Key words
Alzheimer’s disease - acetylcholinesterase inhibitors - drug response -
CHAT
Introduction
Dementia is a progressive neurodegenerative disease with a rising prevalence and societal
burden [1]
[2]. Alzheimer’s disease (AD) is the most common form in dementia and it accounts for
60–70% of all cases [3].
AD is associated with widespread degeneration of cholinergic neurons, and acetylcholinesterase
inhibitor (AChEI) drugs are approved for symptomatic treatment, with the aim of restoring
the cholinergic deficit [4]. However, therapeutic response rates vary from 40–70% [5]. If the response to the drug initially selected is insufficient, a change of drugs
can be considered. However, the recognition of non-response requires prolonged observation.
Thus, an ability to predict response early in the course of AD is an important therapeutic
objective.
One promising approach is pharmacogenomics [6]. Several preliminary pharmacogenomic studies [7]
[8] have reported that the clinical response to donepezil is highest in carriers of
the APOE epsilon4 allele, although a recent large study obtained a negative result
[9].
The pathology of the brain cholinergic system is prominent in AD and AChEI drugs are
widely used. Thus, the cholinergic system is a logical target for pharmacogenomic
studies. There have been several studies on possible associations between genetic
polymorphisms of cholinergic-related genes and the therapeutic effect of AChEIs [10]
[11]. However, there are limitations as follows in those previous studies. First, the
results are inconsistent. Second, the selection of SNPs was limited, so the entire
candidate gene region was not covered. Moreover, ethnic heterogeneity was not explored
in these studies. In the present study we have examined the polymorphic variations
of the genes encoding 3 enzymes involved in the synthesis, transport, and metabolism
of acetylcholine in the cholinergic system ([Fig. 1]) [11]
[12]. Choline acetyltransferase (ChAT) encoded by the gene CHAT synthesizes acetylcholine, using choline and acetyl-CoA as substrates [11]. The choline transporter (ChT) catalyzes the uptake of choline from the extracellular
space to the neuronal cytoplasm, and is encoded by SLC5A7 [12]. Acetylcholine esterase (AChE) is encoded by ACHE and acts to hydrolyze acetylcholine, thereby inactivating the neurotransmitter [11]. This study extends previous reports by the simultaneous coverage of 3 genes important
for function of the brain cholinergic system. We also aimed to assess in our Asian
(Korean) population the replicability of previous reports in Caucasians [10]
[11]. The hypothesis of this exploratory study is that SNPs of genes involved in the
synthesis and movement of acetylcholine may affect the response of AChEIs in AD.
Fig. 1 The function of ChAT, AChE and ChT involved in synthesis and movement of acetylcholine
in the cholinergic system. (from the KEGG database, http://www.genome.jp/kegg/). (Color
figure available online only).
Patients and Methods
Subjects
Subjects were 158 patients diagnosed with AD from the Clinical Trial Program in the
Geropsychiatry Clinic at the Samsung Medical Center. All were of unrelated Korean
ancestry. Patients were registered between November 2001 and January 2012. Subjects
were eligible for this clinical trial only if they satisfied all following criteria:
All patients were diagnosed as AD or probable AD according to the standards of the
NINCDS-ADRDA (National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s
Disease and Related Disorders Association) [13]; They had a score of 26 or less in the Korean version of the Mini-Mental State Examination
(K-MMSE) [14]; They had a history of cognitive decline which was gradual in onset and progressive
for more than 6 months; they had a reliable caregiver who helped them to take their
medication, participate in the assessment, and provide ongoing information about them
[13]. Patients were excluded if any of the following conditions was present: other neurodegenerative
diseases except AD (i. e., Parkinson’s disease or Huntington’s disease), psychiatric
disorder or severe behavioral disturbances requiring psychotropic medications, cerebral
injuries induced by trauma, hypoxia, and/or ischemia, clinically active cerebrovascular
disease, medical history of seizure disorder, and other physical conditions requiring
acute treatments.
All subjects underwent brain magnetic resonance imaging (MRI), neurological evaluation,
and routine laboratory tests prior to this clinical trial in order to screen for other
possible causes of dementia. The Institutional Review Board (IRB) at Samsung Medical
Center approved the protocol. Written informed consent was obtained from both caregiver
and patient. The study is registered (NCT01198093) in ClinicalTrials.gov.
