Sage Journals: Discover world-class research

Abstract

Background

Little is known about the rare missense variants (RMVs) of APOE in East Asians, including the Japanese, and their association with Alzheimer's disease (AD) and lipid metabolism.

Objective

To identify APOE RMVs in the Japanese population and investigate their association with AD and lipid metabolism, including low-density lipoprotein cholesterol levels.

Methods

APOE RMVs were explored in the Niigata (NIG; 2589 subjects) and Tohoku (ToMMo; 3307 subjects) cohorts. A case-control study included 6261 AD cases and 16,331 controls, all of whom were aged 65 or older. Sanger sequencing, whole-exome sequencing, or a combination of both was performed on the NIG subjects. We used the genotype data from the ToMMo cohort. APOE RMV frequencies in the Japanese population were compared with various ethnic populations. Associations between APOE RMV genotypes, AD, and lipoproteins were examined.

Results

Fourteen RMVs were identified (minor allele frequency 0.02–0.73%), with 10 unique to East Asians. Five previously reported RMVs, such as the Christchurch RMV, were absent in Japanese individuals. Two RMVs (rs140808909 and rs190853081), which exhibit complete linkage disequilibrium, were found to have protective effects against AD: p_Bonferroni = 4.28E-02, OR (95% CI) = 0.70 (0.54–0.92). No significant differences in cholesterol levels were observed between RMV carriers and non-carriers.

Conclusions

The two APOE RMVs identified in Japanese individuals may have exhibited potential protective effects against AD. Further large-scale studies are needed to confirm these findings and to explore their roles in AD and lipid metabolism.

Keywords

Alzheimer's disease cholesterol Japanese lipid rare missense variants

Introduction

Alzheimer's disease (AD) is the leading cause of dementia in the elderly. AD is a multifactorial disease resulting from interactions between environmental factors and individual genetic susceptibility. A large twin epidemiological study in Sweden estimated the heritability of AD to be 60–80%.¹ Multiple loci are thought to contribute to the onset of AD.² This indicates that genetic factors play substantial roles in the development and progression of AD. Large-scale genome-wide association studies (GWAS) have identified 90 susceptibility genes/loci for AD.³ Recent studies have focused on the application of polygenic risk^4–6 and resilience⁷ score analyses, which integrate multiple AD susceptibility variants to assess individual risks, with potential clinical applications being explored.

Among the AD susceptibility genes/loci identified to date, APOE remains the most significant genetic risk factor across populations.^8–11 A search for “Alzheimer” on the GWAS Catalog website¹² highlights the APOE genomic region as a multiple-hit locus with a remarkably strong association with AD, consistent with findings reported in our previous publications^13–15 and those of others.^3,16 Recently, rare missense variants (RMVs) of APOE were reported to be associated with AD, shedding light on new links between APOE and AD. For example, RMVs such as p.L46P (p.L28P; rs769452 [Pittsburgh]),¹⁷ p.R154S (p.R136S; rs121918393 [Christchurch]),¹⁸ p.R163C (p.R145C; rs769455),¹⁹ p.V254E (p.V236E; rs199768005 [Jacksonville]),²⁰ and p.R269G (p.R251G; rs267606661)²¹ have been reported in Western populations, suggesting potential resistance to amyloid-β (Aβ) accumulation, phosphorylated tau pathology, or the effects of the APOE ε4 allele. Rasmussen et al.²² identified 20 APOE RMVs in an analysis involving more than 100,000 participants and demonstrated their association with plasma APOE levels and AD onset.

However, few reports have addressed the identification and significance of APOE RMVs in the Japanese population. Therefore, this study represents the first attempt to characterize APOE RMVs in this population. We examined the genetic association between the identified RMVs and AD, as well as differences in peripheral blood lipoprotein levels between carriers and non-carriers of these RMVs.

Methods

Subjects for identifying APOE RMVs

We explored APOE RMVs in two Japanese cohorts (Tables 1 and 2). The first cohort, referred to as the Niigata (NIG), consisted of 2589 participants (1150 females and 1439 males) from medical and research institutions across Japan. This cohort was primarily established for research-level genetic diagnoses, including AD and various other neurological disorders. Of these, 826 subjects were based on clinical diagnoses, and 1763 were based on pathological diagnoses. Additionally, 508 of the 2589 NIG subjects were derived from the Japanese Alzheimer's Disease Neuroimaging Initiative (J-ADNI).²³ Sanger sequencing (SS), whole-exome sequencing (WES), and a combination of both (SS + WES) were performed on 839, 1,509, and 241 subjects, respectively. The second cohort, referred to as Tohoku (ToMMo), consisted of 3307 community-dwelling healthy subjects (1810 females and 1497 males) recruited by the Tohoku Medical Megabank Organization.²⁴ In compliance with the data usage policy (Research ID: 2019-0074), individual genotype data for 176 variants within a genomic region of approximately 7.2 kb, including APOE (positioned at 45,407,254–45,414,449 bp in GRCh37/hg19), were obtained. The genotype data for each variant were previously obtained through whole-genome sequencing.²⁴ The total number of subjects from both cohorts was 5896 (2960 females and 2936 males), with participants aged ≥20 years (mean: 64.1 years; median: 66.0 years; range: 20–111 years).

Table 1.

The information on subjects used in the exploration of APOE RMVs.

Cohort	Group	No. of subjects			Sex		Age (AAE/AAD)				Special mention
Cohort	Group	All	CL	NP	F	M	Mean	Median	SD	Range: Min-Max	Special mention
NIG	SS	839	0	839	348	491	76.8	79.0	13.1	20–103
	WES	1509	826	683	723	786	74.1	75.0	10.9	24–111	Including 508 J-ADNI subjects: AD, 140; MCI, 221; and CN, 147
	SS + WES	241	0	241	79	162	70.9	71.0	12.8	21–103
ToMMo	3.5KJPNv2	3307	3307	0	1810	1497	55.9	61.0	14.1	20–90
NIG + ToMMo	−	5896	4133	1763	2960	2936	64.1	66.0	16.2	20–111

AAD: age at death for neuropathologically-characterized subjects; AAE: age at entry/examination for clinically-verified subjects;

AD: Alzheimer's disease; All: CL + NP; CL: clinically-verified subjects; CN: cognitively normal; F: female;

J-ADNI: Japanese Alzheimer's Disease Neuroimaging Initiative; M: male; Max: maximum; MCI: mild cognitive impairment;

Min: minimum; NIG: Niigata University; NP: neuropathologically-characterized subjects; RMVs: rare missense variants;

SD: standard deviation; SS: Sanger sequencing; ToMMo: Tohoku Medical Megabank Organization; WES: whole exome sequencing.

Table 2.

Genotypic and allelic distribution of APOE in the cohorts used for the exploration of APOE RMVs.

Cohort	No. of subjects	Genotype						Allele
Cohort	No. of subjects	ε22*	ε23*	ε24*	ε33*	ε34*	ε44*	ε2	ε3	ε4
NIG	2589	6	155	27	1685	621	95	194	4146	838
	(100.00%)	(0.23%)	(5.99%)	(1.04%)	(65.08%)	(23.99%)	(3.67%)	(3.75%)	(80.07%)	(16.18%)
ToMMo	3307	6	251	30	2428	564	28	293	5671	650
	(100.00%)	(0.18%)	(7.59%)	(0.91%)	(73.42%)	(17.05%)	(0.85%)	(4.43%)	(85.74%)	(9.83%)
NIG + ToMMo	5896	12	406	57	4113	1185	123	487	9817	1488
	(100.00%)	(0.20%)	(6.89%)	(0.97%)	(69.76%)	(20.10%)	(2.09%)	(4.13%)	(83.25%)	(12.62%)

Frequency (%) of each APOE genotype/allele in parentheses.

