Age-Related Association between Apolipoprotein E ε4 and Cognitive Function in Japanese Patients with Alzheimer's Disease.

AIMS: In the present study, we investigated whether apolipoprotein E (APOE) polymorphisms influenced the cognitive function of Japanese patients with Alzheimer's disease (AD) at certain ages.
METHODS: Among 200 outpatients with dementia and amnestic mild cognitive impairment, 133 Japanese patients with AD were recruited and divided into two genotypic groups: APOE ε4 carriers and noncarriers. Then, we compared several neuropsychological test scores between the two genotypic groups for two different generations: 70s (70-79 years) and 80s (80-89 years).
RESULTS: The total Mini-Mental State Examination score (p < 0.05) and one of its subtest scores, the 3-stage command score (p < 0.01), were significantly lower for the ε4 carriers than for the noncarriers among patients in their 80s, but not among those in their 70s. The duration of illness differed significantly between the ε4 carriers and the noncarriers among subjects in their 80s but was not correlated with cognitive function.
CONCLUSION: The present results suggest that APOE may significantly influence comparatively simple memory processing in certain generations of Japanese patients with AD.

Introduction

Apolipoprotein E (APOE), phosphatidylinositol-binding clathrin assembly protein, and clusterin have been reported as candidate genes for Alzheimer's disease (AD) in a recent genome-wide study [1]. Among these candidate genes, APOE polymorphisms have been reported as being associated with a robust genetic risk of AD since the 1990s, and ε4 carriers are about three times more susceptible to AD than noncarriers [1,2]. The relation between the APOE gene and the risk of AD has been investigated in populations all over the world, and a significant association between ε4 allele carriers and the manifestation of AD has been accepted. In Japan, a significant association between APOE ε4 carriers and the manifestation of AD has been reported in a large-scale study [3].

The APOE gene has functional roles: it influences onset age, induces morphological changes in the hippocampus, and causes neural vulnerability via cerebrovascular mechanisms [2,4,5,6,7,8,9]. Regarding the influences on cognition, previous studies have reported that memory impairment or executive function is poorer in ε4 carriers than in noncarriers among patients with AD [6,7,10]. Moreover, from age-related viewpoints, the APOE gene tends to have a robust influence on the cognition of patients with AD who are in their 70s or 80s [2,7,11].

In Japan, while the APOE ε4 allele is associated with late-onset manifestation, a few reports have shown that it also influences cognitive function, neuroimaging, and neurobiological data [3,8,9,12]. A previous report has shown a significant difference in memory decline among three genotypic groups classified according to ε4; however, the age of these three groups differed significantly [8]. Regarding clinical function in AD patients, cognitive impairment directly influences activities of daily living and might lead to subsequent behavioral and psychological problems (e.g. persecutory delusion and aberrant motor behavior) [13,14]. Therefore, the investigation of risk factors of cognitive decline might have an impact on the prevention of subsequent problems.

The aims of the present study were to examine whether APOE ε4 influences cognitive function and to analyze its relation to the cognitive profiles of Japanese AD patients. Thus, we retrospectively compared two representative screening test scores, the Mini-Mental State Examination (MMSE) reflecting memory or attentional function and the Frontal Assessment Battery (FAB) reflecting executive function, among APOE ε4 carriers (homozygous and heterozygous) and noncarriers in two different generations (70s: 70-79 years old and 80s: 80-89 years old) that tend to be influenced by the APOE gene [7,11,15,16].

