oa Monoallelic Expression in Melanoma

Background

Monoallelic expression (MAE) is a frequent genomic phenomenon in normal tissues, however its role in cancer is yet to be fully understood. MAE is defined as the expression of a gene that is restricted to one allele in the presence of a diploid heterozygous genome. Constitutive MAE occur for imprinted genes, odorant receptors and random X inactivation. Several studies in normal tissues have showed MAE in about 5–20% of the cases. However, little information exists on the MAE rate in cancer. In this study we assessed the presence and rate of MAE in melanoma. The genetic basis of melanoma has been studied in depth over the past decades, leading to the identification of mutations/genetic alterations responsible for melanoma development. To examine the role of MAE in melanoma we used 15 melanoma cell lines and compared their RNA-seq data with genotyping data obtained by the parental TIL (tumor infiltrating lymphocytes).

Methods

Genotyping was performed by Illumina HumanOmni1 beadchip. For the RNA-seq experiment, library preparation and sequencing was performed using the Illumina TruSeq Stranded Total RNA Human Kit and subsequently sequenced using a HiSeq 2500 according to manufacturer's guidelines. Genotype calling and subsequent quality filtering was performed in Illumina's Genome Studio software. IMPUTE2 (University of Oxford) was used for imputation of the genotyped data. RNA-Seq analysis was performed using the Broad Best Practice Workflows for RNA-Seq with the exception that a genotype was created for every available base pair in order to compare to the genotyped array data. In house custom perl scripts were then created to compare the genotyping data to the processed RNA-Seq data. The genotype and the B-allele frequency were calculated the latter creating a range of expression to evaluate.

By comparing genotyping data with RNA-seq data, we identified SNPs in which DNA genotypes were heterozygous and corresponding RNA genotypes were homozygous. All homozygous DNA genotypes were removed prior to the analysis. To confirm the validity to detect MAE, we examined heterozygous DNA genotypes from X chromosome of female samples as well as for imprinted and olfactory receptor genes and confirmed MAE.

Results

MAE was detected in all 15 cell lines although to a different rate (spanning from approximately 17% to 75% MAE). When looking at the B-allele frequencies we found a preferential pattern of complete monoallelic expression rather then differential monoallelic expression across the 15 melanoma cell lines. As some samples showed high differences in the homozygous and heterozygous call rate, we looked at the single chromosomes and showed that MAE may be explained by underlying large copy number imbalances and isodisomy in some instances. Nevertheless, some chromosome regions showed MAE without CN imbalances suggesting that additional mechanisms (including epigenetic silencing) may explain MAE in melanoma.

Conclusion

The biological implications of MAE are yet to be realized. Nevertheless our findings suggest that MAE is a common phenomenon in melanoma cell lines. Further analyses are currently being undertaken to evaluate whether MAE is gene/pathway specific and to understand whether MAE can be employed by cancers to achieve a more aggressive phenotype.