RMVs: rare missense variants; NIG: Niigata University; ToMMo: Tohoku Medical Megabank Organization.

SS for APOE

SS was performed in 1080 subjects from the NIG cohort (839 for SS and 241 for SS + WES) (Table 1). Genomic DNA was extracted from peripheral blood or autopsied brain tissues and quantified using the NanoDrop spectrophotometer (Thermo Fisher Scientific). DNA quality was assessed by calculating the DNA integrity number (Agilent Technologies). The primers used to sequence the entire coding region of APOE is provided in Supplemental Table 1.

WES

WES was performed on 1750 subjects from the NIG (1509 for WES and 241 for SS + WES) (Table 1). Genomic DNA was used to prepare an exome library using the SureSelect Human All Exon V6 Kit (Agilent Technologies). The library was sequenced on a NovaSeq 6000 platform (Illumina) in 151-cycle paired-end mode. Sequence reads were processed with fastp (v.0.19.5) using the default settings for quality control and adapter trimming. The cleaned reads were then mapped to the human reference genome (hg38) using BWA-MEM (v.0.7.15-r1140) with default parameters. Subsequent analyses, including read processing, variant calling, and filtration, were performed following GATK4 Best Practices. The resulting variant call sets were annotated using snpEff (v.4.3t) and stored in an SQLite database using GEMINI (v.0.20.1).

Determination of six common APOE genotypes

We determined six common APOE genotypes (ε2*2, ε2*3, ε2*4, ε3*3, ε3*4, and ε4*4) in all 5896 subjects from the NIG and ToMMo cohorts (Table 1) based on nucleotide combinations (haplotypes) of the common missense variants rs429358 (T > C: p.C130R [p.C112R]) and rs7412 (C > T: p.R176C [p.R158C]) in APOE. For the NIG cohort, these genotypes were determined using both the SS and TaqMan methods. Information regarding the TaqMan assays used (TaqMan assay IDs: 19AyD0001 for rs429358 and 19AyD0002 for rs7412) is provided in Supplemental Table 1. For the ToMMo cohort, genotypes were determined based on digital data (genotype: 0/0, 0/1, and 1/1). The results are presented in Table 2. The effects of APOE genotypes and alleles on AD risk and protection are shown in Supplemental Figure 1 and Supplemental Table 6. Genotypic and allelic data for APOE were obtained from Asanomi et al.²⁵ and used in this analysis.

Case-control study

We conducted a case-control study to examine whether the identified APOE RMVs were genetically associated with AD. Information concerning the subjects used in this analysis is provided in Table 3, including sex, age (age at entry/examination/death), and APOE (genotype and allele). The diagnostic criteria used for subject recruitment are described in detail elsewhere.^25,26

Table 3.

The information on subjects used in the case-control study.

Group	No. of subjects		Sex			Age(AAE/AAD)				APOE
			Sex			Age(AAE/AAD)				Genotype						Allele
			F	M_Ref		Mean	Median	SD		ε22*	ε23*	ε24*	ε33* _Ref	ε34*	ε44*	ε2	ε3 _Ref	ε4
Cases	6261		4048	2213		77.5	77.5	6.2		4	260	62	3222	2294	419	330	8998	3194
										(0.06%)	(4.15%)	(0.99%)	(51.46%)	(36.64%)	(6.69%)	(2.64%)	(71.86%)	(25.51%)
Controls	16,331		8738	7593		73.9	81.0	6.0		43	1319	158	11,944	2715	152	1563	27,922	3177
										(0.26%)	(8.08%)	(0.97%)	(73.14%)	(16.62%)	(0.93%)	(4.79%)	(85.49%)	(9.73%)
Sum	22,592		12,786	9806		74.8	80.0	6.3		47	1579	220	15,166	5009	571	1893	36,920	6371
										(0.21%)	(6.99%)	(0.97%)	(67.13%)	(22.17%)	(2.53%)	(4.19%)	(81.71%)	(14.10%)
		p _FE	<0.0001		p _t	<0.0001			p _FE	0.0313	<0.0001	0.0160	−	<0.0001	<0.0001	<0.0001	−	<0.0001
		OR	1.59						OR	0.34	0.73	1.45	1.00	3.13	10.22	0.66	1.00	3.12
		95% CI	1.50–1.69						95% CI	0.13–0.93	0.64–0.84	1.08–1.95	−	2.93–3.35	8.46–12.37	0.58–0.74	−	2.95–3.29

Frequency (%) of each APOE genotype/allele in parentheses.

p < 0.05 is shown in bold.

AAD: age at death for neuropathologically-characterized subjects; AAE: age at entry/examination for clinically-verified subjects; CI: confidence interval; F: female; M: male; OR: odds ratio (crude); p_FE: p value for Fisher's exact test; p_t: unpaired t-test with Welch's correction; Ref: reference; SD: standard deviation.

Genomic variants were genotyped using the Asian Screening Array (ASA) (Illumina). A total of 6261 AD cases and 16,331 controls (Table 3) were analyzed using the ASA.

None of the 14 APOE RMVs identified in the NIG and ToMMo cohorts were covered by the ASA platform. Therefore, genotypes were inferred through imputation. Imputation accuracy, reflecting the reliability of inferred genotypes, was evaluated using the INFO score. In SNP analysis, imputation is used to infer missing genetic variants from known genotype data and a reference panel. This approach increases the number of analyzed variants without the need for direct genotyping. SNP imputation was performed using Minimac4 with a Japanese reference panel constructed from 1000 Genomes Project Phase 3 (1KGP3 [May 2013, n = 2503]) and 3181 Japanese whole-genome sequences from the National Center for Geriatrics and Gerontology (NCGG) biobank database. To ensure high imputation accuracy, variants with an INFO score ≥ 0.7 were included in the association analysis. Among the 14 APOE RMVs, genotype data were obtained for four variants: rs121918392 (ε5), rs762461580, rs140808909 (ε7), and rs190853081 (ε7) (Table 4, Figures 1 and 2). The INFO scores for each imputed variant were as follows: rs121918392 (ε5), 0.93; rs762461580, 0.83; rs140808909 (ε7), 0.82; and rs190853081 (ε7), 0.82.

Figure 1.

Case-control study on APOE RMVs identified. Of the 14 APOE RMVs identified in this study, four imputed variants were included in the analysis. Two RMVs, rs140808909 and rs190853081, which were in complete LD (Supplemental Figure 2) and define the ε7 allele, remained statistically significant after multiple comparison corrections in Models 1, 2, and 4. The OR was less than 1.0, suggesting a protective effect against AD. However, after adjusting for the ε4 allele in Models 3 and 5, the significant association of the ε7 allele disappeared. This suggests that the ε7 allele is not independent of the ε4 allele but rather is a dependent allele. The detailed numerical data are presented in Table 4. Bold text denotes that the 95% CI of the OR does not contain 1.00.

Figure 2.

