Overcoming Melphalan Resistance By Targeting Crucial DNA Repair Pathways in Multiple Myeloma

Abstract

High dose melphalan (MEL) and autologous stem cell transplant (ASCT) is the standard of care in the treatment of multiple myeloma (MM). Resistance to MEL has been linked to increased DNA repair. Here we sought to identify which DNA repair pathways lead to MEL resistance. As inhibitors of the single strand DNA repair enzyme PARP1 are clinically available, we first examined whether PARP1 gene expression affected clinical outcomes. In a dataset of 414 patients treated with tandem ASCT, high PARP1 expression lead to reduced survival compared to patients with low expression (42.7 months vs. median not reached, p=0.003). We therefore tested in vitro the cytotoxicity of PARPi with veliparib (VEL), olaparib (OLA) or niraparib (NIRA) in RPMI8226 and U266 MM cell lines as well as their MEL resistance counterparts, RPMI8226-LR5 (LR5) and U266-LR6 (LR6). At physiologic doses PARPi alone was not cytotoxic. However, the MEL IC50 in RPMI8226 cells decreased from 27.8 to 23.1, 22.5 and 18.0 µM with the addition of VEL, OLA and NIRA respectively. Similarly, MEL IC50 decreased in U266 cells from 6.2 to 3.2, 3.3 and 3.0 µM. However, in LR5 and LR6 cells, PARPi did not completely reverse MEL resistance. We confirmed this in a NOD/SCID/gamma null xenograft mouse model with either MEL sensitive (RPMI8226) or resistant (LR5) MM. Treatment with a MEL-VEL combination prolonged survival in RPMI8226 mice compared to MEL alone (107 vs. 67.5 days, p=0.0009), but not in LR5 mice (41 vs. 39 days, p=0.09). Next we tested whether the double strand DNA repair mechanisms, homologous recombination (HR) and non-homologous end joining (NHEJ), cause MEL resistance which is not overcome by PARPi. Using an HR assay, LR6 cells had 4.5 fold higher HR ability than parent U226 cells (p = 0.05) showing this to be a possible mechanism of resistance. However, LR5 cells had equivalent HR ability to parent RPMI8226 cells. We therefore hypothesized that NHEJ may be a mediator of MEL resistance in LR5 cells. As DNA-PK is integral to NHEJ and may be a therapeutic target, we treated LR5 cells with the DNA-PK inhibitor NU7026 in combination with MEL. Although NU7026 alone at 2.5 µM had no cytotoxicity, in combination it completely reversed resistance to MEL resulting in a lower IC50 than even parent RPMI8226 cells (46.4 vs. 14.4 µM). Based on our findings, we hypothesized that combined expression of the HR gene BRCA1, the NHEJ gene DNA-PK, and PARP1 may predict survival in patients undergoing ASCT. Using the same clinical dataset, we found that overexpression of 0 (n=101), 1-2 (n=287), or all 3 genes (n=26) strongly predicted median survival (undefined vs. 57.8 vs. 14.8 months, p Here we demonstrated that MEL resistance is not completely reversed by PARPi, while it may be overcome by targeting HR and NHEJ pathways. In addition, we observed that a 3 gene analysis pre-ASCT may help predict patients that are inherently resistant or sensitive to high dose MEL.