- Email: [email protected]
Association between apolipoprotein E gene polymorphism and the risk of multiple sclerosis: A meta-analysis of 6977 subjects
Gene 511 (2012) 12–17
Contents lists available at SciVerse ScienceDirect
Gene journal homepage: www.elsevier.com/locate/gene
Association between apolipoprotein E gene polymorphism and the risk of multiple sclerosis: A meta-analysis of 6977 subjects Yan-Wei Yin a, 1, Yun-Dong Zhang b, 1, Jing-Zhou Wang a, Bing-Hu Li a, Qing-Wu Yang a, Chuan-Qin Fang a, Chang-Yue Gao a, Jing-Cheng Li a,⁎, Li-Li Zhang a,⁎ a b
Department of Neurology, Institute of Surgery Research, Daping Hospital, Third Military Medical University, 10 Changjiang Branch Road, Yuzhong District, Chongqing 400042, PR China Department of Neurosurgery, Institute of Surgery Research, Daping Hospital, Third Military Medical University, 10 Changjiang Branch Road, Yuzhong District, Chongqing 400042, PR China
a r t i c l e
i n f o
Article history: Accepted 6 September 2012 Available online 13 September 2012 Keywords: Multiple sclerosis Apolipoprotein E Polymorphism Meta-analysis
a b s t r a c t Epidemiological studies have evaluated the association between apolipoprotein E (ApoE) gene polymorphism and multiple sclerosis (MS) risk. However, the results remain conflicting. Therefore, in order to derive a more precise association of ApoE gene polymorphism with MS risk, we performed this meta-analysis. Systematic searches of electronic databases PubMed, Embase and Web of Science, as well as hand searching of the references of identified articles were performed. Twenty studies were identified, covering a total of 4080 MS cases and 2897 controls. The results showed evidence for significant association between ApoE ε2 mutation and MS risk (for ε2/ε4 versus ε3/ε3: OR = 1.74, 95% CI = 1.12–2.71, p = 0.01; for ε2 allele versus ε3 allele: OR = 1.16, 95% CI = 1.01–1.35, p = 0.04). In the subgroup analysis by ethnicity, the similar results were obtained among Europeans (for ε2/ε4 versus ε3/ε3: OR = 1.81, 95% CI = 1.14–2.87, p = 0.01; for ε2 allele versus ε3 allele: OR = 1.19, 95% CI = 1.02–1.38, p = 0.03). After excluding the outlier studies by observing Galbraith plot, marginal association was found between ApoE ε3/ε4 genotype and the protective factor for MS (for ε3/ε4 versus ε3/ε3: OR = 0.86, 95% CI = 0.75–0.99, p = 0.04). In summary, the present metaanalysis provides evidence that ApoE ε2 mutation is associated with MS risk. In addition, ApoE ε3/ε4 genotype appears to be a protective factor for MS. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Multiple sclerosis (MS) is a chronic neurological autoimmune disease, which seriously affects the quality of life for patients and national healthcare systems (Olesen et al., 2012). Despite its results from inflammation, demyelination, gliosis and varying degrees of axonal pathology leading to progressive neurological dysfunction (Brück, 2005), a detailed etiology underlying MS is still obscure. Previous studies have reported that the prevalence rates of MS vary substantially throughout the world (Koch-Henriksen and Hyllested, 1998; Noonan et al., 2002; Varadé et al., 2011), and relatives of MS patients are at greater risk for developing the disease than the general population (Ebers et al., 1995). It is assumed that genetic factors may be associated with in the development of MS. Apolipoprotein E (ApoE) plays important role in the lipid and cholesterol transport, and also is involved in brain development and repair (Ignatius et al., 1986; Mahley, 1988). The ApoE gene, located in Abbreviations: ApoE, apolipoprotein E; MS, multiple sclerosis; AD, Alzheimer's disease; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses; OR, odds ratio; CI, confidence interval; NOS, Newcastle–Ottawa Scale; HWE, Hardy–Weinberg equilibrium. ⁎ Corresponding authors. Tel.: +86 23 68757842; fax: +86 23 68757841. E-mail address: [email protected] (L.-L. Zhang). 1 These authors contributed equally to this work. 0378-1119/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gene.2012.09.010
the 19q13 region, is polymorphic with three common alleles: ε2, ε3 and ε4. Consequently, three homozygous (ε2/ε2, ε3/ε3, and ε4/ε4) and three heterozygous (ε2/ε3, ε2/ε4, and ε3/ε4) phenotypes are found in the general population (Lahiri et al., 2004). ApoE ε3 allele is the most (77%) and ε2 allele is the least (8%) common allele. Furthermore, the ApoE ε4 allele frequency is approximately 15% in the general population (Bu, 2009). Since 1993, the association between ApoE ε4 allele and earlier age of onset of both familial and sporadic Alzheimer's disease (AD) has consistently been demonstrated (Corder et al., 1993; Huang, 2006; Strittmatter et al., 1993). MS has the similar pathological properties to AD with inflammation and gliosis. Recently, numerous epidemiological studies have focused on the association between ApoE gene polymorphism and MS risk, and showed the ApoE allele frequencies in MS group (1.3 to 14.4% for ε2 allele, 66.1 to 91.4% for ε3 allele, and 4.5 to 19.4% for ε4 allele) and control group (2.7 to 18.9% for ε2 allele, 74.1 to 91.5% for ε3 allele, and 4.4 to 20.9% for ε4 allele), respectively (Al-Shammri et al., 2005; Ballerini et al., 2000; Carlin et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004). However, despite the intensive research efforts, the
Y.-W. Yin et al. / Gene 511 (2012) 12–17
13
results are inconclusive. In addition, the sample size in each of published studies is relatively small, which limited the credibility of results. To better clarify the association between ApoE gene polymorphism and the risk of MS, we performed this meta-analysis by collecting and sorting the previously published studies.
to the NOS score). Publication bias was examined using both the Begg's funnel plot and the Egger's regression test (p b 0.05 was considered representative of statistically significant publication bias). All above statistical analyses were performed using Review Manager 5.1.2 and Stata 11.0.
2. Materials and methods
3. Results
2.1. Literature search
3.1. Study characteristics
This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) criteria (Moher et al., 2009). We systematically searched the electronic databases PubMed (up to May 26, 2012), Embase, and Web of Science for relevant English language articles. The following keywords were used for searching: (“apolipoprotein E” OR “ApoE”) AND (“polymorphism” OR “mutation” OR “variant” OR “genotype”) AND (“Multiple sclerosis” OR “MS”). The search was limited to human studies, but didn't include unpublished dissertations. Additionally, all studies were retrieved, and their bibliographies were checked and evaluated for other relevant publications that have not been identified as aforementioned.
The study selection process is detailed in Fig. 1. Our initial search identified a total of 568 potentially eligible articles. Of these, 540 papers were excluded after reading the title or abstract because of obvious irrelevance to the aim of our study. In addition, two publications were excluded for poor quality and six for insufficient data for calculation of an OR and 95% CI. Therefore, 20 studies were included in the final meta-analysis (Al-Shammri et al., 2005; Ballerini et al., 2000; Carlin et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004), which contained 4080 MS cases and 2897 controls. Table 1 shows the studies identified and their main characteristics. Eighteen studies were performed in Europeans, and two studies were performed in Asians. The countries of these studies included Denmark, Finland, France, Greece, Hungary, Italy, Japan, Kuwait, Netherlands, Poland, Russia and USA. Controls were population-based in 18 studies (Al-Shammri et al., 2005; Ballerini et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004), and hospital-based in two studies (Carlin et al., 2000; Pirttilä et al., 2000). Two studies didn't follow the HWE (Al-Shammri et al., 2005; Rubinsztein, 1994) and two studies have insufficient data for calculation of the HWE (Carlin et al., 2000; Pirttilä et al., 2000).
2.2. Inclusion criteria The inclusion criteria were as follows: (1) studies on the relationship between ApoE gene polymorphism and MS; (2) published case– control studies; (3) studies with full text articles; (4) sufficient data for estimating an odds ratio (OR) with 95% confidence interval (CI); and (5) not republished data. 2.3. Data extraction Two authors (Yin YW and Zhang LL) of this article independently extracted data from all eligible studies. Any disagreement was resolved by consensus and involved a third party (Li JC). For each study, information was extracted including name of the first author, publication date, country of origin, ethnicity of the studied population, source of controls, total numbers of cases and controls, and frequency of ApoE gene polymorphism in cases and controls, respectively.
