Voxel-Based Acetylcholinesterase PET Study in Early and Late Onset Alzheimer’s Disease

Abstract

Background:

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by chronic progressive cognitive decline and displays underlying brain cholinergic dysfunction, providing a rationale for treatment with cholinomimetic medication. The clinical presentations and courses of AD patients may differ by age of onset.

Objective:

The objective of the present study was to illustrate the regional differences of brain acetylcholinesterase (AChE) activity as quantified by N-[¹¹C]methylpiperidinyl-4-acetate ([¹¹C]MP4A) and PET using parametric whole brain analysis and clarify those differences as a function of age.

Methods:

22 early onset AD (EOAD) with age at onset under 65, the remaining 26 as late onset AD (LOAD), and 16 healthy controls (HC) were enrolled. Voxel-based AChE activity estimation of [¹¹C]MP4A PET images was conducted by arterial input and unconstrained nonlinear least-squares method with subsequent parametrical analyses. Statistical threshold was set as Family Wise Error corrected, p-value <0.05 on cluster-level and cluster extent over 30 voxels.

Results:

Voxel-based group comparison showed that, compared to HC, both EOAD and LOAD showed cortical AChE decrement in parietal, temporal, and occipital cortices, with wider and stringent cortical involvement in the EOAD group, most prominently demonstrated in the temporal region. There was no significant correlation between age and regional cerebral AChE activity except for a small left superior temporal region in the AD group (Brodmann’s area 22, Z_max = 5.13, 396 voxels), whereas no significant cluster was found in the HC counterpart.

Conclusion:

Difference in cortical cholinergic dysfunction between EOAD and LOAD may shed some light on the cholinomimetic drug efficacy in AD.

Keywords

Acetylcholinesterase age Alzheimer’s disease parametric analysis positron emission tomography

INTRODUCTION

Alzheimer’s disease (AD) is the most common neurodegenerative disorder of dementia, characterized by progressive impairment of memory and cognitive functions associated with behavioral disturbances. Based on early biochemical and pathological investigations, alteration of acetylcholine synthesis was associated with cholinergic neuron loss in nucleus basalis of Meynert in the AD brain, which led to the theory of ‘cholinergic hypothesis’ as a pathophysiological basis [1, 2]. Postmortem neuropathological and neurochemical studies have consistently demonstrated selective loss of cholinergic neurons in AD brain, reduced cortical cholineacetyltransferase (ChAT: acetylcholine synthesizing enzyme) and acetylcholinesterase (AChE: acetylcholine degrading enzyme) [2 –5]. These findings have contributed to the development of central cholinesterase inhibitors that ameliorate symptoms of AD by increasing the synaptic acetylcholine level. Early onset AD (EOAD) is arbitrarily defined as first symptoms presenting before 65 years of age. When compared to late onset AD (LOAD) in which age of symptom onset is later than 65, clinical presentation of EOAD is prominent in deficits of non-memory cognitive domains such as visuospatial, executive, attentional, language, and praxis functions [6 –8]. Given the devastating nature of dementia beginning at an early age, further understanding of EOAD is crucial.

We developed an acetylcholine analogue PET radiotracer, N-[¹¹C]methylpiperidinyl-4-acetate ([¹¹C]MP4A), which enables us to estimate local brain AChE activity of human subjects in vivo. AChE exists mainly in presynaptic cholinergic neurons, implying that AChE is a biological marker for the brain cholinergic system. Previous region-of-interest (ROI) studies have shown that brain AChE activity is widely reduced in the neocortex of mild to moderate AD patients, and hippocampal AChE activity is already diminished in mild cognitively impaired subjects [9 –12]. In a subsequent study, 14 EOAD and 14 LOAD were directly compared by ROI method [13]. When compared to healthy controls, EOAD patients showed widespread AChE decline in the cerebral cortex as well as amygdala and hippocampus, while the LOAD group demonstrated abnormalities in temporal and parietal cortices and amygdala.

The objective of the present study was to further characterize cholinergic deficits in EOAD and LOAD patients by quantitative [¹¹C]MP4A PET with arterial input function and less subjective, parametric image analyses. Association of cholinergic deficits with clinical variables, including their age at the time of scans, was investigated by using both a parametric approach and a volume of interest (VOI) approach followed by multiple regression analyses.

MATERIALS AND METHODS

Subjects

Patients were diagnosed with probable AD according to the criteria of the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) [14]. Patients were recruited from the department of neurology, Chiba University Hospital, or Asahi Hospital for Neurological Diseases, Chiba, between 1996 and 2004. Forty-eight patients with AD and 16 healthy controls (HC) were enrolled. Previously reported subjects (30 AD and 14 HC) were included in the present cohort [9, 13]. In the AD group, there were 22 EOAD and 26 LOAD. None of the subjects had a history of taking cholinomimetic or anticholinergic medication. None of the subjects had cerebrovascular disease, past history of head trauma, or substance abuse. HC subjects had no history of neurological or psychiatric disorders, and they were not taking medications that would react with the central nervous system. Mini-Mental State Examination (MMSE) and the Japanese version of Alzheimer’s Disease Assessment Scale - Cognitive subscale (ADAS-Jcog) were conducted [15, 16]. In AD patients, Clinical Dementia Rating (CDR) and Functional Assessment Staging of Alzheimer’s disease (FAST) were evaluated by the caregivers [17, 18]. All participants, and in the case of AD subjects their responsible caregivers, had provided written informed consent for participating in the study. The study was approved by the Institutional Review Board of the National Institute of Radiological Sciences.

