Acute Leukemias

Activation of FGFR1β signaling pathway promotes survival, migration and resistance to chemotherapy in acute myeloid leukemia cells

Article metrics

Abstract

Fibroblast growth factors (FGFs) are important regulators of hematopoiesis and have been implicated in the tumorigenesis of solid tumors. Recent evidence suggests that FGF signaling through FGF receptors (FGFRs) may play a role in the proliferation of subsets of acute myeloid leukemias (AMLs). However, the precise mechanism and specific FGF receptors that support leukemic cell growth are not known. We show that FGF-2, through activation of FGFR1β signaling, promotes survival, proliferation and migration of AML cells. Stimulation of FGFR1β results in phosphoinositide 3-kinase (PI3-K)/Akt activation and inhibits chemotherapy-induced apoptosis of leukemic cells. Neutralizing FGFR1-specific antibody abrogates the physiologic and chemoprotective effects of FGF-2/FGFR1β signaling and inhibits tumor growth in mice xenotransplanted with human AML. These data suggest that activation of FGF-2/FGFR1β supports progression and chemoresistance in subsets of AML. Therefore, FGFR1 targeting may be of therapeutic benefit in subsets of AML.

Introduction

The survival, proliferation and invasiveness of leukemic cells are dependent on a complex network of cytokines, chemokines and angiogenic factors released by leukemic cells themselves and their microenvironment.1 Fibroblast growth factors (FGFs), a family of heparin-binding polypeptides, affect cellular proliferation, motility and survival in a broad spectrum of cells and are important regulators during hematopoiesis.2, 3, 4 To date, more than 20 FGFs have been isolated, including the FGF prototypes FGF-1 (acidic FGF) and FGF-2 (basic FGF). Fibroblast growth factors signal through five different high-affinity tyrosine kinase FGF receptors (FGFR1–5).5 Recent data have shown that FGF-2 is overexpressed in bone marrow samples of acute myeloid leukemia (AML) patients,6 and that AML blasts express FGFRs,6, 7, 8, 9 suggesting a possible role of autocrine or paracrine FGF/FGFR signaling loops in AML. However, the specific type of the FGFR supporting leukemic cell progression is not known. Moreover, the pathophysiological role of FGFR signaling in AML remains poorly defined.

Indeed, in vitro studies have yielded conflicting results in regard to FGFR expression and FGF-mediated growth stimulation in AML.6, 7, 8, 9 To dissect the mechanism of FGFR1 signaling in AML, we have used a novel anti-FGFR1 neutralizing antibody on AML cell lines, which express functional FGFR1β and are responsive to FGF-2 activation. The effects of FGF-2 stimulation on proliferation and migration were completely or partially abrogated by anti-FGFR1 neutralizing antibody, pointing to active FGF-2/FGFR1β signaling loops. We demonstrate that FGF-2/FGFR1β signaling mediates phosphoinositide 3-kinase (PI3-K)/Akt activation and promotes survival, proliferation, migration and protection from chemotherapy in subsets of FGFR1β-positive AML. In addition, we show that these FGF-2-induced effects can be blocked effectively by an FGFR1-neutralizing antibody in vitro, and that FGFR1 blockade is also effective in vivo using a human leukemia xenograft model. These data suggest that FGF-2 signaling through FGFR1β plays a critical role in progression and chemoresistance of AML subsets, and may set the stage for future FGFR-directed therapy to treat human leukemias.

Materials and methods

All reagents were obtained from Sigma (St Louis, MO, USA) unless otherwise stated.

Cell culture, cytokines and neutralizing antibodies

The following myeloid/monocytic (AML) cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA): HL-60, KG-1a (acute myeloid leukemia), HEL (HEL 92.1.7, erythroleukemia), THP-1 (acute monocytic leukemia), K-562 (chronic myeloid leukemia in blast crisis), NB4 (acute promyelocytic leukemia) and U937 (histiocytic lymphoma). One cytogenetically normal AML cell line (designated R81), which was established in our laboratory from primary AML in an 81-year-old male, was used additionally. The immunophenotype of R81 AML cells in culture was CD7+, CD13+, CD33+, CD3, CD19, CD34, CD117 consistent with immature myeloid blasts. All leukemic cell lines were cultured in IMDM or RPMI with L-glutamine and HEPES, supplemented with 10% (v/v) fetal bovine serum. Before all experiments, leukemic cells were starved in serum-free X-VIVO 20 medium (Bio-Whittaker, Walkersville, MA, USA) for 12 h. Recombinant hgfuman FGF-2 and stromal cell-derived factor 1α (SDF-1α) were obtained from PeproTech Inc. (Rocky Hill, NJ, USA). IgG1 human anti-human FGFR1-specific neutralizing antibody (clone H7) was kindly provided by ImClone Systems Inc. (New York, NY, USA). For all experiments, FGF-2 and anti-FGFR1 neutralizing antibody were used at concentrations of 20 ng/ml and 10 μg/ml, respectively, unless stated otherwise.

Proliferation assays

Leukemic cells were seeded in 96-well plates at a concentration of 2.5 × 105 cells/ml. Anti-FGFR1 neutralizing antibody, FGF-2 and/or chemotherapeutic agents were added subsequently. Cells were manually counted using a hemocytometer and viability was determined by Trypan blue exclusion. Each experiment was performed in triplicate and repeated three times.

