Association of the EPHA1 gene polymorphism with idiopathic mild intellectual disability

Aim. To investigate a possible association of the EPHA1 gene polymorphism with mild intellectual disability (ID). Methods. The group of patients with mild (IQ score between 50 and 70) idiopathic intellectual disability consisted of 65 individuals including 41 (63.1 %) males and 24 (36.9 %) females. The control group consisted of 250 healthy volunteers from different regions of Ukraine. The genotyping was performed using PCR followed by RFLP analysis for rs11768549, rs11767557, rs11771145 and ARMS PCR analysis for novel c.1891G>A EPHA1 gene mutation. Results. The data concerning the EPHA1 genotypes and allelic variants distribution in ID patients and control group were obtained. Statistical analysis showed a signifi cant association of minor rs11768549-A allele (OR = 3.96, 95 % CI = 1.13–13.89) and wild-type rs11767557-T (OR = 1.99, 95 % CI = 1.18–3.37) and rs11771145-G (OR= 1.55, 95 % CI = 1.02–2.37) alleles with a higher risk of mild ID development (p < 0.05 for all). Conclusions. Our results suggest that SNPs (rs11768549, rs11767557, rs11771145) in the EPHA1 gene are associated with idiopathic mild intellectual disability. Therefore, we propose the EPHA1gene as a new candidate gene and the polymorphisms rs11768549, rs11767557, rs11771145 as new markers of genetic susceptibility for intellectual disability.


Introduction
In recent reports we have presented our study on the whole exome sequencing (WES) in two affected siblings with non-syndromic intellectual disability from Ukrainian family and their healthy non-consanguineous parents that identifi ed two missense mutations in the coding region of the EPHA1 gene (c.1475 G > A -rs11768549 and novel c.1891G > A) [1][2][3].
The EPHA1 gene is located at chromosome 7q34 and contains 18 exons that span a little over 18 kb [4].The human EPHA1 gene is transcribed as a single 3.5 kb mRNA [5].The EPHA1 protein contains 976 amino acids and is approximately 108 kDa [6].EPHA1 is the fi rst identifi ed member of the erythropoietin-producing hepatocellular (Eph) receptors family [7].EPHA1 is a receptor tyrosine kinase that plays an important role in developmental processes, including nervous system, where it participates in the forward signaling in receptorbearing cells and the reverse signaling in ligandbearing cells by binding to GPI-linked A ephrins, which together facilitate communication between neighboring cell populations and axon guidance [6,[8][9][10][11][12].Ephrin type-A receptor 1 is highly expressed in the adult brain and in the spinal neural tube during neurulation [13,14] and seems to be a good disease gene in ID.
The previous studies have shown that EPHA1 regulates the cell morphology and motility through the ILK-RhoA-ROCK pathway [15].It is therefore important to note that mutations in the genes encoding the regulators and effectors of the Rho GTPases such as OPHN1, PAK3, alpha-PIX (also known as ARHGEF6 or Cool-2), FMR1 and MEGAP have been found to underlie various forms of ID [16,17].Interestingly, the defects in Rho signaling pathways have recently also been linked with Alzheimer disease (AD) as well as the EPHA1 gene SNPs [18,19].EPHA1 was documented to be one of the most strongly associated locus with AD in a few recent genome wide association studies [20][21][22].It was shown that the minor C allele at SNP rs11767557 of EPHA1 was associated with signifi cantly lower odds of being Aβ-positive and protective for late onset AD (LOAD) as well as having the minor A allele in another SNP rs11771145 on EPHA1 was also associated with lower odds of AD [20][21][22].Furthermore, the EPHA1 gene SNP rs11768549 was associated with the rapid AD progression and the missense mutations in the coding region of the EPHA1 gene rs202178565 -with LOAD [23,24].
Intellectual disability is a common neurodevelopmental disorder occurring in 1-3 % of the general population and its burden on the affected individuals, their families and society is enormous.[25].Still, in ~ 60 % of cases of ID the etiology is unknown because of the extensive clinical and genetic heterogeneity [26].The causes of intellectual disability vary with the severity of the condition: moderate-tosevere intellectual disability (IQ less than 50) is much more likely to be due to a single pathological cause (genetic or environmental) whereas mild ID (defi ned as an IQ score between 50 and 70) is rather a complex condition in origin as well as AD [27].