Procedures
Subjects were assigned to receive monotherapy for 26 weeks with an acetylcholinesterase
inhibitor (donepezil, galantamine or rivastigmine) as determined by a clinician. In
this semi-naturalistic clinical trial, the choice of drug was based on the anticipated
side effects in at-risk individuals and on current clinical practice guidelines. Donepezil
was administered to 84 patients, galantamine to 52 patients and rivastigmine to 22
patients.
Doses were titrated into the usual range based on tolerability and side effects. All
subjects were assessed in clinic visits after 1 week and 4 weeks on drug to adjust
the dosage and evaluate adverse events. Psychotropic medications except acetylcholinesterase
inhibitors were not allowed with one exception. Benzodiazepines could be used only
as a short-term adjunctive for insomnia. If the subjects did not show any significant
changes or serious adverse events, the interval for clinic visits was increased to
13 weeks. Experienced geriatric psychiatrists performed the assessment at each visit
for clinical review of cognitive status, to examine physical and neurological status,
and to review adverse events. Vital sign checks, physical examinations, laboratory
tests including complete blood counts, blood chemistry profiles, vitamin B12/folate
levels, syphilis serology, thyroid function tests, and ECG at baseline were carried
out in all subjects.
Selection of SNP markers and genotyping
These SNPs were genotyped using the MassARRAY system (Sequenom, Inc., San Diego, Calif).
25 SNPs were discovered and selected as candidate genes with the computer program
Tagger [15] with criteria of r
2>0.65 and minor allele frequency>0.05 in combined Asian population (JPT/HCB). 21 for
CHAT, 3 for SLC5A7 and one for ACHE were genotyped. The total missing genotype counts were 50 (total call rate: 98.7%),
these genotyping data were not included in the SNP association analyses. All investigators
and raters were blinded to the results of genotyping throughout the study. The laboratory
worker who performed the genotyping was blind to clinical data of the subjects. The
organization and selected SNP locations of CHAT gene are shown in [Fig. 2]. There were no significant differences in genotype distribution of the 25 SNPs according
to drug choice.
Fig. 2 CHAT organization and single-nucleotide polymorphism (SNP) locations (from National Center
for Biotechnology Information Gene Database, http://www.ncbi.nlm.nih.gov/gene/). The
horizontal line represents the genomic sequence and vertical bars represent exons.
Plus signs and minus signs denote SNPs with significant association and SNPs with
negative results, respectively.
Measures
The response rate was assessed and compared at 26 weeks of treatment. Response was
defined as no change (i. e., no deterioration) or improvement on the score of the
Korean version of the Mini-Mental State Examination (K-MMSE) [16]
[17]. Global severity of disease was assessed according to the Clinical Dementia Rating
(CDR) [18]. These research assessments of cognitive outcome were performed by a single, trained
rater.
Data analysis
Continuous variables were presented as mean±standard deviation (SD) or as median and
interquartile range. Categorical variables were summarized as frequencies and proportions.
Wilcoxon rank-sum test or Student’s t test was performed to compare continuous variables
between 2 groups according to the normality of the distribution. The association of
categorical variables was determined based on the chi-squared test in all subjects.
We assessed the associations between each SNP and responsiveness by using the exact
Cochran-Armitage test for trend (a genotypic trend model) [19]. Chi-squared testing was used to examine deviation from Hardy-Weinberg equilibrium
[20]. The four-gamete rule by Haploview was used to check linkage disequilibrium (LD)
structure [21]. Phasing haplotypes were conducted using PHASE 2.1.1 for each of the haplotype blocks
individually [22]. The exact Cochran-Armitage test for a trend was used to examine the associations
between a haplotype allele and response. For the significance of association of a
SNP or haplotype allele, the false discovery rate (FDR) control was used to correct
each P-value [23].
The associated SNPs and haplotype alleles were entered into a multiple logistic regression
model to evaluate the impact of each genetic variable on response, adjusting for other
variables. In this model, the genetic variable represented the minor allele count
for a subject (0, 1 or 2) and the dependent variable represented the treatment outcome
(1=response and 0=non-response). Results were considered as significant with a threshold
of P<0.05. All statistical tests were performed using SAS 9.1 (SAS Institute, Inc.,
Cary, North Carolina).
