Knock-in of an internal tandem duplication mutation into murine FLT3 confers myeloproliferative disease in a mouse model

Within the hematopoietic system, FMS-like tyrosine kinase 3 (FLT3; also known as FLK2, STK-1, or CD135), together with its ligand FL, play important roles in the proliferation, survival, and differentiation of hematopoietic stem/progenitor cells (HSPCs), particularly myeloid and lymphoid progenitors.^1-4  Aberrant FLT3 expression has been reported in a substantial fraction of leukemias. In acute myeloid leukemia (AML), about 30% to 35% of patients have either internal tandem duplication (ITD) mutations in the juxtamembrane domain or point mutations in the kinase domain of FLT3, making it the most frequent alteration found in AML.^5-7  FLT3/ITD mutations, in particular, confer poor prognosis in most studies of both pediatric and adult AML,^5,6,8,9  and are considered a target for AML treatment drug discovery.^10-14 

Unlike the wild-type FLT3, whose activation is dependent on stimulation with FL,^15-17  FLT3/ITD mutations result in constitutive dimerization and activation of the receptor independent of FL.^10,18,19  The downstream signaling pathways elicited by constitutive FLT3 activation have not been fully elucidated, but the STAT5, RAS/MAPK, and PI3K/AKT pathways have all been shown to be involved.^19-22  Gene expression profiling studies based on cell lines and primary cells carrying activated FLT3 mutations reveal a number of genes to be dysregulated by FLT3/ITD signaling.^23,24 

One way to further our understanding of the role FLT3 mutations play in leukemia development is to develop animal models of leukemia in which activated FLT3 contributes. Several models have been previously generated and provided us with important insights into the in vivo biological consequences of this activated mutation. Injection of FLT3/ITD-transfected Ba/F3 or 32D cells into syngeneic recipient mice results in leukemia-like disease.^10,19  Retroviral transduction of FLT3/ITD mutations into primary mouse bone marrow (BM) cells followed by BM transplantation leads to a rapidly fatal myeloid proliferative disease (MPD).^25,26  In a transgenic model, expression of FLT3/ITD driven by the vav promoter results in nontransplantable MPD and B- or T-lymphoid disorders.²⁷  Constitutively activated FLT3 leads to MPD in another transgenic mouse model with the expression of Tel-FLT3 fusion protein.²⁸  These findings support the idea that mutant FLT3 plays an important role in the pathogenesis of leukemia. However, in these models, the immortal nature of the cell lines makes it likely that additional genetic alterations pre-exist in the Ba/F3 or 32D cells. The use of genetically engineered exogenous promoters (ie, the long terminal repeat [LTR] promoter in the case of retrovirus/BM transplantation models and the panhematopoietic vav or ubiquitously expressed cytomegalovirus [CMV] promoters in the transgenic models) are unlikely to accurately reproduce the level, developmental distribution and timing of the aberrant FLT3 expression in hematopoiesis.^26-28  In addition, random integration of the retrovirus might result in the activation of proto-oncogenes or inactivation of tumor suppressor genes that cooperate with FLT3 signaling to result in the observed fatal MPD or leukemia. To overcome these limitations, we generated a FLT3/ITD knock-in mouse model targeting murine Flt3 genomic DNA. This model closely simulates and reflects the events that occur when a normal mouse hematopoietic cell acquires an activating FLT3 mutation and hence provides a useful platform to study the role FLT3/ITD mutations play in the development and maintenance of leukemia. It might also provide a valuable tool for screening new drugs targeting FLT3 mutations.

A 12-kb Flt3 genomic DNA fragment spanning exons 10 to 16 of murine Flt3, obtained from the RPCI-22 mouse 129S6/SvEvTac bacterial artificial chromosome (BAC) library (Roswell Park Cancer Institute, Buffalo, NY), was isolated to construct the targeting vector. An 18-bp ITD mutation isolated from a patient with AML (AGGACTGATTTCAGAGAA)²²  was inserted into exon 14 (Figure 1A). The linearized targeting construct was then electroporated into 129 embryonic stem (ES) cells followed by G418 and ganciclovir selection. Homologously recombinant ES cells were injected into blastocysts and reimplanted into the uterine horns of 2.5-day postcoital pseudopregnant foster females to allow embryos to develop to term. Potential chimeras were crossed with the wild-type C57BL/6 mice (National Cancer Insitute, Frederick, MD) to facilitate the detection of germ-line transmission. Mice with an ITD mutation on one allele were bred with C57BL/6 mice for 7 generations to generate inbred strains in this background. All mice were housed in a pathogen-free animal facility in microisolator cages. Southern blotting and polymerase chain reaction (PCR) analyses were performed to genotype the mice. For a detailed record of primers and probes used; see “Primers and probe used for genotyping analysis and RT-PCR analysis” (available on the Blood website; see the Supplemental Materials link at the top of the online article).