3.2. Quantitative synthesis 2.4. Quality score assessment Two authors (Wang JZ and Li BH) of this article independently assessed the qualities of included studies using the Newcastle–Ottawa Scale (NOS) (Wells et al., 2011). The NOS ranges between zero (worst) up to nine stars (best). Studies with a score of seven stars or greater were considered to be of high quality. Disagreement was settled as described above. 2.5. Statistical analysis The strength of association between ApoE gene polymorphism and MS risk was estimated by calculating a pooled OR and 95% CI under seven genetic models (ε2/ε2 versus ε3/ε3, ε2/ε3 versus ε3/ε3, ε2/ε4 versus ε3/ε3, ε3/ε4 versus ε3/ε3, ε4/ε4 versus ε3/ε3, ε2 allele versus ε3 allele and ε4 allele versus ε3 allele), respectively. The ORs were pooled through a fixed effects model, using the Mantel– Haenszel approach when no heterogeneity was observed among studies. Otherwise, a random effects model was adopted. Heterogeneity was measured using Cochran's Q statistic and the I 2 statistic, and Galbraith plot was used to detect the potential sources of heterogeneity. To evaluate ethnicity-specific effects, subgroup analysis was performed based on the ethnicity of study population. Sensitivity analysis was carried out by limiting the meta-analysis to studies conforming to Hardy–Weinberg equilibrium (HWE, p b 0.05 of HWE was considered significant) and the high quality studies (according
The main results of meta-analysis were shown in Table 2. The overall results showed evidence of significant association between ApoE gene polymorphism and MS risk in the genetic model of ε2/ε4 versus ε3/ε3 (for OR= 1.74, 95% CI= 1.12–2.71, p = 0.01) and in the genetic model of ε2 allele versus ε3 allele (for OR= 1.16, 95% CI= 1.01–1.35, p = 0.04) (Figs. 2A, B). However, no obvious evidence for association was found in the genetic model of ε2/ε2 versus ε3/ε3 (for OR= 1.44, 95% CI= 0.72–2.87, p = 0.30), ε2/ε3 versus ε3/ε3 (for OR= 1.08, 95% CI = 0.91–1.28, p = 0.39), ε3/ε4 versus ε3/ε3 (for OR= 0.93, 95% CI= 0.77– 1.12, p = 0.45), ε4/ε4 versus ε3/ε3 (for OR= 1.18, 95% CI = 0.81–1.70, p = 0.39) and ε4 allele versus ε3 allele (for OR= 1.01, 95% CI= 0.87– 1.17, p = 0.94). In the subgroup analysis by ethnicity, we only analyzed the Europeans due to rare studies on Asians. Significant associations were found in two genetic models (for ε2/ε4 versus ε3/ε3: OR= 1.81, 95% CI = 1.14–2.87, p = 0.01; for ε2 allele versus ε3 allele: OR= 1.19, 95% CI= 1.02–1.38, p = 0.03) (Figs. 2C, D). 3.3. Sensitivity analysis Sensitivity analyses were conducted to determine whether modification of the inclusion criteria of the meta-analysis affected the final results. The included studies were limited to those conforming to HWE and those with high NOS score (≥7). Two studies without HWE (p b 0.05) (Al-Shammri et al., 2005; Rubinsztein, 1994), two studies with insufficient data for testing the HWE (Carlin et al.,
14
Y.-W. Yin et al. / Gene 511 (2012) 12–17
Fig. 1. Flow diagram of the study selection process.
2000; Pirttilä et al., 2000) and four studies with relatively low NOS score (b 7) (Carlin et al., 2000; Gaillard et al., 1998; Mehta et al., 2003; Rubinsztein, 1994) were excluded from the sensitivity analysis. The corresponding pooled ORs were not materially altered, indicating that our results were statistically robust. Also, HWE and NOS score should not be considered as a factor influencing the overall results. The results of sensitivity analysis were shown in Table 2. 3.4. Heterogeneity analysis Significant heterogeneity existed among the genetic models of ε3/ε4 versus ε3/ε3 (for PQ =0.06, I2 =36%) and ε4 allele versus ε3 allele (for
PQ =0.09, I2 =31%). In contrast, the other five genetic models didn't present significant heterogeneity (for ε2/ε2 versus ε3/ε3: PQ =0.89, I2 =0%; for ε2/ε3 versus ε3/ε3: PQ =0.67, I2 =0%; for ε2/ε4 versus ε3/ε3: PQ = 0.89, I2 =0%; for ε4/ε4 versus ε3/ε3: PQ =0.22, I2 =20%; for ε2 allele versus ε3 allele: PQ =0.89, I2 =0%). To clarify the source of heterogeneity, we firstly performed the subgroup analysis and sensitivity analysis. However, we failed to remove the heterogeneity (Table 2). We next created a Galbraith plot to graphically assess the source of heterogeneity. Two studies (Losonczi et al., 2010; Savettieri et al., 2003) were identified as the main contributor to heterogeneity on ε3/ε4 versus ε3/ε3 (Fig. S1A), and one study (Losonczi et al., 2010) was identified as the main contributor to heterogeneity on ε4 allele versus ε3 allele (Fig. S1B). After excluding
Table 1 Characteristics of studies included in this meta-analysis. First author
Rubinsztein Gaillard Ferri Weatherby Ballerini Høgh Carlin Pirttilä Savettieri Niino Mehta Zwemmer Zakrzewska-Pniewska Cocco Al-Shammri Ramsaransing Pinholt Koutsis Mustafina Losonczi
Year
1994 1998 1999 2000 2000 2000 2000 2000 2003 2003 2003 2004 2004 2005 2005 2005 2005 2007 2008 2010
Country
UK France Italy UK Italy Denmark UK Finland Italy Japan USA Netherlands Poland Italy Kuwait Netherlands Denmark Greece Russia Hungary
Ethnicity
Source Sample size (case/ of controls control)
European PB European PB European PB European PB European PB European PB European HB European HB European PB Asian PB European PB European PB European PB European PB Asian PB European PB European PB European PB European PB European PB
PB: population-based, HB: hospital-based. HWE: Hardy–Weinberg equilibrium, Y: yes, N: no.
36/91 129/100 161/153 370/159 66/67 240/361 71/41 105/41 428/107 135/134 15/31 408/144 117/100 871/348 39/106 82/29 385/361 212/216 120/263 90/45
Genotype distribution (case/control) ε2/ ε2
ε2/ε3
ε3/ε3
0/2 0/0 0/0 7/1 0/0 3/3
1/4 26/51 11/13 89/65 17/16 129/119 51/16 224/99 15/11 45/41 37/40 124/203
1/1 0/0 0/0 5/0 1/0 3/1 0/2 0/0 5/3 0/0 0/0 0/0
50/13 10/8 3/5 55/14 10/8 51/16 0/7 9/6 48/40 26/23 7/32 20/17
315/84 109/109 8/20 233/78 83/71 729/294 31/78 49/14 223/203 151/167 79/170 35/24
ε2/ ε4 1/1 2/2 1/0 7/3 0/0 7/5
ε3/ε4
ε4/ε4
8/29 0/4 25/16 2/4 14/17 0/1 74/38 7/2 6/15 0/0 54/100 15/10
ε2
ε3
ε4
HWE Y/N(P)
2/9 61/135 9/38 N(0.0013) 13/15 214/159 31/26 Y(0.1774) 18/16 289/271 15/19 N(0.5842) 72/21 573/252 95/45 Y(0.8586) 15/11 111/108 6/15 Y(0.2744) 50/51 339/546 91/125 Y(0.3996) 9/5 114/63 19/14 7/4 170/66 33/12 2/1 57/7 3/1 54/16 737/188 65/10 Y(0.3428) 1/2 15/13 0/2 11/10 243/239 16/19 Y(0.1015) 0/0 4/5 0/1 3/5 23/50 4/7 Y(0.5726) 7/2 93/45 15/5 72/16 614/215 130/57 Y(0.6811) 5/1 15/19 3/1 17/9 191/169 26/22 Y(0.9654) 5/1 83/33 0/3 62/19 1592/637 88/40 Y(0.1333) 1/1 5/18 2/0 1/12 67/181 10/19 N(0.0169) 2/0 19/9 3/0 11/6 126/43 27/9 Y(0.3170) 10/5 87/100 12/10 68/51 581/546 121/125 Y(0.3996) 0/0 35/24 0/2 26/23 363/381 35/28 Y(0.2830) 5/2 22/54 7/5 12/34 187/426 41/66 Y(0.3552) 6/0 29/4 0/0 26/17 119/69 35/4 Y(0.2439)
Score
6 6 7 9 9 7 6 7 8 8 6 8 7 7 8 8 7 9 8 8
0.07 0.08 1.03[0.87,1.21]a 1.04[0.89,1.22]a 0.70 0.64
3.5. Publication bias Visual inspection of the funnel plot (Fig. 3A for ε2/ε3 vs. ε3/ε3 and Fig. 3B for ε2 allele vs. ε3 allele) displays asymmetrical distribution of OR estimations, suggesting publication bias, especially for the ε2 allele. In addition, the results of Egger's regression test provided evidence for publication bias (p = 0.013 for ε2/ε3 vs. ε3/ε3 and p = 0.004 for ε2 allele vs. ε3 allele, respectively). 4. Discussion
1.17[0.80,1.71] 1.32[0.90,1.95] 0.05 0.05 0.96[0.78,1.17]a 0.93[0.76,1.14]a 0.79 0.87 BH: based on HWE (Studies without HWE were excluded.). BS: based on score (Studies with score ≤ 6 were excluded.). a Significant heterogeneity: the random-effects model was chosen to summarize the result. b p values for heterogeneity from Q-test.