Results
Subject characteristics
Clinical and demographic characteristics are shown in [Table 1]. Mean age of the subjects was 72.66 (SD=8.31) years and most were in the early stage
of Alzheimer’s disease. The rate of response to acetylcholinesterase inhibitors was
102 of 158 (64.6%). There was no significant difference between responders and non-responders
with respect to gender, age, education level and baseline global severity (CDR). The
rate of response was not affected by choice of drug (donepezil, galantamine and rivastigmine).
However, there was a marginally significant difference between responders and non-responders
in baseline K-MMSE score (P=0.04).
Table 1 Clinical and demographic characteristics (n=158).
|
Total
|
Responder (n=102)
|
Non-Responder (n=56)
|
Statistics
|
P
|
|
Gender, male (%)
|
64 (40.5%)
|
39 (38.2%)
|
25 (44.6%)
|
Χ2
1=0.62
|
0.43a.
|
|
Age (year, mean±SD)
|
72.66±8.31
|
73.47±8.18
|
71.18±8.41
|
t156=−1.67
|
0.10b.
|
|
Education (year, median and interquartile)
|
8 (6, 12)
|
6 (6, 12)
|
9 (6, 12)
|
Z=1.19
|
0.23c.
|
|
Drug (%)
|
|
|
|
|
|
|
Donepezil
|
84 (53.2%)
|
57 (55.9%)
|
27 (48.2%)
|
Χ2
2=1.39
|
0.50a.
|
|
Galantamine
|
52 (32.9%)
|
33 (32.4%)
|
19 (33.9%)
|
|
|
|
Rivastigmine
|
22 (13.9%)
|
12 (11.8%)
|
10 (17.9%)
|
|
|
|
Baseline Dementia Severity
|
|
|
|
|
|
|
K-MMSE score (mean±SD)
|
19.11±4.73
|
18.55±4.70
|
20.13±4.64
|
t156=2.02
|
0.04
b.
|
|
CDR (%)
|
|
|
|
|
|
|
0.5
|
54 (34.2%)
|
32 (31.4%)
|
22 (39.3%)
|
Χ2
2=1.30
|
0.52a.
|
|
1
|
74 (46.8%)
|
51 (50.0%)
|
23 (41.1%)
|
|
|
|
2
|
30 (19.0%)
|
19 (18.6%)
|
11 (19.6%)
|
|
|
SD, standard deviation; K-MMSE score, Korean Mini Mental State Examination score;
CDR, Clinical Dementia Rating
a. Chi-squared test was used
b. Student’s t test was used
c. Wilcoxon rank-sum test was used
SNP association analysis with responder of acetylcholinesterase inhibitors
The results of SNP association analysis are shown in [Table 2]. The observed genotype frequencies in each case fitted the ones expected according
to the Hardy-Weinberg equilibrium, except one SNP, rs12246528 (P=2.80×10−24). However, we did not exclude this SNP, because this pharmacogenetic study was conducted
in AD patients and did not have a normal control group [19]
[24]. Moreover, 2 adjacent SNPs (rs2177370 and rs3793790) were significantly associated
with response.
Table 2 SNP association analysis with responsiveness.
|
SNP by Group
|
Genotype Count
|
|
|
Location a.
|
Statistics for HWE b.
|
P c.
