• Open Access

Isolation of human mAbs that directly modulate FMS-related tyrosine kinase 3 signaling

Authors


To whom correspondence should be addressed.
E-mail: yyukiya@fujita-hu.ac.jp

Abstract

FMS-related tyrosine kinase 3 (FLT3) is a class III receptor tyrosine kinase that plays important roles in hematopoiesis, including early progenitors and dendritic cell development. FLT3 is expressed at high levels in 70–100% of cases of AML and in virtually all cases of B-lineage acute lymphoblastic leukemia. FLT3 is regarded as a molecular target in the development of novel therapies for acute leukemia patients. Currently, many small-molecule FLT3 inhibitors have been developed, but clinical trials have resulted in limited antileukemia effects because of off-target toxicities and drug resistance. The development of anti-FLT3 Abs might overcome these difficulties and enhance the antileukemia efficacy of FLT3 inhibitors. In the present study, we demonstrate the isolation of novel human mAbs against FLT3 with antagonistic or agonistic activities. An antagonistic Ab, designated A2, continuously inhibits FLT3 ligand (FL)-induced phosphorylation of FLT3 and MAPK. A2 cooperatively induces apoptosis with daunorubicin, even in the presence of FL. An agonistic Ab, designated 3E6, surprisingly induces the phosphorylation of FLT3 and MAPK, and supports the growth of a factor-dependent cell line independently of FL addition. In addition, A2 showed complement-dependent cytotoxicity activity, but was devoid of Ab-dependent cell mediated cytotoxicity. Finally, we evaluated Ab internalization in a cell line. Immunofluorescence and flow cytometry analyses revealed that A2 is efficiently internalized. Collectively, these data demonstrate that A2 is a potent human Ab that might be capable of delivering cytotoxic reagents and that has antagonistic effects on FLT3 signaling. In addition, 3E6 might be a potential scaffold for novel dendritic cell-based immunotherapies. (Cancer Sci 2012; 103: 350–359)

FMS-related tyrosine kinase 3 (FLT3) is a class III receptor tyrosine kinase expressed on early hematopoietic progenitor cells that plays important roles in hematopoiesis.(1–4) The FLT3 ligand (FL) is active in both soluble and membrane-bound forms, and is produced by bone marrow stromal cells, T cells, and endothelial cells. The FLT3 ligand acts in synergy with other cytokines in promoting hematopoietic expansion.(1,5) Upon stimulation with FL, FLT3 dimerizes and undergoes autophosphorylation, which results in an upregulation of its tyrosine kinase activity. This increase in activity triggers signaling through an array of downstream pathways, including the phosphatidylinositol-3 kinase and MAPK cascades, thereby promoting cell proliferation and inhibiting apoptosis.(2,3) FLT3 is expressed at high levels in 70–100% of cases of AML and in virtually all cases of B-lineage acute lymphoblastic leukemia (ALL).(2,3,6,7) Therefore, FLT3 is expected to be a potent molecular target in the development of efficient AML therapies.(8) Several inhibitors are now in preclinical investigation stages. Available clinical trial data show that there is a strong decline in peripheral blood blasts in response to current FLT3 inhibition-based therapies, but bone marrow responses are less common. Thus, molecular-targeting therapies against FLT3 are urgently sought as combination therapies that could potentially overcome drug resistance to FLT3 inhibitors. One of reasons for the drug resistance has been considered as the contribution of the leukemia microenvironment, harboring an abundance of FL.(9,10)

In the present study, we aimed to isolate human mAbs against FLT3 that could be useful in clinical use. Recently, our group established a procedure for the comprehensive identification of tumor-associated antigens through the extensive isolation of human mAbs that might be therapeutically advantageous.(11,12) Using this strategy, several mAbs with agonist and antagonist activities were isolated. Two of the identified mAbs are antagonists and continuously inhibit the FL-induced phosphorylation of FLT3 and MAPK. In addition, the A2 mAb induces complement-dependent cytotoxicity (CDC) and was shown to be efficiently internalized. Therefore, A2 is a potent human Ab able to deliver cytotoxic reagents and has an antagonistic effect on FLT3 signaling.

Materials and Methods

Cell lines.  The human leukemia cell lines, EOL-1, was purchased from RIKEN (Tsukuba, Japan). KOCL58 and KOPB26 were kindly provided by Dr Kanji Sugita (University of Yamanashi, Yamanashi, Japan).(9) These cell lines, BaF3/FLT3–internal tandem duplication (ITD), and BaF3/FLT3–D835Y(7) were maintained in RPMI-1640 medium (Sigma, St Louis, MO, USA) supplemented with 10% FBS (Equitech-Bio, Kerrville, TX, USA) at 37°C in a humidified atmosphere of 5% CO2. OCI-AML5 cells were purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ, Braunschweig, Germany) and maintained in alpha-minimum essential medium (MEM) with 20% FBS and 10% (v/v) conditioned medium from cell line 5637 (DSMZ).