1.72[1.09,2.71] 1.97[1.21,3.19] 0.70 0.64 1.10[0.92,1.31] 1.11[0.93,1.32] 1.71[0.80,3.68] 1.58[0.76,3.25] Sensitivity analysis BH 3668/2465 BS 3803/2634
0.87 0.89
1.44[0.72,2.87] 1.54[0.75,3.17] 4080/2897 3906/2657 Overall Europeans
15
the outlier studies, the heterogeneity was effectively removed (for ε3/ε4 versus ε3/ε3: PQ =0.56, I2 =0%; for ε4 allele versus ε3 allele: PQ =0.42, I2 =3%) (Figs. S2A, B).
0.29 0.24
1.21[1.04,1.40] 1.21[1.03,1.41]
0.09 0.07 1.01[0.87,1.17]a 1.00[0.86,1.18]a 0.64 0.68 0.06 0.04 1.74[1.12,2.71] 1.81[1.14,2.87] 0.67 0.66 1.08[0.91,1.28] 1.09[0.91,1.30]
OR(95% CI)
0.89 0.86
0.93[0.77,1.12]a 0.93[0.76,1.14]a
1.18[0.81,1.70] 1.16[0.79,1.69]
0.22 0.29
1.16[1.01,1.35] 1.19[1.02,1.38]
pb OR(95% CI) pb OR(95% CI) OR(95% CI) pb OR(95% CI)
0.89 0.86
ε2/ε3 vs. ε3/ε3
OR(95% CI) pb
ε2/ε2 vs. ε3/ε3
OR(95% CI)
Sample size (case/control) Category
Table 2 Meta-analyses of ApoE gene polymorphism and risk of MS in each subgroup.
pb
ε2/ε4 vs. ε3/ε3
pb
ε3/ε4 vs. ε3/ε3
ε4/ε4 vs. ε3/ε3
pb
ε2 allele vs. ε3 allele
ε4 allele vs. ε3 allele
Y.-W. Yin et al. / Gene 511 (2012) 12–17
The association between ApoE gene polymorphism and the risk of MS has been intensively studied. However, the results remain controversial. Some studies reported a positive association between ApoE gene polymorphism and the risk of MS progression (Al-Shammri et al., 2005; Carlin et al., 2000; Cocco et al., 2005; Gaillard et al., 1998; Høgh et al., 2000; Losonczi et al., 2010; Mustafina et al., 2008; Pinholt et al., 2005; Zwemmer et al., 2004), whereas other studies generated negative results (Ballerini et al., 2000; Ferri et al., 1999; Koutsis et al., 2007; Mehta et al., 2003; Niino et al., 2003; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004). Also, the credibility of results from a single case–control study is questionable due to too small sample size of the study populations. As suggested, to generate robust data, a much larger sample size in each group might be required. Meta-analysis has the benefit to overcome this limitation by increasing the sample size and may generate more precise results. To better explain the association between MS risk and the ApoE gene polymorphism, we conducted a comprehensive genetic meta-analysis. In this meta-analysis, we examined ApoE gene polymorphism and its relationship with the risk of MS on the basis of data from 6977 genotyped cases and controls. The results showed that the risk of developing MS in ε2 allele carriers was 1.16-fold higher than those without. Moreover, the risk of developing MS in individuals with ε2/ε4 genotype was 1.74-fold higher than individuals with ε3/ε3 genotype. The above results suggested that the ε2 allele of ApoE was an independent risk factor for the development of MS, although this notion is inconsistent with previous meta-analysis which developed a negative result regarding the association between ApoE gene polymorphism and MS risk (Burwick et al., 2006). Moreover, the previous meta-analysis investigating the association between ApoE gene polymorphism and AD risk showed that ApoE ε4 allele was associated with an increased risk of developing AD (Farrer et al., 1997). Despite MS and AD have the similar pathological properties, we found that ApoE ε2 allele was an independent risk factor for developing MS in the present meta-analysis. We therefore assumed that genetic factors may play different roles in the pathogeneses of MS and AD. In the subgroup analysis by ethnicity, significant association between ApoE ε2 mutation and MS risk was observed among Europeans. As for the Asians, we didn't perform subgroup analysis due to insufficient studies. Consider that there is potential ethnic difference in the distribution of genotypes and the allele frequencies, it is a defect in this study for the failure to analyze the association between ApoE gene polymorphism and MS risk among Asians. Further studies with large sample size, especially among Asians, will be needed to confirm our findings. Significant heterogeneity between studies was identified for the genetic models of ε2/ε3 versus ε3/ε3 and ε2 allele versus ε3 allele. Heterogeneity is a potential problem that may affect the interpretation of the results. Heterogeneity may be attribute to the potential confounding resulted from diversity in design, sample-sizes, methods of genotyping, difference of ethnicity, or the interaction with other
16
Y.-W. Yin et al. / Gene 511 (2012) 12–17
Fig. 2. A, B. Forest plots for ApoE gene polymorphism and MS risk in the overall study. A: ε2/ε4 versus ε3/ε3; B: ε2 allele versus ε3 allele. C, D. Forest plots for ApoE gene polymorphism and MS risk in the subgroup analysis of Europeans.
risk factors. By using Galbraith plot, two studies were identified as the main contributor to heterogeneity. However, the pooled OR was materially altered on ε3/ε4 versus ε3/ε3 when the outlier study was excluded (Fig. S2A), indicating that individuals with ε3/ε4 genotype
have more obvious effect on an individual's phenotype than those with ε3/ε3 genotype, and may have lower risk of MS. Publication bias is one of the most important sources of bias in meta-analysis, which can cause the false positive results. Asymmetrical
Fig. 3. Funnel plots for ApoE gene polymorphism and MS risk. A: ε2/ε3 versus ε3/ε3; B: ε2 allele versus ε3 allele.