Positron emission tomography

After 10 min of transmission scan, emission data were acquired with a dynamic sequence of 16 PET scans over a 40-min period, using an EXACT47 scanner (Siemens/CTI, Knoxville, TN) with eyes covered and head immobilized. [¹¹C]MP4A (19.1±1.7 mCi in 5 mL) was intravenously injected over 60 s. Twenty-seven timed arterial blood samples were drawn from the radial artery. All emission scans were reconstructed by Hanning filter with a cutoff frequency of one-half of the maximum. After reconstruction, spatial resolution was 9×9×6 mm full-width at half-maximum (FWHM). Metabolite-corrected arterial plasma input function was obtained by fitting an exponential function to plasma [¹¹C]MP4A radioactivity as described before [13, 19].

Data analysis

Image preprocessing was performed using Statistical Parametric Mapping (SPM) software (SPM2, Wellcome Department of Cognitive Neurology, London, UK) implemented in MATLAB 7.1 (Mathworks, Sherborn, MA). Each frame was realigned and individual mean images were created. [¹¹C]MP4A template image was created using PET images and T1-weighted MRI in 11 HC according to the standard method [20]. Each realigned image was individually normalized to the [¹¹C]MP4A template image by using the mean [¹¹C]MP4A PET image as target. These images were finally smoothed with a 12-mm FWHM Gaussian kernel filter.

The k₃ value [min^–1] was estimated in a two-tissue three-parameter model by fitting brain theoretical curves to observed brain PET data and arterial input function, using an unconstrained nonlinear least-squares method and weighted time-activity curve, as described elsewhere [21]. In this model, the k₃ values are not influenced by local atrophy or blood flow alteration. AChE activity in the cerebellum and striatum is much higher than in the cerebral cortex, and estimates of k₃ values in the cerebellum and striatum are highly variable and not reliable [22, 23]. Custom software operating with the IDL image analysis software (Research Systems Inc., Boulder, CO) was employed to calculate the kinetic parameters in each voxel and, subsequently, to generate individual parametric images. For parametric whole brain analysis, we created a brain mask to exclude the cerebellum, white matter and striatum by using the Talairach Daemon template implemented into the WFU Pickatlas [24, 25]. VOI analysis was performed by WFU Pickatlas to extract the cerebral cortex (frontal, temporal, parietal, occipital), 42 Brodmann’s area, hippocampus, amygdala, and thalamus [25]. Right and left were averaged in all VOI analyses. Mean cortical AChE activity was calculated by multiplying voxel number and mean k₃ values in each Brodmann’s area, summing up all those values and dividing by whole cortex voxels.

Statistical analysis

All SPM analyses of k₃ parametric images were conducted with absolute values. Comparison between EOAD and LOAD was conducted by unpaired t-test option. Comparisons between EOAD, LOAD and HC group were performed with one-way analysis of variance (ANOVA). For whole brain k₃ correlation analysis, simple regression option was utilized in both AD groups, and independent variables such as age at scan, MMSE total score, CDR, FAST score, CDR, and ADAS-J cog scores were each incorporated into each model as a covariate. Correlation analysis in the HC group was also conducted with age at scan as a covariate. Clusters were accepted as significant if p-values corrected for Family-Wise Error rate (FWE) were less than 0.05 on a voxel-level with extent threshold above 30 voxels (p < 0.008 cluster corrected). MRIcroN version 7 and BSPMVIEW version 20161108 (https://zenodo.org/badge/latestdoi/21612/spunt/bspmview) were used for representing significant clusters in parametric analyses (http://www.nitrc.org/projects/mricron/). All reported coordinates were expressed in the Montreal Neurological Institute standard space.

Group comparisons in demographic variables and VOI analyses were performed by either unpaired t-test or one-way ANOVA followed by Bonferroni correction, using SPSS software (SPSS version 11, SPSS Inc., Chicago, IL). Correlation analyses between mean cortical k₃ values and age at scan, disease duration, MMSE total score, FAST, CDR, and ADAS J-cog scores were performed in each group by Spearman’s rank correlation test with subsequent multiple comparisons correction. To further identify which factor predicts cortical AChE activity, stepwise multiple regression analysis (P_in = 0.05, P_out = 0.10) was conducted. Durbin-Watson test were performed to check autocorrelation. Cortical mean AChE activity was considered as dependent variable, and independent (explanatory) variables were age at scan, MMSE total score, and disease duration. ADAS-Jcog score, CDR, and FAST scores were not included because of collinearity with MMSE score. Multicollinearity was not detected (variance inflation factor = 1.030). P-value <0.05 was considered statistically significant.

RESULTS

Participants

Age did not differ between HC and all 48 AD (p = 0.15, unpaired t-test) but was high in the LOAD group when compared with HC and the EOAD group [F(2,61) = 26.5, p < 0.001, one-way ANOVA] (Table 1). Gender did not differ among the 3 groups (p = 0.12, chi-square). Educational period was shorter in the LOAD group compared to both control and EOAD groups (F(2,48) = 59.9, p = 0.003). MMSE score was significantly lower in the AD groups [F(2,61) = 59.2, p < 0.001, one-way ANOVA], and post-hoc analysis showed reduction of MMSE scores in the EOAD and LOAD groups compared to HC (p < 0.001), but there was no difference between the EOAD and LOAD groups (p = 0.69). ADAS-Jcog, CDR score, and FAST score did not show any differences between EOAD and LOAD (p > 0.53, unpaired t-test). Disease duration was longer in the EOAD group (p < 0.001, unpaired t-test).

