Abstract
BACKGROUND: Inoculation of human normal or leukemic myeloblasts into sublethaly irradiated NOD/SCID mice often results in persistent low-level engraftment (< 5%), but significant proliferation (≥ 5-fold expansion) rarely occurs. Most malignant samples that engraft and proliferate are of FAB M4 subtype and exhibit rapid extramedullary growth at the site of injection without significant marrow or spleen involvement. We hypothesized that low engraftment and proliferation of less mature FAB subtypes results from an increased requirement of these cells for a marrow environment of cytokine and contact-dependent growth and survival factors not adequately provided across species by the mouse bone marrow stroma. Here we show that the subcutaneous injection of minimally-differentiated human mesenchymal stem cells (MSC) in a Matrigel matrix creates an artificial human marrow environment resulting in improved survival and proliferation of human myeloblasts.
METHODS: Human leukemic myeloblasts were obtained from the marrow or peripheral blood of 14 newly diagnosed pediatric patients under an IRB-approved collection and banking protocol. MSC were obtained from sterile filters following processing of human marrow from healthy donors or from the NIH-funded MSC bank at Tulane University. 6-to-12 week old NOD-SCID mice were injected IV with 5x106 AML blasts via the retro-orbital sinus (N=38), subcutaneously in 0.5mL Matrigel (N=18), or subcutaneously with 5x105 MSC in 0.5mL of Matrigel (N=14). Mice were euthanized when evidence of tumor burden was present. Peripheral blood, bone marrow, spleen, and subcutaneous nodules were obtained for flow immunophenotyping, FISH, and histopathology. Percent engraftment was determined by flow cytometry for human CD33-APC and mouse H2Kd-PE.
RESULTS: Median time from injection to necropsy was 12.5 weeks. 18% died of spontaneous murine thymomas. No animals died of progressive human AML if myeloblasts were injected IV or subcutaneously with Matrigel, and all had < 5% involvement of bone marrow, spleen, and blood. Six animals injected with AML and MSC (43%) developed visible tumors at a median of 8.5 weeks. These tumors were easily reduced to single cell suspensions of > 98% CD33+ by flow cytometry, with mean estimated recovery of 1.3x108 human myeloblasts per mouse tumor (mean 36-fold expansion, range 4 to 52-fold). For cases in which the AML and MSC were derived from subjects of disparate gender, the origin of the cells (leukemic donor vs. MSC donor) was validated by FISH for human X/Y chromosomes. Histopathology of the resulting mass revealed the central development of a stromal chondroid matrix similar to trabecular bone. Marrow, spleen, and blood for all these animals contained < 5% human myeloblasts.
CONCLUSIONS: Here we describe an effective method for expanding immature human leukemic myeloblasts in the NOD/SCID mouse. These findings suggest that less mature myeloblasts require human MSC for survival and proliferation and appear to lack significant homing to or expansion in mouse marrow even in the presence of a significant ectopic tumor burden. This is a useful technique for expanding human AML cells for research, may be a model for more broad-based patient-oriented testing of chemotherapeutic and biologic therapies for AML, and represents a novel animal model for studying the stromal interactions and growth requirements of malignant and non-malignant myeloid precursors.
Author notes
Corresponding author