Mice with the ITD mutation on one allele were bred with Mx1-Cre transgenic mice (Jackson Laboratory, Bar Harbor, ME) to generate the FLT3^wt/ITD/Mx1-Cre mice. Expression of Mx1-Cre and excision of PGK-Neo cassette were induced by polyinosinic-polycytidylic acid (PIPC) treatment in vivo. In brief, 2- or 3-day-old FLT3^wt/ITD pups were injected intraperitoneally with PIPC (Sigma-Aldrich, St Louis, MO) at 300 μg/dose every other day for 3 times. All the experiments were conducted with the PIPC-treated FLT3^wt/ITD/Mx1-Cre mice, referred to as FLT3^wt/ITD.

Quantitative reverse transcription (RT)–PCR was performed using an iCycler iQ multicolor real-time PCR system (Bio-Rad, Hercules, CA) as described previously.²⁹  In brief, 200 ng RNA extracted from BM cells were reverse-transcribed and amplified using primers P3F and P3R. The level of FLT3 expression was normalized based on that of mS16. Densitometry analysis was used to measure the relative level of ITD versus wild-type FLT3 expression after electrophoresed in agarose gel using QuantityOne densitometry analysis software (Bio-Rad).

Spleen cells were washed with ice-cold PBS (Quality Biologicals, Gaithersburg, MD) and lysed with NP-40 lysis buffer (150 mM NaCl, 20 mM Tris-HCl [pH 7.4], 10% glycerol, 1% NP40, 10 mM EDTA, and 100 mM NaF) containing 1 mM sodium orthovanadate (Sigma) and protease inhibitors (Roche, Mannheim, Germany), followed by immunoprecipitation with anti-FLT3 antibody (EB-10; Imclone, New York, NY). Immunoblotting was performed using either 4G-10 (Millipore, Billerica, MA) or anti-murine FLT3 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Protein bands were detected by enhanced chemiluminescence (Amersham, Piscataway, NJ).

A total of 50 μL peripheral blood was collected from murine retrooccular vessels and subjected to complete blood cell counting and white blood cell (WBC) differential counting using a Hemavet950 Hematology system (Drew Scientific, Oxford, CT).

Murine tissues were prepared and stained with hematoxylin and eosin solution (H&E) as described previously.²⁸  Representative H&E-stain images were acquired using a Zeiss Axioskop upright microscope system (Zeiss, Peabody, MA). BM cytospins were prepared and stained with Wright-Giemsa solution (Sigma-Aldrich), and images were acquired at room temperature using a Nikon Eclipse E600 microscope system (Nikon).

Flow cytometry analysis was performed as described previously.²⁸  For a detailed record of all antibodies used, refer to the supplemental information in “Antibodies used for flow cytometry.” Cell acquisition and analysis were performed on a 2-laser FACSCalibur (Becton Dickinson, San Jose, CA) using CellQuest software (Becton Dickinson).

Determination of mouse serum FL concentration was performed using Quantikine mouse FL immunoassay kit (R&D Systems, Minneapolis, MN) as per the manufacturer's instructions. Optical density was determined using a microplate reader (Bio-Rad) set at 450 nm with wavelength correction of 570 nm.

Murine BM was cultured in methylcellulose medium (Methocult M3434: StemCell Technologies, Vancouver, BC) supplemented with recombinant murine SCF (50 ng/mL), IL-3 (10 ng/mL), IL-6 (10 ng/mL), GM-CSF (10 ng/mL), and EPO (3 U/mL). For the clonogenic assay on Lin⁻ populations, total BM cells were stained using negative selection mouse hematopoietic progenitor kit (StemCell Technologies). Cells flowing through the magnetic columns were collected and plated in Methocult M3434 medium. The secondary clonogenic assay was conducted by pooling cells from the primary culture and replating in Methocult M3434 medium. Tertiary assay was performed in the same way. All colonies were scored 11 days after plating. Immortalized progenitor cell lines were generated as described by Lavau et al.³⁰  Cells pooled from the tertiary colonies were cultured in RPMI medium (Invitrogen, Gaithersburg, MD) supplemented with SCF (50 ng/mL), IL-3 (10 ng/mL), and IL-6 (10 ng/mL; Peprotech, Rocky Hill, NJ). IL-6 was withdrawn from the medium after 3 passages.

