Leukemia stem cells in a genetically defined murine model of blast-crisis CML

Abstract

Myeloid leukemia arises from leukemia stem cells (LSCs), which are resistant to standard chemotherapy agents and likely to be a major cause of drug-resistant disease and relapse. To investigate the in vivo properties of LSCs, we developed a mouse model in which the biologic features of human LSCs are closely mimicked. Primitive normal hematopoietic cells were modified to express the BCR/ABL and Nup98/HoxA9 translocation products, and a distinct LSC population, with the aberrant immunophenotype of lineage(-), Kit(+/-), Flt3(+), Sca(+), CD34(+), and CD150(-), was identified. In vivo studies were then performed to assess the response of LSCs to therapeutic insult. Treatment of animals with the ABL kinase inhibitor imatinib mesylate induced specific modulation of blasts and progenitor cells but not stem- cell populations, thereby recapitulating events inferred to occur in human chronic myelogenous leukemia (CML) patients. In addition, challenge of leukemic mice with total body irradiation was selectively toxic to normal hematopoietic stem cells (HSCs), suggesting that LSCs are resistant to apoptosis and/or senescence in vivo. Taken together, the system provides a powerful means by which the in vivo behavior of LSCs versus HSCs can be characterized and candidate treatment regimens can be optimized for maximal specificity toward primitive leukemia cells.

Figure 1

Characterization of leukemia models. (A) Example of bone marrow derived from a primary recipient of coinfected cells. The BCR/ABL-transduced cells are detected by GFP expression and the Nup98/HoxA9-infected cells are detected by YFP expression. Panels I and II indicate singly infected populations, whereas panel III shows cells that are successfully transduced with both the BCR/ABL and Nup98/HoxA9 vectors. (B) Example of bone marrow derived from a secondary recipient of coinfected cells. The doubly transduced cells typically outgrow the singly infected populations shown in panel A. (C) GFP⁻/YFP⁻ cells from the secondary recipient in panel B were analyzed with respect to Ly5.1 versus Ly5.2 expression as a means to verify host origin and to establish that potential donor leukemic cells have not lost GFP/YFP expression (host cells in this experiment were of Ly5.2 origin). Fidelity of the system is demonstrated by greater than 99% purity of the host marker.

Figure 2

Kaplan-Meier analysis of survival for CML and bcCML models. (A) Donor populations of purified HSCs, CMPs, or GMPs were infected with the BCR/ABL retrovirus and transplanted into primary mice. (B) Donor populations of purified HSCs, CMPs, or GMPs were infected with the BCR/ABL + Nup98/HoxA9 retroviruses and transplanted into primary mice. Numbers of recipients indicated in parentheses for each donor cell type. Days of survival after transplantation are indicated for each cell type on the x-axis.

Figure 3

Southern-blot analysis of bcCML. Bone marrow DNA from 5 primary recipient mice (left panel) was digested with EcoRI and hybridized with a GFP probe, which detects bands of unique size for each independent viral integrant. For primary animals nos. 2, 3, and 4, 3 secondary recipients (a, b, and c) were established. The right panel shows EcoRI digestion and GFP probe analysis of the secondary recipients in comparison to each primary donor.

Figure 4

Phenotypic analysis of normal versus leukemic stem-cell populations. Examples of stem-cell phenotyping are shown for GFP control (top left), CML (top right), and bcCML (bottom panels) populations. In each case, cells were first gated on the Lin⁻ population. GFP and CML populations show typical labeling for stem-cell markers c-kit and sca-1. Labeling of bcCML shows reduced c-kit expression (bottom left). The CD150⁻ and Flt3⁺ subpopulation of bcCML cells (bottom right) is highly enriched for LSCs (1 in 7 cells by limiting dilution). Percentages shown in each panel represent the overall frequency in the marrow population.

Figure 5

Phenotypic and in vitro analyses of irradiated normal versus leukemic animals. (A) Bone marrow cells were isolated from control versus irradiated (5.5 Gy) animals and labeled with antibodies to detect stem-cell populations (indicated by the box in each panel). The top panels show GFP⁻ (ie, normal) cells and the bottoms panels show GFP⁺ (ie, leukemic) cells. The data shown for each panel are gated on the Lin⁻ population. (B) Colony-forming units (CFUs) were determined for sorted GFP⁺ (leukemic) and GFP⁻ (normal) populations. Data shown are the absolute number of colonies measured per 5000 cells plated (triplicate assays). *Statistical significance(P = .009) as determined by Student t test. Error bars are SD.