Table 1

Demographics of subjects [mean (SD)]

	Control	EOAD	LOAD
n	16	22	26
Male:Female	8:8	12:10	7:19
Family history	–	3/22 (13.6%)	3/26 (11.5%)
Age	63.3 (11.0)	59.5 (5.3)	74.0 (5.2)^a,b
Education (y)	14.7 (4.1)	12.8 (2.4)	10.3 (3.3)^c,d
Duration (y)	–	2.8 (1.6)^e	1.5 (0.8)
CDR	–	1.0 (0.4)	1.0 (0.5)
FAST	–	4.0 (0.6)	3.9 (0.7)
MMSE	29.3 (1.1)	16.9 (4.2)^f	18.3 (4.2)^g
ADAS-Jcog	–	23.4 (13.7)	21.0 (10.0)

EOAD, early onset Alzheimer’s disease; LOAD, late onset Alzheimer’s disease; CDR, Clinical Dementia Rating; FAST, Functional Assessment Staging of Alzheimer’s disease; MMSE, Mini-Mental State Examination; ADAS-Jcog, Japanese version of Alzheimer’s Disease Assessment Scale – Cognitive. Group comparisons were performed by one-way ANOVA with post hoc Bonferroni correction for 3 groups and unpaired t-test for 2 groups for continuous variables and Chi-square test for categorical variables. ^acontrol <LOAD (p < 0.001), ^bEOAD <LOAD (p < 0.001), ^ccontrol >LOAD (p < 0.01), ^dEOAD >LOAD (p < 0.05), ^eEOAD >LOAD (p < 0.001), ^fcontrol >EOAD (p < 0.001), ^gcontrol >LOAD (p < 0.001).

AChE activity in disease group compared with healthy controls

Parametric analysis

First, we compared the difference between all 48 AD patients and the 16 HC of [¹¹C]MP4A PET k₃ images by parametric analysis. AChE activity was reduced in a wide neocortical region (Fig. 1A, Table 2A), with the temporal cortex being the most affected region (P_FWE < 0.05, extent >30 voxels).

Fig.1

Acetylcholinesterase (AChE) reduction in Alzheimer’s disease (AD) and its subtype, measured by [¹¹C]MP4A PET. Statistical parametric T-score maps, with blue-green color representing AChE reduction, are superimposed on a single subject high-resolution T1 image by using MRIcroN, thresholded at voxel-level p < 0.05; Family-wise error (FWE) corrected with cluster above 30 voxels. Coordinates are shown in Table 2. A) Regional AChE reduction in total AD group compared with healthy control group. B) Regional AChE reduction in early onset AD group compared with healthy control group. C) Regional AChE reduction in late onset AD group compared with healthy control group. D) Early onset AD group showing lower AChE activity compared with late onset AD group. Note that the early onset AD group has widely reduced AChE activity in neocortical regions, compared to the relatively restricted reduced region in the late onset AD group.

Table 2

Coordinate table of statistical parametric mapping group analyses

Regions	Brodmann’s area	Coordinates^a
		x	y	z	p– value^b	Extent	Z_max
A. Control >Alzheimer’s disease
Lt Superior &Middle Temporal, Lt Insula	2–4, 6, 7, 9, 10, 13, 17–23, 29–31, 36–47	–52	–36	–14	<0.001	18511	7.24
Rt Inferior Parietal, Rt Mid Temporal, Rt Insula	1–4, 6–10, 13, 18–22, 29, 31, 36–45, 47	54	–32	22	<0.001	13788	6.9
Lt Medial Frontal, Bil Anterior cingulate gyrus, Lt Parahippocampus	11, 24, 25, 28, 32, 34, 35, 38, 47	–6	28	–16	<0.001	724	5.64
Rt Dorsolateral prefrontal cortex	9, 10	14	58	16	<0.001	164	5.2
Rt Precuneus	7	16	–62	38	<0.001	137	4.8
B. Control >Early onset Alzheimer’s disease
Bil Parieto-Temporal	1–11, 13, 17–24, 29–31, 36–47	–50	–22	–18	<0.001	41032	7.65
Lt Medial Frontal, anterior cingulate cortex	11, 24, 25, 32, 47	–10	20	–20	<0.001	419	5.56
Rt Precuneus	7	34	–66	58	<0.001	135	5.26
Lt Dorsolateral Prefrontal cortex	11	–16	52	–16	<0.001	91	5.25
Rt Superior Parietal	6	14	18	62	<0.001	85	5.09
Rt Medial Frontal	10	16	62	10	<0.001	70	5.08
Lt Parahippocamus	28, 34, 35	–16	–10	–24	<0.001	81	5.02
C. Control >Late onset Alzheimer’s disease
Lt Middle Temporal	2, 6, 13, 19–22, 29, 36–41, 43, 44	–50	–34	–14	<0.001	4334	6.23
Rt Middle Temporal	20, 21, 37	54	–34	–26	<0.002	377	5.44
Rt Insula, Inferior Parietal	2, 3, 13, 22, 29, 40, 42	44	–28	18	<0.003	838	5.4
Lt Precuneus	18, 19, 31	–26	–80	22	<0.004	136	5.16
Rt Medial Frontal	13, 45, 47	10	26	–14	0.001	98	5.1
Rt Inferior Occipital	19	32	–78	–4	0.003	49	5.08
Rt Inferior Parietal	7, 19, 39	–38	–68	42	0.002	63	4.99
Rt Precentral gyrus	8, 9	38	22	42	0.002	70	4.99
Lt Inferior Frontal, Insula	11, 25	–36	26	6	0.001	98	4.99
Lt Occipital lobe	17, 18	–14	–86	14	0.001	72	4.95
Rt Insula	13, 22, 44	44	6	2	<0.001	151	4.93
Rt Middle Temporal	21, 38	58	2	–16	0.002	59	4.87
Rt Inferior Frontal	44, 45	58	14	14	0.004	41	4.84
Lt Occipital lobe	17, 18	–6	–96	0	<0.001	118	4.74
Rt Temporal	37	50	–48	–10	0.002	58	4.69
D. Early onset Alzheimer’s disease <Late onset Alzheimer’s disease
Lt Middle Temporal	39	–38	–64	28	0.001	98	5.29
Rt Inferior Parietal	40	52	–42	42	0.003	58	5.15
Rt Middle Temporal	37	56	–62	2	0.005	40	4.89
Lt Superior Temporal	41	–52	–22	6	0.003	58	4.72

