Ectopic expression of fibroblast growth factor receptor 3 promotes myeloma cell proliferation and prevents apoptosis

Two pcDNA3 plasmids containing the full-length human complementary DNA (cDNA) of wild-type or TD II mutant form of FGFR3 were obtained from Daniel. J. Donoghue (San Diego, CA). FGFR3 was released by restriction cutting with HindIII and SpeI, also removing the 3′ untranslated region. The recessed 3′ termini ends of the cDNA fragments were filled, and the fragments ligated into an HpaI/SnaBI-digested MINV retrovirus28 containing the neomycin-resistance gene as the selectable marker. The resulting vectors were termed MFINV-WT (wild-type FGFR3) and MFINV-TD (TD K650E mutant form of FGFR3). The parental MINV vector (neo^r only) was used as a negative vector control.

Twenty micrograms of the constructed retroviral vectors and the parental MINV (neo^r only) vector were electroporated into 1.5 × 10⁶ PA317 amphotropic packaging cells.29 After 2 days of culture, the supernatant was collected, passed through a 0.45-μm filter, and overlaid on subconfluent monolayers of GP+E ecotropic packaging cells30 in the presence of 8 μg/mL of polybrene (Sigma, St. Louis, MO). On day 3 and 4 of culture, this process was repeated. Transduced GP+E cells were selected in fresh medium containing 1.2 mg/mL of G418 (Gibco, Grand Island, NY) over a period of 14 days. Vector supernatant was collected from GP+E producer cells 24 hours after the medium was changed to Iscove's Modified Dulbecco's Medium (IMDM) plus 5% fetal calf serum (FCS) (Gibco).

Supernatant from the G418 selected GP+E producer lines was added to B9 cells along with 2% IL-6 conditioned media from the SP2/mIL-6 cell line31 and 8 μg/mL of polybrene (Sigma). After 2 days, fresh retroviral supernatant was added. After an additional 72 hours of incubation, B9 cells were selected in 2% IL-6 conditioned medium and 1.2 mg/mL G418 for 2 weeks, generating the cell lines B9-MINV (empty vector), B9-WT (wild-type FGFR3), and B9-TD (TD II mutant form of FGFR3). After selecting for stably transduced cells, a limiting cell dilution was performed to generate single-cell clones of B9-WT and B9-TD. Subsequent Western blot analysis used individual high- or low-expressing clones of B9-WT and B9-TD.

Cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. B9 cells were grown in IMDM supplemented with 5% FCS, 2% IL-6 conditioned medium, and penicillin-streptomycin (Gibco). KMS11, a human myeloma cell line containing an FGFR3 translocation, was kindly provided by Masayoshi Namba (Okayama, Japan). KMS11 was grown in RPMI Medium 1640 supplemented with 10% FCS and penicillin-streptomycin. U266, a human myeloma cell line, was grown in IMDM with 10% FCS. GP+E and PA317 cells were grown in Dulbecco modified Eagle medium supplemented with 10% FCS and penicillin-streptomycin.

To examine cell viability and proliferation, B9, B9-MINV, B9-WT, and B9-TD were washed 4 times in IMDM without FCS and then plated in triplicate in a 96-well plate at a density of 1 × 10⁴ cells/well in varying concentrations of IL-6 and cultured for 48-96 hours. Total cell number and percent viability were determined every 24 hours by trypan blue staining and enumeration with a hemocytometer. For proliferation assays, 0.5 μCi of ³H-thymidine (Amersham, Arlington Heights, IL) was added to each well in the last 18 hours of the assay. Cells were harvested onto glass filter paper (Gelman Science, Ann Arbor, MI), and³H-thymidine counts determined by liquid scintillation.

B9-WT cells were plated at a concentration of 2 × 10⁵ cells/mL in a 96-well plate in the presence of increasing concentrations of acidic fibroblast growth factor (aFGF) (R&D Labs, Minneapolis, MN) and heparin (Sigma). Cells were subsequently incubated for 2 or 3 days and then analyzed with the use of the chromogenic dye 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Boehringer Mannheim, Mannheim, Germany). Absorbance at 570 nm-650 nm was read on an enzyme-linked immunosorbent assay plate reader (Molecular Devices, Sunnyvale, CA). After optimal concentrations of aFGF (40 ng/mL) and heparin (30 μg/mL) for the support of cell proliferation were determined, B9, B9-MINV, B9-WT, and B9-TD were washed four times in IMDM lacking FCS and plated in triplicate at a concentration of 2 × 10⁵ cells/mL in 96-well plates. Each cell line was analyzed under 4 different conditions: 40 ng/mL aFGF and 30 μg/mL of heparin, 40 ng/mL aFGF and 30 μg/mL of heparin plus 2% IL-6, 2% IL-6 without ligand, and 0% IL-6 without ligand. Cells were incubated for 48-72 hours and then analyzed by MTT.

