A robust and rapid xenograft model to assess efficacy of chemotherapeutic agents for human acute myeloid leukemia

Relevant preclinical mouse models are crucial to screen new therapeutic agents for acute myeloid leukemia (AML). Current in vivo models based on the use of patient samples are not easy to establish and manipulate in the laboratory. Our objective was to develop robust xenograft models of human AML using well-characterized cell lines as a more accessible and faster alternative to those incorporating the use of patient-derived AML cells. Five widely used AML cell lines representing various AML subtypes were transplanted and expanded into highly immunodeficient non-obese diabetic/LtSz-severe combined immunodeficiency IL2Rγcnull mice (for example, cell line-derived xenografts). We show here that bone marrow sublethal conditioning with busulfan or irradiation has equal efficiency for the xenotransplantation of AML cell lines. Although higher number of injected AML cells did not change tumor engraftment in bone marrow and spleen, it significantly reduced the overall survival in mice for all tested AML cell lines. On the basis of AML cell characteristics, these models also exhibited a broad range of overall mouse survival, engraftment, tissue infiltration and aggressiveness. Thus, we have established a robust, rapid and straightforward in vivo model based on engraftment behavior of AML cell lines, all vital prerequisites for testing new therapeutic agents in preclinical studies.


INTRODUCTION
Acute myeloid leukemia (AML) is the most common adult acute leukemia and is characterized by clonal expansion of immature myeloblasts, initiating from rare leukemic stem or progenitor cells. In Europe and the USA, the incidence and mortality rates of AML are about 5 to 8/100.000 and 4 to 6/100.000 per year, respectively. 1 Despite a high rate of complete remission after treatment with genotoxic agents, the relapse rate remains very high and the prognosis very poor. Overall survival at 5 years is 30-40% in patients younger than 60 years, and o 20% in patients over 60 years. Front-line chemotherapy is highly effective in ablating leukemic cells, but distant relapses are observed in the majority of patients, characterized by a refractory phase during which no other treatment has shown any efficacy thus far. Relapses are caused by malignant cell regrowth initiated by chemoresistant leukemic clones. 2,3 This unfavorable situation leads to a strong need for new therapeutic strategies, as well as relevant preclinical mouse models in which to test them. We have previously established a robust xenotransplantation model to study primary human AML biology and stem cell function based on the engraftment of primary AML samples into non-obese diabetic (NOD)/LtSz-severe combined immunodeficiency (SCID) IL2Rγ c null (NSG) mice. 4,5 These highly immunodeficient mice have the advantages of a longer life span and higher levels of engraftment of human AML cells compared with other immunocompromised mouse strains such as SCID, NOD-SCID. 4,6,7 However, this model is based on three-to-six month-long experiments with a large excess of cells and requires access to a biobank of diverse primary AML patient samples. Moreover, extensive characterization of this xenograft model with more commonly used and easily accessible AML cell lines has not been reported to date in the same setting to compare models in vivo. Herein, we report experimental conditions (bone marrow preconditioning and injected cell number) for a faster and easier preclinical mouse model of AML based on the xenotransplantation of a panel of six adult and childhood human AML cell lines, representing different FAB types, genetic variants and chromosomal abnormalities. In addition, we have also shown that our cell line-derived xenograft (CLDX as mirrored to patient-derived xenograft) models exhibited a broad range of overall mice survival, engraftment, tissue infiltration and aggressiveness based on their AML cell line characteristics.

Cell lines and culture conditions
Human AML cell lines U937, MV4-11, MOLM-14, HL-60 (DSMZ, Braunschweig, Germany) and KG1a (ATCC, Manassas, VA, USA) were maintained in minimum essential medium-α medium supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA). All cell lines were grown in the presence of 100 units per ml of penicillin and 100 μg/ml of streptomycin, and were incubated at 37°C with 5% CO 2 . The cultured cells were split every 2-3 days and maintained in an exponential growth phase. All cell lines were annually re-ordered to DSMZ or ATCC stocks and  Xenotransplantation of human leukemic cells Animals were used in accordance with a protocol reviewed and approved by the Institutional Animal Care and User Ethical Committee of the UMS006 and Région Midi-Pyrénées (Approval#13-U1037-JES08). NSG mice were produced at the Genotoul Anexplo platform at Toulouse (France) using breeders obtained from The Charles River Laboratory. Mice were housed in sterile conditions using high-efficiency particulate arrestance filtered micro-isolators and fed with irradiated food and acidified water.
Adult mice (6-8 weeks old) were sublethally irradiated with 250 cGy of total body irradiation or treated with 20 mg/kg busulfan (Busilvex, Pierre

Hematopoietic cells count
Peripheral blood was obtained with retro-orbital bleeding. Fifty microliter of blood were collected in heparin-coated collection tube for analyzed with hematology counter (ABX Micros 60, Horiba, Montpellier, France).

Assessment of leukemic engraftment
NSG mice were humanely killed in accordance with the Institutional Animal Care and User Ethical Committee of the UMS006 and Région Midi-Pyrénées Protocols. Bone marrow (mixed from tibias and femurs) and spleen were dissected, crushed in PBS and made into single cell suspensions for analysis by flow cytometry (FACS Calibur, FACS Canto, FACS LSR II-BD Biosciences, San Jose, CA, USA).

