Several studies have evaluated the association of Siglec-3(CD33) rs3865444 polymorphism and ATP-binding cassette transporter A7(ABCA7) rs3764650 polymorphism with susceptibility to Alzheimer’s disease. However, these studies have yielded contradictory results. Therefore, to resolve this issue, a meta-analysis was undertaken to examine 12 previously published studies. The pooled effect of CD33 rs3865444 showed no significant relationship with susceptibility to Alzheimer’s disease under various genetic models. The pooled effect of ABCA7 rs3764650 also lacked association with susceptibility to Alzheimer’s disease in the allele model (p = 0.06, OR = 1.06, 95% CI, 1.00-1.13), while significant associations were revealed for the dominant model (p
Alzheimer’s disease (AD) is a complex disease involving the interaction of genetic and environmental factors [1]. It has been shown that Aβ-amyloid (Aβ) deposition, related neuronal apoptosis, and neuroinflammation are important factors leading to cognitive deficits [2]. However, AD has a strong genetic component. Genome wide associated studies (GWAS) have identified polymorphisms in or near several genes that are associated with AD risk: ABCA7, CLU, CR1, CD33, CD2AP, EPHA1, BIN1, PICALM, MS4A [3, 4]. Genes such as Siglec-3(CD33) and ATP-binding cassette transporter A7 (ABCA7) are considered to play significant roles in AD progression. Siglec-3(CD33) is located on chromosome 19q13.3 [5]. It is a member of the sialic acid-binding Ig-like lectin family of receptors and is expressed on myeloid cells and microglia [6]. CD33 plays an important role in immunological regulation of Siglecs. Additionally, CD33 activates sialic acid binding, leading to monocyte inhibition via immunoreceptor tyrosine-based inhibitory motif domains [4]. Furthermore, CD33 rs3865444, which is proximal to CD33, was shown to reduce the risk of AD in GWAS [7]. The minor allele of CD33 rs3865444 is correlated with reduced CD33 mRNA expression and insoluble Aβ-amyloid(Aβ) in AD brains [8]. Moreover, CD33 positive immunoreactive microglia have a positive association with insoluble Aβ 42 and plaque burden in AD brains [9]. Thus, CD33 may play an important role in Aβ clearance and other neuroinflammatory pathways mediated by microglia in the brain. ATP-binding cassette transporter A7 (ABCA7) is a member of the ABC transporter superfamily and is encoded by a gene located on chromosome 19p13.3 [10], It assists both transport of substrate across the cell membrane and intracellular lipid into lipoprotein particles [11]. ABCA7 not only stimulates cholesterol efflux and inhibits Aβ secretion [12], but also regulates phagocytosis of apoptotic cells by macrophage through the C1q complement pathway [13]. It may have an effect on AD risk via cholesterol transfer to APOE or by clearance of A aggregates [14, 15]. ABCA7 rs3764650 is associated with neuritic plaque burden in AD brains [16]. ABCA7 mRNA expression in autopsy brain tissue is also associated with advanced cognitive decline [17, 18]. It is expressed in hippocampal CA1 neurons but is 10-fold lower than that in microglia[19]. Thus, we considered that ABCA7 rs3764650 may play an important role in the risk of AD disease.
To date, a number of studies have been conducted to investigate the association between CD33 polymorphism and ABCA7 polymorphism and AD risk. But the number of studies concerned with AD is relatively small. The evidence for the role of single nucleotide polymorphisms in CD33 and ABCA7 as a genetic marker for AD risk is inconsistent. Relatively small sample size, weak effect, or weak penetrance in published studies may provide some of the reasons for conflicting results. Thus, a meta-analysis examining all published data was conducted to determine the statistical evidence for an association between CD33 rs3865444 and ABCA7 rs3764650 polymorphisms and AD susceptibility.
Articles (to February 11, 2017) were electronically searched and screened a from databases including PubMed, Embase, Cochrane Library, China National Knowledge Infrastructure database (CNKI), and Wanfang database (an affiliate of the Chinese Ministry of Science & Technology). Mesh terms with their corresponding synonyms were combined to form the search strategy: “Single nucleotide polymorphisms”, “Alzheimer’s disease”, “CD33” and “ABCA7”. Additional articles were also screened manually from the references in each eligible study.
