Predicting the response of multiple myeloma to the proteasome inhibitor Bortezomib by evaluation of the unfolded protein response

Multiple myeloma (MM) remains a predominantly incurable malignancy despite high-dose chemotherapy, autologous stem cell transplant and novel agents. Proteasome inhibitors (PI) such as Bortezomib have increased the response rate and survival of patients with MM. The overall patient response rate of newly diagnosed MM to Bortezomib and Dexamethasone is about 67%. In relapsed refractory MM, the response rate is reduced to about 40–60%. Therefore, there are a significant number of MM patients who are resistant to Bortezomib. Combining Bortezomib with another class of novel drugs, for example, immunomodulatory drugs (IMIDs), such as Lenalidomide, is associated with increased overall response rate of 94% in newly diagnosed myeloma patients and 64% in relapsed or refractory myeloma. Although the combination of a proteasome inhibitor and an IMID may yield an improved response rate, it is not always possible due to the cost, availability, local regulatory policies, side-effects profile, convenience and personal preference. Therefore the choice of novel agents (PI or IMID) is predominantly empirical and based on other factors such as side effects and tolerability, making it difficult to choose the best therapy. Currently there is no way of predetermining if a patient will respond to Bortezomib treatment. However, previous studies have shown that XBP1, a key regulator of the unfolded protein response (UPR), predicts sensitivity to Bortezomib, and its level correlates proportionally with sensitivity to Bortezomib. We therefore aimed to assess if the sensitivity to Bortezomib is dependent on the UPR, and that the expression level of ATF6 mRNA and the size of the endoplasmic reticulum can predict sensitivity to the drug. ATF6 is a regulator of the UPR and is capable of activating XBP1, which is a regulator of the UPR and correlates with Bortezomib sensitivity. Previous studies have shown that the protein expression of ATF6 is reduced in MM cell lines resistant to Bortezomib compared with their parent cell line. We therefore analysed ATF6 gene expression in Bortezomib sensitive and resistant KMS11 cells (Supplementary Information). Our results showed that ATF6 gene expression decreased with increasing Bortezomib resistance. KMS11 cells resistant to Bortezomib were seen to have substantially lower ATF6 mRNA levels compared with parent sensitive cells (Figures 1a, P= 0.06), resembling the same trend as seen in protein expression. These results were also seen in 45 MM patients with various levels of resistance (Supplementary Information). Patient responses after completion of cycle 2 of therapy with Bortezomib were categorized according to the International Myeloma Working Group (IMWG) uniform response criteria. Complete response (CR), very good partial response (VGPR) or partial response (PR) patients had significantly higher levels of ATF6 mRNA compared with patients with stable disease (SD) or of progressive disease (PD; Figure 1b; P= 0.007). The mean ATF6 expression of the CR, VGPR and PR patient groups were 3.92-fold higher compared with patients of SD and PD groups. On an individual group basis, there was no significant difference between each group, however there was a significant difference between PR patients vs SD+PD patients (Figure 1b; P = 0.01). Nonetheless, gene expression levels of ATF6 in SD+PD patients were significantly lower than those seen in patients with CR+VGPR