^aMontreal Neurological Institute coordinates; ^bFamily-wise error, cluster corrected p-value.

Second, [¹¹C]MP4A images of the 16 HC, 22 EOAD, and 26 LOAD were incorporated into a single model and compared by parametric ANOVA (P_FWE <0.05, extent >30 voxels). Compared with the HC group, reduced AChE activity in EOAD was noted in a widespread neocortical region (Fig. 1B, Table 2B), whereas that of LOAD was observed in relatively limited regions in the temporal, occipital, and frontal cortices (Fig. 1C, Table 2C). There were no increased AChE regions in the AD group. When compared between EOAD and LOAD, AChE activity of the EOAD group was more reduced in the left angular gyrus, right supramarginal gyrus, left superior temporal area, right middle temporal area, left dorsal prefrontal area, and right superior parietal regions (Fig. 1D, Table 2D). Compared with the EOAD group, no decreased AChE activity was detected in the LOAD group.

Third, voxel-based regression analysis of [¹¹C]MP4A PET k₃ image was carried out in all 48 AD patients. There was no significant correlation between age and regional cerebral AChE activity except for a small left superior temporal region (BA 13, 22, 40, 41, 42, 43) (Fig. 2A). No brain region was noted with negative correlation with function by age in the AD group. The HC group demonstrated no brain regional AChE activity when correlation analysis was conducted with age at scan as an independent variable. We also investigated the regional k₃ value difference between male and female subjects, but there were no significant clusters in either HC or AD groups, even with a less stringent threshold (voxel-level p < 0.001 uncorrected, cluster-level p < 0.05). Brain regions where AChE activity was positively correlated with MMSE total score were exhibited in the bilateral supplementary motor area, bilateral middle temporal area, right superior area, and right middle occipital cortex (Fig. 2B, Supplementary Table 1). No negative correlation with MMSE total score was detected. There were also no significant clusters when FAST score and CDR score were each analyzed as independent variables in the AD group.

Fig.2

Localization of correlation by [¹¹C]MP4A k₃ value (acetylcholinesterase activity) in Alzheimer’s disease group (voxel-level p < 0.05; Family-wise error (FWE) corrected, extent threshold >30 voxels). A) Positive correlation with age at scan as a function. No significant correlation between age and regional cerebral AChE activity were found except for a small left superior temporal region (Brodmann’s area 13, 22, 40, 41, 42, 43) using Montreal Neurological Institute coordinates [x, y, z] [–52, –24, 8], Zmax = 5.13, extent = 396 voxels. B) Positive correlation with Mini-Mental State Examination total score as a function. Surface projections of statistical parametric map were rendered on a template image created by bspmview.

Volume of interest analysis

In LOAD, the AChE activities of the neocortical areas were lower than those of HC (p < 0.001, post-hoc Bonferrroni correction) (Table 3). EOAD showed consistent AChE reduction in each cortical region compared to the HC group (F(2,61) >18.7, p < 0.001, one-way ANOVA, p < 0.001, post-hoc Bonferrroni correction) and even lower than the LOAD group (p < 0.03, post-hoc Bonferrroni correction). In the hippocampus, AChE activity was lower in the EOAD group, and the LOAD group showed a trend of decrement compared with the HC group [F(2,61) = 7.65, p = 0.001, one-way ANOVA, HC versus EOAD p = 0.001, HC versus LOAD p = 0.055, post-hoc Bonferrroni correction)]. In the amygdala, AChE activity was lower in both EOAD and LOAD compared to the HC group [F(2,61) = 8.19, p = 0.001, one-way ANOVA, HC versus EOAD p < 0.001, HC versus LOAD p < 0.02, post-hoc Bonferrroni correction]. No overall difference in AChE activity among the groups was found in the thalamus [F(2,61) = 0.49, p = 0.62, one-way ANOVA].