Y.-W. Yin et al. / Gene 511 (2012) 12–17
“missing” data, which was in the upper part of the funnel plot (Figs. 3A, B) of the genetic models of ε2/ε3 versus ε3/ε3 and ε2 allele versus ε3 allele, suggests the possibility of missed or unpublished studies. So, projections from the literature of who is at risk for ApoE gene attributable to MS and who would benefit from ApoE gene-targeted therapies should therefore be approached with caution. For better interpreting the results, some limitations of this metaanalysis should be acknowledged. Firstly, this meta-analysis was based predominantly on European research. Only two studies involving Asians and no studies from other parts of the world were found, which may develop a partial result. Indeed, MS occurs worldwide and affects more than 2.5 million people. Secondly, we did not perform subgroup analyses based on the subtype of MS (relapsing remitting MS, primary progressive MS and secondary progressive MS) due to only five studies clearly described the subtype of MS (Høgh et al., 2000; Losonczi et al., 2010; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005), the sample sizes of which are really small (476 relapsing remitting MS cases, 164 primary progressive MS cases and 138 secondary progressive MS cases) and underpowered, being unable to provide a definite answer even in the case where a true association exists. Thirdly, publication bias may bias the present results. This meta-analysis only focused on papers published in English language and studies with full text articles, missing some eligible studies which were unpublished or reported in other languages. Thus, some inevitable publication bias may exist in the results. Furthermore, MS is a complex disease, involving potential interactions among gene–gene and gene–environment. However, many eligible studies included in this meta-analysis didn't consider the environmental factors. In conclusion, our meta-analysis of 20 studies suggests that ApoE ε2 mutation is associated with increased MS risk. In addition, ApoE ε3/ε4 genotype appears to be a protective factor for MS. However, the result should be interpreted with caution because of its limitations. Further studies with large sample size, especially with the consideration of gene–gene and gene–environment interactions, will be needed to confirm our findings. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.gene.2012.09.010. Acknowledgments This study was supported by Chongqing Natural Science Foundation (CSTC2011BB5031 to Lili Zhang). References Al-Shammri, S., Fatania, H., Al-Radwan, R., Akanji, A.O., 2005. The relationship of APOE genetic polymorphism with susceptibility to multiple sclerosis and its clinical phenotypes in Kuwaiti Arab subjects. Clin. Chim. Acta 351, 203–207. Ballerini, C., et al., 2000. Association of apolipoprotein E polymorphism to clinical heterogeneity of multiple sclerosis. Neurosci. Lett. 296, 174–176. Brück, W., 2005. The pathology of multiple sclerosis is the result of focal inflammatory demyelination with axonal damage. J. Neurol. 252 (Suppl. 5), v3–v9. Bu, G., 2009. Apolipoprotein E and its receptors in Alzheimer's disease: pathways, pathogenesis and therapy. Nat. Rev. Neurosci. 10, 333–344. Burwick, R.M., et al., 2006. APOE epsilon variation in multiple sclerosis susceptibility and disease severity: some answers. Neurology 66, 1373–1383. Carlin, C., Murray, L., Graham, D., Doyle, D., Nicoll, J., 2000. Involvement of apolipoprotein E in multiple sclerosis: absence of remyelination associated with possession of the APOE epsilon2 allele. J. Neuropathol. Exp. Neurol. 59, 361–367. Cocco, E., et al., 2005. HLA-DR, DQ and APOE genotypes and gender influence in Sardinian primary progressive MS. Neurology 64, 564–566.
17
Corder, E.H., et al., 1993. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 261, 921–923. Ebers, G.C., Sadovnick, A.D., Risch, N.J., 1995. A genetic basis for familial aggregation in multiple sclerosis. Canadian Collaborative Study Group. Nature 377, 150–151. Farrer, L.A., et al., 1997. 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 278, 1349–1356. Ferri, C., et al., 1999. APOE epsilon2-4 and ‐491 polymorphisms are not associated with MS. Neurology 53, 888–889. Gaillard, O., et al., 1998. Apolipoprotein E and multiple sclerosis: a biochemical and genetic investigation. J. Neurol. Sci. 158, 180–186. Høgh, P., et al., 2000. Apoliprotein E and multiple sclerosis: impact of the epsilon-4 allele on susceptibility, clinical type and progression rate. Mult. Scler. 6, 226–230. Huang, Y., 2006. Apolipoprotein E and Alzheimer disease. Neurology 66 (2 Suppl. 1), 79–85. Ignatius, M.J., et al., 1986. Expression of apolipoprotein E during nerve degeneration and regeneration. Proc. Natl. Acad. Sci. U. S. A. 83, 1125–1129. Koch-Henriksen, N., Hyllested, K., 1998. Epidemiology of multiple sclerosis: incidence and prevalence rates in Denmark 1948–64 based on the Danish Multiple Sclerosis Registry. Acta Neurol. Scand. 78, 369–380. Koutsis, G., et al., 2007. APOE genotypes in Greek multiple sclerosis patients: no effect on the MS Severity Score. J. Neurol. 254, 394–395. Lahiri, D.K., Sambamurti, K., Bennett, D.A., 2004. Apolipoprotein gene and its interaction with the environmentally driven risk factors: molecular, genetic and epidemiological studies of Alzheimer's disease. Neurobiol. Aging 25, 651–660. Losonczi, E., et al., 2010. APOE epsilon status in Hungarian patients with primary progressive multiple sclerosis. Swiss Med. Wkly. 140, w13119. Mahley, R.W., 1988. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science 240, 622–630. Mehta, P.D., Patrick, B.A., Pirttila, T., Coyle, P.K., Aisen, P.S., 2003. Detection of apolipoprotein E phenotype in unconcentrated cerebrospinal fluid. J. Clin. Lab. Anal. 17, 18–21. Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., PRISMA Group, 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann. Intern. Med. 151, 264–269. Mustafina, O.E., et al., 2008. Analysis of the association of allelic variants of apolypoprotein E and interleukin 1 beta genes with multiple sclerosis in ethnic Tatars. Genetika 44, 407–413. Niino, M., Kikuchi, S., Fukazawa, T., Yabe, I., Tashiro, K., 2003. Polymorphisms of apolipoprotein E and Japanese patients with multiple sclerosis. Mult. Scler. 9, 382–386. Noonan, C.W., Kathman, S.J., White, M.C., 2002. Prevalence estimates for MS in the United States and evidence of an increasing trend for women. Neurology 58, 136–138. Olesen, J., et al., 2012. The economic cost of brain disorders in Europe. Eur. J. Neurol. 19, 155–162. Pinholt, M., Frederiksen, J.L., Andersen, P.S., Christiansen, M., 2005. Apo E in multiple sclerosis and optic neuritis: the apo E-epsilon4 allele is associated with progression of multiple sclerosis. Mult. Scler. 11, 511–515. Pirttilä, T., Haanpää, M., Mehta, P.D., Lehtimäki, T., 2000. Apolipoprotein E (APOE) phenotype and APOE concentrations in multiple sclerosis and acute herpes zoster. Acta Neurol. Scand. 102, 94–98. Ramsaransing, G.S., Heersema, D.J., De Keyser, J., 2005. Serum uric acid, dehydroepiandrosterone sulphate, and apolipoprotein E genotype in benign vs. progressive multiple sclerosis. Eur. J. Neurol. 12, 514–518. Rubinsztein, D.C., 1994. Apo E genotypes in multiple sclerosis, Parkinson's disease, schwannomas and late-onset Alzheimer's disease. Mol. Cell. Probes 8, 519–525. Savettieri, G., et al., 2003. Apolipoprotein E genotype does not influence the progression of multiple sclerosis. J. Neurol. 250, 1094–1098. Strittmatter, W.J., et al., 1993. Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc. Natl. Acad. Sci. U. S. A. 90, 1977–1981. Varadé, J., et al., 2011. Replication study of 10 genes showing evidence for association with multiple sclerosis: validation of TMEM39A, IL12B and CLBL genes. Mult. Scler. 18, 959–965. Weatherby, S.J., et al., 2000. Polymorphisms of apolipoprotein E; outcome and susceptibility in multiple sclerosis. Mult. Scler. 6, 32–36. Wells, G.A., Shea, B., O'Connell, D., 2011. The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Health Research Institute. www.ohri.ca/programs/clinical_epidemiology/oxford.asp (accessed Oct 20, 2011). Zakrzewska-Pniewska, B., et al., 2004. Association of apolipoprotein E and myeloperoxidase genotypes to clinical course of familial and sporadic multiple sclerosis. Mult. Scler. 10, 266–271. Zwemmer, J.N., et al., 2004. No major association of ApoE genotype with disease characteristics and MRI findings in multiple sclerosis. Mult. Scler. 10, 272–277.