MTT assay

Leukemic cells were seeded in 96-well plates at a concentration of 2 × 105 cells/ml. Anti-FGFR1 neutralizing antibody and/or FGF-2 were added subsequently. Cells were incubated for 48 h and MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) reagents used as per the manufacturer's instructions. Extinction was measured at 540 nm and reference extinction subtracted. Each experiment was performed in triplicate and repeated three times.

Transwell migration assay

Leukemic cell migration was assayed using a Boyden chamber system with 5 μm pore size inserts (Corning, Corning, NY, USA). Serum-starved leukemic cells were incubated with or without anti-FGFR1 neutralizing antibody. Leukemic cells (1 × 106 per insert) were then placed into the upper chamber of the transwell system in X-VIVO 20 medium. The lower chamber was filled with X-VIVO 20 medium in the presence or absence of 20 ng/ml of FGF-2. Recombinant human SDF-1α (30 ng/ml) was used as a control chemoattractant. The plates were then incubated at 37°C for 8 h, transmigrated cells in the lower chamber were harvested and the number of viable migrated cells was assessed by Trypan blue exclusion. The transwell migration assay was performed three times in triplicate.

Apoptosis assay

To evaluate anti-apoptotic effects of FGF-2/FGFR1 signaling in leukemic cells, serum-starved leukemic cells were incubated with or without FGF-2 and/or anti-FGFR1 neutralizing antibody and the chemotherapeutic agents etoposide (0.1–0.5 μ M), cytarabine (1–10 μg/ml) and daunorubicin (1–10 μg/ml) in X-VIVO 20 medium, collected after 24 h and analyzed for the presence of apoptotic cells using the ApoAlert annexin V-FITC propidium iodide (PI) apoptosis kit (Becton Dickinson, Palo Alto, CA, USA), following the manufacturer's instructions. All experiments were performed three times.

RNA extraction, cDNA synthesis and reverse transcription-PCR

Total RNA was extracted using the RNeasy mini kit (Qiagen, Valencia, CA, USA) and cDNA was synthesized from total RNA using the SuperScript III first-strand synthesis system (Invitrogen, Carlsbad, CA, USA) as per the manufacturers’ instructions. A cDNA fragment encoding the kinase region of FGFR1 was amplified with the primers 5′-IndexTermTACAAGATGAAGAGTGGTA-3′ (forward) and 5′-IndexTermTCCACGATGACATACAAG-3′ (reverse) using the FGFR1 SingleGene PCR kit (SuperArray, Frederick, MD, USA), including a gylceraldehyde-3-phosphate dehydrogenase (GAPDH) internal normalizer control, as per the manufacturer's instructions. Expected specific PCR product sizes were 496 bp for FGFR1 and 226 bp for GAPDH. To determine the FGFR1 isotype expressed by R81 AML cells, a cDNA fragment encoding the extracellular domains of FGFR1 was amplified using the primers 5′-IndexTermTGTCACCAACCTCTAACTGC-3′ (forward) and 5′-IndexTermCTTGTAGACGATGACCGACC-3′ (reverse) located in the 5′ untranslated and transmembrane coding regions of FGFR1, respectively. The PCR products were subsequently analyzed by direct DNA sequencing.

Western blot analysis

Acute myeloid leukemia cells were serum starved overnight in serum-free medium and subsequently incubated with or without the addition of FGF-2 (20 ng/ml) for 30 min. Cells were spun down and lysed in radioimmunoprecipitation assay (RIPA) buffer. For immunoprecipitation and Western blot analysis of FGFR1, 500 μg of total protein from each sample and 50 μg of total cell lysate from L6 cells (negative control) and L6R1c cells (positive control expressing full-length FGFR1α receptor (FGFR1-IIIc), kindly provided by ImClone Systems Inc., New York, NY, USA), were incubated with 100 μg/ml rabbit polyclonal anti-FGFR1 antibody (sc-121; Santa Cruz Biotechnology, Santa Cruz, CA, USA) followed by addition of a 1:2 slurry of Protein A/Protein G beads (Amersham Biosciences, Piscataway, NJ, USA) in RIPA buffer and overnight incubation at 4°C. The supernatant was loaded onto a 4–12% Bis-Tris Gel (Invitrogen, Carlsbad, CA, USA), electrophoresed and transferred to a Trans-Blot Nitrocellulose Membrane (Bio-Rad Laboratories, Hercules, CA, USA). Fibroblast growth factor receptor 1 protein was detected using the polyclonal rabbit anti-FGFR1 antibody (sc-121) at a 1:500 dilution. Phosphorylated FGFR1 was detected using rabbit anti-phosphotyrosine antibody (CalBiochem, San Diego, CA, USA) at a 1:1000 dilution. Western blot analysis of Akt/phospho-Akt was performed as previously described.10 Intensity of Western blot bands was quantified by densitometry using the NIH ImageJ software, version 1.34j (NIH, Bethesda, MD, USA).

