Characterization of Bortezomib Resistance in Multiple Myeloma Cell Lines

Bortezomib is the first proteasome inhibitor in clinical use. It is particularly effective in myeloma compared with other cancers; as a single agent, its response rate in relapsed myeloma is about 40%. Nevertheless, a large proportion of patients are primarily resistant and acquired bortezomib resistance is also an emerging issue. The mechanism of bortezomib resistance in myeloma remains elusive. We have previously shown that bortezomib sensitivity correlates inversely with the levels of XBP-1 mRNA in vitro and in myeloma patients. XBP-1 is a major regulator of the unfolded protein response (UPR), which is essential for immunoglobulin folding and assembly, and plasma cell development. Previous studies suggested that immunoglobulin production sensitises myeloma cell lines to proteasome inhibitors (S Meister, et al., 2007, Cancer Res 67:1783). We hypothesize that dependence on XBP-1 and the UPR mediate sensitivity to bortezomib. The goal was to develop bortezomib resistant myeloma cell lines, study their phenotype, characterise the mechanism of resistance and, in particular, study the changes in the UPR. Bortezomib resistant sublines (9 in total) were derived from 4 myeloma cell lines (KMS-11, H929, U266 and OPM2), by long term exposure to bortezomib. The fold resistance varies from 3 to 12 and is stable for at least 4 passages without bortezomib. During the evolution of resistance, the myelomas all adopted adherent growth, independent of the morphology of the parent cell lines. When adherent versions of the parent cell lines were derived without exposure to bortezomib, they were, surprisingly, resistant to bortezomib. All bortezomib-resistant myeloma sublines downregulated XBP-1 mRNA levels and the reduction was correlated with the fold resistance. Moreover, splicing of XBP-1 mRNA to the active form was also reduced. The adherent versions of the parent cell lines that showed de novo resistance to bortezomib also had lower levels of XBP-1. Bortezomib normally induces XBP-1 expression and splicing acutely within 12–24 hours (AH Lee, et al., 2003. PNAS 100: 9946) but the resistant sublines were less responsive than the sensitive parents in this respect. Protein levels of other components of the UPR were evaluated by immunoblotting; BiP, phosphorylated-eIF2α, ATF6α and p58IPK were found to be unchanged in the resistant sublines. Cytoplasmic and secreted immunoglobulins were assayed by immunoblotting and ELISA respectively but no differences between the parent and resistant sublines were observed. In a small cohort of relapsed/refractory patients whose bone marrow biopsy was collected prior to bortezomib treatment, the quantity of intracytoplasmic light chain in the myeloma cells, as detected by flow cytometry, did not predict clinical response. The bortezomib-resistant myeloma sublines showed significant cross-resistance to doxorubicin and, to a lesser extent, vincristine, but none to melphalan. The cross-resistance is associated with adherent growth, as it was also observed in the adherent versions of the parent cell lines which were never exposed to bortezomib. Moreover, functional analysis of the multi-drug transporter activity, by drug accumulation assay, showed no difference between parent and resistant sub-lines. These results suggest that resistance to bortezomib is associated with downregulation of the IRE-XBP-1 pathway, but not necessarily other components of the UPR, consistent with observations by others concerning BiP (DT Rutkowski, et al., 2006. Plos Biology 4, e374). The association between adherence, bortezomib resistance and low XBP-1 requires further study.