ERN1/IRE1 and XBP1 Are Dispensable for Myeloma Cell Survival but Are Required for the Anti-Myeloma Activity of Bortezomib

Abstract 986

We have previously conducted kinome- (Tiedemann et al., Blood 2010) and genome- (Zhu et al., Blood 2011) wide small interfering RNA (siRNA) studies in human myeloma cells in the absence and presence of titrated bortezomib (BTZ) in order to identify genes that modulate BTZ sensitivity or resistance. From high throughput RNAi screening, endoplasmic reticulum-to-nucleus-signaling-1 (ERN1), the human homolog of yeast IRE-1, was identified with multiple siRNA as the kinase most potently required for BTZ activity, with genetic silencing associated with BTZ resistance. ERN1/IRE1 is an endoplasmic reticulum (ER)-resident transmembrane protein and a core component of the unfolded protein response (UPR). By gene expression profiling and western blot IRE1 is ubiquitously expressed across primary myeloma tumors and a panel of 14 human myeloma tumor lines (HMCL). Significantly, silencing of IRE1 with lentiviral shRNA induced BTZ resistance in 9/9 HMCL. IRE1 silencing in the absence of BTZ was surprisingly well tolerated in the majority of HMCL tested, with no loss of viability; however in 2/9 HMCL (JJN3, 8226) silencing of IRE1 resulted in reduced proliferation. Multiple RNAi yielded identical results. In response to unfolded protein accumulation within the ER lumen, IRE1 oligomerizes and auto-phosphorylates to reveal an endonuclease activity that splices X-box protein 1 (XBP1) mRNA to activate the potent bZIP transcription factor, XBP1s. XBP1 is essential for physiologic differentiation of B cells to plasma cells and its absence results in profound plasma cell deficiency (Reimold et al., 2001). Despite it's essential role in plasma cell differentiation, we find that XBP1, like IRE1, is redundant for myeloma cell survival under standard growth conditions. Lentiviral knockdown of XBP1 using multiple distinct RNAi failed to induce HMCL cytotoxicity; and resulted in reduced growth in only 2/9 HMCL, closely mirroring results observed with IRE1 inhibition. Significantly, XBP1 silencing also closely reproduced the BTZ resistance observed with IRE1 silencing. As plasmacytoma in patients grow under hypoxic conditions we next evaluated myeloma cell growth in response to hypoxia (0.2% O2) and concurrent UPR signaling disruption. At 48hrs, viability was comparable in IRE1 knockouts and parental cells. In contrast, HMCL with IRE1 silencing showed marked vulnerability to protein synthesis inhibition induced by puromycin, with selection against IRE1 silenced cells in mixed cultures. Notably, ER stress is signaled by 3 different transmembrane transducers (IRE1, PERK and ATF6), each contributing to distinct UPR pathways. As induced BTZ resistance from IRE1 or XBP1 inhibition might conceivably result either from loss of a death signal transmitted via IRE1-XBP1 following PI-induced ER stress, or from a compensatory increase in PERK or ATF6 homeostasis signaling in response to IRE1-XBP1 silencing, we generated single, double and triple knockdown HMCL with suppressed expression of 1, 2 or all 3 ER stress transducers (IRE1, PERK, ATF6). Individual silencing of either IRE1 or PERK in HMCL had no effect on levels of other ER stress transducers by western blot, however silencing of ATF6 caused compensatory increases in PERK while XBP1 silencing resulted in reduced IRE1. Importantly, HMCL lacking IRE1 and PERK or ATF6, showed no loss of BTZ resistance induced by either IRE1 or XBP1 knockdown, demonstrating that secondary increases in signaling via these alternative ER stress pathways do not account for the BTZ resistance induced by IRE1-XBP1 pathway inhibition. Consistent with the hypothesis that increased UPR signaling (via IRE1-XBP1) causes myeloma cell death, knock down of the ER luminal chaperone BiP (HSPA5), which negatively regulates all 3 stress transducers, proved cytotoxic in multiple HMCL. Overall, these data suggest that the anti-myeloma activity of BTZ involves a specific IRE1-XBP1 death signal generated in response to proteasome inhibition and unfolded protein ER stress, rather than ER failure from unfolded protein. As IRE1-XBP1 signaling is dispensable for myeloma cell survival under many growth conditions, targeted small molecule inhibitors of IRE1 currently under development may prove ineffective as single agent therapy for myeloma; moreover such agents may prove antagonist to PI therapy. Finally, disruption of the IRE1-XBP1 axis in primary tumors may mediate bortezomib resistance in patients.