect

1

70

F

Right

7

18

7

5

×

2

69

M

Right

3

16

2

0

×

3

76

M

Rght

3

24

7

0

4

69

F

Right

1

23

9

4

5

69

F

Right

2

20

6

2

×

6

81

M

Right

0.5

25

11

4

7

74

F

Right

2

22

4

0

8

57

M

Right

1

25

12

0

9

75

M

Right

2

26

10

0

10

75

M

Right

3

15

3

0

×

11

62

F

Right

2

23

8

3

Mean ± SD

70.6 ± 6.4

  

2.4 ± 1.7

21.5 ± 3.5

   

aThe duration between disease onset and PET examination

bWord generation test: counting as many words as possible within 1 min from a category of animals. In this table, the word “correct” means the number of words from an appropriate category and “incorrect” means an inappropriate one

cCubic copy: correct, × incorrect

MRI scanning

MRI was performed to determine the areas of the regions to be used for setting regions of interest (ROIs) using a static magnet (0.3 T MRP7000AD, Hitachi, Tokyo, Japan) with the following acquisition parameters; three-dimensional mode sampling, repetition time (TR)/echo time (TE) (200/23), 75° flip angle, 2-mm slice thickness with no gap and 256 × 256 matrices. The MRI measures and a mobile PET gantry allowed us to reconstruct the PET images parallel to the intercommissural (ACPC) line without reslicing; using this approach, we were able to allocate ROIs on the target regions of the original PET images [18].

PET data acquisition

Patients underwent a series of PET measurements after a battery of neuropsychological tests and MRI. We used a high-resolution brain PET scanner (SHR12000, Hamamatsu Photonics K.K., Hamamatsu, Japan), which was capable of yielding 47 tomographic images [19]. A 10-min transmission scan for attenuation correction using a 68Ge/68Ga source was conducted with the subject’s head fixed by a radiosurgery-purpose thermoplastic face mask under resting conditions. After backprojection and filtering (Hanning filter, cutoff frequency 0.2 cycles per pixel), the image resolution was 2.9 × 2.9 × 3.4 mm full-width at half-maximum (FWHM). The voxel of each reconstructed image measured 1.3 × 1.3 × 3.4 mm. After a bolus intravenous injection of 280 MBq dose of [11C](R)PK11195, 32 serial PET scans (time frames: 4 × 30, 20 × 60 and 8 × 300 s) were performed within 62 min. After taking a rest for 2.5 h, 40 serial PET scans (time frames: 12 × 10, 18 × 60 and 10 × 300 s) were performed during 70 min after injection of 280 MBq [11C]PIB. Later, a static 15-min PET scan was done 45 min after injection of 100 MBq dose of [18F]FDG. No arterial sampling was performed in each measurement.

Image data processing

The binding potential (BP) of [11C](R)PK11195 was estimated based on a simplified reference tissue model [20, 21], in which we used a normalized mean time-activity curve created from control subjects as a reference tissue curve. This procedure has been described elsewhere [22]. Briefly, a normalized input curve was first created by averaging the ROIs placed over the bilateral frontal cortex, temporal cortex, thalamus, basal ganglia, cerebellar hemisphere and brain stem in the control group. The normalized mean tissue-activity curve was then used as the reference input function, because the desirable reference region free from specific binding is not present in patients with neurodegenerative disorders. The normalized input curve derived from the control group was used as the time-activity curve for the reference region of both the control subjects and AD patients.

To evaluate Aβ accumulation, [11C]PIB BP was estimated based on a noninvasive Logan plot method using an input curve created from the cerebellar hemisphere in each subject [23, 24]. All BP parametric PET images were generated using PMOD 2.95 software (PMOD Technologies Ltd., Zurich, Switzerland).

To evaluate neuronal function (glucose metabolism), a semi-quantitative ratio index of [18F]FDG was calculated as standardized uptake value ratio (SUVR), where the SUV showed tracer activity per injected dose normalized to body weight and then the SUVs of each region in the AD subjects were divided by the SUV of the cerebellar hemisphere in the same subjects and expressed as the SUVRs.