Case-control study on common APOE genotypes and alleles in carriers with APOE RMVs identified. For the four APOE RMVs analyzed in the case-control study (Table 4, Figure 1), we conducted a case-control analysis focusing on subjects carrying alternative alleles, examining common APOE genotypes (A) and alleles (B). Forest plots showing OR and 95% CI are provided. For genotypes (A), the OR (95% CI) of ε3*4 relative to ε3*3 was calculated. For alleles (B), the OR (95% CI) of the ε4 allele relative to the ε3 allele was calculated. Bold text denotes that the 95% CI of the OR does not contain 1.00. “All” indicates the OR and its 95% CI for all subjects (6261 AD cases and 16,331 controls) as presented in Table 3. Importantly, both the ε5 (rs121918392) and ε7 alleles (rs140808909-rs190853081) tended to attenuate the AD risk conferred by the ε4 allele slightly. Ref: reference.

Table 4.

Case-control study on APOE RMVs identified.

	dbSNP (rs ID): ASA (imputed)^#
	rs121918392 (ε5)	rs762461580	rs140808909 (ε7)	rs190853081 (ε7)
Allele
Ref	G	G	G	G
Alt	a	a	a	a
No. of subjects
Cases	6206	6235	6194	6194
Controls	16,164	16,270	16,084	16,084
No. of RMV carriers
Cases	55	26	67	67
Controls	167	61	247	247
MAF
Cases	4.39E-03	2.06E-03	5.35E-03	5.35E-03
Controls	5.11E-03	1.90E-03	7.59E-03	7.59E-03
Model 1: Without covariate adjustment
p_Unadj	3.26E-01	7.04E-01	1.07E-02	1.07E-02
p_Bonferroni	1.00E + 00	1.00E + 00	4.28E-02	4.28E-02
OR	0.86	1.09	0.70	0.70
95% CI	0.63–1.17	0.69–1.72	0.54–0.92	0.54–0.92
Model 2: Sex + Age
p_Unadj	4.37E-01	4.12E-01	7.32E-03	7.32E-03
p_Bonferroni	1.00E + 00	1.00E + 00	2.93E-02	2.93E-02
OR	0.88	1.22	0.68	0.68
95% CI	0.64–1.21	0.76–1.95	0.51–0.90	0.51–0.90
Model 3: APOE ε4
p_Unadj	7.63E-01	1.80E-01	3.18E-01	3.18E-01
p_Bonferroni	1.00E + 00	7.22E-01	1.00E + 00	1.00E + 00
OR	1.05	1.37	0.87	0.87
95% CI	0.77–1.43	0.86–2.18	0.66–1.15	0.66–1.15
Model 4: APOE ε2
p_Unadj	2.52E-01	7.48E-01	9.09E-03	9.09E-03
p_Bonferroni	1.00E + 00	1.00E + 00	3.64E-02	3.64E-02
OR	0.84	1.08	0.70	0.70
95% CI	0.62–1.14	0.68–1.70	0.53–0.91	0.53–0.91
Model 5: Sex + Age + APOE ε4 + APOE ε2
p_Unadj	6.05E-01	7.33E-02	2.81E-01	2.81E-01
p_Bonferroni	1.00E + 00	2.93E-01	1.00E + 00	1.00E + 00
OR	1.09	1.56	0.85	0.85
95% CI	0.79–1.51	0.96–2.53	0.64–1.14	0.64–1.14

The OR and its 95% CI of each model (Models 1–5) are illustrated visually in Figure 1.

Boldface denotes a p-value less than 0.05.

, The INFO scores for each RMV obtained through imputation are as follows: rs121918392 (ε5), 0.93; rs762461580, 0.83; rs140808909 (ε7), 0.82; and rs190853081 (ε7), 0.82.

Alt: alternative; ASA: asian screening array (Illumina); CI: confidence interval; MAF: minor allele frequency; OR: odds ratio; p_Bonferroni: Bonferroni-corrected p value, calculated by setting the number of tests to four; p_Unadj: unadjusted p value without multiple comparison correction; Ref: reference; RMV: rare missense variant.

Table 5.

APOE RMVs identified.

dbSNP(rs ID)	Genomic position(GRCh38.p14)	Allele		Substitution				NIG						ToMMo						Special mention
		Ref	Alt	Nucleotide	a.a.	a.a. position		Genotype			Allele			Genotype			Allele
		(0)	(1)	(Triplet)	a.a.	Immature (317 a.a.)^b	Mature (299 a.a.)^c	0/0	0/1	1/1	0	1	MAF	0/0	0/1	1/1	0	1	MAF
rs1969812567^a	44,906,643	G	c	GCG > cCG	Ala(A) > Pro(P)	7	Cleaved	2588	1	0	5177	1	1.93E-04	3304	2	0	6610	2	3.02E-04
rs121918392	44,907,777	G	a	GAG > aAG	Glu(E) > Lys(K)	21	3	2571	18	0	5160	18	3.48E-03	3284	23	0	6591	23	3.48E-03	ε5
rs201672011	44,907,807	G	a	GAG > aAG	Glu(E) > Lys(K)	31	13	2589	0	0	5178	0	0.00E + 00	3306	1	0	6613	1	1.51E-04
rs762461580	44,907,865	G	a	CGC > CaC	Arg(R) > His(H)	50	32	2586	3	0	5175	3	5.79E-04	3304	3	0	6611	3	4.54E-04
rs370594287	44,907,908	G	c	CAG > CAc	Gln(Q) > His(H)	64	46	2587	2	0	5176	2	3.86E-04	3306	1	0	6613	1	1.51E-04
rs748703149	44,908,729	G	a	GGC > aGC	Gly(G) > Ser(S)	145	127	2589	0	0	5178	0	0.00E + 00	3305	2	0	6612	2	3.02E-04
rs954186991	44,908,762	G	a	GCC > aCC	Ala(A) > Thr(T)	156	138	2589	0	0	5178	0	0.00E + 00	3306	1	0	6613	1	1.51E-04
rs1339651557	44,908,850	G	a	CGC > CaC	Arg(R) > His(H)	185	167	2589	0	0	5178	0	0.00E + 00	3306	1	0	6613	1	1.51E-04
rs1969871875	44,908,852	G	a	GAG > aAG	Glu(E) > Lys(K)	186	168	2589	0	0	5178	0	0.00E + 00	3305	2	0	6612	2	3.02E-04
rs1385100920	44,908,913	G	a	GGC > GaC	Gly(G) > Asp(D)	206	188	2589	0	0	5178	0	0.00E + 00	3306	1	0	6613	1	1.51E-04
rs267606663^a	44,909,021	G	a	CGG > CaG	Arg(R) > Gln(Q)	242	224	2586	3	0	5175	3	5.79E-04	3299	7	0	6605	7	1.06E-03
rs140808909	44,909,080	G	a	GAG > aAG	Glu(E) > Lys(K)	262	244	2555	34	0	5144	34	6.57E-03	3260	46	1	6566	48	7.26E-03	ε7
rs190853081	44,909,083	G	a	GAG > aAG	Glu(E) > Lys(K)	263	245	2555	34	0	5144	34	6.57E-03	3260	46	1	6566	48	7.26E-03	ε7
rs767339630	44,909,129	G	a	CGC > CaC	Arg(R) > His(H)	278	260	2589	0	0	5178	0	0.00E + 00	3305	2	0	6612	2	3.02E-04

One subject without genotypes for two SNPs, rs1969812567 and rs267606663, was excluded in ToMMo.

Number containing the signal peptide consisting of 18 amino acids on the N-terminal side.

Number excluding the signal peptide consisting of 18 amino acids on the N-terminal side.

a.a.: amino acid; Alt: alternative (assigned the number one: 1); MAF: minor allele (= alt. allele 1) frequency; NIG: Niigata University; Ref: reference (assigned the number zero: 0); ToMMo: Tohoku Medical Megabank Organization.