Screening of the phage Ab library with the FLT3 antigen.  The AIMS5 phage Ab library was used as a source of human mAbs.(11,12) The antigen used in the screenings of the Ab library was prepared as follows: cDNA encoding an myc-tagged extracellular portion of human FLT3 (myc-rEC-FLT3) was inserted into a pcDNA3 expression vector.(13) The resultant plasmid DNA was transfected into 293T cells with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA), and the transformants were grown. The myc-rEC-FLT3 protein was purified from the supernatant of the cell culture with an anti-myc tag Ab-conjugated column. The myc-rEC-FLT3 protein was mixed with protein G Dynabeads (Veritas, Tokyo, Japan) and anti-myc tag Ab (MBL, Nagoya, Japan), and the mixture was incubated for 2 h at 4°C. Phages (1 × 1013 c.f.u.) from the AIMS5 library were then added to the mixture and incubated for 2 h at 4°C. Following several PBS washes using a magnetic device, phages that bound to the beads were directly infected into Escherichiacoli (E. coli) DH12S cells (Invitrogen). Following phage preparation, this panning procedure was performed three times. The phages prepared after three pannings were subjected to isolation of antigen/antibody complexes through organic solvent (ICOS) screening.(11) The EOL-1 cells expressing FLT3 were used as antigens. ICOS screenings were performed twice.

Selection of clones that bound to FLT3.  Following the final selection rounds, E. coli infected with the recovered phages were spread on plates, and several hundred colonies were picked. When E. coli with phagemid was grown without helper phages, the scFv-CL molecule fused with cp3 was secreted into the medium.(14) The supernatant of the culture was subjected to ELISA with anti-cp3 Abs (MBL) and with the myc-rEC-FLT3 protein. Clones that reacted with both the anti-cp3 Ab and myc-rEC-FLT3 protein were selected. The DNA sequence was determined using an ABI Prism 3100 genetic analyzer and a BigDye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA, USA).

Preparation of Ab forms.  The scFv-CL-cp3 levels were determined with ELISA assay. After the scFv-CL-cp3 molecules were converted into the scFv-CL-PP (P denotes a single Fc-binding domain of protein A) form of the Ab, they were purified with an IgG-conjugated column. IgG1 was prepared using a high-expression vector, and purified with a protein G-conjugated column.

Flow cytometry.  Cells (5 × 105) were incubated with 10 μg scFv-pp or IgG in 100 μL PBS for 45 min on ice. After the cells were washed with PBS, ALEXA488-conjugated antihuman IgG (Invitrogen) was incubated with the cells for 45 min on ice. After two washes, the cells were subjected to flow cytometry (FCM) analysis with a FACSCalibur system (Becton Dickinson, San Jose, CA, USA).

Immunoprecipitation and immunoblotting analysis.  Cells were first cultured in serum-free X-VIVO 10 medium (LONZA, Basel, Switzerland) overnight, and then treated with FL and/or in serum-free medium under various conditions. Following treatments, cells were lysed in lysis buffer: 50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5 mM EDTA, 10 mM NaF, 1 mM PMSF, 0.5 mM NaVO4, and 0.5 mM PhosSTOP (Roche, Basel, Switzerland). Equal amounts of cell lysate from each sample were mixed with anti-FLT3 Ab (C-20; Santa Cruz Biotechnology, Santa Cruz, CA, USA), agitated for 1 h at 4°C, and incubated with protein G agarose (Santa Cruz Biotechnology) for 2 h at 4°C. Protein mixtures were separated by SDS-PAGE, blotted onto a PVDF membrane, and probed with an anti-phosphotyrosine Ab (4G10; Upstate Biotechnology, Lake Placid, NY, USA) or anti-FLT3 Ab.(7) Proteins were visualized with chemiluminescence (Santa Cruz Biotechnology). To detect phospho-MAPK or the signal transducer and activator of transcription 5 (STAT5) protein, equal amounts of cell lysate were separated by SDS-PAGE, transferred onto a PVDF membrane (Immobilon-P; Millipore, Billerica, MA, USA) and immunoblotted with anti-phospho-MAPK or STAT5 Ab (Cell Signaling, Danvers, MA, USA), respectively. To detect total MAPK or the STAT5 protein, stripped membranes were reprobed with anti-MAPK or STAT5 Ab (Cell Signaling), respectively.(15)

Cell proliferation assay.  OCI-AML5 cells were first grown in the presence of FL (1.7 nM),(16) and were then split into a 96-well plate at a concentration of 2 × 104/100 μL. IgG Abs (67 nM) were added to the medium. Cells were then cultured with or without FL for 7 days. Cell growth was measured using CellTiter96 (Promega, Madison, WI, USA). Optical absorbance was determined using an ARVO X spectrophotometer (Perkin Elmer, Waltham, MA, USA). Experiments were performed in triplicate.