|
FDR Corrected P
|
|
CHAT (chromosome 10)
|
|
|
|
|
|
|
|
|
rs3810950
|
GG
|
GA
|
AA
|
|
|
|
|
|
Responder
|
75
|
24
|
1
|
50824619
|
Χ
2
1=0.34
|
0.73
|
1
|
|
Non-Responder
|
40
|
14
|
1
|
|
P=0.56
|
|
|
|
rs4838391
|
CC
|
TC
|
TT
|
|
|
|
|
|
Responder
|
58
|
39
|
4
|
50832109
|
Χ
2
1=0.09
|
0.14
|
1
|
|
Non-Responder
|
27
|
23
|
6
|
|
P=0.77
|
|
|
|
rs4838392
|
AA
|
GA
|
GG
|
|
|
|
|
|
Responder
|
34
|
49
|
15
|
50834978
|
Χ
2
1=0.46
|
0.61
|
1
|
|
Non-Responder
|
19
|
26
|
6
|
|
P=0.50
|
|
|
|
rs12246528
|
GA
|
GG
|
AA
|
|
|
|
|
|
Responder
|
12
|
89
|
0
|
50835264
|
Χ
2
1=103.35
|
0.42
|
1
|
|
Non-Responder
|
4
|
51
|
0
|
|
P=2.80×10-24
|
|
|
|
rs11101187
|
CC
|
CT
|
TT
|
|
|
|
|
|
Responder
|
93
|
9
|
0
|
50837034
|
Χ
2
1=2.18
|
0.79
|
1
|
|
Non-Responder
|
53
|
2
|
1
|
|
P=0.14
|
|
|
|
rs2177370
|
CC
|
TC
|
TT
|
|
|
|
|
|
Responder
|
48
|
47
|
6
|
50838874
|
Χ
2
1=0.29
|
0.003
|
0.03
|
|
Non-Responder
|
44
|
7
|
4
|
|
P=0.59
|
|
|
|
rs3793790
|
AA
|
GA
|
GG
|
|
|
|
|
|
Responder
|
46
|
52
|
4
|
50840736
|
Χ
2
1=0.91
|
0.002
|
0.03
|
|
Non-Responder
|
42
|
10
|
3
|
|
P=0.34
|
|
|
|
rs3793791
|
CC
|
TC
|
CC
|
|
|
|
|
|
Responder
|
49
|
45
|
8
|
50841704
|
Χ
2
1=0.25
|
0.90
|
1
|
|
Non-Responder
|
31
|
18
|
7
|
|
P=0.61
|
|
|
|
rs12266458
|
CC
|
TC
|
TT
|
|
|
|
|
|
Responder
|
36
|
47
|
19
|
50847997
|
Χ
2
1=0.81
|
0.21
|
1
|
|
Non-Responder
|
15
|
25
|
15
|
|
P=0.37
|
|
|
|
rs1917818
|
AA
|
CA
|
CC
|
|
|
|
|
|
Responder
|
62
|
30
|
10
|
50849342
|
Χ
2
1=6.36
|
0.20
|
1
|
|
Non-Responder
|
39
|
13
|
3
|
|
P=0.01
|
|
|
|
rs11101192
|
GG
|
GA
|
AA
|
|
|
|
|
|
Responder
|
60
|
32
|
8
|
50854767
|
Χ
2
1=4.66
|
1
|
1
|
|
Non-Responder
|
34
|
15
|
6
|
|
P=0.03
|
|
|
|
rs7094248
|
CC
|
GC
|
GG
|
|
|
|
|
|
Responder
|
52
|
36
|
12
|
50855368
|
Χ
2
1=4.34
|
0.55
|
1
|
|
Non-Responder
|
33
|
17
|
6
|
|
P=0.04
|
|
|
|
rs11101193
|
GG
|
GT
|
TT
|
|
|
|
|
|
Responder
|
71
|
24
|
7
|
50856138
|
Χ
2
1=7.56
|
0.68
|
1
|
|
Non-Responder
|
41
|
12
|
3
|
|
P=0.01
|
|
|
|
rs3793797
|
TT
|
CT
|
CC
|
|
|
|
|
|
Responder
|
64
|
27
|
11
|
50857849
|
Χ
2
1=3.15
|
1
|
1
|
|
Non-Responder
|
31
|
22
|
2
|
|
P=0.08
|
|
|
|
rs10776586
|
TT
|
TC
|
CC
|
|
|
|
|
|
Responder
|
53
|
36
|
8
|
50858346
|
Χ
2
1=0.08
|
0.89
|
1
|
|
Non-Responder
|
27
|
20
|
3
|
|
P=0.78
|
|
|
|
rs12264845
|
CC
|
CA
|
AA
|
|
|
|
|
|
Responder
|
37
|
54
|
11
|
50863083
|
Χ
2
1=0.69
|
0.90
|
1
|
|
Non-Responder
|
23
|
24
|
8
|
|
P=0.41
|
|
|
|
rs7076926
|
TT
|
CT
|
CC
|
|
|
|
|
|
Responder
|
55
|
43
|
4
|
50863565