Image data analysis and statistics

SPM analysis

To explore the brain regions showing characteristic accumulation of the radiotracers, we examined the whole brain using a voxel-wise analytic method. For this purpose, SPM2 was used (Wellcome Department of Cognitive Neurology, London, UK, http://www.fil.ion.ucl.ac.uk/spm/software/spm2/). All [11C](R)PK11195 BP, [11C]PIB BP and [18F]FDG SUVR parametric images were first normalized to the Montreal Neurological Institute space and smoothed with an isotropic Gaussian kernel of 8 mm. The between-group comparisons (AD vs healthy control) for each parameter were performed using t statistics on a voxel-by-voxel basis with a statistical threshold set at p < 0.05 corrected for multiple comparisons (FDR), where the MMSE scores served as confounding covariates.

ROI analysis

Referring to the knowledge from the results of SPM analysis, multiple ROIs (44–474 mm2) were drawn bilaterally over the anterior cingulate cortex (ACC), posterior cingulate cortex (PCC), precuneus (Prec), superior frontal cortex (SFC) (Brodmann area or BA9), middle frontal cortex (MFC) (BA9), parahippocampal cortex (PHC) and hippocampus (Hip) on the ACPC-aligned MR images of each subject [25]. Using an image processing system (Dr. View, Asahi Kasei Co., Tokyo, Japan), we determined ROIs on the MR images reformatted to the size of PET images. While PET images were displayed side by side together with the MR images, the delineated ROIs were placed on the same area on both the MR and the corresponding PET images. Because both reconstructed PET and MRI images were obtained parallel to the intercommissural line, they theoretically required no reorientation procedure [26]. In this way, these ROIs were transferred onto the corresponding quantitative PET parametric images. We examined correlations between MMSE scores and PET parameters in each delineated brain region. We also made direct comparisons among PET parameters obtained across multiple ROIs. The Pearson correlation and simple regression analyses were used with statistical significance set at p < 0.05 corrected for multiple comparisons (SPSS version 17 J; SPSS Japan Inc., Tokyo, Japan).

Results

Levels of each PET parameter in AD patients vs healthy controls

The SPM analysis showed significant increases in [11C](R)PK11195 BP in the medial frontal and parietal cortex and left temporal cortex in AD (p < 0.05, corrected) (Fig. 1a, Supplementary Table 1). This voxel-wise analysis was confirmed by the ROI analysis showing that the quantitative value of [11C](R)PK11195 BP was significantly higher in these regions in the AD group (Table 2). The analysis also showed that the level of [11C]PIB BP was significantly higher in the medial and lateral parietal cortex, temporal cortex and frontal cortex in the AD group (p < 0.05, corrected) (Fig. 1b, Supplementary Table 1). In addition, there were significant reductions in [18F]FDG SUVR in the medial and lateral parietal cortex, temporal cortex and frontal cortex in the AD group compared to the healthy control group (p < 0.05, corrected) (Fig. 1c, Supplementary Table 1). These findings in AD patients in comparison with healthy subjects were all consistent with the previous literature. Figure 2 shows a sample of PET parametric images of [11C](R)PK11195 BP (Fig. 2a), [11C]PIB BP (Fig. 2b) and [18F]FDG SUVR (Fig. 2c) in AD patient 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1612-0/MediaObjects/259_2010_1612_Fig1_HTML.gif
Fig. 1

Results of SPM analyses. The regions with statistically significant increases in [11C](R)PK11195 BP (a) and [11C]PIB BP (b) and decrease in [18F]FDG SUVR (c) in AD patients compared with healthy control subjects (p < 0.05, corrected). The upper images show glass brains and lower images denote the rendering of PET and MRI fusions. The colour bar denotes levels of [11C](R)PK11195 BP, [11C]PIB BP and [18F]FDG SUVR

Table 2

Levels of all PET measures in the AD and control groups

Region

 

