gms | German Medical Science

Genomically imprinted genes contribute to the broader class of parent-of-origin effects (POEs) as their alleles are epigenetically inactivated depending on their parental origin. The existence of imprinting effects on the susceptibility to rheumatoid arthritis (RA) and type 1 diabetes (T1D) has been investigated earlier [1], [2]. However, findings have been contradicting and other POEs including shared household effects (environmental effects of the mother) and maternal genetic effects may have biased the results.

We used a special type of linear mixed model (imprinting model) to investigate the existence of imprinting effects on the susceptibility to T1D and RA. The advantage of this model is that it is able to simultaneously consider all kinds of imprinting (maternal, paternal, full, and partial imprinting) [3], [4] and to separate maternal imprinting effects from other POEs. This was not possible with previous population-scale imprinting analyses models. The imprinting model utilizes a gametic relationship matrix and predicts two random gametic effects (one as father and one as mother). The imprinting effect is defined as difference of these effects and the variation of this difference represents the imprinting variance. Where no imprinting is observed, the imprinting variance is zero. The inclusion of fixed effects, such as sex, birth year, and social economic index also allowed the investigation of environmental effects. Through linkage of the Swedish Hospital Discharge Register and the Multigeneration Register, a suitable population database was available [5], [6]. For RA the dataset contained 15,850 patients with 60,684 ancestors; for T1D 27,255 patients with 208,114 ancestors were analyzed.

With regard to RA the imprinting variance did not turn out to be significant but significant variances were found for maternal effects. They could not clearly be determined to be due to the mother's genetic influence or due to the maternal environment. Whereas a considerable ratio of the phenotypic variance could be explained by the first (14%; ±4.14%), 5% (±6.81%) could be explained by the latter. With regard to T1D the imprinting variance was not significant; however, the significance of the variance due to maternal environmental effects indicated the importance of a shared household effect explaining 18.8% (±1.81%) of the phenotypic variance. Fixed effects such as sex and environmental factors including social economic index, number of offspring, birth year and their interactions with sex showed large effects on the susceptibility to RA and T1D.

The results indicated that imprinting effects on the development of RA and T1D can be neglected but that other POEs including maternal genetic and maternal environmental effects have a considerable impact. The results implicate that these POEs must be taken into account in future studies. Furthermore, in light of the increasing digitization of human pedigree and medical data [7], the imprinting model constitutes an important tool to fully exploit these sources of information and fully elucidate the importance of POEs.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.