Contents lists available at SciVerse ScienceDirect
Gene journal homepage: www.elsevier.com/locate/gene
Association between apolipoprotein E gene polymorphism and the risk of multiple sclerosis: A meta-analysis of 6977 subjects Yan-Wei Yin a, 1, Yun-Dong Zhang b, 1, Jing-Zhou Wang a, Bing-Hu Li a, Qing-Wu Yang a, Chuan-Qin Fang a, Chang-Yue Gao a, Jing-Cheng Li a,⁎, Li-Li Zhang a,⁎ a b
Department of Neurology, Institute of Surgery Research, Daping Hospital, Third Military Medical University, 10 Changjiang Branch Road, Yuzhong District, Chongqing 400042, PR China Department of Neurosurgery, Institute of Surgery Research, Daping Hospital, Third Military Medical University, 10 Changjiang Branch Road, Yuzhong District, Chongqing 400042, PR China
a r t i c l e
i n f o
Article history: Accepted 6 September 2012 Available online 13 September 2012 Keywords: Multiple sclerosis Apolipoprotein E Polymorphism Meta-analysis
a b s t r a c t Epidemiological studies have evaluated the association between apolipoprotein E (ApoE) gene polymorphism and multiple sclerosis (MS) risk. However, the results remain conflicting. Therefore, in order to derive a more precise association of ApoE gene polymorphism with MS risk, we performed this meta-analysis. Systematic searches of electronic databases PubMed, Embase and Web of Science, as well as hand searching of the references of identified articles were performed. Twenty studies were identified, covering a total of 4080 MS cases and 2897 controls. The results showed evidence for significant association between ApoE ε2 mutation and MS risk (for ε2/ε4 versus ε3/ε3: OR = 1.74, 95% CI = 1.12–2.71, p = 0.01; for ε2 allele versus ε3 allele: OR = 1.16, 95% CI = 1.01–1.35, p = 0.04). In the subgroup analysis by ethnicity, the similar results were obtained among Europeans (for ε2/ε4 versus ε3/ε3: OR = 1.81, 95% CI = 1.14–2.87, p = 0.01; for ε2 allele versus ε3 allele: OR = 1.19, 95% CI = 1.02–1.38, p = 0.03). After excluding the outlier studies by observing Galbraith plot, marginal association was found between ApoE ε3/ε4 genotype and the protective factor for MS (for ε3/ε4 versus ε3/ε3: OR = 0.86, 95% CI = 0.75–0.99, p = 0.04). In summary, the present metaanalysis provides evidence that ApoE ε2 mutation is associated with MS risk. In addition, ApoE ε3/ε4 genotype appears to be a protective factor for MS. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Multiple sclerosis (MS) is a chronic neurological autoimmune disease, which seriously affects the quality of life for patients and national healthcare systems (Olesen et al., 2012). Despite its results from inflammation, demyelination, gliosis and varying degrees of axonal pathology leading to progressive neurological dysfunction (Brück, 2005), a detailed etiology underlying MS is still obscure. Previous studies have reported that the prevalence rates of MS vary substantially throughout the world (Koch-Henriksen and Hyllested, 1998; Noonan et al., 2002; Varadé et al., 2011), and relatives of MS patients are at greater risk for developing the disease than the general population (Ebers et al., 1995). It is assumed that genetic factors may be associated with in the development of MS. Apolipoprotein E (ApoE) plays important role in the lipid and cholesterol transport, and also is involved in brain development and repair (Ignatius et al., 1986; Mahley, 1988). The ApoE gene, located in Abbreviations: ApoE, apolipoprotein E; MS, multiple sclerosis; AD, Alzheimer's disease; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses; OR, odds ratio; CI, confidence interval; NOS, Newcastle–Ottawa Scale; HWE, Hardy–Weinberg equilibrium. ⁎ Corresponding authors. Tel.: +86 23 68757842; fax: +86 23 68757841. E-mail address: [email protected] (L.-L. Zhang). 1 These authors contributed equally to this work. 0378-1119/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.gene.2012.09.010
the 19q13 region, is polymorphic with three common alleles: ε2, ε3 and ε4. Consequently, three homozygous (ε2/ε2, ε3/ε3, and ε4/ε4) and three heterozygous (ε2/ε3, ε2/ε4, and ε3/ε4) phenotypes are found in the general population (Lahiri et al., 2004). ApoE ε3 allele is the most (77%) and ε2 allele is the least (8%) common allele. Furthermore, the ApoE ε4 allele frequency is approximately 15% in the general population (Bu, 2009). Since 1993, the association between ApoE ε4 allele and earlier age of onset of both familial and sporadic Alzheimer's disease (AD) has consistently been demonstrated (Corder et al., 1993; Huang, 2006; Strittmatter et al., 1993). MS has the similar pathological properties to AD with inflammation and gliosis. Recently, numerous epidemiological studies have focused on the association between ApoE gene polymorphism and MS risk, and showed the ApoE allele frequencies in MS group (1.3 to 14.4% for ε2 allele, 66.1 to 91.4% for ε3 allele, and 4.5 to 19.4% for ε4 allele) and control group (2.7 to 18.9% for ε2 allele, 74.1 to 91.5% for ε3 allele, and 4.4 to 20.9% for ε4 allele), respectively (Al-Shammri et al., 2005; Ballerini et al., 2000; Carlin et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004). However, despite the intensive research efforts, the
Y.-W. Yin et al. / Gene 511 (2012) 12–17
13
results are inconclusive. In addition, the sample size in each of published studies is relatively small, which limited the credibility of results. To better clarify the association between ApoE gene polymorphism and the risk of MS, we performed this meta-analysis by collecting and sorting the previously published studies.
to the NOS score). Publication bias was examined using both the Begg's funnel plot and the Egger's regression test (p b 0.05 was considered representative of statistically significant publication bias). All above statistical analyses were performed using Review Manager 5.1.2 and Stata 11.0.
2. Materials and methods
3. Results
2.1. Literature search
3.1. Study characteristics
This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) criteria (Moher et al., 2009). We systematically searched the electronic databases PubMed (up to May 26, 2012), Embase, and Web of Science for relevant English language articles. The following keywords were used for searching: (“apolipoprotein E” OR “ApoE”) AND (“polymorphism” OR “mutation” OR “variant” OR “genotype”) AND (“Multiple sclerosis” OR “MS”). The search was limited to human studies, but didn't include unpublished dissertations. Additionally, all studies were retrieved, and their bibliographies were checked and evaluated for other relevant publications that have not been identified as aforementioned.
The study selection process is detailed in Fig. 1. Our initial search identified a total of 568 potentially eligible articles. Of these, 540 papers were excluded after reading the title or abstract because of obvious irrelevance to the aim of our study. In addition, two publications were excluded for poor quality and six for insufficient data for calculation of an OR and 95% CI. Therefore, 20 studies were included in the final meta-analysis (Al-Shammri et al., 2005; Ballerini et al., 2000; Carlin et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004), which contained 4080 MS cases and 2897 controls. Table 1 shows the studies identified and their main characteristics. Eighteen studies were performed in Europeans, and two studies were performed in Asians. The countries of these studies included Denmark, Finland, France, Greece, Hungary, Italy, Japan, Kuwait, Netherlands, Poland, Russia and USA. Controls were population-based in 18 studies (Al-Shammri et al., 2005; Ballerini et al., 2000; Cocco et al., 2005; Ferri et al., 1999; Gaillard et al., 1998; Høgh et al., 2000; Koutsis et al., 2007; Losonczi et al., 2010; Mehta et al., 2003; Mustafina et al., 2008; Niino et al., 2003; Pinholt et al., 2005; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004; Zwemmer et al., 2004), and hospital-based in two studies (Carlin et al., 2000; Pirttilä et al., 2000). Two studies didn't follow the HWE (Al-Shammri et al., 2005; Rubinsztein, 1994) and two studies have insufficient data for calculation of the HWE (Carlin et al., 2000; Pirttilä et al., 2000).
2.2. Inclusion criteria The inclusion criteria were as follows: (1) studies on the relationship between ApoE gene polymorphism and MS; (2) published case– control studies; (3) studies with full text articles; (4) sufficient data for estimating an odds ratio (OR) with 95% confidence interval (CI); and (5) not republished data. 2.3. Data extraction Two authors (Yin YW and Zhang LL) of this article independently extracted data from all eligible studies. Any disagreement was resolved by consensus and involved a third party (Li JC). For each study, information was extracted including name of the first author, publication date, country of origin, ethnicity of the studied population, source of controls, total numbers of cases and controls, and frequency of ApoE gene polymorphism in cases and controls, respectively.