Table 3

Brain acetylcholinesterase activity (k₃ value) by volume-of-interest analysis in early onset and late onset Alzheimer’s disease and healthy controls

	Healthy control k₃ value (min^–1) mean (SD)	Early onset AD k₃ value (min^–1) mean (SD), % reduction	Late onset AD k₃ value (min^–1) mean (SD), % reduction	Statistics F-value, p-value
Mean cortex	0.080 (0.004)	0.064 (0.007), –19.0% ^a,f	0.070 (0.007), –11.5% ^d	F_(2,61) = 27.0, p < 0.001
Frontal cortex	0.082 (0.004)	0.068 (0.007), –16.5% ^a,g	0.074 (0.008), –10.0% ^e	F_(2,61) = 18.7, p < 0.001
Occipital cortex	0.066 (0.004)	0.053 (0.006), –19.3% ^a,f	0.057 (0.006), –12.4% ^e	F_(2,61) = 24.2, p < 0.001
Parietal cortex	0.073 (0.003)	0.058 (0.007), –19.9% ^a,f	0.065 (0.005), –10.5% ^c	F_(2,61) = 30.3, p < 0.001
Temporal cortex	0.077 (0.005)	0.057 (0.007), –25.6% ^a,f	0.065 (0.007), –15.0% ^c	F_(2,61) = 42.1, p < 0.001
Thalamus	0.182 (0.038)	0.174 (0.037), –4.1%	0.171 (0.031), –6.0%	F_(2,61) = 0.49, p > 0.05
Hippocampus	0.113 (0.013)	0.095 (0.014), –15.8% ^a	0.102 (0.015), –9.4%	F_(2,61) = 7.65, p = 0.001
Amygdala	0.165 (0.037)	0.122 (0.032), –26.5% ^b,h	0.135 (0.032), –18.1%	F_(2,61) = 8.19, p = 0.001

AD, Alzheimer’s disease. Brain regions are right and left averaged. % reduction is calculated by (mean AD value – mean control value)/mean control value x 100. One-way analysis of variance was performed in each brain region with post-hoc Bonferroni correction. Group comparisons were performed by one-way ANOVA with post hoc Bonferroni correction. Early onset Alzheimer’s disease <healthy control: ^ap < 0.001, ^bp = 0.001. Late onset Alzheimer’s disease <healthy control: ^cp < 0.001, ^dp = 0.005, ^ep < 0.05. Early onset Alzheimer’s disease <late onset Alzheimer’s disease: ^fp < 0.001, ^gp = 0.001, ^hp < 0.05.

Correlation analysis was conducted in the whole AD group between mean cortical AChE value and age at scan, disease duration, MMSE total score, CDR, FAST, and ADAS J-cog. The mean cortical k₃ value correlated with age at scan (rho = 0.417, p < 0 02), MMSE (rho = 0.507, p < 0.01), and FAST (rho = –0.417, p < 0.03), but not with duration (rho = –0.338, p = 0.114), CDR (rho = –0.368, p = 0.06), or ADAS-J cog (rho = –0.315, p = 0.252). In line with the parametric result, VOI analysis of mean cortical AChE activity in the HC group did not correlate with age (rho = 0.107, p = 0.70). Regression model analysis showed that age at scan (standardized coefficient = 0.46, t = 3.84, p < 0.001) and MMSE score (standardized coefficient = 0.332, t = 2.78, p < 0.01) were predictors of mean cortical AChE activity, but disease duration was not significant (F (2,45) = 13.4, R = 0.611, p < 0.001, Durbin-Watson test 1.88). Voxel-wise whole brain search by multiple regression analysis was performed to examine the association between regional k₃ value as a dependent value and age, MMSE total score and disease duration as independent variables. In line with the VOI result, age showed robust positive regression with widespread cortical regions (cluster size: 99155 voxels) and 24 cortical clusters (cluster size: 30 to 1360 voxels) showed positive regression with MMSE total score, but the remaining regression did not display any significant clusters (p < 0.05, cluster corrected, data not shown).

DISCUSSION

Brain AChE activity measured by PET demonstrated widespread decline in the neocortical areas and limbic areas (hippocampus and amygdala) in AD subjects by parametric whole brain analysis. Brain AChE activity in AD patients was further diminished in the EOAD patients compared to their late-onset counterparts, while cortical AChE activity was independent from normal aging in the HC group as we previously reported [23]. The pattern of AChE activity as a function of age showed positive correlation (i.e., lower AChE activity when younger) in the left superior temporal cortex. Multivariate analysis showed that the variables associated with cortical AChE reduction were both age and global cognitive function, measured by MMSE, but not disease duration. In the present study, to our knowledge, the number of AD subjects participating in an AChE PET study was the largest to date.

ChAT activity correlates with the brain distribution of AChE in human [26]. Therefore, both ChAT and AChE can be utilized as an index of presynaptic cholinergic neurons, although no PET ligand is available for ChAT, and its assessment is limited to postmortem brain tissues. Cognitive dysfunction correlates with brain ChAT activity in AD patients [27]. Younger AD (age at death <79) showed widespread and severe presynaptic cholinergic deficit, as measured by ChAT activity, compared with older AD (age at death >79) [28]. Cortical ChAT activity was lower in younger AD, and both cortical AChE and ChAT activities positively correlated with age at death in AD [29 –31]. Along with these observations, neuronal loss in nucleus basalis of Meynert was more pronounced both in the younger age at onset and younger age at death in AD, and was not correlated with disease duration or brain volume, or the distribution of cortical atrophy [32, 33]. Brain AChE activity in autopsied healthy controls has shown no alteration by age [34]. The present neuroimaging data are in accord with these previous neuropathological observations in AD and aging. Cognitive reserve, represented by education and occupation, may influence AChE activity in early AD [35]. Garibotto and colleagues have reported that higher AChE activity was associated with higher educational status in bilateral hippocampus, and with higher occupational status in posterior cingulate cortex. EOAD patients were more educated than LOAD patients in this study. When higher AChE activity associated with cognitive reserve was taken into account, hippocampal AChE activity might be impaired further in EOAD patients than LOAD patients. Low educational status was found in the LOAD group, probably related to socio-economical background.