Animal experiments

Approval for performing animal experiments was obtained from the institutional animal review board at Weill-Cornell Medical College. Age- and sex-matched non-obese diabetic severe combined immunodeficient (NOD-SCID) mice were used for in vivo experiments. For the first experiment (3 days treatment duration), R81 AML cells (1 × 107/mouse) were injected subcutaneously into 12 mice. Seven days after injection, when the subcutaneous tumor size had reached a diameter of approximately 3 mm, the animals were divided into two groups of six mice each. One group was treated intraperitoneally with 100 μg of anti-FGFR1 neutralizing antibody diluted in 100 μl phosphate-buffered saline daily for 3 days, and the control group received 100 μg of a sterile human IgG preparation (Bayer, Elkhart, IN, USA) daily. All animals were killed after 3 days of treatment and the subcutaneous tumors removed. For the second experiment (25 days treatment duration), the animals were divided into two groups of six mice each 3 days after subcutaneous injection. One group was treated intraperitoneally with 25 μg of anti-FGFR1 neutralizing antibody every other day, and the control group received 25 μg of a sterile human IgG preparation every other day. All animals were killed after 25 days of treatment (i.e. 28 days after injection) and the subcutaneous tumors removed. Tumor volume was measured weekly beginning at week 2 after injection, and calculated using the formula mm3=length × (width)2 × π/6.

Histology, immunohistochemistry and in situ cell death detection

To assess the extent of intratumor vascularization, paraffin-embedded tumor sections were incubated overnight at 4°C with a biotinylated rat antibody against mouse endothelial cell antigen MECA-32 (BD Biosciences Pharmingen, San Diego, CA, USA).11 The microvessel density (MVD) was determined as previously described.12 To assess the proliferative index, immunostaining for Ki-67 was used. Briefly, tumor sections were antigen retrieved in a steamer and sections were incubated with rabbit anti-human polyclonal antibody against Ki-67 (LabVision, Fremont, CA, USA) for 1 h, followed by secondary biotinylated anti-rabbit antibody (Jackson ImmunoResearch, West Grove, PA, USA) and streptavidin-horseradish peroxidase and DAB+ as per the manufacturer's instructions (DAKO, Carpinteria, CA, USA). Proliferative index, that is, total number of cells divided by Ki-67-positive cells, was averaged from three microscopic fields per slide at × 250 magnification in a blinded fashion using the NIH ImageJ software, version 1.34j with cell counter plugin (NIH, Bethesda, MD, USA). Detection of cell death in the leukemia xenografts was performed by TdT-mediated dUTP nick end labelling assay (Roche, Indianapolis, IN, USA) according to the manufacturer's protocol. Three stained slides per mouse were evaluated for statistical analysis. On each slide, three microscopic fields were examined at × 250 magnification (representing an area of 0.72 mm2 per field), and the mean number of TUNEL-positive cells was recorded. The number of apoptotic cells, that is, TUNEL-positive cells, is presented as mean number and standard deviation per microscopic field. To identify stromal cells/fibroblasts, the Sigma Accustain Trichrome Stain Kit was used as per the manufacturer's instructions.

Statistical analysis

Statistical significance was determined using an unpaired Student's t-test. The minimal level of significance was P<0.05.

Results

FGFR1 mRNA is expressed in subsets of AML

Fibroblast growth factor receptor 1 expression was found on the mRNA level in four out of eight AML lines using reverse transcription-PCR (RT-PCR) for the receptor kinase region (Figure 1).

Figure 1
figure1

Acute myeloid leukemia subset cells expressing fibroblast growth factor receptor 1 (FGFR1) mRNA. Acute myeloid leukemia cell lines expressing FGFR1 mRNA transcripts. Upper band represents FGFR1-specific reverse transcription-PCR amplificates (496 bp) corresponding to the kinase region of the receptor. Lower band represents gylceraldehyde-3-phosphate dehydrogenase (GAPDH) internal control (226 bp). HEL: FGFR1-negative erythroleukemia cell line (negative control).

FGF-2 signaling through FGFR1 promotes proliferation and migration in subsets of FGFR1-positive AML

To assess the effects of FGF-2 on growth and migration in FGFR1-positive leukemias, cells were plated in 96-well plates using serum-free medium as described in Materials and methods. Subsequently, FGF-2 (20 ng/ml) was added daily, and viable cells were counted at 72 h. A significant increase in proliferation (approximately twofold, P<0.05) was detected only in R81 AML cells. Remarkably, this increase in proliferation was completely abrogated by a novel anti-FGFR1 antibody, which specifically blocks the activation of FGFR1 signaling (Figure 2a). The proliferation assay results were confirmed by MTT assay (Figure 2b). Because leukemic invasiveness has been implicated in the spread of leukemias to various organs, we evaluated the capacity of FGF-2 to induce motility of FGFR1-positive AML cells. Cell migration was assayed by measuring the transmembrane migration activity across transwell inserts (8 μm pore size). Notably, FGF-2 induced a significant increase in both HL-60 and R81 cell migration as compared to control, whereas coincubation with anti-FGFR1 neutralizing antibody decreased the FGF-2-induced transmigration to baseline (P<0.05; Figure 2c). In summary, these data suggested that FGFR1 signaling in AML can induce a variety of effects that are biologically relevant to leukemic pathophysiology.

Figure 2
figure2

Fibroblast growth factor-2/fibroblast growth factor receptor 1 (FGF-2/FGFR1) signaling supports proliferation and migration of acute myeloid leukemia (AML) subsets. (a) Fibroblast growth factor-2-induced leukemic cell proliferation is observed in R81 AML cells only, but is abrogated by blocking ligand binding to FGFR1 using an anti-FGFR1 neutralizing antibody in (*P<0.05, significant difference compared to no stimulation). (b) Proliferation results are confirmed by MTT assay showing increase in proliferation upon FGF-2 stimulation and reversal of this effect by neutralizing anti-FGFR1 antibodies. (c) The FGF-2/FGFR1 axis induces migration of AML subsets in serum-free medium. Fibroblast growth factor-2 stimulates migration of HL60 and R81 leukemia cells through 8 μm transwell inserts, and this effect is abrogated by anti-FGFR1 neutralizing antibody. *P<0.05 (significant difference compared to medium alone).