|
Χ
2
1=1.42
|
0.49
|
1
|
|
Non-Responder
|
28
|
24
|
4
|
|
P=0.23
|
|
|
|
rs7094421
|
AA
|
GA
|
GG
|
|
|
|
|
|
Responder
|
76
|
23
|
1
|
50863623
|
Χ
2
1=0.49
|
0.25
|
1
|
|
Non-Responder
|
47
|
9
|
0
|
|
P=0.48
|
|
|
|
rs3793798
|
TT
|
AT
|
AA
|
|
|
|
|
|
Responder
|
51
|
39
|
10
|
50871466
|
Χ
2
1=0.21
|
0.53
|
1
|
|
Non-Responder
|
25
|
25
|
6
|
|
P=0.64
|
|
|
|
rs3793800
|
AA
|
AG
|
GG
|
|
|
|
|
|
Responder
|
80
|
21
|
1
|
50871716
|
Χ
2
1=0.29
|
0.43
|
1
|
|
Non-Responder
|
47
|
9
|
0
|
|
P=0.59
|
|
|
|
rs3793801
|
CC
|
TC
|
TT
|
|
|
|
|
|
Responder
|
44
|
47
|
10
|
50872912
|
Χ
2
1=0.67
|
0.61
|
1
|
|
Non-Responder
|
26
|
26
|
4
|
|
P=0.41
|
|
|
|
SLC5A7
(chromosome 2)
|
|
|
|
|
|
|
|
|
rs6542746
|
CC
|
TC
|
TT
|
|
|
|
|
|
Responder
|
50
|
41
|
10
|
13279665
|
Χ
2
1=0.02
|
0.37
|
1
|
|
Non-Responder
|
31
|
22
|
3
|
|
P=0.88
|
|
|
|
rs6720783
|
GG
|
GT
|
TT
|
|
|
|
|
|
Responder
|
48
|
46
|
6
|
13297151
|
Χ
2
1=2.64
|
1
|
1
|
|
Non-Responder
|
25
|
26
|
3
|
|
P=0.11
|
|
|
|
rs11685873
|
GG
|
AG
|
AA
|
|
|
|
|
|
Responder
|
73
|
27
|
1
|
13285348
|
Χ
2
1=0.86
|
0.75
|
1
|
|
Non-Responder
|
42
|
10
|
4
|
|
P=0.35
|
|
|
|
ACHE (chromosome 7)
|
|
|
|
|
|
|
|
|
rs6942609
|
GG
|
AG
|
AA
|
|
|
|
|
|
Responder
|
40
|
50
|
11
|
38928323
|
Χ
2
1=1.06
|
0.70
|
1
|
|
Non-Responder
|
24
|
27
|
5
|
|
P=0.30
|
|
|
SNP, single-nucleotide polymorphism; HWE, Hardy-Weinberg equilibrium; FDR, false discovery
rate
a. Genomic position (NCBI Build 37)
b. Chi-squared test was used
c. Exact Cochran-Armitage test for trend was used
The rs2177370 in the intronic region of CHAT gene was significantly associated with response (uncorrected P=0.0025, FDR controlled
P=0.026). The rs3793790 located in the same intron of the rs2177370 showed a significant
association with responsiveness (uncorrected P=0.0024, FDR controlled P=0.026). These
associations were preserved after controlling for gender, age, education year, drug
and baseline K-MMSE score (for rs2177370, P=0.0065, odds ratio=2.45, 95% confidence
interval=1.28–4.68; for rs3793790, P=0.0039, odds ratio=2.73, 95% confidence interval=1.38–5.38).
Haplotype association analysis with responder of acetylcholinesterase inhibitors
We discovered 4 haplotype blocks in the CHAT gene ([Fig. 3]). Among the 13 haplotype allele, 2 alleles in block 2 that included rs2177370 had
significant associations with response (for haplotype CC, uncorrected P=0.004, FDR
controlled P=0.023; for haplotype CT, uncorrected P=0.003, FDR controlled P=0.023).
