Abstract

Genomewide association studies (GWAS) have identified seven independent regions associated with Multiple Myeloma (MM) risk with an additional locus linked to the t11;14 subtype. Inherited variation is an important determinant of gene expression, such that the majority of published GWAS risk loci across all diseases can be linked to gene regulation. To understand whether the functional mechanisms that confer MM risk are related to allelic differences within regulatory regions, we sought to identify expression quantitative trait loci (eQTLs) in MM plasma cells. Recent studies have also suggested that eQTLs can be specific to cell type. In a heterogeneous disease such as MM, we might expect there to be variation in eQTLs between cytogenetic subgroups. To address this, we performed a genomewide analysis to identify MM related eQTLs, as well as potential subgroup specific MM eQTLs. The identification of eQTLs specific to MM plasma cells provides a means to link regulatory function to the powerful hypothesis-free tool of GWAS.

To generate MM related eQTLs, we combined orthogonal mRNA expression data (Affymetrix U133+2 arrays) from CD138+ selected plasma cells with genotyping data (Illumina Omni Express BeadChips) from germline DNA in the same individual. Two independent datasets comprising 183 MM patients from UK and 662 from Germany were analysed in parallel. Genotype data was filtered by standard quality control parameters. Single nucleotide polymorphisms (SNPs) showing deviation from Hardy-Weinberg equilibrium with P <1 × 10−6, having a call rate <95% or a minor allele frequency <1% were excluded. Samples were removed if closely related or if they had a non-Northern and Western European descent (CEU) ancestry. German and UK expression data were normalised independently using GC-RMA and a custom chip definition file (v17) mapping to Entrez genes. Genes showing a variance of less than 0.1 in expression between the analysed cases or a log2expression value of less than 5 in at least 95% of cases or genes located on the X or Y chromosome were excluded from the analysis. Known batch effects and hidden co-founders due to experimental and tumour-related factors were accounted for using a Bayesian Framework model. eQTLs were identified by performing a linear regression between residual expression levels and genotypes. A cis-eQTL analysis was performed that included SNPs located within 1 Mb of the transcript start site of the proximal gene. Results from the two independent studies were combined by meta-analysis.

We report in a cis-eQTL analysis, that there is evidence that 6 out of the 8 MM risk alleles have an impact on regulation of a proximal gene. We also show that the eQTLs can be replicated on contrasting technologies i.e. competitive allele-specific PCR (genotyping) and real-time PCR (expression). In a global cis-eQTL analysis, we found that the expression of >600 genes was significantly influenced by proximal SNPs (P < 5 x 10-8). By comparing these results with previous described eQTLs in other tissue types, we estimate some 10% of these eQTLs to be specific to MM plasma cells.

We conclude that informative regulatory regions important to myeloma biology can be identified by the combination of global gene expression and genomewide genotyping data. A number of these eQTLs can be shown to be MM specific and even specific to a cytogenetic subgroup. This can give us a greater understanding of the regulatory mechanisms underpinning genetic associations linked to MM risk and clinical outcomes following MM treatments.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.