Multiple myeloma (MM) is the second most prevalent hematological malignancy, and it remains incurable with a median survival of 3 to 5 years. Despite the introduction of several novel drugs, only 60% of the patients respond to therapy, and more than 75% of patients relapse after 5 years. The interaction of MM cells with extracellular matrix (ECM) and stromal cells in the bone marrow (BM) milieu was shown to play a crucial role in MM progression and drug resistance. Therefore, an appropriate model that demonstrates the complex interactions between MM cells and their environment and predicts the response to treatment of MM patients is warranted. The goal of this study is to produce an in vitro model of BM microenvironment in MM, which will mimic the BM in the patient and have closest prediction of cell-behavior in vivo.


Scaffolds for the 3D tissue-engineered BM (3DBM) were prepared by crosslinking BM supernatants from MM patients with calcium chloride in the presence of the antifibrinolytic tranexamic acid. The concentration of the different components for scaffold formation and physico-chemical properties of the scaffold were optimized. Then, we incorporated cellular components (MM cell lines and stromal cell lines) and ECM components (fibronectin) to 3DBM scaffold. We tested the effect of different MM cell densities, fibronectin concentration, and presence of different densities of stromal cells in the scaffold on the proliferation of MM cells, which was analyzed by flow cytometry. After optimization of the cellular and ECM components, we tested the sensitivity of MM cells to chemotherapeutic agents used in MM, including melphalan, bortezomib and carfilzomib, in 3DBM scaffolds with and without the presence of stromal cells. We assessed the sensitivity of MM cells to the drugs in the 3DBM scaffolds compared with the sensitivity in two-dimensional (2D) classic in vitro tissue culture system, with or without the presence of stromal cells.


We optimized the conditions for the production of a 3DBM from BM supernatants from MM patients. We found that calcium concentration was a critical factor for the formation of the 3DBM scaffold; and 1mg/ml calcium chloride was the optimal concentration for fastest and most stable scaffold formation. Moreover, we found that the presence of an antifibrinolytic agent was essential for the stabilization of the scaffold, and 4mg/ml of tranexamic acid was sufficient to stabilize the scaffold.

We optimized the conditions for seeding of the cellular components and found that MM cells proliferated in the 3DBM scaffolds better than in regular 2D tissue culture plates, in which the expansion rate over 3 days was about 5-folds in the 3DBM, compared to about 2.4-folds in 2D classic in vitro tissue culture system. In addition, we optimized the co-culture condition of MM cells with stromal cells and found that the ratio of MM/stroma of 3/1 generated the highest proliferation of MM cells in the 3DBM scaffolds. Similarly, the addition of ECM components (fibronectin) increased the proliferation of MM in the 3DBM by about 30%.

In addition, we tested the effect of the 3DBM scaffold on drug resistance of MM cells. In a 2D classic in vitro tissue culture system approximately about 40-50% of the MM cells were killed using the different chemotherapies, the culture of MM cells in 3DBM decreased the sensitivity of MM cells to treatment, and addition of stromal cells to the co-culture decreased the sensitivity of MM cells to treatment in both 3D and 2D. We found that the co-culture of MM cells with stromal cells in the 3DBM treatment and using the same concentrations of the different chemotherapies killed only 5-15% of the MM cells.


We produced and characterized a 3D scaffold made from supernatant of BM aspirates from MM patients. This 3DBM is all made from patient material (autologous), no synthetic scaffolds were used, and it could represent the closest in vitro model to the BM in MM. The 3DBM induced drug resistance in MM cells, especially when it contained stromal cells. The autologous 3DBM seems to be closer to the real BM microenvironment, therefore, a better drug screening method than 2D classic in vitro tissue culture systems.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.