[11C](R)PK11195 BP

[11C]PIB BP

[18F]FDG SUVR

AD

HC

AD

HC

AD

HC

ACC

Mean

0.39*

0.13

0.81*

0.20

1.00

1.06

SD

0.17

0.09

0.24

0.11

0.08

0.06

 

R: 0.50, L: 0.29

R: 0.14, L: 0.11

R: 0.84, L: 0.78

R: 0.20, L: 0.19

R: 0.93, L: 1.06

R: 1.05, L: 1.06

PCC

Mean

0.30*

0.04

1.04*

0.25

1.03*

1.50

SD

0.21

0.03

0.26

0.11

0.13

0.11

 

R: 0.26, L: 0.34

R: 0.01, L: 0.07

R: 1.03, L: 1.05

R: 0.27, L: 0.22

R: 1.04, L: 1.02

R: 1.51, L: 1.49

PC

Mean

0.24*

0.02

0.99*

0.18

1.17*

1.49

SD

0.11

0.03

0.32

0.07

0.13

0.11

 

R: 0.20, L: 0.28

R: 0.03, L: 0.00

R: 1.00, L: 0.98

R: 0.21, L: 0.14

R: 1.14, L: 1.19

R: 1.51, L: 1.46

Hip

Mean

0.34

0.20

0.31*

0.17

0.81*

0.99

SD

0.11

0.13

0.16

0.13

0.16

0.06

 

R: 0.33, L: 0.34

R: 0.22, L: 0.17

R: 0.33, L: 0.29

R: 0.14, L: 0.18

R: 0.81, L: 0.80

R: 0.96, L: 1.02

PHC

Mean

0.32*

0.04

0.40*

0.16

0.76*

1.06

SD

0.12

0.06

0.16

0.08

0.11

0.07

 

R: 0.32, L: 0.33

R: 0.01, L: 0.07

R: 0.37, L: 0.43

R: 0.15, L: 0.16

R: 0.76, L: 0.76

R: 1.06, L: 1.07

SFC

Mean

0.24

0.11

0.89*

0.16

1.04*

1.18

SD

0.14

0.07

0.27

0.08

0.10

0.08

 

R: 0.20, L: 0.27

R: 0.13, L: 0.09

R: 0.88, L: 0.91

R: 0.16, L: 0.17

R: 1.05, L: 1.03

R: 1.19, L: 1.17

MFC

Mean

0.28*

0.07

0.78*

0.16

1.09*

1.31

SD

0.12

0.04

0.23

0.08

0.14

0.10

 

R: 0.30, L: 0.25

R: 0.06, L: 0.07

R: 0.76, L: 0.79

R: 0.16, L: 0.17

R: 1.10, L: 1.08

R: 1.31, L: 1.31

The data on both hemispheres were averaged in each region and expressed as mean ± SD

AD Alzheimer's disease, HC healthy control, ACC anterior cingulate cortex, PCC posterior cingulate cortex, Prec precuneus, SFC superior frontal cortex, MFC middle frontal cortex, Hip hippocampus, PHC parahippocampal cortex, R right, L left

*p < 0.05, corrected, vs healthy controls

https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1612-0/MediaObjects/259_2010_1612_Fig2_HTML.gif
Fig. 2

Quantitative PET parametric images of an AD patient. PET parametric images of [11C](R)PK11195 BP (a), [11C]PIB BP (b) and [18F]FDG SUVR (c) are superimposed on MRI. The colour bars denote levels of [11C](R)PK11195 BP, [11C]PIB BP and [18F]FDG SUVR

Clinico-biotracer correlation in the AD group

Based on the present SPM results and previous literature [813, 27], we focused on the levels of tracer accumulation in the ACC, PCC, Prec, SFC (BA9), MFC (BA9), PHC and Hip, all of which are particularly affected in the AD brain. Pearson correlation analysis showed that there were significant negative correlations between MMSE scores and [11C](R)PK11195 BPs in the left ACC (r = −0.90, p < 0.05, corrected), left Prec (r = −0.80, p < 0.05, corrected), left Hip (r = −0.79, p < 0.05, corrected) and left MFC (BA9) (r = −0.94, p < 0.05, corrected) in the AD group (Fig. 3, Supplementary Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1612-0/MediaObjects/259_2010_1612_Fig3_HTML.gif
Fig. 3