Dye-exclusion assay.  KOCL58 (2 × 104/mL) cells were precultured in the presence or absence of FL (1.7 nM) and the IgG Abs (67 nM) overnight. Cells were treated with daunorubicin (30 ng/mL; Meiji Seika, Tokyo, Japan) for 48 h at 37°C. Living and dead cells were counted in triplicate using a dye-exclusion assay.(9) The percentage of viability was calculated as (living cell number)/(total cell number) × 100.

Immunofluorescence analysis and microscopy.  The EOL-1 or BaF3 cells were incubated with the IgG Abs (67 nM) under various conditions. Cells were fixed for 30 min in methanol at −20°C. The slides were then incubated with 200 μL Image-iT FX signal enhancer (Molecular Probes, Carlsbad, CA, USA) for 30 min at room temperature. After washing three times with PBS, the cells were incubated with the primary Ab at a 1:50 dilution for 1 h at room temperature. Following the three PBS washes, cells were incubated for 1 h at room temperature with antihuman IgG–Alexa Fluor 488 Ab (Molecular Probes) at a 1:200 dilution. After three PBS washes, the slides were mounted with Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA, USA). Confocal laser images were captured with an Olympus BX51 microscope (Olympus, Tokyo, Japan). Original magnification was ×400 for all panels.(17)

Complement-dependent cytotoxicity and Ab-dependent, cell-mediated cytotoxicity.  Complement-dependent cytotoxicity and Ab-dependent, cell-mediated cytotoxicity (ADCC) were measured with a cytotoxicity detection kit (Roche). The assay measures lactate dehydrogenase (LDH) activity released from damaged cells. MOLM14 and KOCL58 were used as the target cells. Peripheral blood monocytes and, complement were isolated from human whole blood from healthy donors. Target cells, prepared as 1 × 106 cells/well, and containing several concentration of anti-FLT3 Abs, were incubated with effector cells (effector-to-target ratio: 50:1) or complement. To determine the percentage of cell-mediated cytotoxity, we calculated the average absorbance of the triplicates and substracted the background. These values were substituted as follows: Percentage of cytotoxity = ([effector − target cell mix − effector cell control] − low control)/(high control − low control) × 100. “High control” provides information about detergent lysed target cells; “low control” provides information about the untreated target cell.

Colony-forming cell assay.  Methocult (H4435; STEMCELL Technologies, Vancouver, Canada) was used for the colony-forming cell (CFC) assay according to the manufacturer’s instructions. CD34-positive bone marrow cells from healthy donors under an institutional protocol were isolated using the CD34-MicroBeads kit (Miltenyi Biotec, Bergisch Gladbach, Germany). The cells were cultured at 37°C in a humidified atmosphere of 5% CO2. Colony-forming cells were counted at day 14.

Results

Isolation of human mAbs specifically bound to FLT3.  The AIMS5 phage Ab library was used as a source of human mAbs. To isolate mAbs able to bind native FLT3 on the cell surface, a two-step screening technique was used, as described in the Materials and Methods. Of the several hundred clones picked at the final screening stage, 265 positively reacted with both anti-cp3 Ab and myc-rEC-FLT3 in the ELISA assays. These clones were classified by sequencing. Five unique clones were isolated and designated as A2, G3, 1B8, 1E3, and 3E6. The total numbers obtained for each clone were 239, 13, 10, 1, and 2 for A2, G3, 1B8, 1E3, and 3E6, respectively.

The binding ability of these clones to FLT3 on the cell surface was directly examined using FCM. As shown in Figure 1(a), these clones bound to the molecule on FLT3-expressing 293T cells, but not on 293T cells not expressing the receptor. We then examined interactions between FLT3, the anti-FLT3 Ab, and FL. In the first experiment, shown in the upper lane of Figure 1(b), FL was added following the incubation of EOL-1 cells with the mAb. In the second experiment, the anti-FLT3 Ab was added following the incubation of EOL-1 cells and FL (Fig. 1b, lower lane, green line). When FL was added following the formation of FLT3/anti-FLT3 Ab complexes, the complexes appeared to remain stable. Alternatively, after FL bound to the FLT3 expressed by EOL-1 cells, the Abs could no longer bind to the cell surface. These results suggest that the binding of the Abs to FLT3 on the cell surface clearly competes with that of FL to FLT3, and therefore, the Abs should potently intervene in FLT3 signaling.

