It is well established that the bone marrow microenvironment contributes to the growth and survival of multiple myeloma plasma cells, however the molecular mechanisms that control these changes within the myeloma cell have not been completely elucidated. We have recently demonstrated that myeloma cell lines can be dependent on the anti-apoptotic protein MCL-1 or can be co-dependent on MCL-1 and Bcl-2/xL. The primary distinction between MCL-1 dependence and co-dependence is the distribution of the pro-apoptotic BH3-only protein Bim. In MCL-1-dependent lines, Bim is primarily associated with MCL-1. In contrast in co-dependent lines Bim is either predominantly associated with Bcl-2/xL or when it is released from Bcl-2/xL it can not bind to MCL-1 because of the presence of the MCL-1 inhibitor, Noxa. This renders these cells sensitive to the Bcl-2/xL inhibitor ABT-737. We have confirmed these findings in freshly isolated myeloma cells and demonstrated that among the 4 patient samples tested, Bim was associated with both MCL-1 and Bcl-xL and the cells were sensitive to ABT-737. This suggests that co-dependence on MCL-1 and Bcl-xL may be a common phenomena in myeloma. We now demonstrate that in about 50% of patient samples tested, CD138 column-purified MM cells are more sensitive to ABT-737 than cells that were not purified following ficoll separation of the bone marrow aspirate. This suggests that stromal components in the marrow alter the dependence on Bcl-2 proteins resulting in a cell that is more MCL-1 dependent when in the presence of stroma. To directly test this possibility we determined the effect of co-culturing the co-dependent line, MM.1s with the human stromal line Hs-5 on ABT-737 (0.6 mM) sensitivity and Bim binding. MM1.s cells were incubated with Hs-5 cells for 30 min prior to addition of drug or were cultured in medium containing 50% conditioned medium (48 h growth) from Hs-5 cells to determine if any effects observed were due to a soluble factor. Following a 24 h culture, cells were subjected to CD138 column purification and apoptosis analyzed by Annexin V-FITC/PI flow cytometry. While mono-cultured MM1.s were very sensitive to ABT-737 (88.5% +/− 3.6% control apoptosis, mean +/− SE), the addition of Hs-5 cells or conditioned medium had a profound effect on cell survival resulting in significant resistance to Bcl-2/xL inhibition (23% +/− 4.4% and 24.5% +/− 2.3% control apoptosis respectively). To determine the molecular basis in this shift from co-dependence to MCL-1 dependence, western blot and co-immunoprecipitation analyses were performed. While neither co-culture nor conditioned medium had any significant effect on the expression of levels of Bim, MCL-1 or Bcl-xL, both conditions resulted in a shift of Bim binding from being equally split amongst MCL-1 and Bcl-xL to a binding pattern where Bim was preferentially bound to MCL-1. This is consistent with loss of sensitivity to ABT-737 and a shift to MCL-1 dependence. We next tested the effects of Hs5 co-culture and conditioned medium on other inducers of cell death. We focused on bortezomib (5 nM) and arsenic trioxide (2 mM) as both can induce the MCL-1 inhibitor Noxa, therefore sensitivity would not be expected to be significantly influenced by a change from co-dependence to MCL-1 dependence. Consistent with this possibility we found that bortezomib induced cell death was only modestly affected by co-culture (79% +/− 2.3% vs. 66.5% +/− 1.9% control apoptosis) and no protection was observed with Hs5 conditioned medium. Similar results were observed with arsenic trioxide (monoculture – 31% +/− 2.3%, Hs5 co-culture −15% +/− 1.0%, conditioned medium – 30% +/− 4.2% control apoptosis). Taken together these data demonstrate that a soluble factor produced by stromal support can alter Bim binding to anti-apoptotic Bcl-2 family members, rendering myeloma cells primarily dependent on MCL-1 and protecting them from cell death signals that function through inhibition of Bcl-2/xL. Moreover these data in part explain why bortezomib can overcome survival signaling provided by stromal support, as proteasome inhibition is capable of inducing the MCL-1 inhibitor Noxa.
Kaufman:Millenium: Consultancy; Onyx Pharmaceuticals: Consultancy; Novartis: Consultancy; Keryx: Consultancy; Merck: Research Funding; Celgene: Research Funding. Boise:University of Chicago: Patents & Royalties.
Asterisk with author names denotes non-ASH members.