The basal forebrain cholinergic system does not have strict anatomical boundaries, and its projections overlap [36]. Medial septal nucleus (Ch1) and vertical limb of diagonal band nucleus (Ch2) project to hippocampal complex, and nucleus basalis of Meynert (Ch4) projects to cortex and amygdala [36, 37]. Ch4 is further divided into anterior subregion projecting to limbic and medial cortical regions, frontoparietal opercular regions and amygdala, and posterior subregion projecting to superior temporal and adjacent regions [36, 37]. Lower AChE activity of superior temporal region demonstrated in younger AD may indicate that younger AD subjects are susceptible in posterior subregion of nucleus basalis of Meynert. Pathways of cholinergic neurons originating from nucleus basalis of Meynert are largely classified into the medial pathway, which innervates occipital lobe, and the lateral pathway, which supplies frontal, parietal and temporal lobes [38]. By deduction from the present result, both medial and lateral cholinergic pathways are affected in AD, and the perisylvian branch of the lateral pathway may be more involved in the younger onset AD.

[¹²³I]-iodobenzovesamicol SPECT, a marker for vesicular acetylcholine transporter that also reflects the function of presynaptic cholinergic neurons, was assessed in AD patients, and it was found that binding was severely decreased in the whole cortex and hippocampus in EOAD, whereas the LOAD group showed milder decrement restricted to the temporal lobe and hippocampus [39]. Measurement of AChE activity in cerebrospinal fluid of AD patients demonstrated significant reduction in EOAD, but no difference was found in the LOAD patients when compared with HC [40]. These observations are also in line with our findings. Younger patients with AD represent relatively pure cases of AD. However, older AD patients are more likely to have less salient AD pathology in combination with overlapping of other forms of age-related disease, such as vascular impairment, microbleeds, and coexisting Lewy body pathology [41 –43]. Of note, in this study, we excluded patients with apparent cerebrovascular lesions and motor impairment.

Brain AChE activity in frontotemporal dementia (FTD) is comparable to healthy subjects, while patients with dementia with Lewy bodies (DLB) show profound reductions [19, 44]. Patients with DLB are generally old, and those with FTD are relatively young. Previously, we showed that [¹¹C]MP4A PET is useful in differentiating between patients with LOAD and DLB [44]. Collectively, AChE activity is lowest in DLB, intact in FTD, and intermediate in AD.

From the current findings, EOAD may benefit more from AChE inhibitor than LOAD. Indeed, in a retrospective observation, donepezil and rivastigmine showed significant improvement in younger AD while older AD did not change by the treatment [45]. It is of interest that the region where positive correlation between AChE activity and age was found in the superior temporal lobe (BA13, 22, 40, 41, 42, 43), which is where Wernicke’s receptive language area is located. In sum, this observation may imply that cholinergic enhancement may ameliorate language function in AD patients, especially in young onset subtype. Donepezil treatment with 23 mg for 24 weeks against moderate to severe AD was effective in the language domain, although its effective difference according to age was not stated [46].

Some limitations of the present study should be noted. First, apolipoprotein E (ApoE) genotype, diagnostic biomarkers (cerebrospinal fluid, amyloid PET), and autopsy findings are lacking. Second, the HC group was relatively young compared to the LOAD group, although there was no effect of age on AChE activity in HC. Thirdly, the cross-sectional design of the study precludes us from drawing conclusions in regard to the disease course of early and late stage AD. Fourthly, the EOAD and LOAD groups were matched by MMSE total score. One could argue that equally impaired cognitive function, evaluated by MMSE scores, may in fact have different cognitive dysfunction when normal cognition of aging was taken into account. However, the MMSE total score declines only very slightly by normal aging [47]. Although disease duration was longer in the EOAD group compared to the LOAD group, it did not correlate with cortical AChE loss in the present study.

In conclusion, the cholinergic deficit observed in LOAD is more prominent in EOAD despite comparable levels of global cognitive impairment between the two groups. Age is positively correlated with neocortical AChE activity in AD patients, and therefore age should be taken into account when interpreting the clinical trials of cholinomimetic agents against AD. Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/17-0749r2)

Footnotes

ACKNOWLEDGMENTS

The authors wish to thank Mr. Shuichi Sendoda, Medical Information Systems Division, Toshiba Information Systems Corporation, Japan, for technical support with the software programming. This work was partly supported by grants from Grants-in-Aid for the Research and Development Grants for Dementia (16768966) to M.H. from the Japan Agency for Medical Research and Development.

The supplementary material is available in the electronic version of this article: .

References

Bartus

, Dean

3rd , Beer

, Lippa

(1982) The cholinergic hypothesis of geriatric memory dysfunction. Science 217, 408–414.

Whitehouse

, Price

, Struble

, Clark

, Coyle

, Delon

(1982) Alzheimer’s disease and senile dementia: Loss of neurons in the basal forebrain. Science 215, 1237–1239.