Correlations between MMSE scores and regional [11C](R)PK11195 BPs in the AD group. Scatter grams of correlations between MMSE scores and [11C](R)PK11195 BPs in the ACC (a), Prec (b), MFC (c) and Hip (d). Filled circles indicate the left side of each region and open circles the right. Straight lines show significant correlations (p < 0.05, corrected). ACC anterior cingulate cortex, Prec precuneus, MFC middle frontal cortex, Hip hippocampus

Other clinical parameters failed to correlate with the levels of [11C](R)PK11195 and [11C]PIB binding in any region (data not shown). The MMSE score tended to correlate with [18F]FDG SUVR in the hypometabolic regions (Supplementary Table 2)

Relationship between microglia activation and Aβ accumulation in the AD group

Direct comparisons showed a significant negative correlation between [11C](R)PK11195 BPs and [11C]PIB BPs in the PCC (left: r = −0.84, right: r = −0.90, p < 0.05, corrected) (Fig. 4, Supplementary Table 2). There was a tendency toward negative correlation between [11C](R)PK11195 BPs and [11C]PIB BPs in the left PHC, right SFC (BA9) and right MFC (BA9). No significant correlations between [11C](R)PK11195 BPs and FDG SUVR or between [11C]PIB BPs and FDG SUVR were found in any ROIs (Supplementary Table 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00259-010-1612-0/MediaObjects/259_2010_1612_Fig4_HTML.gif
Fig. 4

Correlations between [11C](R)PK11195 and [11C]PIB BPs in the AD group. Scatter grams of correlations between [11C](R)PK11195 BP and [11C]PIB BP in the PCC. Filled circles indicate the left side of each region and open circles the right. A straight line denotes a significant correlation (p < 0.05, corrected). PCC posterior cingulate cortex

Discussion

To our knowledge, this is the first study to examine the pathophysiological conspirators in the living AD brain using PET with three tracers ([11C](R)PK11195, [11C]PIB and [18F]FDG). We showed a significant increase in [11C](R)PK11195 BP in the ACC and PCC, where there was a more robust increase in [11C]PIB accumulation and marked decrease in [18F]FDG uptake. This elevation of [11C](R)PK11195 binding was consistent with the previous reports [8, 27]. In contrast to the relatively confined region of [11C](R)PK11195 accumulation, an elevation of [11C]PIB binding was found in broader brain regions such as the parietal, temporal and frontal cortices, which was also compatible with the previous results [912, 23, 24, 2729]. It is generally accepted that after the reduction in [18F]FDG uptake is seen in the medial parietal cortex at a very early disease stage, it then progresses to the lateral parietal and temporal cortices, and finally to the frontal cortex [3032]. Our patients had clinically mild-to-moderate AD (MMSE scores 21.5 ± 3.6), and the present [18F]FDG uptake pattern indicated that our patients were pathophysiologically in mid-stage disease. The findings of each tracer in the current study might be consistent with those already reported, but these pathophysiological events were measured for the first time without any delay, i.e. we completed a whole set of neuropsychological and imaging measurements within the same day for an individual patient. Thus, the present protocol offers the great advantage of depicting pathological events occurring simultaneously in vivo in AD.

Among pathological events such as Aβ deposits, microglial activation and glucose metabolism in the AD brain, only microglial activation in the left limbic region was found to correlate with MMSE scores in the present study, which was in line with the previous report (see the supplementary data and [27]). Although glucose metabolic reduction tended to correlate with MMSE scores, no tendency of correlation was found for Aβ accumulation. This suggests that not Aβ accumulation but microglia activation is critical in the damage of neurons involved in the orchestration of cognitive function in AD. Laterality of the correlation between microglial activation and MMSE scores was observed in the present study. Lateralized [11C](R)PK11195 binding was also reported in the previous results from Cagnin et al. [8]. In line with their report, we found a weak but significant [11C](R)PK11195 accumulation displayed in the left temporal cortex in the SPM analysis (see Fig. 1a). The reason for this was unclear but there might have been a selection bias; all patients were right-handed. We also considered that this laterality was related to a greater vulnerability of language function in such a group, as suggested in the previous report [8]. Indeed, the present patients performed poorly on the word generation test.

