Research Article: Alkylator-Induced and Patient-Derived Xenograft Mouse Models of Therapy-Related Myeloid Neoplasms Model Clinical Disease and Suggest the Presence of Multiple Cell Subpopulations with Leukemia Stem Cell Activity

Date Published: July 18, 2016

Publisher: Public Library of Science

Author(s): Brian A. Jonas, Carl Johnson, Dita Gratzinger, Ravindra Majeti, Connie J Eaves.

http://doi.org/10.1371/journal.pone.0159189

Abstract

Acute myeloid leukemia (AML) is a heterogeneous group of aggressive bone marrow cancers arising from transformed hematopoietic stem and progenitor cells (HSPC). Therapy-related AML and MDS (t-AML/MDS) comprise a subset of AML cases occurring after exposure to alkylating chemotherapy and/or radiation and are associated with a very poor prognosis. Less is known about the pathogenesis and disease-initiating/leukemia stem cell (LSC) subpopulations of t-AML/MDS compared to their de novo counterparts. Here, we report the development of mouse models of t-AML/MDS. First, we modeled alkylator-induced t-AML/MDS by exposing wild type adult mice to N-ethyl-N-nitrosurea (ENU), resulting in several models of AML and MDS that have clinical and pathologic characteristics consistent with human t-AML/MDS including cytopenia, myelodysplasia, and shortened overall survival. These models were limited by their inability to transplant clinically aggressive disease. Second, we established three patient-derived xenograft models of human t-AML. These models led to rapidly fatal disease in recipient immunodeficient xenografted mice. LSC activity was identified in multiple HSPC subpopulations suggesting there is no canonical LSC immunophenotype in human t-AML. Overall, we report several new t-AML/MDS mouse models that could potentially be used to further define disease pathogenesis and test novel therapeutics.

Partial Text

Acute myeloid leukemia (AML) is an aggressive bone marrow malignancy characterized by the accumulation of immature myeloid cells with defective maturation and function. AML is a heterogeneous disease and is classified by the World Health Organization into several subtypes on the basis of cytogenetic, molecular, and phenotypic characteristics [1]. Therapy-related myeloid neoplasms (t-MNs), consisting of therapy-related AML (t-AML) and therapy-related myelodysplastic syndrome (t-MDS), are one such subtype accounting for 10–20% of AML cases and occur in patients previously treated with radiation and/or chemotherapy for other diseases [2]. t-AML/MDS is typically diagnosed 5–7 years after previous treatment, and the t-AML phase can be preceded by a t-MDS phase characterized by cytopenias related to bone marrow failure and less than 20% bone marrow blasts [3, 4]. t-AML/MDS is clinically characterized by deletions in chromosomes 5 and/or 7 in nearly 70% of cases and by a distinct set of recurrent molecular mutations, including TP53 [3, 5–8]. TP53 mutations are likely an early event in the pathogenesis of these diseases [6, 9, 10]. While de novo AML is associated with a 30–40% 5-year overall survival (OS) with current standard therapies, t-AML/MDS has an even worse prognosis, with a 5-year OS of less than 10% [3, 4].

t-AML/MDS is a distinct subgroup of AML that develops years after exposure to chemotherapy and ionizing radiation. These diseases are characterized clinically by recurrent chromosomal alterations and gene mutations, and carry a dismal prognosis with current treatments. Overall, less is known about t-AML/MDS pathogenesis compared to de novo disease, and relatively few t-AML/MDS mouse models have been described. Here, we describe multiple mouse models of t-AML/MDS with features consistent with clinical disease. We first developed models of alkylator-induced t-AML/MDS by treating wild type adult DBA/2J and SWR/J mice with ENU and established a diagnosis of AML or MDS in nineteen mice using the rigorous diagnostic criteria proposed by the MMHCC. We demonstrated that these models have clinical and pathologic characteristics consistent with human t-AML/MDS and are therefore relevant models to study human disease. We also established xenograft models derived from three primary human t-AML samples. These models were characterized by lethality to recipient mice and LSC immunophenotype that varied among the primary t-AML cases.

 

Source:

http://doi.org/10.1371/journal.pone.0159189