TA cloning for confirmation of APOE RMV alleles

In one case (Subject ID #35: female, age at death 99, APOE ε3*3) harboring three APOE RMVs (all heterozygous: rs267606663-G/a, rs140808909 [ε7]-G/a, rs190853081 [ε7]-G/a) (Supplemental Table 5), we examined whether these RMVs were co-located on the same chromosome. All three RMVs mapped to exon 4 of APOE. PCR was performed using genomic DNA as a template, followed by TA cloning with the TaKaRa Mighty TA-cloning Kit. The PCR primers used were as follows: forward, 5’-TGCCGATGACCTGCAGAA-3’; reverse, 5’-GCTGCATGTCTTCCACCAG-3’.

Lipoproteins

Lipoprotein measurements, including total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides (TG), were obtained from the J-ADNI subjects in the NIG (508 subjects: baseline) and ToMMo (TC: 234 subjects; non-TC: 1630 subjects) cohorts. The ToMMo cohort included subjects assessed under both fasting and non-fasting conditions, as well as those with missing fasting status data. In contrast, all blood samples from the J-ADNI cohort were collected under fasting conditions. We evaluated the quantitative differences in lipoprotein levels (TC, LDL-C, HDL-C, and TG) between APOE RMV carriers and non-carriers.

Statistical analysis

Differences in sex distribution and APOE genotypic and allelic distributions between cases (AD) and controls were assessed using Fisher's exact test based on 2 × 2 contingency tables (Table 3, Supplemental Tables 4 and 6). Additionally, unadjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated (Table 3, Supplemental Tables 4 and 6). For age, an unpaired Student's t-test with Welch's correction was used (Table 3).

Logistic regression analysis was conducted in the case-control study to calculate p-values and OR (95% CI) (Table 4, Figure 1). Five models were used in the analysis: Model 1, no covariate adjustment; Model 2, adjusted for sex and age; Model 3, adjusted for the APOE ε4 allele; Model 4, adjusted for the APOE ε2 allele; and Model 5, adjusted for all four covariates used in Models 2, 3, and 4 (i.e., sex, age, the APOE ε4 allele, and the APOE ε2 allele). The Bonferroni method was applied to correct for multiple comparisons for the APOE RMVs analyzed.

A meta-analysis was conducted to integrate and assess the association between APOE genotypes/alleles and AD derived from the two cohorts, as described in the study by Asanomi et al.:²⁵ the NCGG and the Japanese Genetic Study Consortium for Alzheimer's Disease (JGSCAD)²⁷ (Supplemental Table 6, Supplemental Figure 1). Both fixed-effects and random-effects models were assumed, and Chi-square p-values and pooled OR (95% CI) were calculated. Cochran's Q and I² statistics were also computed to evaluate the heterogeneity between the two datasets.

Comparisons of TC, LDL-C, HDL-C, TG, and the LDL-C/HDL-C (LH) ratio between the two groups (APOE RMV carriers versus non-carriers) were conducted using the non-parametric Mann-Whitney U-test (Supplemental Figure 4).

A series of genetic statistical analyses related to the case-control study (Table 4, Figure 1) was conducted using PLINK (v.1.90; https://www.cog-genomics.org/plink/).²⁸ A meta-analysis (Supplemental Table 6, Supplemental Figure 1) was conducted using StatsDirect software (v.4.0.4; https://www.statsdirect.com). Other statistical analyses (Tables 1 and 3, Supplemental Tables 4 and 6, Figure 2, Supplemental Figure 4) were performed using GraphPad Prism software (v.10.3.1; https://www.graphpad.com). p-values of less than 0.05 were considered statistically significant.

Databases

Databases and analysis websites used in this study are listed below in alphabetical order.

Alzforum MUTATIONS APOE: https://www.alzforum.org/mutations/apoe

ClinVar: https://www.ncbi.nlm.nih.gov/clinvar/

dbSNP: https://www.ncbi.nlm.nih.gov/snp/

Ensemble Human (GRCh38.p14): https://asia.ensembl.org/Homo_sapiens/Info/Index

gnomAD (v4.1.0): https://gnomad.broadinstitute.org

GWAS Catalog: https://www.ebi.ac.uk/gwas/home

jMorp: https://jmorp.megabank.tohoku.ac.jp

LDpair: https://ldlink.nih.gov/?tab=ldpair

UCSC genome: https://genome.ucsc.edu

Results

Identification of APOE RMVs in the Japanese population

Using 2589 subjects from the NIG cohort (Table 1), we identified seven APOE RMVs (Table 5, Figure 3): rs1969812567 (p.A7P), rs121918392 (p.E21 K), rs762461580 (p.R50H), rs370594287 (p.Q64H), rs267606663 (p.R242Q), rs140808909 (p.E262K), and rs190853081 (p.E263K). All carriers were heterozygous, and no homozygotes were identified. A summary of the clinical and pathological diagnostic information of these patients is provided in Supplemental Table 5. Among these seven RMVs, the two variants with the highest alternative allele frequencies were rs140808909 (p.E262K) and rs190853081 (p.E263K), both with a minor allele frequency (MAF) of 0.00657 (Table 5). These two variants were in complete linkage disequilibrium (LD) (LD coefficient R² = 1), indicating that they were fully co-inherited (Supplemental Figure 2). In other words, knowing the genotype of one variant makes it possible to predict the genotype of the other with 100% accuracy. Both variants have been previously identified in Japanese patients with dyslipidemia and atherosclerosis and is defined as the APOE ε7 allele.^29–31 rs121918392 (p.E21K) was defined as the APOE ε5 allele^29,32–34 and was the third-most frequent variant after the APOE ε7-defining rs140808909 (p.E262K) and rs190853081 (p.E263K), with an MAF of 0.00348 (Table 5). rs121918392 (p.E21K) was also associated with dyslipidemia and atherosclerosis.^29,32–34

Figure 3.

APOE RMVs identified in this study. The 14 APOE RMVs identified in this study were added to the Alzforum MUTATIONS APOE diagram (https://www.alzforum.org/mutations/apoe). RMVs identified in both ToMMo and NIG cohorts are marked in red, whereas those identified only in ToMMo are marked in magenta. The five RMVs previously reported are labeled with the letters “a” to “e”. The ε7 allele, which is the primary focus of this paper, is defined by two RMVs that replace two glutamic acid (Glu, E) residues with lysine (Lys, K) at the entrance of the lipid-binding domain (colors are visible in the online version).

We also examined the presence of loss-of-function variants in the NIG cohort (1750 subjects; Table 1) that underwent WES. No loss-of-function mutations were detected.

In the analysis of 3307 subjects from the ToMMo cohort (Table 1), we identified 14 APOE RMVs (Table 5, Figure 3). Seven variants were identified in both the NIG and ToMMo cohorts. The following seven other variants were found in the ToMMo cohort but not in the NIG cohort: rs201672011 (p.E31K), rs748703149 (p.G145S), rs954186991 (p.A156T), rs1339651557 (p.R185H), rs1969871875 (p.E186K), rs1385100920 (p.G206D), and rs767339630 (p.R278H). All carriers were heterozygous, and no homozygotes were identified. We identified one homozygous individual for the two variants in complete LD, rs140808909 (p.E262K) and rs190853081 (p.E263K): male sex; age at entry, 65 years; APOE genotype, ε3*3; and body mass index, 23.0. This individual exhibited LDL-C, HDL-C, and TG levels of 96, 49, and 67, respectively, all within the normal ranges (TC data were not available).