Inclusion criteria were defined as:
• Studies must assess the association between polymorphism of CD33 rs3865444 or ABCA7 rs3764650 and susceptibility to AD;
• Case-control studies must be based on humans;
• Studies containing sufficient information to obtain the odds ratio (OR) and 95% confidence intervals (CI);
• AD diagnosis should meet clinical criteria set by the NINCDS-ADRDA Alzheimer’s Criteria (National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association);
• Genotype distribution of controls conform to the Hardy-Weinberg equilibrium (HWE).
Exclusion criteria were defined as follows:
• Abstracts, reviews, meta-analysis, case reports, comments, and editorials;
• Duplications, grey literatures, unpublished articles;
• Studies without sufficient genotyping information.
Allele, dominant, recessive, and additive models were employed for the meta-analysis. The CD33 rs3865444 polymorphism has two alleles, A and C, A is the minor allele. The ABCA7 rs3764650 polymorphism has two alleles, G and T, G is the minor allele. The four models were described as: CD33 rs3865444: allele model (A allele versus C allele), dominant model (AA + AC versus CC), recessive model (AA versus AC + CC), and additive model (AA versus CC); ABCA7 rs3764650: allele model (G allele versus T allele), dominant model (GG + GT versus TT), recessive model (GG versus GT + TT), and additive model (GG versus TT).
Cochran’s Q-statistic and I2 test were used to evaluate statistical heterogeneity. A p value of Cochran’s Q test greater than 0.05 and I2 statistic less than 50% (p
The meta-analysis started with 173 articles. Initial review removed 46 duplicate articles to give 127 articles which were then subjected to abstract and keyword review. Of the 127 articles, 109 unrelated articles were discarded to give only 18 articles suited for full-text and data assessment, 6 articles lacked genotyping information, thus 12 eligible articles were finally available for data extraction [22-33]. The detailed information of each study is given in Table 1 and Table 2.
First author | Year | Country | Ethnicity | Case | Control | HWE |
---|---|---|---|---|---|---|
CD33 rs3865444 | ||||||
Hollingworth(GERAD1) [22] | 2011 | Europe | Caucasian | 3333 | 1225 | yes |
Hollingworth(EADI1) [22] | 2011 | Europe | Caucasian | 2025 | 5328 | yes |
Hollingworth(deCODE) [22] | 2011 | Europe | Caucasian | 925 | 612 | yes |
Carrasquillo(Jacksonville) [29] | 2011 | USA | Caucasian | 492 | 920 | yes |
Carrasquillo(Rochester) [29] | 2011 | USA | Caucasian | 312 | 1577 | yes |
Carrasquillo(Autopsy) [29] | 2011 | USA | Caucasian | 298 | 97 | yes |
Carrasquillo(Norway) [29] | 2011 | Europe | Caucasian | 327 | 541 | yes |
Carrasquillo(Poland) [29] | 2011 | Europe | Caucasian | 467 | 187 | yes |
Carrasquillo(ARUK) [29] | 2011 | Europe | Caucasian | 642 | 730 | yes |
Deng [28] | 2012 | China | Asian | 190 | 193 | yes |
Chung [27] | 2013 | Korea | Asian | 290 | 554 | yes |
Tan [25] | 2013 | China | Asian | 612 | 612 | yes |
Omoumi [23] | 2014 | Canada | Caucasian | 580 | 524 | yes |
Walker [31] | 2014 | USA | Caucasian | 97 | 96 | yes |