Davies

, Maloney

(1976) Selective loss of central cholinergic neurons in Alzheimer’s disease. Lancet 2, 1403.

McGeer

, McGeer

, Suzuki

, Dolman

, Nagai

(1984) Aging, Alzheimer’s disease, and the cholinergic system of the basal forebrain. Neurology 34, 741–745.

Rossor

, Garrett

, Johnson

, Mountjoy

, Roth

, Iversen

(1982) A post-mortem study of the cholinergic and GABA systems in senile dementia. Brain 105, 313–330.

Warren

, Fletcher

, Golden

(2012) The paradox of syndromic diversity in Alzheimer disease. Nat Rev Neurol 8, 451–464.

Koedam

, Lauffer

, van der Vlies

, van der Flier

, Scheltens

, Pijnenburg

(2010) Early-versus late-onset Alzheimer’s disease: More than age alone. J Alzheimers Dis 19, 1401–1408.

Smits

, Pijnenburg

, Koedam

, van der Vlies

, Reuling

, Koene

, Teunissen

, Scheltens

, van der Flier

(2012) Early onset Alzheimer’s disease is associated with a distinct neuropsychological profile. J Alzheimers Dis 30, 101–108.

Iyo

, Namba

, Fukushi

, Shinotoh

, Nagatsuka

, Suhara

, Sudo

, Suzuki

, Irie

(1997) Measurement of acetylcholinesterase by positron emission tomography in the brains of healthy controls and patients with Alzheimer’s disease. Lancet 349, 1805–1809.

10.

Rinne

, Kaasinen

, Järvenpää

, Någren

, Roivainen

, Yu

, Oikonen

, Kurki

(2003) Brain acetylcholinesterase activity in mild cognitive impairment and early Alzheimer’s disease. J Neurol Neurosurg Psychiatry 74, 113–115.

11.

Herholz

, Weisenbach

, Zündorf

, Lenz

, Schröder

, Bauer

, Kalbe

, Heiss

(2004) In vivo study of acetylcholine esterase in basal forebrain, amygdala, and cortex in mild to moderate Alzheimer disease. Neuroimage 21, 136–143.

12.

Marcone

, Garibotto

, Moresco

, Florea

, Panzacchi

, Carpinelli

, Virta

, Tettamanti

, Borroni

, Padovani

, Bertoldo

, Herholz

, Rinne

, Cappa

, Perani

(2012) [11C]-MP4A PET cholinergic measurements in amnestic mild cognitive impairment, probable Alzheimer’s disease, and dementia with Lewy bodies: A Bayesian method and voxel-based analysis. J Alzheimers Dis 31, 387–399.

13.

Shinotoh

, Namba

, Fukushi

, Nagatsuka

, Tanaka

, Aotsuka

, Ota

, Tanada

, Irie

(2000) Progressive loss of cortical acetylcholinesterase activity in association with cognitive decline in Alzheimer’s disease: A positron emission tomography study. Ann Neurol 48, 194–200.

14.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.

15.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

16.

Rosen

, Mohs

, Davis

(1984) A new rating scale for Alzheimer’s disease. Am J Psychiatry 141, 1356–1364.

17.

Hughes

, Berg

, Danziger

, Coben

, Martin

(1982) A new clinical scale for the staging of dementia. Br J Psychiatry 140, 566–572.

18.

Sclan

, Reisberg

(1992) Functional assessment staging (FAST) in Alzheimer’s disease: Reliability, validity, and ordinality. Int Psychogeriatr 4(Suppl 1), 55–69.

19.

Hirano

, Shinotoh

, Shimada

, Aotsuka

, Tanaka

, Ota

, Sato

, Ito

, Kuwabara

, Fukushi

, Irie

, Suhara

(2010) Cholinergic imaging in corticobasal syndrome, progressive supranuclear palsy and frontotemporal dementia. Brain 133, 2058–2068.

20.

Meyer

, Gunn

, Myers

, Grasby

(1999) Assessment of spatial normalization of PET ligand images using ligand-specific templates. Neuroimage 9, 545–553.

21.

Sato

, Fukushi

, Shinotoh

, Shimada

, Tanaka

, Hirano

, Irie

(2012) A short-scan method for k(3) estimation with moderately reversible PET ligands: Application of irreversible model to early-phase PET data. Neuroimage 59, 3149–3158.

22.

Atack

, Perry

, Bonham

, Candy

, Perry

(1986) Molecular forms of acetylcholinesterase and butyrylcholinesterase in the aged human central nervous system. J Neurochem 47, 263–277.

23.

Namba

, Iyo

, Fukushi

, Shinotoh

, Nagatsuka

, Suhara

, Sudo

, Suzuki

, Irie

(1999) Human cerebral acetylcholinesterase activity measured with positron emission tomography: Procedure, normal values and effect of age. Eur J Nucl Med 26, 135–143.

24.

Lancaster

, Woldorff

, Parsons

, Liotti

, Freitas

, Rainey

, Kochunov

, Nickerson

, Mikiten

, Fox

(2000) Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp 10, 120–131.

25.

Maldjian

, Laurienti

, Kraft

, Burdette

(2003) An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets. Neuroimage 19, 1233–1239.

26.

Mesulam

, Geula

(1992) Overlap between acetylcholinesterase-rich and choline acetyltransferase-positive (cholinergic) axons in human cerebral cortex. Brain Res 577, 112–120.