FGFR1β mRNA and protein are predominantly expressed in FGFR1-positive AML cells, and FGF-2 induces phosphorylation of FGFR1β

To confirm that FGF-2-induced proliferation and migration in R81 AML cells is mediated through FGFR1 signaling, we detected FGFR1 protein by immunoprecipitation with polyclonal FGFR1 antibody followed by Western blot with the same antibody. An FGFR1-specific band with a molecular weight of approximately 100 kDa was detected, corresponding to the FGFR1β isoform (Figure 3a). Subsequently, increased phosphorylation of FGFR1β by FGF-2 was demonstrated by Western blot (Figure 3b). The predominant expression of FGFR1β by FGFR1-positive AMLs was further confirmed by RT-PCR amplifying the coding regions for the extracellular domains of FGFR1 (Figure 3c) and subsequent DNA sequencing. Taken together, these data show that FGF-2/FGFR1β signaling is functional in supporting growth and migration in subsets of FGFR1-positive AML cells.

Figure 3
figure3

Fibroblast growth factor receptor 1 (FGFR1)-positive acute myeloid leukemia (AML) cells predominantly express functional FGFR1β. (a) R81 AML cells express FGFR1β protein. Immunopreciptation followed by Western blot with polyclonal anti-FGFR1 antibody shows a band at approximately 120 kDa in L6 cells transfected with a full-length FGFR1α construct (positive control). A band with a molecular weight of approximately 100 kDa, corresponding to FGFR1β, is detected in R81 AML cells. (b) Fibroblast growth factor-2 (FGF-2) induces FGFR1 phosphorylation in R81 AML cells. Probing the stripped blot from (b) with an anti-phospho-tyrosine antibody shows that FGF-2 induces phosphorylation of FGFR1 protein. (c) Reverse transcription-PCR for FGFR1 isotype determination in AML cells showing predominant amplificate of approximately 900 bp consistent with FGFR1β isotype with two extracellular immunoglobulin domains, confirmed by direct DNA sequencing of the PCR product.

FGF-2 induces phosphorylation of Akt in R81 AML cells, which can be abrogated by anti-FGFR1 neutralizing antibody

Given the positive effects of FGFR1β signaling on proliferation and migration, we hypothesized that downstream activation of Akt may occur in AML and that this activation may be blocked by anti-FGFR1 neutralizing antibody. Using Western blot analysis, we showed that FGF-2 induced increased phosphorylation of Akt in FGFR1-positive AML cells except for NB4 (threefold increase in HL-60 and U937 cells, 16-fold in R81 cells; Figure 4). Interestingly, the increase in phosphorylation was highest in R81 AML cells, which also showed a relatively low level of constitutive activation. Furthermore, except for NB4 cells, pretreatment with anti-FGFR1 antibody resulted in decreased Akt phosphorylation (Figure 4).

Figure 4
figure4

Fibroblast growth factor-2 (FGF-2) induces phosphorylation of Akt in fibroblast growth factor receptor 1 (FGFR1)-positive acute myeloid leukemias (AMLs), which can be blocked by anti-FGFR1 neutralizing antibody. Western blot for total and phosphorylated Akt (Akt, p-Akt) in FGFR1-positive AMLs is shown. Fibroblast growth factor-2 induces phosphorylation of Akt. Western blot for total and phosphorylated Akt (Akt, p-Akt) shows induction of phospho-Akt in HL-60, U937 and R81 cells after exposure to FGF-2 and reduced Akt phosphorylation by coincubation with anti-FGFR1 neutralizing antibody in the same cell lines.

FGF-2 stimulation protects subsets of FGFR1-positive AML cells from chemotherapy-induced apoptosis

Fibroblast growth factor receptor 1-positive AML cells were exposed to previously titrated concentrations (approximately 50% cell kill) of the chemotherapeutic agent cytarabine with or without addition of FGF-2 under serum-free conditions. After 24 h, viable cells were counted with Trypan blue exclusion. Fibroblast growth factor-2 exposure resulted in a statistically significant increase in viable cells as compared to chemotherapy alone in HL-60, U937 and R81 cells (Figure 5a), but not NB4 cells. Remarkably, sensitivity to chemotherapy was restored by blocking FGF-2-mediated signaling using anti-FGFR1 neutralizing antibody (P<0.05 for FGF-2-stimulated cells compared to medium alone; Figure 5a). In support of this observation, propidium iodide/annexin apoptosis assay revealed a significant decrease in early/late apoptotic and dead cells when chemotherapy-treated AML were coincubated with FGF-2. Results are shown for cytarabine (Figure 5b) in R81 AML cells. Similar results for survival and apoptosis were obtained when cytarabine was substituted with etoposide or daunorubicin (data not shown). These data suggest that FGF-2 can increase resistance of leukemic cells to chemotherapeutic agents, which can be reversed by blocking FGFR1 activation.

