Biology Of Mixed Phenotype Acute Leukemia In Successful Long-Term Cytokine-Free Three-Dimensional (3D) Static and Perfused 3D Hollow-Fibre Bioreactor Culture

In vitro Acute Myeloid Leukemia (AML) models are currently based on cell lines or, if primary cells are used, require co-culture or supplementation with abnormally high concentrations of exogenous cytokines (EC). These conditions introduce bias to the culture and fail to recapitulate the 3D environment integral to leukemic growth in bone marrow (BM). We have previously shown that human AML cell lines can be grown in 3D-polyurethane scaffolds (3D-PU) best when collagen-coated for 8 weeks in the absence of EC. This same cytokine-free platform supported proliferation and clonogenic capacity of human cord blood mononuclear cells (MNCs) for up to 6 weeks in static conditions (SC), with more rapid growth when cultured in a novel perfused 3D hollow-fibre bioreactor (PHFB), which utilizes permeable membranes embedded within the coated 3D-PU for the selective mass transport of nutrients/metabolites. Herein, we evaluated the potential for human primary de novo Mixed Phenotype Acute Leukemia (MPAL) cells to survive in culture in the absence of EC over 35d, within both collagen-coated 3D-PU scaffolds in SC and PHFB. MPAL cells were obtained from peripheral blood (PB) and BM aspirate samples after informed patient consent. Two distinct blast populations were found: (1) positive for CD33/13/MPO and negative forCD19/10/34/117/79a/TDT and (2) positive for CD19/10/34/79a/TDT and negative for CD33/MPO. MNCs were seeded at a density of 1.6 x 10⁷ cells/cm³ within: (1) single scaffolds (5x5x5 mm³) performing full medium exchange every other day in SC, (2) PHFB (LxD: 141.5x9.7mm) with continuous perfusion of nutrients and removal of metabolites (VFR= 0.12 ml/min), both cultivated without EC for 35d. A standard two-dimensional (2D) Dexter culture was used in parallel as a control. Both PB-PU and BM-PU SC were run in parallel and monitored weekly over 35d for viability, proliferative capacity, and cellular phenotype; the PHFB was also run at the same time and sacrificed at d35 for analysis. Although cell viability decreased in SC over the first 2 weeks in both PB (59.2%) and BM (54.6%), viability increased again from d21 to d35, PB-PU (74.4%) > BM-PU (64.1%) = PB-PHFB (65%). In contrast, leukemic cells (LCs) cultured in 2D-flasks did not survive beyond day 21. In situ growth kinetics of both PB and BM LCs showed a similar trend in 3D-PU, peaking at d14-21 with no statistical differences along the 35d. In situ scanning electron microscopy showed that cellular density increased with culture time in SC and that LCs at d35 appeared organized in dense colonies and clusters within the pores of the scaffolds, particularly in the PB-PHBR. At d35, leukemic blasts similar in appearance to those of the original patient BM sample were observed in BM-PU, PB-PU scaffolds, and the PB-HFB core. Immunophentoyping by flow cytometry on cells extracted from all culture conditions showed that a leukemic blast population (90.9% in PB-HBR, 84.3% in PB-PU and 85.1% in BM-PU) was present in each, which was positive for CD19/CD10/CD34 and negative for CD13/TDT. Interestingly, this specific phenotype was not displayed in the original input cell populations of multilineage blasts, indicating potential clonal progression or selection of a previously non-dominant clone occurring within both SC and PHBR platforms simultaneously. To evaluate HFB performance and AML biology, bioprocess parameters (glucose, lactate, glutamine, glutamate, ammonia and pH) and cytokines were measured. Glucose was not completely consumed whereas lactate, glutamate and ammonia were produced at low and non-toxic levels throughout the 35 day culture; pH was physiologic and maintained at 7.38 (STD 0.07). IL-6 and G-CSF concentrations were high throughout the culture peaking at day 21-28 whereas IFNα2, IFN-γ, IL-10, TNFα and TGFα were all detected at low concentrations. Flt-3L was only produced at the end of culture whereas the cytokines GM-CSF, IL-1β, IL-3, IL-4, IL-12(p70), TNFβ, VEGF and EPO were not detected. In conclusion, these novel 3D in vitro static and perfusion bioreactor platforms successfully supported human MPAL cultivation without the use of EC or allogeneic cells, removing the introduction of bias in the study of leukemogenesis and its microenvironment. These platforms may also be used to evaluate culture fidelity and clonal evolution, to provide insight into AML biology, and as a biomimetic tool for drug discovery and chemotherapy.