These haplotypes were also associated with response after controlling for gender,
age, education year, drug and baseline K-MMSE score (for haplotype CC, P=0.006, odds
ratio=0.44, 95% confidence interval=0.24–0.79; for haplotype CT, P=0.006, odds ratio=2.47,
95% confidence interval=1.29–4.72). However, no haplotype blocks were found to be
significantly associated with response in the SLC5A7 gene or the ACHE gene ([Table 3]).
Fig. 3 Linkage disequilibrium (LD) and haplotype structure of CHAT. Pairwise SNP |D′| values (×100) of linkage (|D′|=1 not shown) are shown together
with haplotype blocks. Black squares represent less than 4 distinct 2-marker haplotypes
and white squares represent 4 distinct 2-marker haplotypes by the 4 gamete rule. Triangles
surrounding the markers represent haplotype blocks identified using the default 4-gamete
rule algorithm of Haploview 4.2.
Table 3 Haplotype association analysis with responsiveness in CHAT gene.
|
Haplotype by Group
|
Allele count
|
|
|
P a.
|
FDR Corrected P
|
|
0
|
1
|
2
|
|
|
|
Block 1 (rs4838391-rs4838392-rs12246528)
|
|
|
|
|
|
|
CGG
|
|
|
|
|
|
|
Responder
|
44
|
46
|
12
|
0.70
|
0.83
|
|
Non-Responder
|
23
|
30
|
3
|
|
|
|
CAG
|
|
|
|
|
|
|
Responder
|
41
|
48
|
13
|
0.61
|
0.83
|
|
Non-Responder
|
23
|
29
|
4
|
|
|
|
TAG
|
|
|
|
|
|
|
Responder
|
58
|
40
|
4
|
0.14
|
0.45
|
|
Non-Responder
|
27
|
23
|
6
|
|
|
|
Block 2 (rs11101187-rs2177370)
|
|
|
|
|
|
|
CC
|
|
|
|
|
|
|
Responder
|
10
|
48
|
44
|
0.004
|
0.023
|
|
Non-Responder
|
5
|
9
|
42
|
|
|
|
CT
|
|
|
|
|
|
|
Responder
|
49
|
47
|
6
|
0.003
|
0.023
|
|
Non-Responder
|
45
|
7
|
4
|
|
|
|
Block 3 (rs1917818-rs11101192-rs7094248-rs11101193-rs3793797)
|
|
|
|
|
|
|
CGCTT
|
|
|
|
|
|
|
Responder
|
72
|
23
|
7
|
0.78
|
0.85
|
|
Non-Responder
|
41
|
12
|
3
|
|
|
|
AGCGC
|
|
|
|
|
|
|
Responder
|
64
|
28
|
10
|
1
|
1
|
|
Non-Responder
|
31
|
23
|
2
|
|
|
|
AAGGT
|
|
|
|
|
|
|
Responder
|
61
|
32
|
9
|
0.62
|
0.83
|
|
Non-Responder
|
34
|
16
|
6
|
|
|
|
AGCGT
|
|
|
|
|
|
|
Responder
|
60
|
38
|
4
|
0.07
|
0.32
|
|
Non-Responder
|
25
|
26
|
5
|
|
|
|
Block 4 (rs12264845-rs7076926-rs7094421-rs3793798-rs3793800-rs3793801)
|
|
|
|
|
|
|
ATGTGC
|
|
|
|
|
|
|
Responder
|
80
|
21
|
1
|
0.43
|
0.83
|
|
Non-Responder
|
47
|
9
|
0
|
|
|
|
ACATAC
|
|
|
|
|
|
|
Responder
|
55
|
43
|
4
|
0.49
|
0.83
|
|
Non-Responder
|
28
|
24
|
4
|
|
|
|
CTAAAC
|
|
|
|
|
|
|
Responder
|
51
|
41
|
10
|
0.62
|
0.83
|
|
Non-Responder
|
25
|
25
|
6
|
|
|
|
CTATAT
|
|
|
|
|
|
|
Responder
|
45
|
47
|
10
|
0.70
|
0.83
|
|
Non-Responder
|
26
|
26
|
4
|
|
|
FDR, False discovery rate
a. Exact Cochran-Armitage test for trend was used
Discussion
In this study we assessed 25 SNPs of 3 cholinergic system genes (CHAT, SLC5A7 and ACHE) for association with response to AChEI drugs in AD. We found that 2 SNPs in the
intronic region of CHAT, rs2177370 and rs3793790 had a significant association with drug response. Haplotype
association analysis which was additionally performed showed that block 2 including
rs2177370 among 4 haplotype blocks of the CHAT gene had a significant association with drug response. However, the ACHE and SLC5A7 genes did not contain SNPs or haplotypes that were significantly associated with
response. From this we conclude that the brain’s ability to synthesize ACh in AD is
a critical factor for response to AChEIs, whereas transport and inactivation of the
transmitter are less important factors.