Cytokine-response colony-forming assay was performed by plating total BM cells into Methocult M3234 medium (StemCell Technologies) supplemented with gradient concentrations of murine IL-3 or GM-CSF. Colony images were acquired at room temperature using a Nikon Eclipse TE300 microscope system (Nikon).

For the colony forming unit–spleen (CFU-S) assay, 5 × 10⁴ BM cells from donor BM were resuspended in phosphate-buffered saline (PBS) and injected into lethally irradiated C57BL/6 recipient mice (8.5 Gy). Recipients were killed 14 days after injection and colonies counted as described previously.²⁸  A total of 5 × 10⁴ total spleen cells from primary and secondary recipients were collected and injected into lethally irradiated secondary and tertiary recipients for the secondary and tertiary assays.

A total of 5 × 10⁵ CD45.1⁺ competitor BM cells obtained from the wild-type B6-Ly5.2 (National Cancer Institute) mice were mixed with 5 × 10⁵ CD45.2⁺ test cells (wild-type or FLT3^wt/ITD C57BL/6 mice) and injected into the lateral tail veins of lethally irradiated (11 Gy) B6-Ly5.2 recipient mice. Complete peripheral blood counts and expression of CD45.1 and CD45.2 were assessed every 6 weeks.

A FLT3/ITD knock-in mouse model was generated in which an 18-bp ITD mutation was inserted into the genomic DNA sequence coding the juxtamembrane domain of murine Flt3 (Figure 1A). Homologous recombination was confirmed by Southern blotting analysis and PCR (Figure 1B). Mice with a single-copy ITD mutation were bred with Mx1-Cre transgenic mice to generate the FLT3^wt/ITD/Mx1-Cre mice. PIPC, an interferon-inducing agent, was injected into newborn pups to induce Mx1-Cre expression and PGK-Neo cassette excision.

FLT3^wt/ITD mice were viable and fertile. The frequency of newborn mutant mice was in accordance with the predicted Mendelian frequency (data not shown). PIPC-induced Mx1-Cre expression and PGK-Neo cassette excision were confirmed by PCR (data not shown). Increased expression and activity of FLT3 in FLT3^wt/ITD mice were detected by quantitative RT-PCR and Western blotting analysis (Figure 1C,D).

FLT3^wt/ITD mice died at 6 to 20 months, with a median survival time of 10 months (Figure 2A). Splenomegaly was observed in FLT3^wt/ITD mutants, with an average of 2.5-fold larger spleens in 2-month-old and 5-fold larger spleens in 12-month-old FLT3^wt/ITD mice relative to the wild-type mice (Figure 2B).

Compared with the wild-type mice, which showed a well-defined normal architecture with distinctive red and white pulp (Figure 3A), spleens from 2-month-old FLT3^wt/ITD mutants displayed notable myeloproliferative changes as evidenced by large areas of mature and immature myeloid-appearing cells in the red pulp (Figure 3B). Spleens from 12-month-old FLT3^wt/ITD mice showed progressive myeloproliferation with expansion of mostly immature myeloid cells in the red pulp. The white pulp was not well defined and also contains more immature myeloid cells with fewer mature lymphocytes compared with control spleens (Figure 3D,F; compare with Figure 3C,E). Enhanced expansion of granulocytes/monocytes and reduced B-lymphoid fraction in the spleen from FLT3^wt/ITD mice were confirmed by flow cytometry analysis (Figure 4A-D).

BM sections of 2-month-old mice failed to reveal significant morphologic differences between the wild-type and FLT3^wt/ITD mice. Total number of mononuclear cells per femur in 2-month-old FLT3^wt/ITD mice was 20.9 plus or minus 1.3 × 10⁶ (n = 11), which was not different from that in the wild-type mice of the same age (20.5 ± 2.5 × 10⁶, n = 20; P = .84). However, flow cytometry analysis showed enhanced expansion of granulocytes and monocytes, and reduced expansion of erythrocytes, megakaryocytes and B-lymphoid cells in BM from 2-month-old FLT3^wt/ITD mice (Figure 4E-H).