3.2. Quantitative synthesis 2.4. Quality score assessment Two authors (Wang JZ and Li BH) of this article independently assessed the qualities of included studies using the Newcastle–Ottawa Scale (NOS) (Wells et al., 2011). The NOS ranges between zero (worst) up to nine stars (best). Studies with a score of seven stars or greater were considered to be of high quality. Disagreement was settled as described above. 2.5. Statistical analysis The strength of association between ApoE gene polymorphism and MS risk was estimated by calculating a pooled OR and 95% CI under seven genetic models (ε2/ε2 versus ε3/ε3, ε2/ε3 versus ε3/ε3, ε2/ε4 versus ε3/ε3, ε3/ε4 versus ε3/ε3, ε4/ε4 versus ε3/ε3, ε2 allele versus ε3 allele and ε4 allele versus ε3 allele), respectively. The ORs were pooled through a fixed effects model, using the Mantel– Haenszel approach when no heterogeneity was observed among studies. Otherwise, a random effects model was adopted. Heterogeneity was measured using Cochran's Q statistic and the I 2 statistic, and Galbraith plot was used to detect the potential sources of heterogeneity. To evaluate ethnicity-specific effects, subgroup analysis was performed based on the ethnicity of study population. Sensitivity analysis was carried out by limiting the meta-analysis to studies conforming to Hardy–Weinberg equilibrium (HWE, p b 0.05 of HWE was considered significant) and the high quality studies (according
The main results of meta-analysis were shown in Table 2. The overall results showed evidence of significant association between ApoE gene polymorphism and MS risk in the genetic model of ε2/ε4 versus ε3/ε3 (for OR= 1.74, 95% CI= 1.12–2.71, p = 0.01) and in the genetic model of ε2 allele versus ε3 allele (for OR= 1.16, 95% CI= 1.01–1.35, p = 0.04) (Figs. 2A, B). However, no obvious evidence for association was found in the genetic model of ε2/ε2 versus ε3/ε3 (for OR= 1.44, 95% CI= 0.72–2.87, p = 0.30), ε2/ε3 versus ε3/ε3 (for OR= 1.08, 95% CI = 0.91–1.28, p = 0.39), ε3/ε4 versus ε3/ε3 (for OR= 0.93, 95% CI= 0.77– 1.12, p = 0.45), ε4/ε4 versus ε3/ε3 (for OR= 1.18, 95% CI = 0.81–1.70, p = 0.39) and ε4 allele versus ε3 allele (for OR= 1.01, 95% CI= 0.87– 1.17, p = 0.94). In the subgroup analysis by ethnicity, we only analyzed the Europeans due to rare studies on Asians. Significant associations were found in two genetic models (for ε2/ε4 versus ε3/ε3: OR= 1.81, 95% CI = 1.14–2.87, p = 0.01; for ε2 allele versus ε3 allele: OR= 1.19, 95% CI= 1.02–1.38, p = 0.03) (Figs. 2C, D). 3.3. Sensitivity analysis Sensitivity analyses were conducted to determine whether modification of the inclusion criteria of the meta-analysis affected the final results. The included studies were limited to those conforming to HWE and those with high NOS score (≥7). Two studies without HWE (p b 0.05) (Al-Shammri et al., 2005; Rubinsztein, 1994), two studies with insufficient data for testing the HWE (Carlin et al.,
14
Y.-W. Yin et al. / Gene 511 (2012) 12–17
Fig. 1. Flow diagram of the study selection process.
2000; Pirttilä et al., 2000) and four studies with relatively low NOS score (b 7) (Carlin et al., 2000; Gaillard et al., 1998; Mehta et al., 2003; Rubinsztein, 1994) were excluded from the sensitivity analysis. The corresponding pooled ORs were not materially altered, indicating that our results were statistically robust. Also, HWE and NOS score should not be considered as a factor influencing the overall results. The results of sensitivity analysis were shown in Table 2. 3.4. Heterogeneity analysis Significant heterogeneity existed among the genetic models of ε3/ε4 versus ε3/ε3 (for PQ =0.06, I2 =36%) and ε4 allele versus ε3 allele (for
PQ =0.09, I2 =31%). In contrast, the other five genetic models didn't present significant heterogeneity (for ε2/ε2 versus ε3/ε3: PQ =0.89, I2 =0%; for ε2/ε3 versus ε3/ε3: PQ =0.67, I2 =0%; for ε2/ε4 versus ε3/ε3: PQ = 0.89, I2 =0%; for ε4/ε4 versus ε3/ε3: PQ =0.22, I2 =20%; for ε2 allele versus ε3 allele: PQ =0.89, I2 =0%). To clarify the source of heterogeneity, we firstly performed the subgroup analysis and sensitivity analysis. However, we failed to remove the heterogeneity (Table 2). We next created a Galbraith plot to graphically assess the source of heterogeneity. Two studies (Losonczi et al., 2010; Savettieri et al., 2003) were identified as the main contributor to heterogeneity on ε3/ε4 versus ε3/ε3 (Fig. S1A), and one study (Losonczi et al., 2010) was identified as the main contributor to heterogeneity on ε4 allele versus ε3 allele (Fig. S1B). After excluding
Table 1 Characteristics of studies included in this meta-analysis. First author
Rubinsztein Gaillard Ferri Weatherby Ballerini Høgh Carlin Pirttilä Savettieri Niino Mehta Zwemmer Zakrzewska-Pniewska Cocco Al-Shammri Ramsaransing Pinholt Koutsis Mustafina Losonczi
Year
1994 1998 1999 2000 2000 2000 2000 2000 2003 2003 2003 2004 2004 2005 2005 2005 2005 2007 2008 2010
Country
UK France Italy UK Italy Denmark UK Finland Italy Japan USA Netherlands Poland Italy Kuwait Netherlands Denmark Greece Russia Hungary
Ethnicity
Source Sample size (case/ of controls control)
European PB European PB European PB European PB European PB European PB European HB European HB European PB Asian PB European PB European PB European PB European PB Asian PB European PB European PB European PB European PB European PB
PB: population-based, HB: hospital-based. HWE: Hardy–Weinberg equilibrium, Y: yes, N: no.
36/91 129/100 161/153 370/159 66/67 240/361 71/41 105/41 428/107 135/134 15/31 408/144 117/100 871/348 39/106 82/29 385/361 212/216 120/263 90/45
Genotype distribution (case/control) ε2/ ε2
ε2/ε3
ε3/ε3
0/2 0/0 0/0 7/1 0/0 3/3
1/4 26/51 11/13 89/65 17/16 129/119 51/16 224/99 15/11 45/41 37/40 124/203
1/1 0/0 0/0 5/0 1/0 3/1 0/2 0/0 5/3 0/0 0/0 0/0
50/13 10/8 3/5 55/14 10/8 51/16 0/7 9/6 48/40 26/23 7/32 20/17
315/84 109/109 8/20 233/78 83/71 729/294 31/78 49/14 223/203 151/167 79/170 35/24
ε2/ ε4 1/1 2/2 1/0 7/3 0/0 7/5
ε3/ε4
ε4/ε4
8/29 0/4 25/16 2/4 14/17 0/1 74/38 7/2 6/15 0/0 54/100 15/10
ε2
ε3
ε4
HWE Y/N(P)
2/9 61/135 9/38 N(0.0013) 13/15 214/159 31/26 Y(0.1774) 18/16 289/271 15/19 N(0.5842) 72/21 573/252 95/45 Y(0.8586) 15/11 111/108 6/15 Y(0.2744) 50/51 339/546 91/125 Y(0.3996) 9/5 114/63 19/14 7/4 170/66 33/12 2/1 57/7 3/1 54/16 737/188 65/10 Y(0.3428) 1/2 15/13 0/2 11/10 243/239 16/19 Y(0.1015) 0/0 4/5 0/1 3/5 23/50 4/7 Y(0.5726) 7/2 93/45 15/5 72/16 614/215 130/57 Y(0.6811) 5/1 15/19 3/1 17/9 191/169 26/22 Y(0.9654) 5/1 83/33 0/3 62/19 1592/637 88/40 Y(0.1333) 1/1 5/18 2/0 1/12 67/181 10/19 N(0.0169) 2/0 19/9 3/0 11/6 126/43 27/9 Y(0.3170) 10/5 87/100 12/10 68/51 581/546 121/125 Y(0.3996) 0/0 35/24 0/2 26/23 363/381 35/28 Y(0.2830) 5/2 22/54 7/5 12/34 187/426 41/66 Y(0.3552) 6/0 29/4 0/0 26/17 119/69 35/4 Y(0.2439)
Score
6 6 7 9 9 7 6 7 8 8 6 8 7 7 8 8 7 9 8 8
0.07 0.08 1.03[0.87,1.21]a 1.04[0.89,1.22]a 0.70 0.64
3.5. Publication bias Visual inspection of the funnel plot (Fig. 3A for ε2/ε3 vs. ε3/ε3 and Fig. 3B for ε2 allele vs. ε3 allele) displays asymmetrical distribution of OR estimations, suggesting publication bias, especially for the ε2 allele. In addition, the results of Egger's regression test provided evidence for publication bias (p = 0.013 for ε2/ε3 vs. ε3/ε3 and p = 0.004 for ε2 allele vs. ε3 allele, respectively). 4. Discussion
1.17[0.80,1.71] 1.32[0.90,1.95] 0.05 0.05 0.96[0.78,1.17]a 0.93[0.76,1.14]a 0.79 0.87 BH: based on HWE (Studies without HWE were excluded.). BS: based on score (Studies with score ≤ 6 were excluded.). a Significant heterogeneity: the random-effects model was chosen to summarize the result. b p values for heterogeneity from Q-test.
