MAGE-a Mediate Resistance to Chemotherapy in Multiple Myeloma through Regulation of Bcl-2 Proteins

Type I Melanoma Antigen Genes of the MAGE-A family are commonly expressed in multiple myeloma (MM). Their frequency of expression is higher in patients who have relapsed after chemotherapy compared to newly diagnosed patients, with a significant association with progression of disease. MAGE were also associated with high-risk disease in gene expression profiling studies. They belong to the Cancer-Testis group of oncofetal genes that are normally restricted to germ cells and trophoblastic tissue, and are aberrantly expressed in a broad range of human cancers. Type I MAGE proteins partner with the RING domain protein Kap1 to form ubiquitin ligase complexes, and in laboratory models of solid tumors they ubiquitinylate cancer-associated targets involved in metabolism, autophagy, and metastasis. We demonstrated that expression of MAGE-A genes, especially MAGE-A3, was correlated with progression of disease in MM and that MAGE-A inhibit p53-dependent and independent apoptosis in MM cells. These results lead us to investigate the hypothesis that MAGE-A promote resistance to chemotherapy-induced apoptosis in MM cells. We performed gene expression profiling by RNAseq of primary tumor specimens from an open label phase 2 clinical trial of panobinostat, a pan-HDAC inhibitor approved for treatment of relapsed MM, in combination with lenalidomide and dexamethasone (NCT01651039). These GEP data were analyzed for differential expression based on progression free survival (PFS) < 90 days (short PFS) versus PFS > 90 days (long PFS). MAGE-A1 was the most overexpressed gene in the short PFS group compared to long (p<0.005), and this correlation was confirmed at the protein level (p<0.05). We then silenced MAGE-A expression in human myeloma cell lines (HMCL) by transduction of a MAGE-A-specific shRNA lentivirus, which resulted in nearly complete loss of MAGE-A protein at 24-48 hrs. HMCL were transduced with MAGE-A shRNA lenti or a negative control construct containing a non-target shRNA sequence, and untreated cells served as an additional negative control. Silencing of MAGE-A increased sensitivity to cell death induced by panobinostat. Similar results were seen with melphalan, an alkylating agent, and bortezomib, a proteasome inhibitor, both of which are approved for treatment of MM. These results indicate that MAGE-A play a critical role in resistance to chemotherapy-induced apoptosis in relapsed myeloma patients. To investigate the mechanisms of inhibition of apoptosis, we analyzed the impact of MAGE-A on Bcl-2 family members, including Bim, Bid, and PUMA, BH3-only family members that initiate apoptosis, Mcl-1, which plays a role in chemotherapy-induced apoptosis in myeloma cells, and the prototypical survival proteins Bcl-2 an Bcl-xl. HMCL were transduced with MAGE-A shRNA lenti or controls and cells were harvested after disappearance of MAGE-A protein. Heavy membrane preparations were made from these cells and probed by Western blot for Bcl-2 proteins. Loss of MAGE-A resulted in higher levels of Bim compared to controls. Of note, the short isoform of Bim (Bim S), which is the most potent activator of apoptosis, demonstrated relatively higher expression compared to the long and extra long isoforms (Bim L and EL). Phosphorylation of Bim at Ser69 and 77 by Erk1/2 and other kinases leads to ubiquitinylation and proteasomal degradation. In the absence of MAGE-A, phosphorylated Bim is stabilized, suggesting MAGE-A inhibits apoptosis by promoting ubiquitinylation and degradation of phospho-Ser^69/77-Bim. These results demonstrate that MAGE-A/Kap1 is a key regulator of Bim that promotes survival of MM cells. Loss of MAGE-A also resulted in increased protein levels of Mcl-1 and decreased protein expression of Bcl-2. No significant changes in Bid or Bcl-xl proteins were observed and PUMA was not detected under these conditions. These results support a model in which MAGE-A mediates resistance to panobinostat and other chemotherapeutic agents through differential regulation of protein expression of Bim, Mcl-1, and Bcl-2. Therefore, MAGE-A is a promising novel therapeutic target, and inhibition of its anti-apoptotic activity may directly induce apoptosis or increase sensitivity to chemotherapy-induced apoptosis.