Introduction: The bone marrow microenvironment (BMM) is essential for hematopoiesis, including granulopoiesis which mostly comprises the production of neutrophils. Due to the relatively short lifespan (1-3 days) of neutrophils in the circulation, bone marrow (BM) failure is manifested first by neutropenia. As such, models which recapitulate human granulopoiesis can provide useful tools in the drug development pipeline to assess efficacy or toxicity. Most BM models to study granulopoiesis include animal models and simple 2D tissue models which do not fully recapitulate the mechanisms of granulopoiesis. In this work we have developed a human microphysiological system (MPS) of the human BMM that recapitulates granulopoiesis. Our model incorporates endothelial cells (ECs), bone marrow stromal cells (BMSCs), and hematopoietic stem/progenitor cells (HSPC, CD34+) cells to mimic the BMM and produce neutrophils for at least 21 days. Our model provides a tool for studying granulopoiesis with applications in drug discovery and personalized medicine.
Materials and Methods: ECs were isolated from cord blood and cultured on 1% gelatin in EGM-2 (Lonza). CD34+ (Lonza) were thawed the day of use. BMSCs (Lonza) were cultured in Myelocult5100 supplemented with antibiotics. All cells were maintained at 37.5 oC in 5% CO2. ECs (1×107 cells/mL), BMSCs (1×107 cells/mL) and CD34+ (1×106 cells/mL) cells were mixed in fibrinogen (10 mg/mL) and then thrombin (5 U/mL) prior to loading into the central chamber of the microfluidic device. Devices were cultured for 21 days and fed a 1:1 blend of SFEM StemSpan/CD34+ Expansion Supplement (1X) + EGM-2. Cells exiting the central chamber (egressed cells) were collected every two days through days 7-21, fixed in 1% formalin and stained (DAPI). Devices were fixed and stained on day 21 (CD31).
Results, Conclusions, and Discussions: Our microfluidic model of the BMM achieved a perfusable vascular network within 7 days post-loading (Fig 1A), which contributes to the microenvironment necessary for granulopoiesis. We observed cells egressing from the central chamber (BMM) (Fig 1B), with a nuclear morphology consistent with band and mature neutrophils (Fig. 1C). We collected and counted the egressed cells from the bone marrow model through days 7-21 in the presence and absence of IL-1b (Fig 1D). IL-1b initially reduced PMN egress, but later (day >17) enhanced PMN egress. Under control conditions, our BMM produced 2.5 PMNs/day per HSPC seeded into the BMM, and approximately 2.5x105 PMNs/mL tissue/day.
Conclusions: We have created a 3D in vitro human MPS of the BMM that recapitulates human granulopoiesis. The model produces neutrophils for an extended period of 21 days without exhausting the HSPC pool. The volume of our model is a scant 200 nl, approximately 7 orders of magnitude smaller than in vivo bone marrow in an adult human. The MPS model responds to a pro-inflammatory stimulus and has the potential for enhanced production of PMNs by seeding more CD34+ cells. Our MPS model recapitulates granulopoiesis, a key feature of the BMM, and has the potential to investigate mechanisms of granulopoiesis and enhance efficiency of drug development (efficacy and toxicity).
Acknowledgements (Optional): NIH grant R01 EB030410 and R01 EB030410S1.