Five previously reported variants (Figure 3), rs769452 (p.L46P: Pittsburgh)^,¹⁷ rs121918393 (p.R154S: Christchurch)^,¹⁸ rs769455 (p.R163C)^,¹⁹ rs199768005 (p.V254E: Jacksonville)^,²⁰ and rs267606661 (p.R269G)^,²¹ were not detected in either the NIG or ToMMo cohorts. Furthermore, these variants were not observed in the Japanese multi-omics database jMorp (version 60KJPN)³⁵ (Supplemental Table 3). These variants were more frequently observed in non-East Asian populations (Supplemental Table 3).

Relationship between identified APOE RMVs and the common APOE alleles, ε2, ε3, and ε4

Clarifying the common APOE alleles (ε2, ε3, and ε4) on which each identified APOE RMV is located is crucial for understanding the potential genetic relationship between AD and APOE. Therefore, for each APOE RMV, we categorized subjects from the NIG and ToMMo cohorts (Tables 1 and 2) according to the common APOE genotypes: ε2*2, ε2*3, ε2*4, ε3*3, ε3*4, and ε4*4 (Supplemental Table 2). The most common APOE genotype among APOE RMV carriers was ε3*3. Several individuals were also identified with ε3*4 and ε2*3, and one individual with ε4*4. None of the APOE RMV carriers had ε2*2 or ε2*4. Variant rs121918392 (p.E21K) that defines the ε5 allele, and two variants, rs140808909 (p.E262K) and rs190853081 (p.E263K), that are in complete LD and define the ε7 allele, appeared to co-segregate with the ε3 allele (Supplemental Table 2). The rs767339630 (p. R278H) variant was observed in a subject with ε4*4, suggesting that this variant is likely to co-segregate with the ε4 allele (Supplemental Table 2).

Ethnicity dependent frequencies of APOE RMVs

To investigate whether the 14 APOE RMVs identified in the NIG and ToMMo cohorts were present in other populations and to assess their frequencies, we explored public databases (Supplemental Table 3). We initially examined the jMorp database.³⁵ Alternative alleles were observed in all 14 variants in this database. For two variants of complete LD, rs140808909 (p.E262K) and rs190853081 (p.E263K), 13 homozygous individuals were identified in the database (version 60KJPN). This finding was consistent with our analysis, which showed that these two APOE RMVs exhibited the highest alternative allele frequencies. The remaining 12 variants were exclusively observed in heterozygous individuals.

Four variants, rs201672011 (p.E31K), rs762461580 (p.R50H), rs370594287 (p.Q64H), and rs748703149 (p.G145S), were also observed in populations other than East Asian populations (Supplemental Table 3). In contrast, six variants, rs121918392 (p.E21K), rs954186991 (p.A156T), rs1339651557 (p.R185H), rs267606663 (p.R242Q), rs140808909 (p.E262K), and rs190853081 (p.E263K), were highly ethnicity-dependent and observed only in the East Asian population (Supplemental Table 3). Information regarding the frequencies of the four variants rs1969812567 (p.A7P), rs1969871875 (p.E186K), rs1385100920 (p.G206D), and rs767339630 (p.R278H) was not available for other populations (Supplemental Table 3).

Effect of APOE RMVs on AD

To investigate whether the 14 identified RMVs (Table 5, Figure 3) were genetically associated with AD, we conducted a case-control study involving 6261 AD cases and 16,331 controls (Table 3). Among the 14 RMVs examined, genotype data for four variants were obtained by imputation. Two variants, rs140808909 (p.E262K) and rs190853081 (p.E263K), which define the ε7 allele, exhibited a statistically significant association with AD in Models 1 (without covariate adjustment: p_Bonferroni = 0.0428, and OR [95% CI] = 0.70 [0.54–0.92]), 2 (adjusted for sex and age: p_Bonferroni = 0.0293, and OR [95% CI] = 0.68 [0.51–0.90]), and 4 (adjusted for the ε2 allele: p_Bonferroni = 0.0364, and OR [95% CI] = 0.70 [0.53–0.91]) (Table 4, Figure 1). These results suggest that the ε7 allele may confer protection against AD (OR < 1.00). However, in Models 3 and 5, where the ε4 allele was included as a covariate, the association was no longer significant: p_Bonferroni = 1.000 and OR (95% CI) = 0.87 (0.66–1.15) in Model 3; and p_Bonferroni = 1.000 and OR (95% CI) = 0.85 (0.64–1.14) in Model 5 (Table 4, Figure 1). Since the ε7 allele co-segregates with the ε3 allele and never with the AD risk-associated ε4 allele (Supplemental Table 2), its protective effect is likely attributable to the ε3 allele. Furthermore, since these two RMVs were in complete LD (Supplemental Figure 2), they exhibited identical p-values and OR (95% CI) across all models examined (Table 4, Figure 1).

Association between APOE genotypes/alleles and AD in APOE RMV carriers

By focusing on carriers of the four RMVs analyzed in the case-control study (Table 4, Figure 1), we investigated whether APOE genotypes and alleles were associated with AD. As no carriers of APOE RMVs with the ε2*2, ε2*4, or ε4*4 genotypes were identified, they were excluded from the analysis. Additionally, due to the low frequencies of the ε2*3 genotype and ε2 allele, they were excluded from the analysis. As a result, the comparison of genotypes focused on ε3*3 and ε3*4, and we compared ε3 and ε4 for alleles. The results of the analyses are presented in Figure 2 and Supplemental Table 4. For reference, the OR (95% CI) for the entire cohort, including both carriers and non-carriers of APOE RMVs and denoted as “All”, was also provided, derived from the data in Table 3. The OR (95% CI) for both APOE genotypes and alleles exhibited similar trends. Specifically, for the three RMVs defining the ε5 (rs121918392) and ε7 (rs140808909 and rs190853081) alleles, the risk associated with the ε4 allele showed a tendency to decrease. For example, in a comparison between the ε3*3 and ε3*4 genotypes, the OR (95% CI) for All was 3.13 (2.93–3.35), whereas it was 1.79 (0.78–4.26) for ε5 (rs121918392). In contrast, for rs762461580, there was a trend towards an increase in risk associated with the ε4 allele. For example, in the same genotype comparison, the OR (95% CI) for All was 3.13 (2.93–3.35), but was 4.18 (1.15–14.00) for rs762461580.

Comparison of lipoprotein levels between APOE RMV carriers and non-carriers

We compared blood lipoprotein levels between carriers and non-carriers of APOE RMVs in J-ADNI participants (508 subjects) from the NIG cohort and in participants from the ToMMo cohort (TC, 234 subjects; and non-TC, 1630 subjects). No statistically significant differences were observed between APOE RMV carriers and non-carriers in terms of TC, LDL-C, HDL-C, TG, or LH ratios (Supplemental Figure 4).