Figure 5
figure5

Fibroblast growth factor-2 (FGF-2) signaling through fibroblast growth factor receptor 1 (FGFR1) protects FGFR1-positive acute myeloid leukemia (AML) cells from chemotherapy-induced cell death and blocking of FGFR1 restores chemosensitivity. (a) Fibroblast growth factor-2 stimulation resulted in a significant increase in viable cells exposed to cytarabine, demonstrating a chemoprotective effect of FGF-2 stimulation in HL-60, U937 and R81 cells, *P<0.05 (significant difference compared to medium alone or medium with chemotherapy alone, respectively). The FGF-2-mediated increase in survival is reversed by preincubation with anti-FGFR1 neutralizing antibody in the same cell lines, *P<0.05 (significant difference compared to chemotherapy alone or chemotherapy with FGF-2, respectively). (b) Blocking of FGFR1 in R81 AML cells exposed to FGF-2 increases chemotherapy-induced apoptosis. Results are shown as the percentage of viable cells (annexin V PI), early apoptotic cells (annexin V+ PI), late apoptotic cells (annexin V+ PI+) and dead cells (annexin V PI+). When cells were exposed to cytarabine, FGF-2 decreased the number of cells undergoing early and late apoptosis and cell death (fourth panel) compared to medium alone (first panel), and this effect is abrogated by addition of anti-FGFR1 neutralizing antibody (third panel). Addition of anti-FGFR1 neutralizing antibody alone (second panel) has no effect on apoptosis compared to medium with chemotherapy alone (first panel).

Anti-FGFR1 neutralizing antibody decreases leukemic cell growth and increases apoptosis in vivo (3-day treatment model)

To evaluate the potential of FGF-2 in initiating leukemic growth in vivo, the effect of anti-FGFR1 neutralizing antibody on the growth of R81 leukemic cells, xenotransplanted into NOD-SCID mice, was determined shortly after xenotransplantation. Treatment was begun at day 7 after inoculation of leukemic cells with daily injections of anti-human FGFR1 neutralizing antibody and the animals were killed after 3 days of treatment. As the anti-human FGFR1 neutralizing antibody may crossreact with mouse FGFR1, we evaluated the tumor sections for possible effects of the antibody on MVD. After staining for mouse endothelial cells by immunohistochemistry for the endothelial-specific marker, MECA-32 (Figure 6a), no significant difference in MVD was found between the anti-FGFR1 neutralizing antibody group and the human IgG control group (Figure 6b). However, a significantly higher numbers of apoptotic cells were found in the treatment vs the control group by TUNEL staining (mean 20.8 vs 10.9 per field, P<0.001; Figure 6a and b). In addition, the proliferative index was significantly lower in the anti-FGFR1 neutralizing antibody-treated group compared to the IgG control group (mean 0.38 vs 0.53, P<0.001; Figure 6a and b) as determined by Ki-67 staining.

Figure 6
figure6

Anti-fibroblast growth factor receptor 1 (FGFR1) neutralizing antibody increases apoptosis and inhibits proliferation in vivo in a murine R81 acute myeloid leukemia (AML) xenograft model (3-day treatment model). (a) Mouse endothelial cells (brown staining, left column), apoptotic cells (red staining, middle column) and proliferating cells (brown staining, right column) were analyzed to compare IgG control and anti-FGFR1 neutralizing antibody-treated groups; magnification × 400. (b) No significant difference in microvessel density was detected between the treatment and control groups, suggesting no effect of anti-FGFR1 neutralizing antibody on angiogenesis. The rate of apoptosis, however, was significantly higher in the anti-FGFR1 neutralizing antibody-treated group with 10.9 TUNEL+ cells per field (control group) vs 20.8 (treatment group), ***P<0.001. Proliferative index (right column) was also significantly different between the groups, with a mean of 0.53 (control group) vs 0.38 (treatment group), ***P<0.001.

Anti-FGFR1 neutralizing antibody blocks tumor growth in vivo (25-day treatment model)

To determine the long-term effect of blocking FGF-2/FGFR1 signaling, subcutaneous AML xenograft tumors were treated for 25 days with anti-FGFR1 neutralizing antibody, beginning at day 3 after inoculation of the leukemic cells. Measured at 11, 18 and 25 days of treatment (i.e. 14, 21 and 28 days afterinjection), the R81 AML tumors in NOD-SCID mice were significantly smaller in the anti-FGFR1 neutralizing antibody-treated group as compared to the IgG control group (n=6, P<0.05; Figure 7a). The effects of acute treatment of the tumors with anti-FGFR1 neutralizing antibody for 3 days were evaluated by quantifying MVD using immunohistochemistry for murine endothelial-specific marker MECA-32 (Figure 7c). No significant difference in MVD was found between the anti-FGFR1 neutralizing antibody- and human IgG-treated groups (Figure 7d). Similarly, equal amounts of tumor-associated fibroblast-like stromal cells were found in both groups after staining with Masson trichrome stain (Figure 7b). Although tumors isolated from mice after anti-FGFR1 neutralizing antibody treatment for 3 days showed increase in apoptosis, there was no difference in the number of apoptotic cells between the treatment and control groups in mice that were treated with anti-FGFR1 for 25 days, as detected by TUNEL staining (Figure 7c and d). The proliferative index quantified by Ki-67 staining, however, was significantly lower in the anti-FGFR1 neutralizing antibody-treated group as compared to the IgG control group (mean 0.50 vs 0.42, P<0.05; Figure 7c and d). These data suggest that inhibition of FGFR1 signaling targets leukemic cells, increasing their rate of apoptosis in the initial phase of tumor growth.