1.72[1.09,2.71] 1.97[1.21,3.19] 0.70 0.64 1.10[0.92,1.31] 1.11[0.93,1.32] 1.71[0.80,3.68] 1.58[0.76,3.25] Sensitivity analysis BH 3668/2465 BS 3803/2634
0.87 0.89
1.44[0.72,2.87] 1.54[0.75,3.17] 4080/2897 3906/2657 Overall Europeans
15
the outlier studies, the heterogeneity was effectively removed (for ε3/ε4 versus ε3/ε3: PQ =0.56, I2 =0%; for ε4 allele versus ε3 allele: PQ =0.42, I2 =3%) (Figs. S2A, B).
0.29 0.24
1.21[1.04,1.40] 1.21[1.03,1.41]
0.09 0.07 1.01[0.87,1.17]a 1.00[0.86,1.18]a 0.64 0.68 0.06 0.04 1.74[1.12,2.71] 1.81[1.14,2.87] 0.67 0.66 1.08[0.91,1.28] 1.09[0.91,1.30]
OR(95% CI)
0.89 0.86
0.93[0.77,1.12]a 0.93[0.76,1.14]a
1.18[0.81,1.70] 1.16[0.79,1.69]
0.22 0.29
1.16[1.01,1.35] 1.19[1.02,1.38]
pb OR(95% CI) pb OR(95% CI) OR(95% CI) pb OR(95% CI)
0.89 0.86
ε2/ε3 vs. ε3/ε3
OR(95% CI) pb
ε2/ε2 vs. ε3/ε3
OR(95% CI)
Sample size (case/control) Category
Table 2 Meta-analyses of ApoE gene polymorphism and risk of MS in each subgroup.
pb
ε2/ε4 vs. ε3/ε3
pb
ε3/ε4 vs. ε3/ε3
ε4/ε4 vs. ε3/ε3
pb
ε2 allele vs. ε3 allele
ε4 allele vs. ε3 allele
Y.-W. Yin et al. / Gene 511 (2012) 12–17
The association between ApoE gene polymorphism and the risk of MS has been intensively studied. However, the results remain controversial. Some studies reported a positive association between ApoE gene polymorphism and the risk of MS progression (Al-Shammri et al., 2005; Carlin et al., 2000; Cocco et al., 2005; Gaillard et al., 1998; Høgh et al., 2000; Losonczi et al., 2010; Mustafina et al., 2008; Pinholt et al., 2005; Zwemmer et al., 2004), whereas other studies generated negative results (Ballerini et al., 2000; Ferri et al., 1999; Koutsis et al., 2007; Mehta et al., 2003; Niino et al., 2003; Pirttilä et al., 2000; Ramsaransing et al., 2005; Rubinsztein, 1994; Savettieri et al., 2003; Weatherby et al., 2000; Zakrzewska-Pniewska et al., 2004). Also, the credibility of results from a single case–control study is questionable due to too small sample size of the study populations. As suggested, to generate robust data, a much larger sample size in each group might be required. Meta-analysis has the benefit to overcome this limitation by increasing the sample size and may generate more precise results. To better explain the association between MS risk and the ApoE gene polymorphism, we conducted a comprehensive genetic meta-analysis. In this meta-analysis, we examined ApoE gene polymorphism and its relationship with the risk of MS on the basis of data from 6977 genotyped cases and controls. The results showed that the risk of developing MS in ε2 allele carriers was 1.16-fold higher than those without. Moreover, the risk of developing MS in individuals with ε2/ε4 genotype was 1.74-fold higher than individuals with ε3/ε3 genotype. The above results suggested that the ε2 allele of ApoE was an independent risk factor for the development of MS, although this notion is inconsistent with previous meta-analysis which developed a negative result regarding the association between ApoE gene polymorphism and MS risk (Burwick et al., 2006). Moreover, the previous meta-analysis investigating the association between ApoE gene polymorphism and AD risk showed that ApoE ε4 allele was associated with an increased risk of developing AD (Farrer et al., 1997). Despite MS and AD have the similar pathological properties, we found that ApoE ε2 allele was an independent risk factor for developing MS in the present meta-analysis. We therefore assumed that genetic factors may play different roles in the pathogeneses of MS and AD. In the subgroup analysis by ethnicity, significant association between ApoE ε2 mutation and MS risk was observed among Europeans. As for the Asians, we didn't perform subgroup analysis due to insufficient studies. Consider that there is potential ethnic difference in the distribution of genotypes and the allele frequencies, it is a defect in this study for the failure to analyze the association between ApoE gene polymorphism and MS risk among Asians. Further studies with large sample size, especially among Asians, will be needed to confirm our findings. Significant heterogeneity between studies was identified for the genetic models of ε2/ε3 versus ε3/ε3 and ε2 allele versus ε3 allele. Heterogeneity is a potential problem that may affect the interpretation of the results. Heterogeneity may be attribute to the potential confounding resulted from diversity in design, sample-sizes, methods of genotyping, difference of ethnicity, or the interaction with other
16
Y.-W. Yin et al. / Gene 511 (2012) 12–17
Fig. 2. A, B. Forest plots for ApoE gene polymorphism and MS risk in the overall study. A: ε2/ε4 versus ε3/ε3; B: ε2 allele versus ε3 allele. C, D. Forest plots for ApoE gene polymorphism and MS risk in the subgroup analysis of Europeans.
risk factors. By using Galbraith plot, two studies were identified as the main contributor to heterogeneity. However, the pooled OR was materially altered on ε3/ε4 versus ε3/ε3 when the outlier study was excluded (Fig. S2A), indicating that individuals with ε3/ε4 genotype
have more obvious effect on an individual's phenotype than those with ε3/ε3 genotype, and may have lower risk of MS. Publication bias is one of the most important sources of bias in meta-analysis, which can cause the false positive results. Asymmetrical
Fig. 3. Funnel plots for ApoE gene polymorphism and MS risk. A: ε2/ε3 versus ε3/ε3; B: ε2 allele versus ε3 allele.