27.

Perry

, Tomlinson

, Blessed

, Bergmann

, Gibson

, Perry

(1978) Correlation of cholinergic abnormalities with senile plaques and mental test scores in senile dementia. Br Med J 2, 1457–1459.

28.

Rossor

, Iversen

, Reynolds

, Mountjoy

, Roth

(1984) Neurochemical characteristics of early and late onset types of Alzheimer’s disease. Br Med J (Clin Res Ed) 288, 961–964.

29.

Bowen

, White

, Spillane

, Goodhardt

, Curzon

, Iwangoff

, Meier-Ruge

, Davison

(1979) Accelerated ageing or selective neuronal loss as an important cause of dementia? Lancet 1, 11–14.

30.

Rossor

, Iversen

, Johnson

, Mountjoy

, Roth

(1981) Cholinergic deficit in frontal cerebral cortex in Alzheimer’s disease is age dependent. Lancet 2, 1422.

31.

Hansen

, DeTeresa

, Davies

, Terry

(1988) Neocortical morphometry, lesion counts, and choline acetyltransferase levels in the age spectrum of Alzheimer’s disease. Neurology 38, 48–54.

32.

Tagliavini

, Pilleri

(1983) Basal nucleus of Meynert. A neuropathological study in Alzheimer’s disease, simple senile dementia, Pick’s disease and Huntington’s chorea. J Neurol Sci 62, 243–260.

33.

Mann

, Yates

, Marcyniuk

(1984) A comparison of changes in the nucleus basalis and locus caeruleus in Alzheimer’s disease. J Neurol Neurosurg Psychiatry 47, 201–203.

34.

Perry

, Perry

, Blessed

, Tomlinson

(1978) Changes in brain cholinesterases in senile dementia of Alzheimer type. Neuropathol Appl Neurobiol 4, 273–277.

35.

Garibotto

, Tettamanti

, Marcone

, Florea

, Panzacchi

, Moresco

, Virta

, Rinne

, Cappa

, Perani

(2013) Cholinergic activity correlates with reserve proxies in Alzheimer’s disease. Neurobiol Aging 34, 2694e13–2694.e18.

36.

Mesulam

, Mufson

, Levey

, Wainer

(1983) Cholinergic innervation of cortex by the basal forebrain: Cytochemistry and cortical connections of the setal area, diagonal band nuclei, nucleus basalis (substantia innominata), and hypothalamus in the rhesus monkey. J Comp Neurol 214, 170–197.

37.

Liu

, Chang

, Pearce

, Gentleman

(2015) Nucleus basalis of Meynert revisited: Anatomy, history and differential involvement in Alzheimer’s and Parkinson’s disease. Acta Neuropathol 129, 527–540.

38.

Selden

, Gitelman

, Salamon-Murayama

, Parrish

, Mesulam

(1998) Trajectories of cholinergic pathways within the cerebral hemispheres of the human brain. Brain 121, 2249–2257.

39.

Kuhl

, Minoshima

, Fessler

, Frey

, Foster

, Ficaro

, Wieland

, Koeppe

(1996) In vivo mapping of cholinergic terminals in normal aging, Alzheimer’s disease, and Parkinson’s disease. Ann Neurol 40, 399–410.

40.

Kumar

, Giacobini

, Markwell

(1989) CSF choline and acetylcholinesterase in early-onset vs. late-onset Alzheimer’s disease patients. Acta Neurol Scand 80, 461–466.

41.

Sonnen

, Larson

, Crane

, Haneuse

, Li

, Schellenberg

, Craft

, Leverenz

, Montine

(2007) Pathological correlates of dementia in a longitudinal, population-based sample of aging. Ann Neurol 62, 406–413.

42.

Schneider

, Arvanitakis

, Bang

, Bennett

(2007) Mixed brain pathologies account for most dementia cases in community-dwelling older persons. Neurology 69, 2197–2204.

43.

Middleton

, Grinberg

, Miller

, Kawas

, Yaffe

(2011) Neuropathologic features associated with Alzheimer disease diagnosis: Age matters. Neurology 77, 1737–1744.

44.

Shimada

, Hirano

, Sinotoh

, Ota

, Tanaka

, Sato

, Yamada

, Fukushi

, Irie

, Zhang

, Higuchi

, Kuwabara

, Suhara

(2015) Dementia with Lewy bodies can be well-differentiated from Alzheimer’s disease by measurement of brain acetylcholinesterase activity-a [11C]MP4A PET study. Int J Geriatr Psychiatry 30, 1105–1113.

45.

Bullock

, Bergman

, Touchon

, Gambina

, He

, Nagel

, Lane

(2006) Effect of age on response to rivastigmine or donepezil in patients with Alzheimer’s disease. Curr Med Res Opin 22, 483–494.

46.

Ferris

, Cummings

, Christensen

, Doody

, Farlow

, Sabbagh

, Liu

, Mackell

, Fain

(2013) Effects of donepezil 23mg on Severe Impairment Battery domains in patients with moderate to severe Alzheimer’s disease: Evaluating the impact of baseline severity. Alzheimers Res Ther 5, 12.

47.

Jacqmin-Gadda

, Fabrigoule

, Commenges

, Dartigues

(1997) A 5-year longitudinal study of the Mini-Mental State Examination in normal aging. Am J Epidemiol 145, 498–506.