Investigation of potential gene-gene interactions between APOE and RELN contributing to autism risk.

Background: Several candidate gene studies support RELN as susceptibility gene for autism. Given the complex inheritance pattern of autism, it is expected that gene-gene interactions will exist. A logical starting point for examining potential gene-gene interactions is to evaluate the joint effects of genes involved in a common biological pathway. RELN shares a common biological pathway with APOE, and Persico et al. have observed transmission distortion of the APOE-2 allele in autism families.

Objective: We evaluated RELN and APOE for joint effects in autism susceptibility.

Methods: A total of 470 Caucasian autism families were analyzed (265 multiplex; 168 trios with no family history; 37 positive family history but only one sampled affected). These families were genotyped for 11 RELN polymorphisms, including the 5' untranslated region repeat previously associated with autism, as well as for the APOE functional allele. We evaluated single locus allelic and genotypic association with the pedigree disequilibrium test and geno-PDT, respectively. Multilocus effects were evaluated using the extended version of the multifactorial dimensionality reduction method.

Results: For the single locus analyses, there was no evidence for an effect of APOE in our data set. Evidence for association with RELN (rs2,073,559; trio subset P=0.07 PDT; P=0.001 geno-PDT; overall geno-PDT P=0.05), however, was found. For multilocus geno-PDT analysis, the joint genotype of APOE and RELN rs2,073,559 was highly significant (trio subset, global P=0.0001), probably driven by the RELN single locus effect. Using the extended version of the multifactorial dimensionality reduction method to detect multilocus effects, there were no statistically significant associations for any of the n-locus combinations involving RELN or APOE in the overall or multiplex subset. In the trio subset, 1-locus and 2-locus models selected only markers in RELN as best models for predicting autism case status.

Conclusion: Thus, we conclude that there is no main effect of APOE in our autism data set, nor is there any evidence for a joint effect of APOE with RELN. RELN, however, remains a good candidate for autism susceptibility.