Figure 7
figure7

Anti-fibroblast growth factor receptor 1 (FGFR1) neutralizing antibody delays tumor growth in a murine R81 acute myeloid leukemia (AML) xenograft model by inhibiting proliferation of R81 AML cells in vivo (3-week treatment model). (a) Tumor volume was significantly smaller in the anti-FGFR1 neutralizing antibody-treated group as compared to human IgG control with n=6 in each group. *P<0.05 (significant difference compared to human IgG control group). (b) Masson trichrome stain for collagen/stroma visualization (blue staining) shows similar numbers of tumor-associated stromal cells/fibroblasts in human IgG control group (upper panels) and anti-FGFR1 neutralizing antibody-treated animals (lower panels); magnification × 400. (c) Mouse endothelial cells (brown staining, left column), apoptotic cells (red staining, middle column) and proliferating cells (brown staining, right column) were analyzed to compare IgG control and anti-FGFR1 neutralizing antibody-treated groups; magnification × 400. (d) No significant difference in microvessel density was detected between the treatment and control groups. In contrast to the 3-day treatment experiment, the relative rate of apoptosis was similar in the treatment and control groups. Proliferative index (right column), however, was again significantly different between the groups, with a mean of 0.50 (control group) vs 0.42 (treatment group), **P<0.01, indicating that FGFR-1 neutralizing antibody inhibits proliferation of R81 AML cells in vivo even after long treatment duration.

Discussion

Tyrosine kinase receptors, such as vascular endothelial growth factor receptors (VEGFRs) and FGFRs, have long been implicated in driving the growth of solid tumors through angiogenesis.13 More recently, paracrine (leukemia–stroma) and autocrine (leukemia–leukemia) VEGF/VEGFR signaling loops have been identified in AML.14, 15, 16, 17, 18 However, the role of FGFs and specific FGFRs in the regulation of leukemic cell growth has not been studied, in part because reagents to block FGFR signaling have not been available. Because FGFR1 signaling is critical in hematopoietic stem cells,3, 4, 19 we hypothesized that FGF/FGFR1 signaling may also be active in leukemias. In this paper, we show that functional FGFR1β is expressed on subsets of AMLs, supporting their proliferation, survival, motility and resistance to chemotherapy. Furthermore, we show for the first time that specific inhibition of FGFR1β blocks progression of leukemias in in vitro and in vivo models. These data suggest that FGF-2/anti-FGFR1β signaling may confer chemoresistance and contribute to the growth of subsets of AMLs.

Studying the physiologic role of FGFR signaling in AML in vitro has been challenging. Although one paper demonstrated increased proliferation in response to FGF-2 in three out of three AML cell lines,6 such effect has not been observed in another report.9 Three prior studies have investigated the expression of FGFR mRNA in AML cell lines.6, 7, 8 Although the results were somewhat conflicting in regard to FGFR expression patterns for each cell line, the data suggest that most AML cell lines express at least one of the four FGFRs. Because of the typically low receptor density of tyrosine kinase receptors, detection on a protein level by standard methods such as flow cytometry20 or Western blot without immunoprecipitation has been a daunting task. In addition, FGFRs may be cleaved21 and/or internalized22 as part of the signaling mechanism, further complicating receptor detection. Therefore, none of the prior studies have determined whether FGF signaling through a specific FGF receptor plays a role in regulating AML apoptosis, migration and resistance to chemotherapy.

To better define the role of FGF/FGFR signaling in AML pathophysiology, we focused on the specific effects of FGFR1 stimulation and blockade using a novel anti-FGFR1 antibody. First, we demonstrated an effect of FGF-2 on proliferation and migration of FGFR1-positive leukemic cells in a serum-free system. Next, we showed that signaling is mediated through FGFR1, as the responses to FGF-2 were abrogated by anti-FGFR1 neutralizing antibody. We confirmed expression of FGFR1β mRNA and protein, and showed that FGF-2 induces FGFR1 phosphorylation and activation of Akt, which is blocked by anti-FGFR1 neutralizing antibody. Activation of the serine/threonine kinase Akt, a downstream effector of phosphatidylinositol 3-kinase (PI3K), is known to protect AML blasts from undergoing apoptosis.23 Furthermore, inhibition of Akt signaling has been shown to reduce chemoresistance in AML cells.24, 25, 26 The particularly low constitutive activation of Akt in R81 cells and marked increase in phosphorylation upon stimulation may explain the observed strong proliferative response to FGF-2 in this cell line. Taken together, our results clearly show that FGF-2 signaling through FGFR1 can be physiologically relevant in subsets of AML. Furthermore, FGF-2/FGFR1 signaling confers resistance to chemotherapy by decreasing chemotherapy-induced apoptosis, but chemosensitivity can be restored by blocking FGFR1 activation. Our data suggest that this mechanism may be particularly important in subsets of AML with low constitutive PI3K/Akt activation. FGFR1β, the isoform that is predominantly expressed by the AMLs tested, represents a lower molecular weight isoform lacking one of the three extracellular immunoglobulin domains present in the FGFR1α isoform. Notably, the FGFR1β isoform has been described to exhibit a 10-fold greater ligand affinity compared to FGFR1α,27, 28 and a switch from FGFR1α to FGFR1β isoform expression correlates with transformation to a malignant phenotype and invasiveness in astrocytomas.29, 30