Figure 1.

 FMS-related tyrosine kinase 3 (FLT3)-binding assay analyses and effects of isolated Abs on FLT3 phosphorylation. (a) FLT3-binding activity of the scFv form of A2, G3, 1B8, 1E3, and 3E6 Abs were examined using flow cytometry (FCM). CD135 is a polyclonal Ab against FLT3 and was used as a positive control. Open histogram represents the staining of FLT3-transfected 293T cells, and the shaded histogram represents the staining of mock-transfected 293T cells. (b) Effects of the FLT3 ligand (FL) treatment on FLT3–Ab interactions. Upper lane: Following the incubation of EOL-1 cells with the scFv form of the Ab, FL was added at 3.3 nM, and incubated for 10 min at 4°C. Cells were subjected to FCM with the second Ab. Green line represents binding with each scFv Ab, and the broken line represents the negative control with the scFv form of the clone 4G5. Lower lane: After the incubation of EOL-1 cells with 3.3 nM FL for 10 min at 4°C, the scFv form of the Ab was added and incubated. Cells were then subjected to FCM analysis with the second Ab. Red lines represent binding of the Ab without FL treatment. Green lines represent binding of the Ab following the addition of FL. Broken lines represent negative controls. (c) FLT3 ligand and Abs were separately added to the EOL-1 cells. FLT3 phosphorylation was examined as described in the Materials and Methods. (d) After the EOL-1 cells were incubated with mAb, FL was added.

Effects of scFv Abs on FLT3 phosphorylation.  The binding of FL to FLT3 on the cell surface induces FLT3 dimerization followed by autophosphorylation. The effects of the anti-FLT3 Abs on this process were examined. The Ab was first added to the EOL-1 cells without FL. As shown in Figure 1(c), 3E6 induced the same degree of FLT3 phosphorylation as FL. We then examined the effects of the mAbs on the FL-mediated phosphorylation of FLT3. Three of the isolated mAbs, A2, 1B8, and 1E3, inhibited the FL-mediated phosphorylation of FLT3, but G3 did not show any effect on FLT3 phosphorylation (Fig. 1d). In summary, 3E6 has an agonistic effect on FLT3 signaling, whereas A2, 1B8, and 1E3 have antagonistic effects on FL-induced phosphorylation. The scFv forms of these four mAbs were converted to complete human IgG1 Abs, and subjected to further analyses.

Agonistic effects of 3E6 on FLT3 signaling.  To further analyze the agonistic effects of 3E6 Ab on FLT3 signaling, two leukemia cell lines with abundant FLT3 expression on the cell surface, KOCL58 and KOPB26,(9) were treated with FL; the scFv form and the IgG form of 3E6. FLT3 and MAPK phosphorylation were then examined by Western blotting. As shown in Figure 2(a), the FL, scFv, and IgG forms of 3E6 induced both FLT3 and MAPK phosphorylation. To quantitatively examine the agonistic capability of 3E6, KOPB26 cells were treated with various concentrations of the scFv form of 3E6. As shown in Figure 2(b), MAPK phosphorylation was observed at concentrations >10 nM.

Figure 2.

 Agonistic effects of 3E6 on KOCL58, KOPB26, and OCI-AML5 cells. (a) After the KOCL58 and KOPB26 cell lines were incubated with 3E6 IgG (67 nM), 3E6 scFv (200 nM) or the FMS-related tyrosine kinase 3 (FLT3) ligand (FL) (1.7 nM) for 10 min, FLT3 and MAPK phosphorylation were evaluated by immunoblotting. (b) After the KOPB26 cells were incubated with 3E6 scFv at concentrations ranging from 0.1 to 40 nM for 10 min, MAPK phosphorylation was examined by Western blotting. (c) Growth of the OCI-AML5 cells was examined with either FL alone (1.7 nM), 3E6 scFv (200 nM), 3E6 IgG (67 nM), or controls at 7 days’ incubation. Cell growth was measured using the MTT assay. (d) Signal transducer and activator of transcription 5 (STAT5) phosphorylation on the OCI-AML5 cells with 3E6. MOLM14 cells, harboring the FLT3–ITD mutation, are the positive control.

The growth of OCI-AML5 cells requires specific hematopoietic growth factors, including FL, granulocyte-colony stimulating factor (G-CSF), granulocyte macrophage colony stimulating factor (GM-CSF), and interleukin-3.(16) To further assess the agonistic effects of 3E6, its effects on OCI-AML5 cells were examined (Fig. 2c). The growth of OCI-AML5 cells was supported to a greater degree by both the scFv (200 nM) and the IgG forms of 3E6 (67 nM), rather than by treatment with FL (1.7 nM).