Methods

Participants

Two hundred Japanese outpatients with dementia (AD: n = 133; dementia with Lewy bodies: n = 5; frontotemporal lobar degeneration: n = 8; subcortical vascular dementia or mixed-type dementia: n = 10, and idiopathic normal pressure hydrocephalus: n = 2) and amnestic mild cognitive impairment (A-MCI: n = 42) from among consecutive memory clinic outpatients visiting the Jikei University Hospitals (Tokyo or Kashiwa) participated in the present genomic study. One hundred and thirty-three consecutive Japanese patients with AD were recruited from among the above-mentioned patients. All AD patients were diagnosed as having probable AD based on the National Institute of Neurology and Communicative Disorder and Stroke/Alzheimer's Disease and Related Disorder Association (NINCDS/ADRDA) criteria [17]. Forty-two patients were diagnosed as having A-MCI according to the diagnostic criteria for MCI [18]. APOE genotypes for all patients with AD were identified at the Department of Psychiatry, Juntendo University School of Medicine, based on a method used in previous reports [19,20]. Next, we divided the 133 subjects into two groups according to whether they were ε4 carriers or noncarriers. Among the subjects' detailed demographic variables, behavioral and psychological symptoms were assessed based on information obtained from a structured interview with each patient's caregiver. The interviews were conducted by the same geriatric psychiatrists using the Behavioral Pathology in Alzheimer's disease (Behave-AD) scale [21]. To investigate neuropsychological characteristics, two representative screening tests, the MMSE (ranging from 0 to 30) reflecting memory or attentional function and the FAB (ranging from 0 to 18) reflecting executive function, were administered by a clinical psychologist [15,16]. To determine the severity of each patient's dementia, the geriatric psychiatrists used the Clinical Dementia Rating-Sum of Boxes (CDR-SB: ranging from 0 to 18) based on information provided by each patient's caregiver [22]. The two geriatric psychiatrists and the clinical psychologist were experienced at performing neuropsychological and behavioral examinations, and the interrater validity of the scales was sustained by periodic discussions and exchanges of views. This study was approved by the Ethics Committees of the Jikei University School of Medicine (Tokyo and Kashiwa), and written informed consent was obtained from both the patients and their caregivers.

Comparison of FAB and MMSE Test Scores between the Two Groups according to Two Generations: 70s and 80s

The APOE gene tends to influence the cognition of patients with AD who are in their 70s or 80s [2,7,11]. Thus, we compared the MMSE and FAB scores between ε4 carriers and noncarriers according to two different generations: 70s and 80s.

Statistical Analysis

SPSS 19.0J for Windows (SPSS Japan Inc.) was used for all statistical analyses. To control for type I errors, we used one-way ANOVA with a post hoc Tukey test for assessment of age of the participants in the present study, education years, duration of illness (in months; from age at onset until entry in the present study) as well as MMSE, Behave-AD, CDR-SB, and FAB scores and compared them between ε4 carriers and noncarriers. Seven MMSE subtest scores (with possible scores with a range of 0-5 for ‘time orientation’, 0-5 for ‘place orientation’, 0-3 for ‘registration’, 0-5 for ‘attention and calculation’, 0-3 for ‘recall’, 0-2 for ‘naming’, and 0-3 for ‘3-stage command’) and six FAB subtest scores (with possible scores ranging from 0-3 for ‘similarities’, ‘lexical fluency’, ‘motor series’, ‘conflicting instructions’, ‘go/no go’, and ‘prehension behavior’) were assessed using the Mann-Whitney test. The sex ratio (female/male) and four other MMSE subtest scores (with possible scores with a range of 0-1 for ‘repetition’, ‘reading’, ‘writing’, and ‘copying’) were assessed using the χ2 test. A p value <0.05 was considered statistically significant.

Results

Patient Characteristics

The APOE genotypes of 133 AD patients were examined: 66 subjects were found to be ε4 carriers (ε2/ε4: n = 3, ε3/ε4: n = 48, and ε4/ε4: n = 15) and 67 subjects to be noncarriers (ε2/ε3: n = 5 and ε3/ε3: n = 62). Among the demographic variables of all patients, only age at consultation was significantly lower among the ε4 carriers than among the noncarriers (F = 15.324; d.f. = 1; p < 0.001) (table 1). No significant differences in any other demographic variables were observed between the two genotypic groups in all AD patients (table 1). Regarding the age distribution of the patients in the present study, most of the subjects (88.7%) were in their 70s or 80s (50s: n = 2, 60s: n = 12, 70s: n = 56, 80s: n = 62, and 90s: n = 1).