Discussion

In this study, we identified 14 APOE RMVs in 5896 Japanese individuals (Table 5, Figure 3). Some APOE RMVs were specific to East Asian populations, including the Japanese population (Supplemental Table 3). Notably, variants previously reported in Western populations (e.g., the Christchurch¹⁸ and Jacksonville²⁰ variants) were absent in the Japanese population (Supplemental Table 3). These findings suggested that RMVs exhibit distinct ethnic and regional specificities. To investigate their genetic association with AD, we conducted a case-control study involving 22,592 individuals (Table 3). Two APOE RMVs, rs140808909 and rs190853081, which were in complete LD (Supplemental Figure 2) and define the ε7 allele,^29–31 exhibited a statistically significant association with AD (Table 4, Figure 1). Although this effect may be influenced by the ε3 allele, it suggests a potential protective effect (Table 4, Figure 1). To the best of our knowledge, this is the first report to demonstrate an association between APOE RMVs and AD in East Asians. Future studies with larger sample sizes are strongly recommended to validate these findings as reproducibility is crucial. Given the genetic diversity and ethnic specificity of genomic variants^,³⁶ detailed analyses for each ethnic group are essential.³⁷

In this exploratory study, we identified 14 APOE RMVs (Table 5, Figure 3), with the ε7 allele^29–31 being the most intriguing due to its potential protective effects against AD (Table 4, Figure 1). The two RMVs that define the ε7 allele replace two glutamate (E) residues with lysine (K) at the entry point of the lipid-binding domain on the C-terminal side of APOE (Figure 3). This substitution reduces the binding affinity of the ε7 isoform to the LDL receptor (LDLR) and increases its affinity for heparin.³⁸ No previous studies have reported a genetic association between the ε7 allele and AD. However, Youn et al.³⁹ suggested that the ε7 allele may be involved in cognitive impairment based on a study of 344 amnestic patients and 345 controls in a Korean population. They identified three heterozygous carriers of the ε7 allele, all of whom belonged to the amnestic group. Two of these individuals were clinically diagnosed with AD and small vessel disease, and the remaining one was diagnosed with subjective cognitive impairment. They also predicted the three-dimensional protein structure of the ε7 isoform, indicating that its C-terminal structure resembled that of the ε4 isoform, despite having the ε3 isoform background. They concluded that the ε7 isoform may behave like the ε4 isoform and potentially exert adverse effects on cognitive function. Although our study suggests that the ε7 allele may have a protective effect against AD, the relationship between cognitive function and the ε7 allele remains an important topic for future investigation.

We compared the peripheral blood cholesterol levels (TC, LDL-C, HDL-C, TG, and LH ratios) between carriers and non-carriers of the identified APOE RMVs. Contrary to expectations, no significant differences were observed in cholesterol levels (Supplemental Figure 4). Due to the rarity of APOE RMVs, all carriers were grouped into a single category for analysis in this study. Different APOE RMVs may have varying effects on peripheral blood cholesterol levels; therefore, each APOE RMV should ideally be analyzed individually. For example, a study on the ε5 and ε7 alleles revealed a trend toward higher TC (>220 mg/dL) and TG (>150 mg/dL) levels compared with those in controls.⁴⁰ Considering the low population frequency of APOE RMVs (<1%), comparative analyses with larger sample sizes are required.

This study had several limitations. First, the number of identified RMVs (Table 5, Figure 3) was lower than expected. Therefore, further studies with larger Japanese cohorts are needed, particularly including individuals aged 65 and older for late-onset AD analysis. Second, we have not confirmed the genetic association between AD and the ε7 allele (rs140808909 and rs190853081) in a larger, independent East Asian sample set, highlighting the need for further investigation. Third, as this study was based on cross-sectional data, the longitudinal relationship between RMVs and AD risk remains unclear. Future studies should incorporate longitudinal data such as those from the ADNI project (https://adni.loni.usc.edu) to clarify the temporal relationships among RMVs, AD onset, and disease progression. Finally, no functional assays have been conducted for RMVs, including the ε7 allele, and their impact on AD pathology and the underlying molecular mechanisms remain unclear. Given APOE's role in lipid metabolism, elucidating its relationship with key AD pathologies like Aβ and tau is crucial. Notably, the ε7 allele has been linked to dyslipidemia and atherosclerosis,^29–31 yet few studies have examined its binding to LDLR and heparin.³⁸ As these molecules play critical roles in Aβ clearance and tau aggregation,^41,42 the ε7 isoform may influence AD pathology through these pathways.

APOE plays a crucial role in the pathophysiology of AD, particularly in relation to Aβ and tau.^43,44 Previous research has largely focused on the major missense alleles of APOE, namely ε2, ε3, and ε4. However, interest in APOE RMVs, including the Christchurch variant (rs121918393: p.R154S [p.R136S]),^18,45–49 is growing. Further research in this area is expected to provide valuable insights into their functional significance. Such studies will contribute to a deeper understanding of APOE's role in AD pathophysiology. To elucidate the biological implications of the 14 APOE RMVs identified in this study, additional investigations using in vitro (cultured cells) and in vivo (animal models such as mice) approaches are required.

Supplemental Material

sj-pdf-1-alz-10.1177_13872877251340710 - Supplemental material for Association of rare APOE missense variants with Alzheimer's disease in the Japanese population

Supplemental material, sj-pdf-1-alz-10.1177_13872877251340710 for Association of rare APOE missense variants with Alzheimer's disease in the Japanese population by Akinori Miyashita, Ai Obinata, Norikazu Hara, Risa Mitsumori, Daita Kaneda, Yoshio Hashizume, Terunori Sano, Masaki Takao, Ramil Gabdulkhaev, Mari Tada, Akiyoshi Kakita, Akira Arakawa, Maho Morishima, Shigeo Murayama, Yuko Saito, Hiroyuki Hatsuta, Tomoyasu Matsubara, Akio Akagi, Yuichi Riku, Hiroaki Miyahara, Jun Sone, Mari Yoshida, Haruyasu Yamaguchi, Tamao Tsukie, Mai Hasegawa, Kensaku Kasuga, Masataka Kikuchi, Hiroyasu Akatsu, Ryozo Kuwano, Takeshi Iwatsubo, , Shumpei Niida, Kouichi Ozaki and Takeshi Ikeuchi in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-2-alz-10.1177_13872877251340710 - Supplemental material for Association of rare APOE missense variants with Alzheimer's disease in the Japanese population

Supplemental material, sj-xlsx-2-alz-10.1177_13872877251340710 for Association of rare APOE missense variants with Alzheimer's disease in the Japanese population by Akinori Miyashita, Ai Obinata, Norikazu Hara, Risa Mitsumori, Daita Kaneda, Yoshio Hashizume, Terunori Sano, Masaki Takao, Ramil Gabdulkhaev, Mari Tada, Akiyoshi Kakita, Akira Arakawa, Maho Morishima, Shigeo Murayama, Yuko Saito, Hiroyuki Hatsuta, Tomoyasu Matsubara, Akio Akagi, Yuichi Riku, Hiroaki Miyahara, Jun Sone, Mari Yoshida, Haruyasu Yamaguchi, Tamao Tsukie, Mai Hasegawa, Kensaku Kasuga, Masataka Kikuchi, Hiroyasu Akatsu, Ryozo Kuwano, Takeshi Iwatsubo, , Shumpei Niida, Kouichi Ozaki and Takeshi Ikeuchi in Journal of Alzheimer's Disease

Footnotes

Acknowledgments

We wish to thank all the participants for their involvement in this study. We also gratefully acknowledge the ToMMo cohort (Tables 1 and ) for providing individual genomic variant data regarding APOE and its surrounding regions as well as various datasets from the subjects (research ID: 2019-0074). In preparation for this manuscript, an initial proofreading in English was conducted using ChatGPT-4o and subsequently reviewed by native English speakers.