To evaluate the effectiveness of FGFR1 blockade in AML in vivo, we tested the anti-FGFR1 neutralizing antibody in NOD-SCID mice using a leukemia xenograft model. Consistent with the in vitro data, intraperitoneal short-term administration of the antibody increased apoptosis and reduced proliferation of R81 AML cells in vivo after 3 days at 100 μg/day. Long-term administration of anti-FGFR1 antibody for 25 days at a reduced dose of 25 μg every other day significantly delayed R81 AML growth, and immunohistochemical analysis revealed a sustained effect on the leukemic cell proliferation index. As FGFR1 receptors are also present on fibroblasts and endothelial cells,5 and the anti-FGFR1 neutralizing antibody is crossreactive with murine FGFR1, we investigated whether the observed reduction in tumor growth may have been mediated through inhibition of tumor-induced angiogenesis or stromagenesis. However, we observed no significant difference in the number of tumor-associated fibroblast-like stromal cells or MVD. In line with our in vitro data, these findings suggest that the anti-FGFR1 neutralizing antibody targets R81 AML tumor growth in vivo directly through action on the leukemic cells, rather than indirectly through a reduction of angiogenesis or recruitment of stromal cells. The proliferative index was significantly lower in both the 3-day and long-term (25-day) experiments, and explains the observed difference in tumor size observed in the 25-day treatment experiment. In contrast, the number of apoptotic cells was significantly different between the treatment and control groups in the higher dose/3-day administration experiment only, with an increase in apoptosis in the treatment group. These findings suggest a growth-inhibitory rather than a direct proapoptotic effect of anti-FGFR1 neutralizing antibody treatment during long-term administration at a lower dose. This effect may be dose dependent or could reflect a selection of leukemic subclones more resistant to apoptosis, possibly through constitutive PI3K/Akt activation.

In summary, our study supports the concept that FGF-2 signaling through FGFR1β can regulate growth, apoptosis, migration and resistance to chemotherapy in AML. In addition, we provide proof-of-concept that FGFR1 receptor targeting in AML may delay leukemic growth in vitro and in vivo. This sets the stage for further studies to determine the significance of FGF-2/FGFR1 receptor signaling, as well as signaling involving other FGFs, FGFRs and FGFR isotypes in human AMLs. We propose that targeting of FGFRs can interfere with proliferation and migration in subsets of AMLs and may enhance the efficacy of chemotherapy.

References

  1. 1

    Bradstock KF, Gottlieb DJ . Interaction of acute leukemia cells with the bone marrow microenvironment: implications for control of minimal residual disease. Leuk Lymphoma 1995; 18: 1–16.

  2. 2

    Allouche M, Bikfalvi A . The role of fibroblast growth factor-2 (FGF-2) in hematopoiesis. Prog Growth Factor Res 1995; 6: 35–48.

  3. 3

    Moroni E, Dell’Era P, Rusnati M, Presta M . Fibroblast growth factors and their receptors in hematopoiesis and hematological tumors. J Hematother Stem Cell Res 2002; 11: 19–32.

  4. 4

    Magnusson PU, Ronca R, Dell’Era P, Carlstedt P, Jakobsson L, Partanen J et al. Fibroblast growth factor receptor-1 expression is required for hematopoietic but not endothelial cell development. Arterioscler Thromb Vasc Biol 2005; 25: 944–949.

  5. 5

    Powers CJ, McLeskey SW, Wellstein A . Fibroblast growth factors, their receptors and signaling. Endocr Relat Cancer 2000; 7: 165–197.

  6. 6

    Bieker R, Padro T, Kramer J, Steins M, Kessler T, Retzlaff S et al. Overexpression of basic fibroblast growth factor and autocrine stimulation in acute myeloid leukemia. Cancer Res 2003; 63: 7241–7246.

  7. 7

    Allouche M, Bayard F, Clamens S, Fillola G, Sie P, Amalric F . Expression of basic fibroblast growth factor (bFGF) and FGF-receptors in human leukemic cells. Leukemia 1995; 9: 77–86.

  8. 8

    Armstrong E, Vainikka S, Partanen J, Korhonen J, Alitalo R . Expression of fibroblast growth factor receptors in human leukemia cells. Cancer Res 1992; 52: 2004–2007.

  9. 9

    Krejci P, Faitova J, Laurell H, Hampl A, Dvorak P . FGF-2 expression and its action in human leukemia and lymphoma cell lines. Leukemia 2003; 17: 818–820.

  10. 10

    Zhang F, Cheng J, Hackett NR, Lam G, Shido K, Pergolizzi R et al. Adenovirus E4 gene promotes selective endothelial cell survival and angiogenesis via activation of the vascular endothelial-cadherin/Akt signaling pathway. J Biol Chem 2004; 279: 11760–11766.

  11. 11

    Hallmann R, Mayer DN, Berg EL, Broermann R, Butcher EC . Novel mouse endothelial cell surface marker is suppressed during differentiation of the blood brain barrier. Dev Dyn 1995; 202: 325–332.