Table 1

Comparison of subject characteristics (mean ± SD) between APOE ε4 carriers and noncarriers in patients with AD (n = 133)

	Noncarriers (n = 67)	ε4 carriers (n = 66)b	χ2 or F score	p value
Sex (male/female)	21/46	16/50	0.835c	0.361
Age, years	80.73 ± 5.68	76.39 ± 7.07	15.324	<0.001*
Duration of illness, months	29.76 ± 18.99	35.80 ± 25.07	2.459	0.119
Education, years	11.49 ± 2.52	11.92 ± 3.06	0.734	0.393
MMSE score	19.39 ± 4.86	18.88 ± 4.63	0.389	0.534
FAB score	11.02 ± 3.5	10.68 ± 3.76	0.243	0.623
Behave-AD score	7.43 ± 5.01	8.05 ± 5.95	0.413	0.521
CDR-SBa	6.88 ± 3.11	6.78 ± 2.96	0.036	0.849

Sex was analyzed using the χ2 test. Other variables were analyzed using one-way ANOVA with post hoc tests (Tukey).

Ranging from 0 to 18.

Heterozygotes: 51, homozygotes: 15.

The χ2 score is given.

p < 0.001.

Comparison of Demographic Variables between ε4 Carriers and Noncarriers in Their 70s

We found a significant difference in the sex ratio (χ2 = 6.132; d.f. = 1; p = 0.013) between ε4 carriers (male/female: 5/27) and noncarriers (male/female: 11/13), but no significant differences in any other demographic variables (table 2). To investigate whether the APOE genotypes (ε4 carriers or noncarriers) or the sex ratio (female/male) affected the MMSE or FAB scores, we performed a generalized linear model analysis (ANOVA). As a result, significant interactions between these two factors were not observed for the MMSE and FAB scores in a generalized linear model analysis (former: F = 1.719; d.f. = 1; p = 0.196, latter: F = 0.001; d.f. = 1; p = 0.971).

Table 2

Comparison of subject characteristics (mean ± SD) between APOE ε4 carriers and noncarriers in patients with AD who were 70–79 years old (n = 56)

	Noncarriers (n = 24)	ε4 carriers (n = 32)b	χ2 or F score	p value
Sex (male/female)	11/13	5/27	6.132c	0.013*
Age, years	76.29 ± 2.94	76.25 ± 1.98	0.004	0.95
Duration of illness, months	33.50 ± 19.39	29.03 ± 25.11	0.525	0.472
Education, years	12.08 ± 2.04	11.97 ± 3.18	0.024	0.878
MMSE score	18.13 ± 6.11	19.72 ± 4.43	1.250	0.269
FAB score	11.68 ± 3.15	11.43 ± 3.97	0.055	0.815
Behave-AD score	6.92 ± 5.09	7.53 ± 4.70	0.218	0.642
CDR-SBa	6.81 ± 3.18	5.75 ± 2.71	1.822	0.183

Sex was analyzed using the χ2 test. Other variables were analyzed using one-way ANOVA with post hoc tests (Tukey).

Ranging from 0 to 18.

Heterozygotes: 26, homozygotes: 6.

The χ2 score is given.

p < 0.05.

Comparison of Demographic Variables between ε4 Carriers and Noncarriers in Their 80s

The MMSE score was significantly lower (F = 5.757; d.f. = 1; p = 0.02) in the ε4 carriers than in the noncarriers, and the duration of illness (months) was significantly longer (F = 4.975; d.f. = 1; p = 0.029) in the ε4 carriers than in the noncarriers (table 3).