In addition to MAPK signaling, FLT3–ITD also activates cell proliferation through STAT5 signaling.(18) To exclude the possibility that 3E6 transduces aberrant signaling through STAT5, we investigated the STAT5 phosphorylation of OCI-AML5 cells with 3E6 addition (Fig. 2d). As expected, 3E6 did not induce STAT5 phosphorylation.

Antagonistic effects of A2, 1B8, and 1E3 on FLT3 signaling.  The scFv forms of A2, 1B8, and 1E3 demonstrated inhibitory effects on FL-induced FLT3 phosphorylation in EOL-1 cells (Fig. 1d). The effects on FLT3 and MAPK phosphorylation were also examined using the IgG forms of the Abs on three acute leukemia cell lines: EOL-1, KOCL58, and KOPB26. Although 1.7 nM FL induced strong FLT3 and MAPK phosphorylation, the IgG forms of A2 and 1B8 (67 nM) clearly inhibited FLT3 phosphorylation, and the IgG form of A2 inhibited MAPK phosphorylation in 1-h incubation (Fig. 3a). Furthermore, KOCL cells were incubated for 48 h in the presence of both 1.7 nM FL and 67 nM IgG mAbs (Fig. 3b). The inhibition of the phosphorylation of MAPK by these two mAbs was observed over this extended culture period. We then quantitatively evaluated the antagonistic capability of A2 on MAPK phosphorylation of the cell line induced with FL. After the KOCL58 cells were incubated with A2 IgG at various concentrations, 1.7 nM FL was added, and MAPK phosphorylation was examined (Fig. 3c). Phosphorylation was clearly inhibited by A2 IgG at concentrations >1.7 nM.

Figure 3.

 Antagonistic effects of A2, 1B8, and 1E3 on FMS-related tyrosine kinase 3 (FLT3) signaling. (a) Leukemia cell lines were incubated with IgG Abs against FLT3 (67 nM) for 60 min with or without subsequent incubation with the FLT3 ligand (FL) (1.7 nM). (b) KOCL58 cells were incubated with IgG Abs (67 nM) with or without FL (1.7 nM) for 48 h. (c) KOCL58 cells were incubated with A2 IgG in doses ranging from 0.34 to 67 nM with FL (1.7 nM) for 10 min. (d) Inhibitory effects of the IgG form of A2, 1B8, and 1E3 on the growth of the OCI-AML5 cell line. OCI-AML5 line, which is a factor-dependent leukemia cell line, was grown with the IgG form of the Abs (67 nM) and FL (1.7 nM) for 7 days. Cell growth was measured using the MTT assay. Degree of growth inhibition was calculated by comparing growth with and without FL. Error bar shows the standard deviation. *< 0.05. (e) Dose-dependent inhibition of FL-induced proliferation of OCI-AML5 cells by A2 IgG. Concentration of FL is 1.7 nM. Cell growth was measured using the MTT assay. (f) A2 IgG alone did not induce the proliferation of OCI-AML5 cells. FL (1.7 nM) is the positive control. Cell growth was measured using the MTT assay. (g) Addition of A2 or 1B8 IgG (6.7 nM) did not inhibit the phosphorylation of FLT3 and MAPK in BaF3 cells expressing FLT3–internal tandem duplication (ITD) or FLT3–D835Y.

Furthermore, the effect of the IgG forms of the three antagonistic mAbs on the FL-dependent growth of OCI-AML5 was examined. As indicated in Figure 3(d), treatment with A2, 1B8, and 1E3 resulted in growth inhibitions of 52.8%, 52%, and 17.1% relative to the controls, respectively. These results indicate that A2 and 1B8 possess antagonistic effects on FLT3 signaling. In addition, Figure 3(e,f) shows that A2 has a dose-dependent inhibition of FL-induced proliferation, and A2 alone does not induce the proliferation of OCI-AML5 cells, respectively.

Finally, we evaluated whether A2 antagonizes mutant FLT3 signaling in BaF3 cells. As shown in Figure 3(g), A2 and 1B8 did not have inhibitory effects in mutant FLT3 phosphorylation and MAPK signaling.