The aim of this study is to evaluate a possible association of the EPHA1 gene polymorphisms rs11768549, rs11767557, rs11771145 and novel c.1891G > A mutation with mild intellectual disability.

Materials and Methods
DNA-samples were extracted from peripheral blood leucocytes of unrelated volunteers from different regions of Ukraine and ID patients by the standard phenol-chloroform method.Informed consents were obtained from all the individuals participating in our study.
The group of patients with mild (an IQ score between 50 and 70) idiopathic intellectual disability consisted of 65 individuals including 41 (63.1 %) males and 24 (36.9 %) females, where the previous extensive genetic investigations revealed no abnormalities.All patients underwent physical and neurological examination (test used for IQ: WISC III, WISC-R, WISC) and standard G-banding karyotype analysis.DNA tests to determine Fragile X status (FRAXA, FRAXE, FRAXF loci) and Prader Willi/Angelman syndromes (PW/AS) were performed to rule out the known genetic causes of ID prior to further investigation.Array-CGH analysis (400K resolution) revealed no pathological rearrangements in all patients.
The control group consisted of 250 individuals including 131 (52.4 %) males and 119 (47.6 %) females.This group may be considered representative for the estimation of DNA polymorphism frequency in autosomal genes [28].
Polymorphic variants c.1475G > A (rs11768549) and c.1891G > A (novel) of EPHA1 gene were detected as described in our recent work [2].
The presence of EPHA1 polymorphism rs11767557 was examined by RsaI PCR-RFLP (restriction fragment length polymorphism) analysis.Specifi c oligonucleotides, designed and synthesized in accordance to corresponding sequences of EPHA1 gene that were used as primers are presented in Table 1.The PCR amplifi cation was performed in a fi nal volume of 25 l containing 1 PCR buffer, 1.5 mM MgCl2, 200 M of each dNTP, 1 M of each primer, 0.2 units of Taq-DNA polymerase and 200 ng of the DNA template.The cycling conditions were as follows: initial denaturation at 95 C for 5 min, 30 cycles consisting of denaturation at 94 C for 30 s, annealing at 63 C for 30 s, extension at 72 C for 30 s and a fi nal elongation step at 72 C for 3 min.The amplifi ed fragments were digested with RsaI.Digestion was performed in 15 l reaction volume containing 1 X reaction buffer, 0.5 units of the restriction enzyme and 10 l of purifi ed PCR product, in-cubated at 37 C overnight and analyzed using 2 % standard agarose gelelectrophoreses.
The assessment of the EPHA1 rs11771145 polymorphism was performed by the DraI PCR-RFLP analysis using the specifi c oligonucleotide primers described in Table 1.In the forward primer, the mismatched nucleotides (underlined) were substituted so that amplifi cation of the minor A-allele resulted in the generation of a new DraI restriction cutting site.The PCR amplifi cation was performed in a fi nal volume of 25 l containing 1 PCR buffer, 1.5 mM MgCl2, 200 M of each dNTP, 1 M of each primer, 0.2 units of Taq-DNA polymerase and 200 ng of the DNA template.The cycling conditions were as follows: initial denaturation at 95 C for 5 min, 30 cycles consisting of denaturation at 94 C for 30 s, annealing at 58 C for 30 s, extension at 72 C for 30 s and a fi nal elonga-tion step at 72 C for 3 min.The amplifi ed fragments were digested with DraI.Digestion was performed in 15 l reaction volume containing 1 X reaction buffer, 0.5 units of the restriction enzyme and 10 l of purifi ed PCR product, incubated at 37 C overnight and analyzed in 2 % standard agarose gels.
The results were statistically assessed using Arlequin and OpenEpi software and Fisher's 2 by 2 exact test, as well as odd ratio (OR) calculation; p < 0.05 was considered to be statistically signifi cant test [29,30].

Results and Discussion
We did not identify c.1891G>A (p.Gly631Arg) substitution in any of the investigated individuals from the ID patients (n = 65) and control groups (n = 250).It was found only in the members of afo resaid Uk rai nian family with ID in two siblings.Furthermore, c.1891G > A substitution was not observed in any of Exome Aggregation Consortium (URL: http://exac.broadinstitute.org),International HapMap Project (URL: http://hapmap.ncbi.nlm.nih.gov/) and 1000 Ge nomes Project (URL: http://browser.1000genomes.org/)populations.Therefore, we assume that the EPHA1 gene c.1891G>A (p.Gly631Arg) is the mutation with an extremely low frequency.
