Establishment of a Novel CNS Infiltrated Xenograft Model through Engraftment of Patient-Derived Acute Lymphoblastic Leukemic Cells Into NOD/SCID/γc ^Null Mouse

Acute lymphoblastic leukemia (ALL) is the most common type of childhood hematologic malignancy. Although improvements in treatment regimen have raised the 5-year survival rate as high as 80% for pediatric ALL patients, a minority of patients with various risk factors, including central nervous system (CNS) infiltration continue to have poor prognosis. Recently, bone marrow (BM) microenvironments which support leukemic stem cells have become noticed as an important element which can influence treatment response and relapse of the disease. Although leukemic cells appear to be completely eradicated through treatment, they are thought to survive within bone marrow and/or extramedullary microenvironments, such as CNS, causing disease recurrence. However, little is known about the CNS microenvironment for leukemic cells because of the lack of appropriate animal model. Even though several investigators have tried to establish a CNS infiltrated model of leukemia, major limitation with these studies are the use of leukemic cell lines and the preconditioning of recipient mice, which did not represent CNS leukemia observed in patients. Here we report the establishment of a novel xenograft model for primary human ALL using NOD/SCID/γc null (NOG) mouse. Without irradiation, this model recapitulates CNS as well as extramedullary leukemic infiltration (hereby referred to as the h-leukemic NOG model).

Primary bone marrow samples were collected from 9 children with ALL at the time of diagnosis with informed consent. The leukemic cells (1×10⁶cells) were injected into the tail veins of non-irradiated 8- to 10-week old NOG mice. Primary samples from 8 out of 9 patients were successfully engrafted. Engrafted leukemic cells could be serially transplanted into secondary, tertiary and quaternary recipients. Morphological and FACS analyses revealed as high as 95% BM chimerism and showed that blast phenotypes were conserved through serial transplantations. Of note, extramedullary organs including the CNS, liver, spleen, and kidneys showed the leukemic invasion consistent with those of the donor ALL patients. Liver pathology in the h-leukemic NOG model is identical to that seen in the ALL patients. We also showed the existence of a functional niche in the liver mediated by SDF-1/CXCR4 axis. In terms of the CNS involvement, we observed the progressive infiltration of leukemic cells into the Virchow-Robin space that is consistent with the pathology of human ALL patients. Using this model, we examined the mechanism of dissemination and harboring of leukemic cells in the CNS niche.

NOG mice model for engraftment of human leukemic cells provides useful insights into the biology of ALL and allows us to answer various questions concerning the mechanism of extramedullary invasion and expansion. We have reported that NOG mice have significantly better human hematopoietic cell engraftment in the BM and extramedullary organs than other immunodeficient mice (Hiramatsu H. Blood. 2003), and is capable of supporting the growth of human neoplastic cells (Kato M. Nature. 2009). Here we report that this non-preconditioned mouse xenograft model reproduces leukemic extramedullary involvement, including the CNS, in sustaining leukemic cells. This approach provides a more sophisticated and physiological model suitable for the evaluation of molecular interactions between patient leukemic cells and host niche. Our h-leukemic NOG model will provide a powerful tool to analyze the CNS niche that harbors leukemia initiating cells. Moreover, this model would be a useful platform for developing novel anti-leukemic therapies that target CNS extramedullary niche.

No relevant conflicts of interest to declare.