Antagonistic IgG enhance daunorubicin-induced cytotoxicity of FLT3-positive leukemia cells in the presence of FL.  It is notable that FLT3 is highly expressed in acute leukemias harboring the myeloid/lymphoid or mixed lineage leukemia gene rearrangement,(19) and because of FL stimulation from bone marrow stromal cells, these cells become resistant to chemotherapy.(9,20) In this context, we further examined the chemotherapy-sensitivity of the KOCL58 cell line, which has the MLL gene rearrangement,(9) in the presence of FL and antagonistic anti-FLT3 mAbs. Without FL, daunorubicin treatment permitted only 17.7% of KOCL58 cells to survive (Fig. 4a). However, the addition of FL (1.7 nM) interfered with the daunorubicin-induced cytotoxicity of KOCL58 cells up to 51% viability. When 1.7 nM FL was added together with 67 nM A2 and 1B8 IgG, the daunorubicin-induced cytotoxicity of KOCL58 cells recovered again to 34% and 36% viability, respectively. In addition, Figure 4(b) shows the dose-dependent augmentation of daunorubicin-induced cytotoxicity by the A2 Ab. These results suggest that antagonistic IgG against FLT3 enhance the daunorubicin-induced cytotoxicity of FLT3-positive leukemia cell lines, even in the presence of FL, which presumably is supplied from the bone marrow and peripheral blood environments.(20,21)

Figure 4.

 Effects of the FMS-related tyrosine kinase 3 (FLT3) ligand (FL) and antagonistic anti-FLT3 mAbs on chemotherapy-induced cytotoxicity. (a) KOCL58 cells were precultured in the presence or absence of FL (1.7 nM) and the IgG Abs (67 nM) overnight. Cells were then treated with daunorubicin (DNR) (30 ng/mL) for 48 h at 37°C. In the control experiments, the addition of FL rescued KOCL58 cells from DNR-induced cytotoxicity with FL-treated cells and -untreated cells showing 17.7% and 51% viability, respectively. Incubation with the IgG form of A2 or 1B8 (67 nM) restored DNR-induced cytotoxicity to 34.7% and 37.3% viability, respectively. Cell viability was determined using a dye-exclusion assay. (b) Dose-dependent augmentation of DNR-induced cytotoxicity of KOCL58 cells by A2 IgG. Concentration of FL was 1.7 nM. Error bar shows the standard deviation. *< 0.05.

Internalization of anti-FLT3 mAbs into FLT3-expressing cells.  There are many cases where antireceptor tyrosine kinase (RTK) mAbs are internalized following complex formation with the target RTK. Immunofluorescence analyses were performed to examine whether the isolated anti-FLT3 mAbs could internalize into EOL-1 cells. Cells were incubated with the IgG forms of the four isolated Abs (A2, 1B8, 3E6, and 1E3) at 4°C for 15 min. The IgG forms of A2, 1B8, and 3E6 were localized on the cell membrane (Fig. 5a, upper lane). While the scFV form of all four mAbs bound to the cells, the IgG form of 1E3 showed no binding to the cells. The FCM analyses also indicated that the IgG form of IE3 lost FLT3-binding activity (Fig. 5a). Furthermore, we examined the localization of the Abs in differential conditions. After the incubation was prolonged to 4 h, A2 and 1B8 were localized not only on the cell membrane, but also in the cytoplasm (Fig. 5b, upper lane). This tendency was accelerated when the temperature was increased to 37°C for 4 h (Fig. 5b, middle lane). These results were confirmed by FCM analysis, where that the number of cells showing surface binding to A2, 1B8, and 3E6 was attenuated from 69.%, 93.1%, and 95% to 11.8%, 44.7%, and 49.2%, relative to the controls, respectively (Fig. 5b, lower lane). These results indicate that the A2 IgG efficiently internalized into EOL-1 cells. Finally, we examined the internalization of A2 IgG on BaF3/mutant-FLT3 cells. As shown Figure 5(c, lower), A2 internalized into BaF3/FLT3–D835Y cells, as well as EOL-1 cells. However, A2 did not internalize into BaF3/FLT3–ITD cells (Fig. 5c, upper).

Figure 5.

 Localization of anti-FMS-related tyrosine kinase 3 (FLT3) mAbs in FLT3-expressing cells. (a) EOL-1 cells were incubated with the IgG Abs (67 nM) for 15 min at 4°C. Subcellular localization of the IgG Abs was analyzed by confocal laser microscopy (upper lane). DAPI was used for nuclear staining. Surface binding of the IgG Abs was analyzed by flow cytometry (FCM) without any reagents for permeabilization (lower lane). Shaded histograms are negative controls. Positive cell population for FCM was calculated by gating and is indicated at the bottom of the Figure. (b) EOL-1 cells were incubated with the IgG Abs (67 nM) for 4 h. Incubation temperature was 4°C (upper lane) or 37°C (middle lane). Surface binding of the IgG Abs was analyzed by FCM without any reagents for permeabilization (lower lane). (c) Internalization of mutant FLT3 in BaF3 cells by A2 IgG. Concentration of IgG was 67 nM. Black lines represent incubation for 4 h at 4°C. Red lines represent incubation for 4 h at 37°C. Shaded histograms are negative controls.