Carrasquillo [33] | 2014 | USA | Caucasian | 135 | 2440 | yes |
Mao [30] | 2015 | China | Asian | 126 | 129 | yes |
Moreno [24] | 2017 | Colombia | Caucasian | 280 | 357 | yes |
HWE: Hardy-Weinberg equilibrium |
First author | Year | Country | Ethnicity | Case | Control | HWE |
---|---|---|---|---|---|---|
ABCA7 rs3764650 | ||||||
Harold [32] | 2009 | UK/Ireland | Caucasian | 2226 | 4704 | no |
Harold [32] | 2009 | Germany | Caucasian | 555 | 824 | no |
Harold [32] | 2009 | USA | Caucasian | 551 | 930 | yes |
Hollingworth(ADNI) [22] | 2011 | USA | Caucasian | 151 | 177 | yes |
Hollingworth(GERAD2) [22] | 2011 | UK | Caucasian | 3262 | 3320 | yes |
Hollingworth(deCODE) [22] | 2011 | Iceland | Caucasian | 925 | 612 | yes |
Hollingworth(AD-IG) [22] | 2011 | USA | Caucasian | 709 | 971 | yes |
Hollingworth(CHARGE) [22] | 2011 | Netherlands | Caucasian | 1239 | 10813 | yes |
Hollingworth(MAYO2) [22] | 2011 | USA | Caucasian | 2490 | 4114 | yes |
Hollingworth(EADI1) [22] | 2011 | France | Caucasian | 2751 | 2620 | yes |
Tan [25] | 2013 | China | Asian | 612 | 612 | yes |
Chung [27] | 2013 | Korea | Asian | 290 | 554 | yes |
Carrasqillo [33] | 2014 | America | Caucasian | 132 | 2486 | yes |
Liu [26] | 2014 | China | Asian | 350 | 283 | yes |
Omoumi [23] | 2014 | Canada | Caucasian | 580 | 524 | yes |
Moreno [24] | 2017 | Colombia | Caucasian | 280 | 357 | yes |
HWE: Hardy-Weinberg equilibrium |
CD33 rs3865444. A total of 10 studies that included 22,262 patients with AD and 32,244 controls were suitable for the meta-analysis of CD33 rs3865444 polymorphism. Significant statistical heterogeneity was identified at the allelic level (I2 = 71%, p
Forest plot of CD33 rs3865444 polymorphism association with AD using the allele model. M-H, Mantel-Haenszel, random effect model, confidence interval (CI).
CD33 rs3865444 polymorphism | ||||||
---|---|---|---|---|---|---|
Genetic model | Cases/controls(n/n) | Ethnicity | No. of studies | OR (95% CI) | I |
|
allele (A vs. C) | 22262/32244 | Overall | 10 | 1.00[0.92, 1.09] | 0.99 | 71 |
2436/2976 | Asian | 4 | 1.10[0.66, 1.83] | 0.71 | 92 | |
19826/29268 | Caucasian | 6 | 0.98[0.93, 1.02] | 0.26 | 0 | |
Dominant(AA + AC vs. CC) | 4052/8237 | Overall | 7 | 1.25[0.90, 1.73] | 0.18 | 92 |
928/934 | Asian | 3 | 1.36[0.74, 2.51] | 0.33 | 87 | |
3124/7303 | Caucasian | 4 | 1.22[0.82, 1.81] | 0.32 | 93 | |
Recessive(AA vs. AC + CC) | 4052/8293 | Overall | 7 | 1.21[0.93, 1.58] | 0.15 | 65 |
928/934 | Asian | 3 | 1.11[0.49, 2.50] | 0.8 | 73 | |
3124/7359 | Caucasian | 4 | 1.09[0.87, 1.38] | 0.45 | 50 | |
Additive(AA vs. CC) | 2686/4964 | Overall | 7 | 1.10[0.83, 1.44] | 0.51 | 69 |
601/649 | Asian | 3 | 1.76[0.65, 4.81] | 0.27 | 78 | |
2085/4315 | Caucasian | 4 | 0.90[0.78, 1.03] | 0.13 | 0 |
ABCA7 rs3764650. A total of eight studies including 52,214 patients with AD and 82,948 controls were included for the meta-analysis of ABCA7 rs3764650 polymorphism. There was no significant statistical heterogeneity among these studies using the allele model (I
Forest plot of ABCA7 rs3764650 polymorphism association with AD using the allele model. M-H, Mantel-Haenszel, random effect model, confidence interval (CI).