  12. 12

    Ruan GR, Liu YR, Chen SS, Fu JY, Chang Y, Qin YZ et al. Effect of antisense VEGF cDNA transfection on the growth of chronic myeloid leukemia K562 cells in vitro and in nude mice. Leuk Res 2004; 28: 763–769.

  13. 13

    Folkman J . The role of angiogenesis in tumor growth. Semin Cancer Biol 1992; 3: 65–71.

  14. 14

    Dias S, Hattori K, Heissig B, Zhu Z, Wu Y, Witte L et al. Inhibition of both paracrine and autocrine VEGF/VEGFR-2 signaling pathways is essential to induce long-term remission of xenotransplanted human leukemias. Proc Natl Acad Sci USA 2001; 98: 10857–10862.

  15. 15

    Zhu Z, Hattori K, Zhang H, Jimenez X, Ludwig DL, Dias S et al. Inhibition of human leukemia in an animal model with human antibodies directed against vascular endothelial growth factor receptor 2. Correlation between antibody affinity and biological activity. Leukemia 2003; 17: 604–611.

  16. 16

    Santos SC, Dias S . Internal and external autocrine VEGF/KDR loops regulate survival of subsets of acute leukemia through distinct signaling pathways. Blood 2004; 103: 3883–3889.

  17. 17

    List AF, Glinsmann-Gibson B, Stadheim C, Meuillet EJ, Bellamy W, Powis G . Vascular endothelial growth factor receptor-1 and receptor-2 initiate a phosphatidylinositide 3-kinase-dependent clonogenic response in acute myeloid leukemia cells. Exp Hematol 2004; 32: 526–535.

  18. 18

    Zhang H, Li Y, Li H, Bassi R, Jimenez X, Witte L et al. Inhibition of both the autocrine and the paracrine growth of human leukemia with a fully human antibody directed against vascular endothelial growth factor receptor 2. Leuk Lymphoma 2004; 45: 1887–1897.

  19. 19

    Kashiwakura I, Takahashi TA . Fibroblast growth factor and ex vivo expansion of hematopoietic progenitor cells. Leuk Lymphoma 2005; 46: 329–333.

  20. 20

    Aziz KA, Till KJ, Chen H, Slupsky JR, Campbell F, Cawley JC et al. The role of autocrine FGF-2 in the distinctive bone marrow fibrosis of hairy-cell leukemia (HCL). Blood 2003; 102: 1051–1056.

  21. 21

    Hanneken A . Structural characterization of the circulating soluble FGF receptors reveals multiple isoforms generated by secretion and ectodomain shedding. FEBS Lett 2001; 489: 176–181.

  22. 22

    Reilly JF, Mizukoshi E, Maher PA . Ligand dependent and independent internalization and nuclear translocation of fibroblast growth factor (FGF) receptor 1. DNA Cell Biol 2004; 23: 538–548.

  23. 23

    Xu Q, Simpson SE, Scialla TJ, Bagg A, Carroll M . Survival of acute myeloid leukemia cells requires PI3 kinase activation. Blood 2003; 102: 972–980.

  24. 24

    Martelli AM, Tabellini G, Bortul R, Tazzari PL, Cappellini A, Billi AM et al. Involvement of the phosphoinositide 3-kinase/Akt signaling pathway in the resistance to therapeutic treatments of human leukemias. Histol Histopathol 2005; 20: 239–252.

  25. 25

    Martelli AM, Tazzari PL, Tabellini G, Bortul R, Billi AM, Manzoli L et al. A new selective AKT pharmacological inhibitor reduces resistance to chemotherapeutic drugs, TRAIL, all-trans-retinoic acid, and ionizing radiation of human leukemia cells. Leukemia 2003; 17: 1794–1805.

  26. 26

    Zhao S, Konopleva M, Cabreira-Hansen M, Xie Z, Hu W, Milella M et al. Inhibition of phosphatidylinositol 3-kinase dephosphorylates BAD and promotes apoptosis in myeloid leukemias. Leukemia 2004; 18: 267–275.

  27. 27

    Shi E, Kan M, Xu J, Wang F, Hou J, McKeehan WL . Control of fibroblast growth factor receptor kinase signal transduction by heterodimerization of combinatorial splice variants. Mol Cell Biol 1993; 13: 3907–3918.

  28. 28

    Wang F, Kan M, Yan G, Xu J, McKeehan WL . Alternately spliced NH2-terminal immunoglobulin-like Loop I in the ectodomain of the fibroblast growth factor (FGF) receptor 1 lowers affinity for both heparin and FGF-1. J Biol Chem 1995; 270: 10231–10235.

  29. 29

    Morrison RS, Yamaguchi F, Saya H, Bruner JM, Yahanda AM, Donehower LA et al. Basic fibroblast growth factor and fibroblast growth factor receptor I are implicated in the growth of human astrocytomas. J Neurooncol 1994; 18: 207–216.

  30. 30

    Yamaguchi F, Saya H, Bruner JM, Morrison RS . Differential expression of two fibroblast growth factor-receptor genes is associated with malignant progression in human astrocytomas. Proc Natl Acad Sci USA 1994; 91: 484–488.

Download references

Acknowledgements

MAK is supported in part by grants from the Children's Cancer Research Fund and the Hope Street Kids Cancer Foundation.

Author information

Correspondence to S Rafii.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • acute myeloid leukemia
  • fibroblast growth factor
  • fibroblast growth factor receptor
  • cell signaling
  • chemotherapy
  • animal studies

Further reading