The association of CHAT gene polymorphisms with response is consistent with a previous study [10], but the association with CHAT rs2177370 has not been previously described. In one previous study CHAT rs733722 had a significant association with AChEI drug response in AD patients [10]. In a second study, no association of CHAT rs2177369 with response was reported [11]. The CHAT gene has also been studied for association with AD onset [11], AD risk factor [25], and depression in AD [25] ([Fig. 2]). Although the significantly associated SNPs in our study differ from those in previous
studies, these other results suggest convergent evidence for the importance of the
CHAT gene as a significant gene marker that affects the response of AChEIs. Our results
call attention to the role of ChAT in the synthesis of acetylcholine and to the mechanism
of action of AChEIs in patients with AD.
Acetylcholinesterase inhibitors are drugs that inhibit the acetylcholinesterase enzyme
from breaking down acetylcholine, thereby increasing both the level and duration of
action of the neurotransmitter acetylcholine [26]. However, AChEI drugs depend for their efficacy on an adequate synthesis of ACh.
When ACh synthesis already is impaired by degeneration of cholinergic neurons in AD,
then a genetically determined relatively high synthesis capacity in the remaining
neurons would be expected to favor response to AChEI drugs, and vice versa. We might
infer that haplotype CC, with an odds ratio for response of 0.44, is associated with
a relatively reduced rate of ACh synthesis, whereas haplotype CT, with an OR of 2.47,
is associated with a relatively high rate of ACh synthesis. Additional studies are
needed to establish the functional direction of influence that rs2177370 and rs3793790
exert on the activity of CHAT.
We did not confirm the report of Scacchi et al. that the ACHE rs2571598 had a significant association with drug response in AD patients treated
with rivastigmine [11]. We found no association between the ACHE gene and response to AChEI drugs. The discrepancies between our study and Scacchi’s
may due to differences of SNP selection, ethnicity and the genetic models adopted.
In a previous study examining the CHAT gene, rs3810950 had a significant association with both depression [25] and disease progression [27] in AD. However, there was no association with drug response in this study. Because
both comorbid depression and disease stage can influence cognitive function in AD,
these factors will need to be considered in future pharmacogenetic studies.
We conducted this study in Korean patients. In our previous pharmacogenetic study
of the serotonin transporter in patients with depression, conflicting results were
reported according to ethnicity (Caucasian, Asian) [28]. Most of the existing genetic studies for drug response in AD patients have been
limited to Caucasian populations. Thus, replication studies in different ethnic populations
will be required. Our study is the first haplotype association study of response of
AChEIs and found that haplotype blocks located on CHAT may affect response in both favorable and unfavorable directions. A possible limitation
or a potential advantage in this study was the use of 3 members of the AChEI drug
class. On the one hand we lack statistical power to examine SNP associations with
response to individual drugs. On the other hand, by using all the common members of
the AChEI class our results may be generalizable to the clinical setting.
Author Contributions
The individual authors contributed as follows: Doh Kwan Kim, Woojae Myung, Shin-Won
Lim and Hyeyeon Yoon were involved in study planning and the writing of the manuscript;
Doh Kwan Kim conducted the clinical parts of the study. Hyo Shin Kang was involved
in data acquisition, Seonwoo Kim, Woojae Myung, Hong-Hee Won and Hyeyeon Yoon performed
the statistical analyses; Bernard J. Carroll edited the manuscript and assisted with
interpretation of the data.