ABCA7 rs3764650 polymorphism | ||||||
---|---|---|---|---|---|---|
Genetic model | Cases/controls(n/n) | Ethnicity | No. of studies | OR (95% CI) | I |
|
allele(G vs. T) | 52214/82948 | Overall | 8 | 1.06[1.00, 1.13] | 0.06 | 49 |
2504/2898 | Asian | 3 | 1.06[0.94, 1.19] | 0.33 | 0 | |
49710/80050 | Caucasian | 5 | 1.06[0.99, 1.14] | 0.1 | 59 | |
Dominant(GG + GT vs. TT) | 4995/10333 | Overall | 8 | 1.20[1.10, 1.31] | <0.0001 | 30 |
962/895 | Asian | 3 | 1.00[0.83, 1.19] | 0.96 | 0 | |
4033/9438 | Caucasian | 5 | 1.28[1.16, 1.41] | <0.00001 | 0 | |
Recessive(GG vs. GT + TT) | 4995/10333 | Overall | 8 | 1.59[1.12, 2.28] | 0.01 | 44 |
962/895 | Asian | 3 | 1.34[0.79, 2.26] | 0.28 | 70 | |
4033/9438 | Caucasian | 5 | 1.96[1.27, 3.04] | 0.002 | 9 | |
Additive(GG vs. TT) | 3886/8484 | Overall | 8 | 1.44[1.13, 1.83] | 0.003 | 36 |
606/535 | Asian | 3 | 1.23[0.92, 1.66] | 0.16 | 23 | |
3280/7949 | Caucasian | 5 | 1.96[1.30, 2.94] | 0.001 | 5 | |
CD33: siglec-3; ABCA7: ATP-Bing Cassette, sub-family A; OR, odds ratio; CI, confidence interval |
CD33 rs3865444. The frequency of CD33 rs3865444 polymorphism was variable among two different subgroups (Asian and Caucasian). A total of 10 articles were suitable for further subgroup analysis of the allele model. There was a significant heterogeneity in Asians (I
Forest plot of CD33 rs3865444 polymorphism association with AD using the allele model for different ethnicities. M-H, Mantel-Haenszel, rondom effct model, confidence interval (CI).
ABCA7 rs3764650. The frequency of ABCA7 rs3764650 polymorphism was variable among different subgroups (Asian and Caucasian). A total of eight articles were suited for further subgroup analysis for the allele model. There was no significant heterogeneity for Asians (I
plot of ABCA7 rs3764650 polymorphism association with AD using the dominant model for different ethnicities. M-H, Mantel-Haenszel, random effect model, confidence interval (CI).
Sensitivity analysis was conducted by excluding one study at a time to assess whether any single study had a strong influence on the pooled OR. For CD33 rs3865444, sensitivity analysis indicate that no single study significantly influenced the pooled OR (data not shown). However, for ABCA7 rs3764650, by excluding the study from Harold et al. [32]. (UK/Ireland) using the additive model, G versus T with OR = 1.44, 95% CI, 1.13-1.83, p = 0.003 changed to OR = 1.29, 95% CI, 0.99-1.68, p = 0.06. Furthermore, Begg’s funnel plot and Egger’s linear regression test were performed to assess publication bias. For CD33 rs3865444, the shapes of the funnel plots show no evidence of publication bias under either the allele model (Begg’s test, p = 0.972; Egger’s test, t = 0.04, p = 0.972) (Fig. 5), dominant model (Begg’s test, p = 0.828; Egger’s test, t =
Publication bias for CD33 rs3865444 and ABCA7 rs3764650 under the allele model detected by Egger’s publication bias plot analysis.
SE:Standard error of mean.
In this meta-analysis, a systematic overview of case-control studies for assessing the association between genetic variants and susceptibility to AD was performed. This meta-analysis, including data from different ethnic groups, may provide a more holistic view of associations. However, the current meta-analysis is somewhat limited. Significant heterogeneity between studies was observed for some comparisons, primarily due to limited sample size, variability among populations, and variations in genotyping and experimental design. Additionally, only data published in selected databases were included. Possibly relevant unpublished studies reporting invalid results were missed, leading to the reported results containing inaccuracies. Despite such limitations, the random or fixed effect model was selected according to the heterogeneity. Hardy-Weinberg equilibrium, sensitivity, and publication bias were also tested. Allele model genotype data were provided in only three of ten studies of CD33. Thus, it is unclear how each study might affect the results of the meta-analysis. ABCA7 likely generates relatively little impact due to the inclusion of specific genetic information.
