Exome Sequencing in Myeloma Pedigrees Implicates RAS1 and NOTCH Signaling Are Involved in Inherited Myeloma Risk

Multiple Myeloma (MM) is a cancer of plasma cells with poor prognosis. Although, MM has been shown to be highly heritable in genealogy studies, no inherited risk-alleles have been identified. We hypothesize MM heritability is in part due to rare, germline variation that can be discovered in high-risk pedigrees.

High-risk MM pedigrees were identified using the Utah Population Database which contains both genealogical and state cancer records. High-risk pedigrees were defined to have  statistical excess of MM (p < 0.05). In this study we whole-exome sequenced germline DNA from 42 MM cases from high-risk pedigrees. Best practice variant calling, joint genotyping, and quality control were performed on the cases, a set of background controls from the 1000 Genome Project (1000G), and a small set of local controls (for technical artifacts), and resulted in 607,908 variants. We prioritized variants that were: 1) shared by at least 3 related MM cases, 2) absent in local controls, and 3) rare (frequency ≤ 0.01 in 1000G). This prioritization resulted in 116 variants of interest.

Of the 116 variants, 3 MM cases in one high-risk pedigree shared multiple variants on chromosome 1p36.11-35.1. To formally assess whether these variants were inherited from a common founder we performed shared genome segment analysis using high-density SNP genotyping. We identified a region at 1p36.11-35.1, 8.9 Mb in length (p = 6.0 × 10^-4; 22,000 simulations), providing positive evidence for segregation from a common founder.

The segregating variants identified from exome sequencing in this region are in the genes CNKSR1, ARID1A, and SDC3. All three variants are individually predicted to be moderately deleterious. The variant in CNKSR1 falls in a splice region and has minor allele frequency of 0.004 in the 1000G Europeans. This variant was also observed in 3 additional cases in our sequencing set, indicating a strong enrichment of this variant in our high-risk MM cases (6/42 = 0.143). CNKSR1 is involved in the RAS1 and NOTCH signaling pathways and is a known target for cancer therapy. The variants in ARID1A and SDC3 both result in non-synonymous codon changes. ARID1A is commonly mutated across cancers and has been associated with accelerated tumor growth in hepatocellular carcinomas. SDC3 also interacts with NOTCH signaling, a pathway involved in MM growth (especially in patients with MAF translocations) and osteoclastogenesis. The variants in ARID1A and SDC3 are not carried by other cases in our sequencing set, which may indicate the combination of these variants is important to confer risk in the pedigree, or merely that the variants are hitchhikers on the segregating chromosome.

Germline risk-alleles will shed light on the genetic factors involved in MM susceptibility and ultimately may provide new avenues for screening, diagnosis and treatment. Here we have identified evidence for segregating variants in a high-risk pedigree that implicate the RAS1 and NOTCH signaling pathways as involved in MM risk. Future work includes confirmation sequencing in the pedigree and a broader set of MM cases, and functional follow-up of the variants and their role in disruption of the genes and pathways.