BM sections of 12-month-old FLT3^wt/ITD mice demonstrated hypercellularity with the accumulation of immature myeloid cells (Figure 3H,J; compare with Figure 3G,I). Total number of mononuclear cells per femur was significantly higher than that in the wild-type mice of the same age (48.1 ± 1.7 × 10⁶ cells vs 22.5 ± 0.5 × 10⁶ cells, n = 12; P < .01). Meanwhile, these mice also showed an increased fraction of myeloid cells with undifferentiated or partially differentiated morphology (Figure 3L; compare with Figure 3K). Livers from some of the older FLT3^wt/ITD mice showed extramedullary hematopoiesis with a population of both mature and immature myeloid-appearing cells in the sinusoids (data not shown). No lymphadenopathy was seen in the FLT3^wt/ITD mice.

Peripheral WBC counts from 2-month-old FLT3^wt/ITD mice were slightly lower than the wild-type littermates, with both lower neutrophil and lymphocyte counts. Red blood cell (RBC) and platelet counts were within the normal range. In contrast, 12-month-old FLT3^wt/ITD mice showed leukocytosis with total WBC count increased to 23.0 plus or minus 8.4 × 10⁹/L versus 11.9 plus or minus 5.4 × 10⁹/L in the wild-type littermates (n = 10; P < .05). Higher neutrophil and monocyte counts accounted for most of the differences (Figure 2C,D; Table 1). Lower hemoglobin concentration (111 ± 25 g/L vs 146 ± 10 g/L; n = 10, P < .05) and platelet counts (560.8 ± 208.2 × 10⁹/L vs 704.6 ± 118.1 × 10⁹/L; n = 10, P = .12) were also detected in 12-month-old FLT3^wt/ITD mice, although the difference in platelet count was not statistically significant.

An increased fraction of BM cells from 2-month-old FLT3^wt/ITD mice were Lin^−/low and/or c-KIT⁺, characteristics of immature HSPCs. The immature immunophenotype became even more evident in the 12-month-old FLT3^wt/ITD mice, where an even higher fraction of cells were Lin^−/low (36.7% ± 19.7% vs 10.0% ± 3.6%; n = 12, P < .001) and/or cKIT⁺ (15.6% ± 6.1% vs 7.3% ± 4.2%; n = 12, P < .001; Figure 4E; Table 2).

Compared with the wild-type controls, an increased fraction of BM cells from 12-month-old FLT3^wt/ITD mice were positive for Mac-1 (57.6% ± 14.9% vs 30.9% ± 6.2%; n = 12, P < .001) or Gr-1 (62.2% ± 14.4% vs 31.0% ± 7.6%; n = 12, P < .001), characteristic of monocytes and granulocytes. In contrast, cells expressing Ter119, an erythrocytic marker, were reduced (12.4% ± 7.4% vs 42.0% ± 9.6%; n = 12, P < .001) in the FLT3^wt/ITD mice (Figure 4F; Table 2). No significant differences were observed in expression of megakaryocytic markers. These data indicate an expansion of the granulocytic/monocytic lineages, and a reduction of the erythrocytic lineage in FLT3^wt/ITD mice.

B220 is a marker throughout B-lymphoid development. CD43 appears on pro-B cells, the earliest identifiable B-lymphoid progenitors, and expression ceases as the cells develop to the pre-B stage.³¹  To investigate the possible effect of FLT3/ITD mutation expression on B-cell development, we analyzed the expression of these surface markers in FLT3^wt/ITD mice. In the 12-month-old wild-type mice, 16.7% plus or minus 5.2% of BM cells express B220. In FLT3^wt/ITD mice, only 5.6% plus or minus 4.0% of BM cells were B220⁺ (n = 12; P < .001). In contrast, there was not a significant difference in the B220⁺CD43⁺ pro–B-cell population between these mice (3.1% ± 1.8% for FLT3^wt/ITD vs 3.3% ± 3.2% for wild-type mice; n = 12, P = .84). However, the fraction of more differentiated pre-B (B220⁺CD43⁻IgM⁻, 1.3% ± 0.7% for FLT3^wt/ITD vs 6.0% ± 1.8% for wild-type mice; n = 12, P < .001) and B cells (B220⁺CD43-IgM⁺, 4.4% ± 2.5% for FLT3^wt/ITD vs 10.8% ± 4.0% for wild-type mice; n = 12, P < .001) were greatly decreased in the BM from FLT3^wt/ITD mice (Figure 4G; Table 2). Similar changes in the B-lymphoid compartment were also found in the BM and spleens from 2-month-old FLT3^wt/ITD mice (Figure 4C,G).