Many independent studies are concerned with and report the relationship between rs3865444 and AD susceptibility. For example, Omoumi et al. [23] and Chung et al. [27] suggested that rs3865444 can reduce susceptibility to AD, while Deng et al. [28] and Tan et al. [25] demonstrated rs3865444 as a risk factor for AD. Several meta-analyses have also been reported. For example, in a meta-analysis based on 11 studies, Bao et al. [34] showed that rs3865444 is associated with a decreased risk of AD susceptibility (OR = 0.94, 95% CI, 0.90-0.98, p = 0.003), which is consistent with a meta-analysis based on eight studies reported by Li et al. [35](OR = 0.94, 95% CI, 0.91-0.97, p = 1.22E-11). However, no significant association was observed between rs3865444 and AD susceptibility in the current meta-analysis. There are many reasons for different results. Firstly, meta-analyses reported by Bao et al. [34] and Li et al. [35] provide only allelic patterns Although subgroup analyses were also reported, no dominant, additive, or recessive genetic models were investigated. These latter three model types are very important for meta-analysis. Secondly, recently published data have been included here and will have an impact on the results of this meta-analysis. Finally, racial differences also impact meta-analyses. The association between rs3865444 and AD under various specific genetic models have first been reported here. Sensitivity analysis and publication bias analysis indicates these results are robustand exhibit no publication bias. As only four Asian-based studies were included in this meta-analysis [25, 27, 28, 30], the was a relatively small number of Asians when compared to the Caucasian sample size. Further studies are strongly advised for Asian, African, and Caucasian populations to further confirm any association between the rs3865444 SNP and AD susceptibility in other ethnic groups.
The current meta-analysis is based on eight case-controlled studies, including 52,214 patients with AD and 82,948 controls from Asian and Caucasian populations, for comprehensive analysis of the association between rs3764650 SNP and AD susceptibility. Bamji-Mirza et al. [36] suggested that the rs3764650 SNP is significantly associated with AD risk. However, articles have reported that the rs3764650 SNP lacks any association with AD [23, 25, 26, 27, 32]. A meta-analysis by Liu et al. [37] based on nine studies suggested that rs3865444 indicates increased risk for AD susceptibility (OR = 1.21, 95% CI, 1.17-1.26, p
It is well known that CD33 and ABCA7 play significant roles in AD pathogenesis. Cao & Crocker [38] have previously suggested that increased CD33 expression of microglia may result in significantly reduced amyloid beta (A) peptide phagocytosis, which reduces the risk of developing AD. Griciuc et al. [4] also recognized that the minor allele A of CD33 rs3865444 was associated with reductions in both CD33 expression and insoluble amyloid beta 42 (A 42) levels in AD brain. Pahnke et al. [39] have indicated that the ABCA7 protein is involved in the processing of amyloid precursor protein. Mao et al. [30] have suggested from in vivo imaging that ABCA7 genotypes contribute to AD risk through involvement in amyloid deposition, but not tau pathology. A single nucleotide polymorphism (SNP), rs3764650 in ABCA7, is reportedly associated with neuritic plaque burden in AD brains, Chan et al. [18]. Although no significant association was observed between rs3865444 and AD susceptibility in this meta-analysis, CD33 and ABCA7 play significant roles in AD pathogenesis. Better-designed studies with larger sample sizes are needed to more comprehensively analyze the associations, which may further reveal the relationship between Alzheimer’s disease and genetic polymorphism.
The pooled effect of CD33 rs3865444 showed no significant relationship with susceptibility to AD under the various genetic models investigated in this meta-analysis. ABCA7 rs3764650 was associated with an increased risk in AD for the dominant model, recessive model, and additive model, while no significant association was revealed by the allele model. The results of this meta-analysis indicate that ABCA7 rs3764650 might increase the risk of AD, especially among Caucasian populations, contrary to what found for an Asian population.
Our project was supported by the National Nature Scientific Foundation of China (Nos.81472209) and the research project of Jilin Provincial Science and Technology Department (Nos. 20130727029YY,2018041607FG).
All authors declare no conflict of interest.