ORCID iDs

Akinori Miyashita

Ai Obinata

Norikazu Hara

Risa Mitsumori

Daita Kaneda

Yoshio Hashizume

Terunori Sano

Masaki Takao

Ramil Gabdulkhaev

Mari Tada

Akiyoshi Kakita

Akira Arakawa

Maho Morishima

Shigeo Murayama

Yuko Saito

Tomoyasu Matsubara

Akio Akagi

Yuichi Riku

Hiroaki Miyahara

Jun Sone

Mari Yoshida

Mai Hasegawa

Kensaku Kasuga

Masataka Kikuchi

Hiroyasu Akatsu

Ryozo Kuwano

Takeshi Iwatsubo

Kouichi Ozaki

Takeshi Ikeuchi

Ethical considerations

This study was approved and conducted according to the ethical guidelines of the Ethics Committee of Niigata University and all participating medical and research institutions (approval number of the Genetic Ethics Review Committee, Niigata University: G2017-0012, G2018-0034, 2018-0409, and G2022-0009).

Consent to participate

Written informed consent was obtained from some participants. For the others, an opt-out method was employed with the approval of the Ethics Committee of Niigata University (approval numbers: G2017-0012, G2018-0034, 2018-0409, and G2022-0009).

Author contributions

Akinori Miyashita (Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation; Visualization; Writing – original draft; Writing – review & editing); Ai Obinata (Formal analysis; Investigation; Methodology; Writing – review & editing); Norikazu Hara (Formal analysis; Investigation; Methodology; Validation; Writing – review & editing); Risa Mitsumori (Formal analysis; Investigation; Methodology; Writing – review & editing); Daita Kaneda (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Yoshio Hashizume (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Terunori Sano (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Masaki Takao (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Ramil Gabdulkhaev (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Mari Tada (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Akiyoshi Kakita (Conceptualization; Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Akira Arakawa (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Maho Morishima (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Shigeo Murayama (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Yuko Saito (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Hiroyuki Hatsuta (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Tomoyasu Matsubara (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Akio Akagi (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Yuichi Riku (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Hiroaki Miyahara (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Jun Sone (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Mari Yoshida (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Haruyasu Yamaguchi (Formal analysis; Investigation; Methodology; Resources; Writing – review & editing); Tamao Tsukie (Formal analysis; Investigation; Methodology; Writing – review & editing); Mai Hasegawa (Formal analysis; Investigation; Methodology; Writing – review & editing); Kensaku Kasuga (Formal analysis; Investigation; Methodology; Writing – review & editing); Masataka Kikuchi (Formal analysis; Funding acquisition; Investigation; Methodology; Writing – review & editing); Hiroyasu Akatsu (Resources; Writing – review & editing); Ryozo Kuwano (Resources; Writing – review & editing); Takeshi Iwatsubo (Funding acquisition; Resources; Writing – review & editing); Shumpei Niida (Funding acquisition; Resources; Writing – review & editing); Kouichi Ozaki (Formal analysis; Funding acquisition; Investigation; Methodology; Resources; Writing – review & editing); Takeshi Ikeuchi (Conceptualization; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Writing – original draft; Writing – review & editing).

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by: (1) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (C), 21K07271 (AM); (2) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (B), 21H03537 (AM); (3) AMED, JP21dk0207055 (AM); (4) JSPS Grant-in-Aid for Scientific Research (KAKENHI), 22H04923 (DK); (5) AMED, JP21wm0425019 (MT); (6) Intramural fund from NCNP, Grant Number 6-8 (MT); (7) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (C), 18K06506 (MT); (8) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (C), 21K06417 (MT); (9) JSPS Grant-in-Aid for Transformative Research Areas, 22H04923 (MT); (10) the Research Committee of Prion Disease and Slow Virus Infection, the Ministry of Health, Labour and Welfare Health and Labour Sciences Research Grants, Japan (MT); (11) AMED, JP24wm0425019 (AK) ; (12) AMED, JP21wm0425019 (YS, SM); (13) JSPS Grant-in-Aid for Scientific Research (KAKENHI), 22H04923 (CoBiA) (YS); (14) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (C), 22K07550 (YS); (15) MHLW Research on rare and intractable diseases Program Grant Number JPMH23FC1008 (YS); (16) AMED, JP21wm0425019 (MY); (17) AMED, JP18kk0205009 (SN); (18) AMED, JP21dk0207045 (SN); (19) AMED, JP22dk0207060 (SN); (20) AMED, JP21dk0207045 (KO); (21) AMED, JP22dk0207060 (KO); (22) Research Funding for Longevity Sciences from NCGG (24-15) (KO); (23) a grant from the Japanese Ministry of Health, Labour, and Welfare for Research on Dementia (KO); (24) AMED, JP23wm0525019 (MK); (25) AMED, JP23dk0207060 (MK); (26) JSPS Grant-in-Aid for Scientific Research (KAKENHI) (B), 23K27515 (TIwatsubo); (27) AMED, JP22dk0207060 (TIkeuchi)

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

The data supporting the findings of this study are available upon request from the corresponding authors (AM and TIkeuchi). However, the WES data used in this study are not publicly available due to privacy or ethical restrictions.

Supplemental material

Supplemental material for this article is available online.

Correction (June 2025):

The originally published supplemental material lacked figure numbers. Some figure legends were also presented in Japanese rather than English. Additionally, the file was initially provided in PowerPoint format, which is not ideal for archiving or referencing. These issues have been corrected, and the supplemental material has been updated to a properly formatted PDF version.

References

Gatz

Reynolds

Fratiglioni

, et al. Role of genes and environments for explaining Alzheimer disease. Arch Gen Psychiatry 2006; 63: 168–174.

Daw

Payami

Nemens

, et al. The number of trait loci in late-onset Alzheimer disease. Am J Hum Genet 2000; 66: 196–204.

Lambert

Ramirez

Grenier-Boley

, et al. Step by step: towards a better understanding of the genetic architecture of Alzheimer’s disease. Mol Psychiatry 2023; 28: 2716–2727.

Escott-Price

Sims

Bannister

, et al. Common polygenic variation enhances risk prediction for Alzheimer’s disease. Brain 2015; 138: 3673–3684.

Stocker

Möllers

Perna

, et al. The genetic risk of Alzheimer’s disease beyond APOE ε4: systematic review of Alzheimer’s genetic risk scores. Transl Psychiatry 2018; 8: 166.

Kikuchi

Miyashita

Hara

, et al. Polygenic effects on the risk of Alzheimer’s disease in the Japanese population. Alzheimers Res Ther 2024; 16: 45.

Hou

Hess

Armstrong

, et al. Polygenic resilience scores capture protective genetic effects for Alzheimer’s disease. Transl Psychiatry 2022; 12: 296.

Farrer

Cupples

Haines

, et al. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. A meta-analysis. APOE and Alzheimer Disease Meta Analysis Consortium. JAMA 1997; 278: 1349–1356.

Bertram

McQueen

Mullin

, et al. Systematic meta-analyses of Alzheimer disease genetic association studies: the AlzGene database. Nat Genet 2007; 39: 17–23.

10.

Belloy

Andrews

Le Guen

, et al. APOE Genotype and Alzheimer disease risk across age, sex, and population ancestry. JAMA Neurol 2023; 80: 1284–1294.

11.

Ridge

Hoyt

Boehme

, et al. Assessment of the genetic variance of late-onset Alzheimer’s disease. Neurobiol Aging 2016; 41: 200.e13–200.e20.

12.

Sollis

Mosaku

Abid

, et al. The NHGRI-EBI GWAS catalog: knowledgebase and deposition resource. Nucleic Acids Res 2023; 51: D977–D985.

13.

Takei

Miyashita

Tsukie

, et al. Genetic association study on in and around the APOE in late-onset Alzheimer disease in Japanese. Genomics 2009; 93: 441–448.

14.