A2 and 1B8 IgG induce CDC, but not ADCC.  Complement-dependent cytotoxicity and ADCC are major effects for therapeutic Abs. To evaluate them in FLT3-expressing AML cell lines, we used MOLM14 or KOCL58 cells with either A2 or 1B8 IgG. Figure 6 shows that A2 and 1B8 IgG induce CDC mildly, but do not induce any ADCC.

Figure 6.

 Effects of A2 and 1B8 IgG on complement-dependent cytotoxicity (CDC) and Ab-dependent, cell-mediated cytotoxicity (ADCC). MOLM14 (a,c) or KOCL58 (b,d) cells were incubated with complement (a,b) or effector cells (c,d) with several concentration of anti-FMS-related tyrosine kinase 3 (FLT3) Abs. Error bar shows the standard deviation.

A2 IgG does not have inhibitory effects on normal hematopoietic progenitor cells.  FLT3 is a class III receptor tyrosine kinase expressed on early hematopoietic progenitor cells.(1–4) To evaluate the possible inhibitory effects of FLT3 signaling with A2 IgG on normal CD34-positive progenitor cells, we performed standard methylcellulose culture assays. Figure 7 shows that A2 IgG does not have inhibitory effects in CFC assays.

Figure 7.

 Effects of A2 IgG on normal hematopoietic progenitors in vitro. CD34-positive bone marrow cells were incubated with A2 or control IgG (6.7 nM) in standard methylcellulose culture for 14 days. Each number on the lower lane indicates a sample from a donor. Error bar shows the standard deviation. BFU-E, burst-forming unit–erythroid, CFU-E, colony-forming unit–erythroid, CFU-GEMM, colony-forming unit–granulocyte, erythroid, macrophage, megakaryocyte; CFU-GM, colony-forming unit–granulocyte macrophage.

Discussion

FLT3 is a unique target for the development of molecular-targeting therapy for the treatment of leukemia,(8) and is frequently expressed in AML and ALL cells.(2,3) Recently, it was reported that differentiation into all hematopoietic lineages involves Flt3-expressing, non-self-renewing progenitors.(4) In contrast, hematopoietic stem cell (HSC) origin and maintenance do not include an Flt3-expressing stage.(4) This finding means that eradication of FLT3-expressing cells does not eliminate normal HSC populations. The FLT3 ligand is expressed at high levels as a soluble or membrane-bound form on bone marrow stromal cells.(22) In bone marrow, the FL/FLT3 interaction from bone marrow stromal cells contributes to persistent minimal residual disease in FLT3-expressing leukemia cells.(9) To inhibit FL/FLT3 interaction in leukemia cells, several small-molecule tyrosine kinase inhibitors against FLT3 are in clinical testing, and some have shown limited clinical activity in patients with relapsed or refractory AML. However, the depth and duration of clinical responses to FLT3 inhibitor monotherapies have been modest.(8,23,24) One of reasons why leukemic blasts acquire resistance to the inhibitors is the contribution of the leukemia microenvironment, harboring an abundance of FL, to the drug resistance of leukemic stem cells.(8,10,23,24)

For developing novel Ab-based FLT3 inhibitors, we isolated several anti-FLT3 mAbs with agonist or antagonist activities by phage-Ab library-based screening. Two of the identified mAb (A2 and 1B8) are antagonists, and 3E6 is an agonist. Because of the functional difference between antagonists and agonists, Verstraete et al.(25) recently described the structural base of FL/FLT3 interaction. According to the results, the N-terminal loop of FL binds with the FLT3 D3 domain, in which FL/FLT3 interaction occurs in a single contact site covering <900Å(2) of buried surface area, which is two times less extensive compared with other class III receptor tyrosine kinases. They suggest that FL/FLT3 interaction is reminiscent of a classic lock and key binding mode observed in affinity-matured Ab–antigen interactions. Therefore, we speculate that 3E6 interacts with the FL-binding epitope on the FLT3 D3 domain; A2 and 1B8 interact with the FL/FLT3 binding surface, except the FL-binding epitope; and the G3 Ab interacts with FLT3 outside of the FL/FLT3 binding surface.