To exclude the possibility that reduced B-lymphoid fraction is a consequence of myeloid expansion, we further assessed the expression of B-cell markers in the BM and spleens of 4-week-old pups in which myeloid expansion was not nearly as extensive as in the older mice. The results showed that the fraction of pre-B and B cells were still reduced (Figure 4I). At the same time, the pro–B-cell fraction was significantly increased, which confirmed that the blockage of B-cell development occurs early in the FLT3/ITD mice, and that the reduction in the more differentiated B cells is not the result of myeloid expansion. Thus, constitutively activated FLT3 signaling possibly causes a block in B-lymphoid maturation.

FL stimulation of the wild-type mice is known to greatly stimulate the production of dendritic cells (DCs).³²  To investigate whether constitutive FLT3 activation in FLT3^wt/ITD mice would alter DC development, we assessed these mice for the expression of CD11c, a marker of DCs. Interestingly, a significant increase in DCs was observed in both the BM (14.0% ± 10.5% vs 4.8% ± 3.6%; n = 12, P < .05; Figure 4H; Table 2) and spleen (18.5% ± 13.6% vs 3.6% ± 1.4%; n = 12, P < .05) of FLT3^wt/ITD mice when compared with the wild-type littermates. Immunoassay showed no differences in FL levels in the serum of the FLT3^wt/ITD mice (313.7 ± 156.3 pg/mL; n = 12) compared with the wild-type controls (339.7 ± 70.6 pg/mL; n = 18, P = .87). This excludes the possibility of FL overproduction as the cause of DC expansion in the FLT3^wt/ITD mice. It is likely that constitutive activation of FLT3 signaling appears to be able to substitute for FL stimulation of the wild-type FLT3 in DC expansion. No consistent differences were found in natural killer (NK) or T-cell populations from the spleen or BM of the FLT3^wt/ITD mice (Table 2). No significant changes were observed in thymocyte subsets, as evidenced by the expression of Thy1.1, CD3, CD4, and CD8a (data not shown).

The morphologic and immunophenotypic results imply a growth advantage for the granulocytic/monocytic lineages from FLT3^wt/ITD mice. To directly investigate whether FLT3/ITD confers a growth advantage, we conducted in vitro clonogenic assays.

Plating 10⁴ total BM cells from FLT3^wt/ITD mice in methylcellulose medium generated a 2-fold increase in CFU-GMs compared with the wild-type controls (49.8 ± 21.0 vs 20.5 ± 8.6; Figure 5A). In contrast, plating of an equal number (1000/dish) of Lin^−/low BM cells from FLT3^wt/ITD mice showed no statistically significant increase in CFU-GM colonies compared with the wild-type mice (61.5 ± 22.7 vs 53.6 ± 17.0; Figure 5B). This indicates that it is the expansion of the Lin^−/low fraction in FLT3^wt/ITDmice that results in the increased CFU-GM fraction. No statistically significant differences in the number of burst-forming units–erythroid (BFU-Es) was observed (Figure 5A,B). This suggests that the differentiation potential toward erythrocytic lineages in HSPCs from FLT3^wt/ITD mice was basically unchanged.

Similar to the results from the BM, spleen cells from FLT3^wt/ITD mice also showed increased ability to generate CFU-GMs (104 ± 21/10⁵ cells for FLT3^wt/ITD vs 61 ± 5/10⁵ cells for wild-type spleen, respectively; P < .05), while the ability to generate BFU-Es was not increased.

Flow cytometry analysis of the cells collected from day-11 colonies showed that 67.4% plus or minus 7.3% of the cells generated from FLT3^wt/ITD BM expressed c-KIT, as compared with 46.7% plus or minus 10.2% in the wild-type group (n = 3; P < .05). Meanwhile, a lower fraction of the cells generated were positive for Mac-1 as compared with the wild-type group (29% ± 6.5% for FLT3^wt/ITD vs 38.6% ± 4.2% for the wild-type group; n = 3, P = .18), although not statistically significant (Figure 5D). These data indicate that FLT3^wt/ITD BM cells generated more immature cells in vitro.