Based on the RFLP analysis of rs11768549 (c.1475G>A and p.Arg492Gln) variant, the individuals were classifi ed into three groups: GG, GA and AA.The genotypes and allele frequencies of the rs11768549 polymorphism are presented in Table 2.The T to C transition in the rs11767557 variant, located in EPHA1-AS1 (EPHA1 antisense RNA 1), generates a restriction site for endonuclease RsaI.Thereby three different patterns could be observed after the restriction digestion of the PCR-product: a 263 bp band (for genotype TT); a 263 bp, a 147 bp and a 116 bp bands (for genotype TC); a 147 bp and a 116 bp bands (for genotype CC) (Fig. 1).
In the forward primer, designed for the rs117 71145 (located in EPHA1-AS1) analysis, the mismatched nucleotides (underlined) were substituted so that am plifi cation of the minor A-allele resulted in the generation of a new DraI restriction cutting site.Thereby three different patterns could be observed after the restriction digestion: a 122 bp band (for GG genotype); a 122 bp, a 96 bp and a 26 bp (not visible on electrophoregram) bands (for GA genotype); a 96 bp and a 26 bp bands (for AA genotype) (Fig. 2).The genotype and allele frequencies of the rs11767557 and rs11771145 polymorphisms are summarized in Table 2.
All observed genotype distributions showed no deviations from Hardy-Weinberg expectations in the general population of Ukraine and in the ID patients group (p > 0.05).A signifi cant linkage disequilibrium (LD) was found between rs11768549 and rs11767557 (p = 0.00436), rs11767557 and rs11771145 (p < 0.0001).
The SNP analysis showed the signifi cant differences between the ID patients and the controls for the allele frequencies of all (rs11768549, rs11767557, rs11771145) investigated SNPs (p < 0.05 for all).We found a signifacant association of minor rs11768549-A (OR = 3.96, 95 % CI = 1.13-13.89)and wild-type rs11767557-T (OR= 1.99, 95 % CI = 1.18-3.37)and rs11771145-G (OR= 1.55, 95 % CI = 1.02-2.37)alleles with a higher risk of mild ID development.Three studied polymorphic variants showed a signifi cant association with mild ID for different genetic models: multiplicative for rs11768549 (A vs G: OR = 3.96, 95 % CI = 1.13 -13.89, p = 0.02), recessive for rs11767557 (TT vs TC + CC: OR = 0.50, 95 % CI = 0.27-0.9,p = 0.02) and dominant for rs11771145 (GG+AG vs AA: OR= 0.28, 95 % CI = = 0.08 -0.93, p = 0.03).Genotypes rs11767557-  TC+CC and rs11771145-AA were associated with signifi cantly lower odds of mild ID developing.The non-synonymous SNP rs11768549 is located in the coding region (exons 7) of the EPHA1 receptor gene and results in the mutation of a positive charged Arg492 to an uncharged Gln492 located in the fi bronectin type III repeat of the EphA1 ectodomain.This mutation may directly cause the changes in the domain conformational fl exibility and might result in the decrease of signal transduction and binding with the ligands or the protein-partners.The SNPs rs11767557 and rs11771145 are located in the EPHA1 antisense RNA 1 (long non-coding RNA), near the promoter region and may affect the regulation of the EPHA1 gene expression.Long non-coding RNAs have emerged as key regulators of the gene expression at different levels, including chromatin remodeling, transcriptional control, mRNA stability, mRNA translation, microRNA function, and protein metabolism [31,32].Our results suggest that SNPs (rs11768549, rs11767557, rs11771145) in the EPHA1 gene are associated with idiopathic mild ID.These data can be explained by the changes in the EPHA1 receptor signal transduction effi ciency in the rs11768549 minor allele A carriers and/or in the EPHA1 gene expression levels in individuals with different rs11767557 and rs11771145 genotypes that in turn, may affect the regulation of neurogenesis, axon guidance and synaptic plasticity at least via the EPHA1-ILK-RhoA-ROCK signaling.
Therefore, we propose EPHA1 as a new candidate gene and the polymorphisms (rs11768549, rs11767557, rs11771145) as new markers of the genetic susceptibility for intellectual disability.Further investigati ons are necessary to explain the molecular mechanisms of the EPHA1 involvement in ID pathogenesis.