Table 3

Comparison of subject characteristics (mean ± SD) between APOE ε4 carriers and noncarriers in patients with AD who were 80–89 years old (n = 62)

	Noncarriers (n = 41)	ε4 carriers (n = 21)b	χ2 or F score	p value
Sex (male/female)	9/32	3/18	0.523c	0.473
Age, years	84.22 ± 2.60	82.81 ± 2.89	3.78	0.057
Duration of illness, months	28.44 ± 18.78	41.52 ± 26.98	4.975	0.029*
Education, years	11.02 ± 2.68	11.38 ± 3.15	0.218	0.642
MMSE score	19.98 ± 4.02	17.29 ± 4.47	5.757	0.02*
FAB score	10.58 ± 3.69	9.35 ± 3.66	1.358	0.249
Behave-AD score	7.88 ± 5.07	10.29 ± 7.65	2.199	0.143
CDR-SBa	7.05 ± 3.13	8.52 ± 3.01	3.163	0.08

Sex was analyzed using the χ2 test. Other variables were analyzed using one-way ANOVA with post hoc tests (Tukey).

Ranging from 0 to 18.

Heterozygotes: 18, homozygotes: 3.

The χ2 score is given.

p < 0.05.

MMSE Subtest Scores between ε4 Carriers and Noncarriers in Their 80s

Among the MMSE subtest scores, the 3-stage command score of the ε4 carriers was significantly lower than that of the noncarriers (Z = −3.007; d.f. = 1; p = 0.003) (table 4). To investigate the APOE genotypes (ε4 carriers or noncarriers) and the duration of illness (months) as covariates affecting the MMSE total and 3-stage command scores, we performed a generalized linear model analysis (ANCOVA). As a result, a significant interaction between these two factors was not observed for the MMSE total and 3-stage command scores (former: F = 0.016; d.f. = 1; p = 0.900, latter: F = 0.216; d.f. = 1; p = 0.644).

Table 4

Comparison of MMSE subtest scores (mean ± SD) between APOE ε4 carriers and noncarriers in patients with AD who were 80–89 years old (n = 62)

MMSE subtests (ranges)	Noncarrier (n = 41)	ε4 carrier (n = 21)a	χ2 score or M-W Z score	p value
Orientation to time (0–5)	2.34 ± 1.67	1.71 ± 1.42	-1.391	0.164
Orientation to place (0–5)	2.99 ± 1.16	2.62 ± 1.24	-1.305	0.192
Registration (0–3)	2.90 ± 0.30	2.57 ± 0.87	-1.616	0.106
Attention and calculation (0–5)	2.49 ± 1.81	2.05 ± 1.77	-0.939	0.348
Recall (0–3)	1.15 ± 1.15	0.71 ± 0.96	-1.412	0.158
Naming (0–2)	1.93 ± 0.26	1.86 ± 0.36	-0.871	0.384
Repetition (0–1)	0.88 ± 0.33	0.76 ± 0.44	1.385b	0.239
3-stage command (0–3)	2.95 ± 0.22	2.57 ± 0.75	-3.007	0.003*
Reading (0–1)	0.93 ± 0.26	0.95 ± 0.22	0.15b	0.698
Writing (0–1)	0.66 ± 0.48	0.81 ± 0.40	1.537b	0.215
Copying (0–1)	0.71 ± 0.46	0.67 ± 0.48	0.108b	0.742

M-W = Mann-Whitney test.

Heterozygotes: 18, homozygotes: 3.

The χ2 score is given.

p < 0.01.

FAB Total and Subtest Scores between ε4 Carriers and Noncarriers in All AD Patients and Patients in Their 70s or 80s

No significant difference in the FAB total score was observed between the two genotypic groups among all AD patients or among the patients in their 70s or 80s (tables 1, 2, 3). Moreover, the six subtest scores were not significantly influenced by these genotypic groups among patients in their 70s or 80s (data not shown).

Discussion

As the results of the present study show, APOE ε4 significantly influenced the age at consultation in all AD patients (table 1). Furthermore, while APOE ε4 influenced the MMSE total and its subtest score, the 3-stage command score, among subjects in their 80s without any significant interaction with the duration of illness, such results were not observed in subjects in their 70s (tables 2, 3, 4). A previous study has reported that the onset age of AD in APOE ε4 carriers was younger than that in noncarriers, which might have influenced the significant differences in the age at consultation in all AD patients and the illness duration of the patients in their 80s between the two genetic groups in the present study [2].