The secondary and tertiary colony assays of cells collected from the primary colonies of FLT3^wt/ITD BM yielded more, though not statistically significant, CFU-GMs when compared with those from the wild-type BM (44 ± 8 for FLT3^wt/ITD vs 33 ± 6 for wild-type group in the secondary plating; P = .11; 33 ± 4 for FLT3^wt/ITD vs 27 ± 4 cells for wild-type group in the tertiary plating; P = .05; Figure 5C). However, a significant difference was observed in the behavior of the cells obtained from the BM of the FLT3^wt/ITD mice. These cells could be serially propagated in medium containing SCF and IL-3 for more than 10 months, while cells from the wild-type mice were exhausted after 3 to 5 rounds of serial liquid culture (Figure 5E). Withdrawal cytokines from the culture medium resulted in cell death as assayed by trypan blue exclusion assay so the growth of the cells was still factor dependent. Flow cytometry analysis showed expression of c-KIT in the vast majority of the cells (Figure 5D), while Mac-1 and Gr-1, the myeloid cell markers, and FcϵRIα, a mast cell marker, were not expressed (data not shown).

In the beginning of the liquid culture, the proliferation rate for FLT3^wt/ITD cells was in parallel with that of the wild-type cells. A lag occurred 2 to 4 weeks into culture, which was followed by an enhanced expansion of the FLT3^wt/ITD cells while the wild-type cells died. This suggests that FLT3/ITD itself might not be sufficient to immortalize cells, but creates conditions in which immortalization can occur.

Since BM cells with a FLT3/ITD mutation generated more CFU-GMs when cultured with saturating concentrations of cytokines, we further wanted to investigate whether cells with the FLT3/ITD mutation have a lower requirement for cytokines. When cultured in the presence of lower concentrations of IL-3 or GM-CSF, BM cells from FLT3^wt/ITD mice generated more CFU-GMs in a dose-dependent manner. The resultant colonies were also larger in size than the wild-type colonies (Figure 5F,G). BM cells from FLT3^wt/ITD mice were still dependent on cytokines for colony formation, as they did not produce any colonies in the absence of cytokines.

These data together indicate that HSPCs with FLT3/ITD mutations have reduced cytokine requirements for producing colonies and are capable of initiating immortalization of HSPCs.

To further explore the effect of the FLT3/ITD mutation on HSPCs, the CFU-S assay was performed. In primary recipients of BM cells, the day-14 CFU-S assay detects progenitor activity, while secondary spleen colonies, resulting from the injection of spleen cells from the primary recipients, in secondary recipients reflects more primitive HSPC activity.^33,34 

In the CFU-S assay, injection of 5 × 10⁴ FLT3^wt/ITD total BM cells into recipient mice yielded 21.5 plus or minus 3.1 colonies in recipient spleens, slightly more than the number obtained from the wild-type controls (16.0 ± 2.6; P < .05; Figure 6A). Retransplantation of 5 × 10⁴ total spleen cells from primary recipients of the FLT3^wt/ITD group resulted in twice as many colonies in the spleens of the secondary recipients as those from the wild-type controls (3.3 ± 1.6 colonies vs 1.7 ± 0.5 colonies). CFU-S colonies were depleted by the third serial transplantation of cells from the wild-type controls. However, cells from FLT3^wt/ITD mice still produced spleen colonies in the tertiary recipient mice (3.3 ± 1.1 colonies vs 0.5 ± 0.5 colonies).

The long-term repopulation assay is an assay that better reflects the potential of primitive HSPCs.³⁵  We chose the long-term competitive repopulation assay to compare the ability of HSPCs to expand in vivo.

At 6 weeks after injection, the ratios of test cells (CD45.2⁺, from FLT3^wt/ITD or wild-type C57BL/6 mice) versus competitor cells (CD45.1⁺) in the peripheral blood were approximately 1.0 in recipients receiving FLT3^wt/ITD BM cells (FLT3^wt/ITD group) or the wild-type BM cells (wild-type group). At later times, this ratio remained stable at 0.9 to 1.5 in recipients from the wild-type group. In this group, flow cytometry analysis showed no statistical differences in the fraction of wild-type CD45.2⁺ cells when compared with the wild-type CD45.1⁺ competitor cells in the BM from the same recipients (Figure 6C).

