Genetic Associations with Therapy Response in the HOVON-65/GMMG-HD4 Trial in Patients with Multiple Myeloma.

Abstract 1790

Poster Board I-816

Multiple Myeloma (MM) is a plasma cell malignancy, characterized by significant heterogeneous disease progression and response to therapy. It is very likely that genetic polymorphisms of genes involved in drug absorption, distribution, metabolism and excretion may affect a patient's response to therapy.

To explore whether genetic variation is associated with therapeutic outcome in the Dutch-German HOVON-65/GMMG-HD4 trial, a custom-built molecular inversion probe (MIP) based single nucleotide polymorphism (SNP) chip, designed by “Bank on a Cure” (BOAC)) was used, containing 3404 SNPs selected in “functional regions” within 983 genes representing cellular functions and pathways that may influence disease response, toxicities, complications, and survival. In the HOVON-65/GMMG-HD4 trial patients were treated with either 3 induction cycles of Vincristine/Adriamycin/Dexamethason (VAD, arm A) or 3 cycles of Bortezomib in combination with Doxorubicin and Dexamethason (PAD, arm B) after which response to treatment was determined. In 189 out of the initial 40% of patients (n=350) being included in the HOVON-65/GMMG-HD4 trial germ-line DNA was collected. To explore SNP associations with treatment response, an analysis was performed on both treatment arms separately and a comparison was made between good responders (CR, VGPR) (VAD n=17; PAD n=50) and patients who responded less well (PR or less) (VAD n=67; PAD n=55). A Cochran-Armitage trend test was applied on both data sets using the program PLINK. To account for multiple testing we carried out label swapping procedures.

The association analysis identified highly significant SNPs uniquely associated with VAD or bortezomib treatment response. The 3 most significant SNPs associated with CR/VGPR on VAD are rs2725362 in WRN (p=8.0×10^-4; Odds Ratio (OR)=4.13, 95% confidence interval (CI)=1.71-9.97) rs891507 in ABCB8 (p=1.2×10^-3; OR=4.44, 1.81-10.88) and rs755622 in MIF (p=1.4×10^-3; OR=4.35, 1.79-10.55). Different SNPs were associated with CR/VGPR on PAD: rs2293616 in SLC15A2 (p=5.0×10^-4; OR=0.43, 0.25-0.75), rs1062372 in PMS2 (p=9.0×10^-4; OR=0.31, 0.15-0.66) and rs90925 in LTA (p=1.0×10^-3; OR=0.41, 0.22-0.75). Pathway analyses showed that the two sets of associated SNPs were enriched for genes involved in DNA repair, immune response/inflammation and drug metabolism. In addition, overlapping SNP associations in the ADME genes (drug absorption, distribution, metabolism, and excretion) ABCC1, CYP2C9, ABCC4 and ABCA1 were seen in both treatment arms.

Our findings demonstrate a trend of genetic polymorphisms associated with treatment response. Genetic variations were found that were uniquely associated with each treatment. The associations found will be validated in an independent dataset.