P-052: Identification of novel genetic loci for risk of multiple myeloma by functional annotation

Abstract

<h3>Background</h3> Genetic factors have proven to have an impact on multiple myeloma (MM) susceptibility, with genome-wide association studies (GWAS) discovering 23 loci associated with MM risk. However, much of the heritability of MM remains unexplained. The stringent significance threshold used in GWAS (p<5×10^-8) accounts for the numerous statistical tests being performed but is prone to the risk of false negatives. One strategy for reducing the number of tests is to consider only SNPs with heightened prior probabilities of association, according to meaningful biological criteria. The best candidate SNPs identified with this approach can then be tested in additional MM cases and controls from independent populations. We aimed at surveying the effect of single nucleotide polymorphisms (SNPs), predicted to have a functional role, on MM risk. <h3>Methods</h3> The association study consisted of two GWAS as discovery datasets, namely the InterLymph consortium and the German GWAS and a replication dataset, namely the the International Multiple Myeloma rESEarch (IMMEnSE) consortium, for an overall total of 5442 MM cases and 6174 controls. SNPs were first ranked according to p-value and concordance of the association between the two discovery datasets, then ranked by functional annotation, using bioinformatic tools and databases. We considered the following classes of functional SNPs: missense, synonymous and non-sense SNPs, expression quantitative trait loci (eQTLs), splicing quantitative trait loci (sQTLs),), SNPs in SNPs affecting function of long non-coding RNAs (lncRNA), and SNPs modifying transcription factor binding sites. We prioritized the resulting SNPs for replication in IMMEnSE by p-values for association in InterLymph and German GWAS and by evidence for a functional role. <h3>Results</h3> In the two discovery datasets, 136 SNPs fit the criteria of association with MM risk with p<10^-4 and did not map close to known MM risk loci. After pruning for linkage disequilibrium, four SNPs (rs12038685, rs2664188, rs12652920, rs29794) were chosen for replication in IMMEnSE. Among these, rs2664188 showed to be significantly associated also in the replication dataset (OR=1.30, 95% CI=1.16-1.46, p=0.001). The final meta-analysis including the three datasets, using a random-effect model, confirmed the association (OR=1.18, 95% CI=1.07-1.30, p=0.0007). <h3>Conclusion</h3> The G-allele of rs2664188, that showed a consistent association with increased risk of developing MM in all phases of our analysis is an eQTL, according to GTEx., The G-allele is associated with increased expression of the N4BP2 gene in whole blood. N4BP2 encodes a protein which binds to B-cell leukemia/lymphoma 3 (BCL-3), a well-known proto-oncogene, known to play a role in cell proliferation and apoptosis inhibition in myeloma cell lines.