In contrast, the CD45.2/CD45.1 ratio in the peripheral blood started rising in most (16 of 22) of the recipients in the FLT3^wt/ITD group 12 to 18 weeks after injection and reached a mean of 6.0 to 8.0 by 24 to 36 weeks after transplantation (Figure 6B). In this group, the recipients with higher CD45.2/CD45.1 ratios also had increased WBC, neutrophil, and monocyte counts and reduced hemoglobin, RBC, and platelet counts (data not shown). BM cells from these recipients displayed an even higher CD45.2/CD45.1 ratio (25.0–80.0). Flow cytometry analysis showed an increased fraction of cKIT⁺, Mac-1⁺, and Gr-1⁺ cells and a greatly reduced fraction of B220⁺ cells in the CD45.2⁺ (FLT3^wt/ITD) population as compared with the CD45.1⁺ competitor population in the same recipients (Figure 6C). In the other recipients (6 of 22), the ratio of CD45.2 to CD45.1 cell populations in peripheral blood remained 1.0 to 2.0, which was slightly higher than the wild-type group but was not statistically significant. However, 48 to 72 weeks after transplantation, recipients with the lower CD45.2/CD45.1 ratios also developed signs of mild MPD, which included BM hypercellularity and splenomegaly, as well as moderately increased WBC counts in the peripheral blood. Flow cytometry analysis of the BM showed that the CD45.2/CD45.1 ratio was 6.0 to 20.0. It was higher than that in the peripheral blood, which remained 1.0 to 2.0. They also displayed a moderately higher fraction of Mac-1⁺/Gr-1⁺ cells and a lower fraction of B220⁺ cells in the CD45.2⁺ (FLT3^wt/ITD) population compared with that of the CD45.1⁺ wild-type competitor population in the same recipients (data not shown). FLT3^wt/ITD BM cells from primary recipients, regardless of the CD45.2/CD45.1 ratios, were able to engraft in the secondary recipients, but failed to reproduce the MPD phenotype when transplanted into lethally irradiated secondary recipients.

In summary, this data indicates that targeted insertion of an ITD mutation into the murine Flt3 gene by itself is capable of partially transforming normal HSPCs to a MPD phenotype. This was manifested by expansion of terminally differentiated granulocytes and monocytes and their immature precursors in the BM as well as by extramedullary myeloproliferation in the spleen and liver. FLT3/ITD expression also initiates in vitro immortalization and in vivo myeloid expansion activity. No acute leukemia was observed in either the FLT3^wt/ITD mice or the recipients of BM transplants from these mice.

In this model, we observed a slowly fatal MPD disorder in mice expressing a FLT3/ITD mutation. Our results demonstrate that FLT3/ITD expression alone is sufficient to initiate the expansion and accumulation of mature granulocytes and monocytes, as well as their immature precursors. We did not observe an expansion for erythrocytes or megakaryocytes. This is consistent with the results of the activation of wild-type FLT3 by its ligand in that in vivo administration of FL results in BM hypercellularity with an increased proportion of cells expressing granulocytic/monocytic markers.^36,37  MPD has also been reported previously in BALB/c mice receiving FLT3/ITD-transduced syngeneic BM cells and in a transgenic model with FLT3/ITD mutation driven by the vav promoter.^26,27  Thus, the constitutive activation of FLT3 resulted from FLT3/ITD mutation primarily stimulates granulocytic/monocytic lineage expansion. This may partly explain the clinical observations that the majority of FLT3/ITD mutations are found in AML.

Data from both the FLT3 knock-out model and FL knock-out model suggests a role for FLT3 in the maintenance, expansion, and generation of B-lymphoid cells.^1,38  Deficiencies in primitive B-cell progenitors were observed in both models. B-lymphoid progenitors have also been reported to express high levels of FLT3 receptor.³⁹  In light of this role, it is somewhat surprising that B-lymphoid expansion was not observed in the FLT3/ITD mice. In fact, we observed a statistically significant decrease in the proportion of B220⁺ lymphocytes, including significantly reduced fractions of pre-B, immature, and mature B cells, but not pro-B lymphoid progenitors, in the FLT3^wt/ITD mice. Furthermore, HSPCs with FLT3/ITD mutations also showed decreased ability to develop into B-lymphoid cells in the long-term competitive transplantation assay. It is likely that constitutive activation of FLT3 disturbs B-cell development. Studies on FLT3 signaling suggest that mutation-activated FLT3 might elicit downstream signaling pathways different from the wild-type receptor.^19-22  It is possible that, in HSPCs, constitutively activated FLT3 targets certain signaling events, which eventually lead to the deficiency in B-cell development.

Within the hematopoietic system, FLT3 expression is mostly restricted to more committed hematopoietic stem cells and progenitors. It is possible that FLT3 might be functioning within a subpopulation of primitive HSCs poised for active hematopoiesis and the commitment/differentiation process rather than in HSC self-renewal events.³⁸  In the competitive transplantation study, FLT3^wt/ITD BM cells show increased long-term myeloid expansion ability in all of the recipients, although to a variable extent and at different time points. This result indicates that FLT3/ITD confers HSPCs with myeloid growth advantage. It is possible that some HSPCs acquired additional mutations over time, which provided these cells with a growth advantage. This could account for the long duration required for the ratio of FLT3^wt/ITD/competitor cells to increase. It could also explain the reason why some FLT3^wt/ITD recipients need a longer time to display signs of enhanced myeloid expansion.