To overcome the obstacles for developing FLT3-targeting therapy, combination therapy with antagonizing mAbs against FLT3 is considered. In this study, pharmacokinetics of A2 in vivo was not yet determined, and the addition of more than 1.7 nM A2 IgG was found to interfere with the addition of 1.7 nM FL in vitro. Plasma FL levels in AML patients are reported as being mostly <3 ng/mL (0.1 nM).(10) In addition, membrane-bound FL surrounds leukemia cells in bone marrow.(1,5) Although it is difficult to estimate precisely how many FL molecules interact with leukemia cells in bone marrow, A2 IgG might be enough to overcome the cytoprotective effects of FL in patients because of following reasons. Because the median value of Ab serum levels reported is 554 nM for responders using rituximab pharmacokinetics,(26) it would be possible that A2 IgG antagonizes FL-induced cytoprotective signaling in clinical settings, if the serum levels of A2 IgG achieved are as high as those of rituximab.

Somatic mutations in the FLT3 gene, including ITD and activation loop mutations, are found in up to 30% of AML cases.(2,3,7) Generally, these FLT3 mutations activate its downstream signaling without FL. We have further examined whether the antagonistic Abs inhibit FLT3 signaling in mutant FLT3-expressing cells. As expected, no significant differences of phospho-FLT3 and phospho-MAPK were observed. Two recent reports indicated that mutated FLT3 is anchored in the perinuclear endoplasmic reticulum and initiates aberrant signaling cascades before translocation of FLT3 to the cell membrane.(27,28) Therefore, it is consistent that the administration of single antagonistic mAbs against FLT3 did not have significant inhibition of mutated FLT3 signaling. Unexpectedly, FLT3–ITD-expressing cells did not internalize A2 IgG. This result might imply that the juxtamembrane domain of FLT3 has important roles for FLT3 internalization, but this requires further investigation.

Complement-dependent cytotoxicity and ADCC are the most important effects of therapeutic Abs for cancer treatment.(29,30) We investigated whether the selected IgG against FLT3 demonstrated CDC an ADCC on FLT3-expressing cell lines. The results show that the IgG forms of A2 and 1B8 induce mild CDC, but not ADCC effects. It is notable that panitumumab does not show ADCC, but is useful in patients.(29)

Gemtuzumab ozogamicin is an Ab-targeted chemotherapy agent for CD33-positive AML, and is conjugated to a derivative of the antitumor compound, calicheamicin.(31) The Ab/antigen complex is internalized into target cells, and the cytotoxic calicheamicin is released intracellularly through hydrolysis. We have shown that A2 IgG efficiently internalize into cells. Therefore, A2 could be utilized as a vehicle of cytotoxic reagents, including antibiotics, radioisotopes, apoptosis-inducing molecules,(32,33) and liposomes(34) carrying siRNA against specific oncogenic fusion transcripts.(35)

FLT3 is expressed at not only hematopoietic progenitors, but also dendritic cells in peripheral lymphoid tissue.(36) The FLT3 ligand treatment dramatically increases the number of functionally-mature dendritic cells in mice.(37) Dendritic cells are powerful antigen-presenting cells and show a remarkable capacity to stimulate antigen-specific T-cell responses.(38) Some reports suggest that FL administration inhibits tumor growth in mice models.(39,40) The agonistic Ab 3E6 is particularly interesting, because virtually a single molecule of the scFv form of 3E6 can induce FLT3 signaling. Although the capability of 3E6 to induce signaling is less than that of FL, the half-life time of therapeutic Abs is more than 14 days.(26) Therefore, it might be possible that 3E6 IgG has long-acting agonistic effects in FLT3 signaling in vivo, which lead to novel dendritic cell-based immunotherapies.(38-41)

In summary, A2 is a potent human Ab that has an antagonistic effect, and 3E6 has an agonistic effect on FLT3 signaling. These human mAbs could be potential scaffolds for developing novel FLT3-targeting therapies.

Acknowledgments

The pCAGGS was a gift from Dr Jun-ichi Miyazaki (Department of Geriatric Medicine, Osaka University, Osaka, Japan). The KOCL58 and KOPB26 cells were kindly provided by Dr Kanji Sugita (Department of Pediatrics, University of Yamanashi, Yamanashi, Japan). We thank Ms Akemi Endo for technical assistance. This study was supported in part by a grant-in-aid for the 21st Century Center of Excellence (COE) Program of Fujita Health University from the Ministry of Education, Culture, Sports, Science, and Technology; a grant from the New Energy and Industrial Technology Development Organization (NEDO) to YK; a a grant for Research on Pharmaceutical and Medical Safety from the Ministry of Health, Labor, and Welfare to YK; a grant-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to YY and YK; and a grant from the Takeda Science Foundation to YY.

Disclosure Statement

The authors declare no financial or commercial conflict of interest.

Ancillary