There is evidence that Aβ accumulation itself is not always detrimental to the brain environment. A major component of extracellular Aβ accumulation in the AD brain is insoluble Aβ fibril and [11C]PIB is reported to bind the Aβ accumulation homogeneously [11]. Because [11C]PIB cannot distinguish insoluble from soluble Aβ, we do not currently know how harmful soluble Aβ can be to the brain in vivo. It has been shown that soluble and insoluble Aβ can directly or indirectly bind to receptors expressed by microglia through TLR, CD40 and MHC-II, which leads to activation of microglia [24]. Recent studies showed that soluble Aβ oligomer is involved in neurodegeneration [33, 34] and induces neuronal apoptosis mainly through activation of the sphingomyelinase-ceramide pathway [35]. Because microglial activation can be found earlier, before insoluble Aβ accumulation becomes apparent [15, 16], and soluble APP fragment activates microglia into releasing cytokines [36, 37], there is now an assumption that a change of microglial activity is related to the degree of production of soluble Aβ oligomer as a neurodegenerative trigger. In the present study, we found a significant negative correlation between microglial activation and Aβ accumulation in the PCC, which is affected early in AD [30]. Thus, it might be possible that microglial activity increases in the region at an early stage of Aβ production and subsides during the phase when neurodegeneration is predominant with large insoluble Aβ fibril deposits. Since we are currently unable to detect the amount of soluble Aβ oligomer accumulated in vivo, microglial imaging may be informative in predicting the presence of toxic substances other than Aβ fibril accumulation in AD.

The contention that microglial activation is important as a therapeutic target in the clinical setting may be based on the previous epidemiological observation that the incidence of AD was lower among arthritis patients with daily use of NSAIDs [38]. The efficacy of NSAIDs for AD patients is still controversial, but many reports favour their usefulness by showing that NSAID treatment delays cognitive decline in AD patients and that the treatment suppresses activated microglia in the postmortem AD brain [39]. A series of animal experiments identified many promising drugs such as aminopyridazine which inhibits microglial activation and cytokine production from neuronal damage in rats with Aβ deposits [40] and memantine which suppresses activation of microglia by antagonizing the N-methyl-D-aspartate (NMDA) receptor [41]. In line with these observations, the present finding that microglial activation may play important roles in the early development of degeneration and cognitive decline would support early introduction of drugs with the capacity to suppress microglial activation in AD. With regard to detecting early events occurring in the living AD brain, our 1-day protocol PET study is useful to disclose the level of alterations and mutual relations among pathophysiological substrates in vivo. However, a larger sample size of very early AD patients or patients with mild cognitive impairment will be necessary to clarify the long-term alteration of these substrates. In addition, the relationship between early production of soluble Aβ and microglial activation needs to be determined in the living AD brain in future.

Acknowledgments

We would like to thank Dr. Mitsuo Kaneko (Kaneko Clinic), Dr. Masanobu Sakamoto and Messrs. Toshihiko Kanno and Yasuo Tanizaki (Hamamatsu Medical Center), Yutaka Naito (Japan Environment Research Corporation), Masami Futatsubashi, Akihito Oda and Ms. Tomomi Shinke (Hamamatsu Photonics KK) for their support. This work was supported by Research Grants from the Japanese Ministry of Health, Labor and Welfare, and Ministry of Economy, Trade and Industry, and NEDO and the Takeda Science Foundation.

Conflicts of interest

None.

Supplementary material

259_2010_1612_MOESM1_ESM.doc (62 kb)
Supplementary Table 1Results of SPM analysis (DOC 61 kb)
259_2010_1612_MOESM2_ESM.doc (66 kb)
Supplementary Table 2Pearson correlation analyses in the AD group (DOC 65 kb)

Copyright information

© Springer-Verlag 2010