In pediatric AML with FLT3/ITD mutations, FLT3/ITD can be detected in all CD34⁺CD33⁺ samples, whereas it is heterogeneously involved in CD34⁺CD33⁻ precursors.⁴⁰  The involvement of FLT3/ITD mutations in the more primitive CD34⁺/CD33⁻ precursors is associated with worse clinical outcome.⁴⁰  In the knock-in mice reported here, the mutation is expressed in the earliest presursors that normally express FLT3, and this may explain why the observed disease is fatal.

Although FLT3/ITD mutations in patients with AML occur most frequently in patients with normal karyotype, they have also been observed together with fusion genes, including PML-RARα, AML1-ETO, CBFβ-MYH11, and MLL-related fusion genes.^5,6  In several murine models, when FLT3/ITD mutations have been combined with PML-RARα, AML1-ETO, or MLL-related fusion genes, the result is AML.^41-43  These clinical observations and in vivo models support the hypothesis that FLT3 itself is not sufficient to confer acute leukemia, and additional cooperative events are required. In human acute leukemia, the involvement of FLT3/ITD in HSPCs is heterogeneous. Patients with de novo AML with FLT3/ITD mutations sometimes lose the mutation or acquire new mutations at relapse. FLT3/ITD⁻ patients sometimes gain a FLT3/ITD mutation at relapse.^5,6  These findings imply that the occurrence of FLT3 mutations could be a consequence of pre-existing mutations in HSPCs. Consistent with this, we did not observe any cases of acute leukemia or lymphoma spontaneously developing in the FLT3/ITD knock-in mice, which support the idea that spontaneous FLT3 mutations might be a secondary rather than an initiating event in leukemogenesis.

AML is generally considered a multistep process. It requires at least 2 cooperative events to occur in HSPCs: one confers mostly enhanced cell survival and proliferation through deregulated signal transduction, while the other is associated with impaired differentiation, mostly due to abnormalities in transcription factors. Constitutively activated FLT3 mutations, like mutations in other receptor tyrosine kinases, contributes to leukemic transformation through deregulated signaling pathways. Recently, genetically accurate mouse models of other mutations have also been characterized. Conditional expression of Kras^G12D, a K-Ras mutation, in the hematopoietic system induces a lethal MPD.^44,45  Somatic inactivation of Nf1 in hematopoietic cells results in a progressive MPD.⁴⁶  D61Y and E76K, 2 SHP-2 mutations found in juvenile myelomonocytic leukemia (JMML), lead to JMML-like or lymphoproliferative disorders.⁴⁷  Mutated K-Ras, inactivation of Nf1, and activating SHP-2 mutations, as well as constitutively activated FLT3, all deregulate Ras signaling. These observations together point to the possibility that deregulated Ras signaling in HSPCs confers MPD in vivo. However, like the FLT3/ITD knock-in model, none of these mutations alone confers the phenotype of acute leukemia. These findings, together, support the multistep model for the development of AML.

In summary, the data demonstrate that knock-in of the FLT3/ITD mutation alone is capable of conferring MPD with a long lag, but not AML. FLT3/ITD confers HSPCs with myeloid expansion potential. Additional cooperative mutations are required to progress to AML. The FLT3/ITD knock-in may serve as a good model for the in vivo study of the role FLT3/ITD mutations play in leukemogenesis.

The authors gratefully acknowledge Drs Alan Friedman, Patrick Brown, and Mark Levis for their insightful discussions and valuable comments on the manuscript. We thank members of the Small laboratory for their contributions to the completion of this work, and for their insightful discussions and valuable comments on the manuscript. We also thank ImClone Systems for providing the EB-10 antibody.

This work was supported by grants from the National Cancer Institute (CA90668, CA70970). D.S. is also supported by the Kyle Haydock Professorship in Oncology.

National Institutes of Health

Contribution: L.L. designed and performed experiments, analyzed data, and wrote the first draft; O.P., H.B.N., and K.G. performed experiments; K.T. designed experiments; F.R. and D.H. analyzed data; and D.S. designed experiments, analyzed data, and revised the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Donald Small, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, CRB-I Rm 251, 1650 Orleans St, Baltimore, MD 21231; e-mail: donsmall@jhmi.edu.