RESULTS AND DISCUSSION
The objective of this study was to develop robust xenograft models by establishing the specific experimental conditions (cell dose, dissemination tropism, growth kinetics, symptoms, lethality and stability of cell surface markers) required for the evaluation of therapeutic agents in routine and easy-to-use setting. The advantages of using well-established AML cell lines as compared with primary patient samples are the unlimited access to a large amount of human AML cells and a faster engraftment in immunodeficient mice. We analyzed in vivo the five most commonly used AML cell lines with a range of molecular abnormalities, clinical, biological and immunophenotypical characteristics (Table 1). MOLM-14 and MV4-11 cells are a widely studied model for FLT3-ITD AML (30% of AML patients associated with worst prognosis). U937 cells are a model for monocytic  CLDX models for AML E Saland et al development and translocation t(10;11)(p13;q14) leading to PICALM-MLLZ10/AF10 fusion gene found in 7% of AML patients with good prognosis factor. HL-60 cells are a model for human myeloid promyelocytic cell differentiation and proliferation, and KG1a is a model for immature myeloid progenitor phenotype and FGFR1 kinase fusion genes (Table 1). These cells were transplanted intravenously in adult NSG mice. Importantly, AML cell lines kill mice as does AML in patients, whereas primary AML cells in NSG mice rarely kill the animals.
Busulfan or irradiation conditioning step does not change the xenotransplantation efficacy of AML cells Xenotransplantation of primary human AML or normal CD34 + cells is greatly enhanced if recipients receive a chemical conditioning regimen such as busulfan (a myeloablative alkylating agent, 25-35 mg/kg) or total body irradiation (up to 4 Gy). [8][9][10][11] However, it is not clear whether conditioning improves the engraftment of human AML cell lines especially in the most recently developed NSG mouse model. We first performed a maximally tolerated dose study to define the best conditioning procedures (Figures 1a and b) and then tested the effect of the determined sublethal conditions (20 mg/kg busulfan vs 2.25 Gy irradiation) on NSG mice. Body weight, hemoglobin, platelet number and white blood cell number were monitored over 4 weeks following the conditioning procedure. Irradiation severely impacts mice body weight and blood counts, whereas busulfan injection has a milder effect on these parameters (Figures 1c and f). In both the procedures, normal mouse hematopoiesis is recovered 3 weeks post conditioning.
We observed a significant reduction in overall survival of the xenografted mice in the busulfan conditioning setting as compared with irradiation setting (Figure 2a  HL-60 and MOLM-14 cells are preferentially located in the bone marrow (Figures 3f and g), indicating differential tissue tropism for this panel of AML cell lines.
These results show a diversity of engraftment capacities and aggressiveness in these AML cell lines, as well as a diversity in tissue tropism, consistent with what we observed with AML patient cells transplanted in the same immunodeficient mice strain. 4 Engraftment levels, tissue distribution and overall survival are not associated with any clinical features and cytogenetic abnormalities. Interestingly, we also observed that AML cell lines bearing differentiation markers such as U937, MOLM-14 and MV4-11 appeared to have a more aggressive in vivo behavior than immature like AML cell lines such as HL-60 or KG1a.
The overall immunophenotype of AML cells is conserved in vivo We next addressed whether the immunophenotype of AML cell lines was conserved in vivo. The expression level of various myeloid cell surface markers such as CD45, CD33, CD44, CD34, CD38, CD45RA and CD123 was analyzed for four AML cell lines  Figure 4. Analysis of the expression of major myeloid cell surface markers in AML cell lines before and after xenotransplantation. The immunophenotype of HL-60, MV4-11, U937 and KG1a cells was analyzed on BD LSRII Fortessa Flow Cytometer using human CD45, CD44, CD33, CD34, CD38, CD45RA and CD123 before and after xenotransplantation in our NSG mice model. All AML cell lines are SSC low , CD45, CD44 and CD33 positive in vitro and in vivo.
before and after xenotransplantation (Figure 4). Most of those markers appeared globally unchanged during engraftment, however, we found that the transplantation of HL-60 cells in mice also led to the appearance of a second population CD44 dim (Figure 4, second column). Moreover, the frequency of CD34 + CD38 + was significantly increased in vivo to the detriment of the CD34 + CD38population ( Figure 4, third column). With the sole exception of the CD38 marker, the NSG-based model of AML maintains cell phenotype more consistently than in the NOD-SCID model. 12 Here we provide necessary technical details and conditions that other laboratories may use to quickly and routinely set up in vivo AML experiments. A xenografts based on native or engineered cell lines are not novel, this is the first study that compares the five most well-characterized AML cell lines in the same experimental settings and mouse strain. The establishment of this preclinical AML model is of special relevance and significance to drugsensitivity studies as in vitro cell culture-based screens do not accurately reflect in vivo effects and responses. In conclusion, we have found that the xenotransplantation models of five wellcharacterized AML cell lines (for example, KG1a, HL-60, MOLM-14, MV4-11 and U937) exhibited a broad range of mice survival, engraftment, tissue infiltration and aggressiveness. By increasing or decreasing the number of injected leukemic cells, we can modulate mice survival without changing the tissue distribution of leukemic cells and their engraftment in NSG mice. This work has also shown that AML cell lines kill mice in a manner similar to that as patients in clinics but different from the xenograft of primary AML patient specimens in NSG mice. Moreover, in contrast to patient-derived xenograft models, these CLDX models are a powerful, rapid and straightforward in vivo assay available to all leukemia research laboratories to perform preclinical studies to assess the in vivo efficacy of conventional or targeted therapies in AML. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/ by/4.0/