Autism-associated haplotype affects the regulation of the homeobox gene, ENGRAILED 2

Abstract

Background: Association analysis identified the homeobox transcription factor, ENGRAILED 2 (EN2), as a possible autism spectrum disorder (ASD) susceptibility gene (ASD [MIM 608636]; EN2 [MIM 131310]). The common alleles (underlined) of two intronic single nucleotide polymorphisms (SNPs), rs1861972 (A/G) and rs1861973 (C/T), are over-transmitted to affected individuals both singly and as a haplotype in three separate datasets (518 families total, haplotype p = .00000035).

Methods: Further support that EN2 is a possible ASD susceptibility gene requires the identification of a risk allele, a DNA variant that is consistently associated with ASD but is also functional. To identify possible risk alleles, additional association analysis and linkage disequilibrium (LD) mapping were performed. Candidate polymorphisms were then tested for functional differences by luciferase (Luc) reporter transfections and electrophoretic mobility shift assays (EMSAs).

Results: Association analysis of additional EN2 polymorphisms and LD mapping with Hapmap SNPs identified the rs1861972-rs1861973 haplotype as the most appropriate candidate to test for functional differences. Luciferase reporters for the two common rs1861972-rs1861973 haplotypes (A-C and G-T) were then transfected into human and rat cell lines as well as primary mouse neuronal cultures. In all cases the A-C haplotype resulted in a significant increase in Luc levels (p < .005). The EMSAs were then performed, and nuclear factors were bound specifically to the A and C alleles of both SNPs.

Conclusions: These data indicate that the A-C haplotype is functional and, together with the association and LD mapping results, supports EN2 as a likely ASD susceptibility gene and the A-C haplotype as a possible risk allele.

Figure 1

ENGRAILED 2 LD map. Inter-marker r² values for rs1861973 are shown. The map includes 26 EN2 polymorphisms typed in the AGRE I dataset (167 families not subset on ethnicity) plus 3120 Hapmap SNPs within 2Mb of EN2 (+1Mb 5', -1Mb 3') typed in the CEU dataset. Only rs1861972 and rs2361688 display high r² values (>.75) with rs1861973. However, rs2361688 is not consistently associated with ASD , identifying rs1861972 and rs1861973 as the most appropriate candidates to test for functional allelic differences.

Figure 2

Functional difference between rs1861972-rs1861973 A-C and G-T haplotypes. (a) The functional difference between the A-C and G-T intronic haplotypes was investigated by generating the diagrammed luc reporter constructs: T- SV40 minimal promoter 5' of luc without EN2 intron, P- EN2 promoter (-1 to -5735) 5' of luc without EN2 intron, AC- EN2 intron with rs1861972-rs1861973 A-C haplotype cloned 3' of luciferase but 5' of the SV40 polyadenylation signal to approximate the endogenous locus, GT- EN2 intron with rs1861972-rs1861973 G-T haplotype cloned 3' of luciferase but 5' of the SV40 polyadenylation signal. (b-d) Relative light units of luciferase normalized to Renilla reniformis and expressed as percent of control, pgl3 promoter vector (T), is shown for the SV40 minimal promoter constructs transiently transfected into (b) P6 cerebellar granule neurons (n=6), (c) PC12 cells (n=6) and (d) HEK293T cells (n=6). (e-g) Normalized relative light units of luciferase for luc EN2 promoter constructs expressed as percent of control (P) is shown for (e) P6 cerebellar granule neurons (n=7), (f) PC12 cells (n=6) and (g) HEK293T cells (n=6). * P<.005, ** P<.001, *** P < .00001, two tailed paired Student's T test

Figure 3

Differential binding of nuclear proteins to rs1861972 and rs1861973 associated alleles. (A) To investigate whether the associated SNPs affect the binding of nuclear proteins, EMSAs were conducted with biotinylated 20-mer oligonucleotides and nuclear extract isolated from P6 mouse cerebellar granule cells. Extract was incubated with oligonucleotides specific to each allele, separated on a denaturing acrylamide gel, transferred to a membrane and detected by chemiluminescence. Several protein-DNA complexes were observed for both SNPs. Specificity was determined by competing with 100 molar excess of unlabelled oligonucleotide. Protein-DNA complexes specific to the associated rs1861972 A allele or rs1861973 C allele were observed (arrows) that were not detected for the corresponding rs1861972 G allele or rs1861973 T allele biotinylated oligonucleotides. In addition, protein-DNA complexes common to both alleles for rs1861972 or rs1861973 were observed (arrowheads). Abbreviations: 972-A: 20-mer oligonucleotide specific to the rs1861972 A allele, 972-G: 20-mer oligonucleotide specific to the rs1861972 G allele, 973-C and C: 20-mer oligonucleotide specific to the rs1861973 C allele, 973-T and T: 20-mer oligonucleotide specific to the rs1861973 T allele, + or -: presence or absence respectively of extract or 100 molar excess of unlabelled oligonucleotide. (B) To examine allele-specific binding of nuclear proteins to rs1861972 (left) and rs1861973 (right), additional competitions were performed. 80 molar excess of 3 different unlabelled oligonucleotides were each added individually to the probe and nuclear extract: oligonucleotide with the same sequence as the biotinylated probe (972-A, 973-C), mutant oligonucleotides predicted to disrupt NF1, NFY, C/EBP binding to the A allele of rs1861972 or Sp1 and Ets binding to the C allele of rs1861873, and oligonucleotides for the non-associated G (972-G) and T (973-T) alleles. The sequence for each oligonucleotide is shown. Abbreviation: - absence of competitor

Figure 4

Conservation of transcription factor binding sites for associated and non-associated alleles of rs1861972 and rs1861973. The 20 bp sequence encompassing rs1861982 and rs1861973 and used as probes in our EMSAs is depicted. Conserved transcription factor sites are underlined with the polymorphic allele for each SNP designated with an asterisk.