Y.-W. Yin et al. / Gene 511 (2012) 12–17
“missing” data, which was in the upper part of the funnel plot (Figs. 3A, B) of the genetic models of ε2/ε3 versus ε3/ε3 and ε2 allele versus ε3 allele, suggests the possibility of missed or unpublished studies. So, projections from the literature of who is at risk for ApoE gene attributable to MS and who would benefit from ApoE gene-targeted therapies should therefore be approached with caution. For better interpreting the results, some limitations of this metaanalysis should be acknowledged. Firstly, this meta-analysis was based predominantly on European research. Only two studies involving Asians and no studies from other parts of the world were found, which may develop a partial result. Indeed, MS occurs worldwide and affects more than 2.5 million people. Secondly, we did not perform subgroup analyses based on the subtype of MS (relapsing remitting MS, primary progressive MS and secondary progressive MS) due to only five studies clearly described the subtype of MS (Høgh et al., 2000; Losonczi et al., 2010; Pinholt et al., 2005; Pirttilä et al., 2000; Ramsaransing et al., 2005), the sample sizes of which are really small (476 relapsing remitting MS cases, 164 primary progressive MS cases and 138 secondary progressive MS cases) and underpowered, being unable to provide a definite answer even in the case where a true association exists. Thirdly, publication bias may bias the present results. This meta-analysis only focused on papers published in English language and studies with full text articles, missing some eligible studies which were unpublished or reported in other languages. Thus, some inevitable publication bias may exist in the results. Furthermore, MS is a complex disease, involving potential interactions among gene–gene and gene–environment. However, many eligible studies included in this meta-analysis didn't consider the environmental factors. In conclusion, our meta-analysis of 20 studies suggests that ApoE ε2 mutation is associated with increased MS risk. In addition, ApoE ε3/ε4 genotype appears to be a protective factor for MS. However, the result should be interpreted with caution because of its limitations. Further studies with large sample size, especially with the consideration of gene–gene and gene–environment interactions, will be needed to confirm our findings. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.gene.2012.09.010. Acknowledgments This study was supported by Chongqing Natural Science Foundation (CSTC2011BB5031 to Lili Zhang). References Al-Shammri, S., Fatania, H., Al-Radwan, R., Akanji, A.O., 2005. The relationship of APOE genetic polymorphism with susceptibility to multiple sclerosis and its clinical phenotypes in Kuwaiti Arab subjects. Clin. Chim. Acta 351, 203–207. Ballerini, C., et al., 2000. Association of apolipoprotein E polymorphism to clinical heterogeneity of multiple sclerosis. Neurosci. Lett. 296, 174–176. Brück, W., 2005. The pathology of multiple sclerosis is the result of focal inflammatory demyelination with axonal damage. J. Neurol. 252 (Suppl. 5), v3–v9. Bu, G., 2009. Apolipoprotein E and its receptors in Alzheimer's disease: pathways, pathogenesis and therapy. Nat. Rev. Neurosci. 10, 333–344. Burwick, R.M., et al., 2006. APOE epsilon variation in multiple sclerosis susceptibility and disease severity: some answers. Neurology 66, 1373–1383. Carlin, C., Murray, L., Graham, D., Doyle, D., Nicoll, J., 2000. Involvement of apolipoprotein E in multiple sclerosis: absence of remyelination associated with possession of the APOE epsilon2 allele. J. Neuropathol. Exp. Neurol. 59, 361–367. Cocco, E., et al., 2005. HLA-DR, DQ and APOE genotypes and gender influence in Sardinian primary progressive MS. Neurology 64, 564–566.
17
Corder, E.H., et al., 1993. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 261, 921–923. Ebers, G.C., Sadovnick, A.D., Risch, N.J., 1995. A genetic basis for familial aggregation in multiple sclerosis. Canadian Collaborative Study Group. Nature 377, 150–151. Farrer, L.A., et al., 1997. 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 278, 1349–1356. Ferri, C., et al., 1999. APOE epsilon2-4 and ‐491 polymorphisms are not associated with MS. Neurology 53, 888–889. Gaillard, O., et al., 1998. Apolipoprotein E and multiple sclerosis: a biochemical and genetic investigation. J. Neurol. Sci. 158, 180–186. Høgh, P., et al., 2000. Apoliprotein E and multiple sclerosis: impact of the epsilon-4 allele on susceptibility, clinical type and progression rate. Mult. Scler. 6, 226–230. Huang, Y., 2006. Apolipoprotein E and Alzheimer disease. Neurology 66 (2 Suppl. 1), 79–85. Ignatius, M.J., et al., 1986. Expression of apolipoprotein E during nerve degeneration and regeneration. Proc. Natl. Acad. Sci. U. S. A. 83, 1125–1129. Koch-Henriksen, N., Hyllested, K., 1998. Epidemiology of multiple sclerosis: incidence and prevalence rates in Denmark 1948–64 based on the Danish Multiple Sclerosis Registry. Acta Neurol. Scand. 78, 369–380. Koutsis, G., et al., 2007. APOE genotypes in Greek multiple sclerosis patients: no effect on the MS Severity Score. J. Neurol. 254, 394–395. Lahiri, D.K., Sambamurti, K., Bennett, D.A., 2004. Apolipoprotein gene and its interaction with the environmentally driven risk factors: molecular, genetic and epidemiological studies of Alzheimer's disease. Neurobiol. Aging 25, 651–660. Losonczi, E., et al., 2010. APOE epsilon status in Hungarian patients with primary progressive multiple sclerosis. Swiss Med. Wkly. 140, w13119. Mahley, R.W., 1988. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science 240, 622–630. Mehta, P.D., Patrick, B.A., Pirttila, T., Coyle, P.K., Aisen, P.S., 2003. Detection of apolipoprotein E phenotype in unconcentrated cerebrospinal fluid. J. Clin. Lab. Anal. 17, 18–21. Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., PRISMA Group, 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann. Intern. Med. 151, 264–269. Mustafina, O.E., et al., 2008. Analysis of the association of allelic variants of apolypoprotein E and interleukin 1 beta genes with multiple sclerosis in ethnic Tatars. Genetika 44, 407–413. Niino, M., Kikuchi, S., Fukazawa, T., Yabe, I., Tashiro, K., 2003. Polymorphisms of apolipoprotein E and Japanese patients with multiple sclerosis. Mult. Scler. 9, 382–386. Noonan, C.W., Kathman, S.J., White, M.C., 2002. Prevalence estimates for MS in the United States and evidence of an increasing trend for women. Neurology 58, 136–138. Olesen, J., et al., 2012. The economic cost of brain disorders in Europe. Eur. J. Neurol. 19, 155–162. Pinholt, M., Frederiksen, J.L., Andersen, P.S., Christiansen, M., 2005. Apo E in multiple sclerosis and optic neuritis: the apo E-epsilon4 allele is associated with progression of multiple sclerosis. Mult. Scler. 11, 511–515. Pirttilä, T., Haanpää, M., Mehta, P.D., Lehtimäki, T., 2000. Apolipoprotein E (APOE) phenotype and APOE concentrations in multiple sclerosis and acute herpes zoster. Acta Neurol. Scand. 102, 94–98. Ramsaransing, G.S., Heersema, D.J., De Keyser, J., 2005. Serum uric acid, dehydroepiandrosterone sulphate, and apolipoprotein E genotype in benign vs. progressive multiple sclerosis. Eur. J. Neurol. 12, 514–518. Rubinsztein, D.C., 1994. Apo E genotypes in multiple sclerosis, Parkinson's disease, schwannomas and late-onset Alzheimer's disease. Mol. Cell. Probes 8, 519–525. Savettieri, G., et al., 2003. Apolipoprotein E genotype does not influence the progression of multiple sclerosis. J. Neurol. 250, 1094–1098. Strittmatter, W.J., et al., 1993. Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc. Natl. Acad. Sci. U. S. A. 90, 1977–1981. Varadé, J., et al., 2011. Replication study of 10 genes showing evidence for association with multiple sclerosis: validation of TMEM39A, IL12B and CLBL genes. Mult. Scler. 18, 959–965. Weatherby, S.J., et al., 2000. Polymorphisms of apolipoprotein E; outcome and susceptibility in multiple sclerosis. Mult. Scler. 6, 32–36. Wells, G.A., Shea, B., O'Connell, D., 2011. The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Health Research Institute. www.ohri.ca/programs/clinical_epidemiology/oxford.asp (accessed Oct 20, 2011). Zakrzewska-Pniewska, B., et al., 2004. Association of apolipoprotein E and myeloperoxidase genotypes to clinical course of familial and sporadic multiple sclerosis. Mult. Scler. 10, 266–271. Zwemmer, J.N., et al., 2004. No major association of ApoE genotype with disease characteristics and MRI findings in multiple sclerosis. Mult. Scler. 10, 272–277.