Some previous studies have shown that APOE ε4 has a significant effect on memory retention, rather than on executive function, in other ethnicities or countries, and our results support these reports [6,7,10]. From a cross-sectional viewpoint, Caselli et al. [7] showed that age-related differences in memory decline among ε4 carriers and noncarriers increased with age. This could explain the present result, which indicates that a significant association between APOE ε4 and cognitive function was found only in patients in their 80s, but not in those in their 70s. Wolk et al. [23] found that the APOE gene had double dissociable roles in memory and executive function through cortical morphometric evaluations performed in ε4 carriers and noncarriers with AD. Luczywek et al. [6] reported that APOE ε4 influences memory retention, rather than working memory, in patients with early-stage AD. However, we did not obtain similar results indicating a significant association between the APOE and FAB scores reflecting executive function.

Due to the effects of APOE ε4 on particular cognitive profiles, ε4 carriers with AD have been reported to have greater hippocampal cortical atrophy, possibly leading to memory impairments [4,9,23]. In a previous study evaluating the characteristics of the MMSE subtest using a factor analysis, the registration and 3-stage command tests were identified as involving immediate or simple memory processing, whereas recall was identified as involving complex memory processing [24,25]. In the present study, a significant correlation was observed between the registration and the 3-stage command scores (Spearman correlation: ρ = 0.373, p = 0.003), but not between the recall and the 3-stage command scores (Spearman correlation: ρ = 0.105, p = 0.418) among patients in their 80s. Thus, APOE ε4 might influence these comparatively simple memory processing tasks in Japanese patients with AD who are in their 80s.

The present study has some limitations. First, since the sample size was comparatively small, we could not compare cognitive function among the three genotypic groups (homozygous for ε4, heterozygous for ε4, and noncarriers) or in younger subjects in their 50s or 60s. Therefore, some type I or type II errors might have occurred. For example, among subjects in their 80s, the mean score for ‘recall’ of the ε4 carriers was lower than that of the noncarriers, but the difference was not statistically significant (p = 0.158) (table 4). Furthermore, the statistical value (p = 0.02) in the MMSE total score comparison was comparatively small (table 3). However, the effect sizes or statistical values were almost the same as those in previous reports [7,8,10]. The statistical significance of the 3-stage command score (p = 0.003) might be reasonable if a p value of <0.05/11 (≒0.0045) is regarded as a significant difference according to the Bonferroni correction, since the MMSE scale has 11 subtest scores (table 4). Second, we did not observe a significant association between ε4 carriers and the FAB scores reflecting executive function (tables 1, 2, 3), as described in previous reports [10,23]. We used representative screening tests that could be completed at bedside without requiring any tools or instruments, and we believe that the simplicity of these tests makes them valuable tools. However, the ranges of the subtest scores were comparatively small; thus, other batteries for evaluating executive function might provide useful information. Third, a significant difference in the sex ratio of the subjects in their 70s was observed in the present study, with a higher frequency of female ε4 carriers (table 2). However, Yasuno et al. [12] reported that the proportion of female elderly subjects with ε4 was comparatively high in Japan. Moreover, a significant interaction between the sex ratio and APOE ε4 was not shown for cognitive function in a generalized linear model analysis using ANOVA. Finally, among the subjects in their 80s, a significant difference in the duration of illness was shown, which might reflect differences in the age at consultation for AD among the two genotypic groups (table 3). However, a significant interaction of illness duration (in months) as a covariance was not shown for cognitive function in a generalized linear model analysis using ANCOVA.

In conclusion, despite some limitations, we found for the first time a significant effect of APOE ε4 on the MMSE score reflecting memory or attentional function (particularly for a simple memory processing task, the 3-stage command), rather than executive function, among Japanese AD patients in their 80s, which was a comparatively older generation. Our results support the findings of previous studies in other countries in addition to furthering the elucidation of the functional roles of the APOE gene [6,7,10]. Furthermore, defining the detailed associations between APOE gene diversity and cognitive behavior in patients with AD using an intermediate phenotype in studies involving neuroimaging or physiological methods in a normal elderly population will be an important task in the future.