Miyashita

Koike

Jun

, et al. SORL1 Is genetically associated with late-onset Alzheimer’s disease in Japanese, Koreans and Caucasians. PLoS One 2013; 8: e58618.

15.

Shigemizu

Mitsumori

Akiyama

, et al. Ethnic and trans-ethnic genome-wide association studies identify new loci influencing Japanese Alzheimer’s disease risk. Transl Psychiatry 2021; 11: 151.

16.

Coon

Myers

Craig

, et al. A high-density whole-genome association study reveals that APOE is the major susceptibility gene for sporadic late-onset Alzheimer’s disease. J Clin Psychiatry 2007; 68: 613–618.

17.

Kamboh

Aston

Perez-Tur

, et al. A novel mutation in the apolipoprotein E gene (APOE*4 Pittsburgh) is associated with the risk of late-onset Alzheimer’s disease. Neurosci Lett 1999; 263: 129–132.

18.

Arboleda-Velasquez

Lopera

O’Hare

, et al. Resistance to autosomal dominant Alzheimer’s disease in an APOE3 Christchurch homozygote: a case report. Nat Med 2019; 25: 1680–1683.

19.

Le Guen

Raulin

Logue

, et al. Association of African ancestry-specific APOE missense variant R145C with risk of Alzheimer disease. JAMA 2023; 329: 551–560.

20.

Medway

Abdul-Hay

Mims

, et al. ApoE variant p.V236E is associated with markedly reduced risk of Alzheimer’s disease. Mol Neurodegener 2014; 9: 11.

21.

Le Guen

Belloy

Grenier-Boley

, et al. Association of rare APOE missense variants V236E and R251G with risk of Alzheimer disease. JAMA Neurol 2022; 79: 652–663.

22.

Rasmussen

Tybjaerg-Hansen

Nordestgaard

, et al. APOE And dementia - resequencing and genotyping in 105,597 individuals. Alzheimers Dement 2020; 16: 1624–1637.

23.

Iwatsubo

Iwata

Suzuki

, et al. Japanese and North American Alzheimer’s Disease Neuroimaging Initiative studies: harmonization for international trials. Alzheimers Dement 2018; 14: 1077–1087.

24.

Tadaka

Katsuoka

Ueki

, et al. 3.5KJPNv2: an allele frequency panel of 3552 Japanese individuals including the X chromosome. Hum Genome Var 2019; 6: 28.

25.

Asanomi

Shigemizu

Miyashita

, et al. A rare functional variant of SHARPIN attenuates the inflammatory response and associates with increased risk of late-onset Alzheimer’s disease. Mol Med 2019; 25: 20.

26.

Wen

Miyashita

Kitamura

, et al. SORL1 Is genetically associated with neuropathologically characterized late-onset Alzheimer’s disease. J Alzheimers Dis 2013; 35: 387–394.

27.

Kuwano

Miyashita

Arai

, et al. Dynamin-binding protein gene on chromosome 10q is associated with late-onset Alzheimer’s disease. Hum Mol Genet 2006; 15: 2170–2182.

28.

Purcell

Neale

Todd-Brown

, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575.

29.

Yamamura

Yamamoto

Sumiyoshi

, et al. New mutants of apolipoprotein E associated with atherosclerotic diseases but not to type III hyperlipoproteinemia. J Clin Invest 1984; 74: 1229–1237.

30.

Maeda

Nakamura

Kobori

, et al. Identification of human apolipoprotein E variant gene: apolipoprotein Ε7 (Glu244,245----Lys244,245). J Biochem 1989; 105: 51–54.

31.

Tajima

Yamamura

Menju

, et al. Analysis of apolipoprotein E7 (apolipoprotein E-Suita) gene from a patient with hyperlipoproteinemia. J Biochem 1989; 105: 249–253.

32.

Yamamura

Yamamoto

Hiramori

, et al. A new isoform of apolipoprotein E--apo E-5--associated with hyperlipidemia and atherosclerosis. Atherosclerosis 1984; 50: 159–172.

33.

Tajima

Yamamura

Yamamoto

. Analysis of apolipoprotein E5 gene from a patient with hyperlipoproteinemia. J Biochem 1988; 104: 48–52.

34.

Maeda

Nakamura

Kobori

, et al. Molecular cloning of a human apolipoprotein E variant: E5 (Glu3----Lys3). J Biochem 1989; 105: 491–493.

35.

Tadaka

Kawashima

Hishinuma

, et al. jMorp: Japanese Multi-Omics Reference Panel update report 2023. Nucleic Acids Res 2024; 52: D622–D632.

36.

Bergström

McCarthy

Hui

, et al. Insights into human genetic variation and population history from 929 diverse genomes. Science 2020; 367: eaay5012.

37.

Miyashita

Kikuchi

Hara

, et al. Genetics of Alzheimer’s disease: an East Asian perspective. J Hum Genet 2023; 68: 115–124.

38.

Yamamura

Dong

Yamamoto

. Characterization of apolipoprotein E7 (Glu(244)-->Lys, Glu(245)--->Lys), a mutant apolipoprotein E associated with hyperlipidemia and atherosclerosis. J Lipid Res 1999; 40: 253–259.

39.

Youn

Lim

Han

, et al. Apolipoprotein ε 7 allele in memory complaints: insights through protein structure prediction. Clin Interv Aging 2017; 12: 1095–1102.

40.

Matsunaga

Sasaki

Moriyama

, et al. Population frequency of apolipoprotein E5 (Glu3-->Lys) and E7 (Glu244-->Lys, Glu245-->Lys) variants in western Japan. Clin Genet 1995; 48: 93–99.

41.

Kanekiyo

. ApoE and Aβ in Alzheimer’s disease: accidental encounters or partners? Neuron 2014; 81: 740–754.

42.

Shi

Andhey

Ising

, et al. Overexpressing low-density lipoprotein receptor reduces tau-associated neurodegeneration in relation to apoE-linked mechanisms. Neuron 2021; 109: 2413–2426.

43.

Yamazaki

Zhao

Caulfield

, et al. Apolipoprotein E and Alzheimer disease: pathobiology and targeting strategies. Nat Rev Neurol 2019; 15: 501–518.

44.

Belloy

Napolioni

Greicius

. A quarter century of APOE and Alzheimer’s disease: progress to date and the path forward. Neuron 2019; 101: 820–838.

45.

Wardell

Brennan

Janus

, et al. Apolipoprotein E2-Christchurch (136 Arg----Ser). New variant of human apolipoprotein E in a patient with type III hyperlipoproteinemia. J Clin Invest 1987; 80: 483–490.

46.

Sepulveda-Falla

Sanchez

Almeida

, et al. Distinct tau neuropathology and cellular profiles of an APOE3 Christchurch homozygote protected against autosomal dominant Alzheimer's dementia. Acta Neuropathol 2022; 144: 589–601.

47.

Nelson

Liu

Agrawal

, et al. The APOE-R136S mutation protects against APOE4-driven Tau pathology, neurodegeneration and neuroinflammation. Nat Neurosci 2023; 26: 2104–2121.

48.

Chen

Song

Parhizkar

, et al. APOE3ch Alters microglial response and suppresses Aβ-induced tau seeding and spread. Cell 2024; 187: 428–445.e20.

49.

Murakami

Watanabe

Hashimoto

, et al. Inhibitory roles of apolipoprotein E Christchurch astrocytes in curbing tau propagation using human pluripotent stem cell-derived models. J Neurosci 2024; 44: e1709232024.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

2.86 MB

0.05 MB