Cell lines were washed and then plated in IMDM plus 5% FCS at a density of 0.5 × 10⁶ in 0% IL-6 or 0.1 × 10⁶ in the presence of IL-6. Cells were cultured for 48 or 72 hours. After harvesting the cells, 500 μL of a hypotonic fluorochrome solution (0.1% sodium citrate, 0.1% Triton X-100, 50 μg/mL of propidium iodide) was added to each cell pellet. The samples were incubated overnight in the dark at 4°C. Flow cytometry was performed on a FACScan (Becton Dickinson, Meylan, France) the following day, using Cell Quest software to acquire data and ModFit LT software for analysis. Tunel assays (Promega, Madison, WI) were performed according to the manufacturer's protocol. In brief, 4 × 10⁶ cells were fixed with 1% paraformaldehyde for 20 minutes on ice. Cells were incubated at 37°C for 1 hour with terminal deoxynucleotidyl transferase (TdT) enzyme and nucleotide mix containing fluorescein-12-dUTP to end label the 3′OH end of the fragmented DNA. The reaction was ceased by addition of 20 mM ethylenediaminetetraacetic acid, and cells were washed twice in 0.1% Triton X-100 and stained with propidium iodide for 30 minutes. Analysis was performed by flow cytometry, and apoptotic cells were fluorescein-12-dUTP positive.

Western blot analysis was carried out according to previously published procedures.32 Membranes were probed with antibodies to the C-terminus end of FGFR3, to murine bcl-x_S/L, to murine bcl-2, to murine bax, anti-phosphotyrosine (all Santa Cruz Biotechnology, Santa Cruz, CA), anti-phospho-STAT1 (Upstate Biotechnology, Lake Placid, NY), anti-STAT3 and anti-STAT1 (Transduction Laboratories, Lexington, KY), or with polyclonal rabbit anti-phospho-STAT1/STAT5, anti-phospho-STAT5, anti-phospho-STAT3 (all 3 kindly provided by David A. Frank, Dana-Farber Cancer Institute, Boston, MA), and anti-STAT5 (kindly provided by Dr J. N. Ihle, Memphis, IN). Goat antirabbit IgG horseradish peroxidase (PharMingen, Mississauga, ON) or sheep antimouse IgG horseradish peroxidase (Amersham, Arlington Heights, IL) were used as secondary antibodies, and blots were developed by enhanced chemiluminescence (Amersham) according to manufacturer's instructions.

EMSA was carried out as previously described.33 After generation of nuclear lysates from cells stimulated with and without IL-6, the gel shift assays were performed with the use of double-stranded ³²P-labeled oligonucleotide derived from the IRF-1 promoter (CTGATTTCCCCGAAATGAC).34 Five micrograms of nuclear lysates was incubated with 0.25 ng of³²P-labeled oligonucleotide in 20 μL of binding buffer,13 mmol/L HEPES (pH 7.9; 65 mmol/L NaCl, 1 mmol/L DTT, 15 mmol/L ethylenediaminetetraacetic acid, 8% glycerol, 50 μg of poly (dI-dC)/mL) for 15 minutes at 4°C. For competition experiments, 50-fold molar excess of either unlabeled IRF-1 or an oligonucleotide derived from T14 promoter (TTCTTGAGGTAATGAAAGCC)35 was added to the binding reaction. For supershifting experiments, a peptide-specific STAT3 (Zymed, San Francisco, CA) was added at the completion of the binding reaction and incubated for an additional 30 minutes at 4°C.

Immunoprecipitation and in vitro kinase assay were carried out as previously described.32 Twenty-five microliters/reaction of 20% (v/v) Protein-A Sepharose (Pharmacia, Uppsala, Sweden) was incubated with 12 μL of anti-FGFR3, anti-ERK1 (recognizes ERK1 and ERK2), anti-p38, or anti-JNK1 (SAPK) (all from Santa Cruz Biotechnology) overnight at 4°C; 1250 μg of total cell lysate was incubated with the bead complex for 2 hours at 4°C. Immunoprecipitates were split into equal aliquots. One aliquot was washed 4 times with kinase buffer and subjected to in vitro kinase assay. The other aliquot was subjected to standard Western blotting procedure, probing the membranes with anti-FGFR3, ERK1, p38, or JNK1 (SAPK).

Western blots confirmed varying levels of expression of FGFR3 in the transduced B9-WT (Figure 1A) and B9-TD (Figure 1B) lines and subclones. An in vitro kinase assay demonstrated that both wild-type and FGFR3-TD lines were capable of autophosphorylation (Figure2). Phosphotyrosine Western blots further demonstrated that FGFR3 was phosphorylated only in the absence of ligand within the B9-TD line, thus indicating that the mutant receptor possessed constitutive tyrosine kinase activity. On addition of ligand, aFGF, and heparin, phosphorylation of wild-type FGFR3 was induced in B9-WT (not shown).

We first asked if overexpression of FGFR3 would influence IL-6– induced proliferation of B9 cells. Parental B9, B9-MINV, and pooled B9-WT cells had a similar rate of proliferation at all concentrations of IL-6 as assessed by ³H-thymidine incorporation. However, the proliferative rate of pooled B9-TD cells expressing mutant FGFR3 was higher than that for the other 3 cell lines at all concentrations of IL-6 (Figure 3). Two days after withdrawal of IL-6, B9, B9-MINV, B9-WT, and B9-TD, cells were 30%, 16%, 44%, and 75% viable, respectively. After 4 days of cytokine starvation, there were 6 times more viable cells in B9-TD culture than for the other lines, and, unlike the parental B9 cell line, B9-TD cells could be maintained indefinitely in the absence of IL-6. Thus, activated FGFR3 enhanced cell proliferation in response to IL-6 and could substitute for the absolute dependence of B9 cells on IL-6. Because many myelomas do not immediately acquire FGFR3 mutation, we next examined the response of cells expressing the wild-type or mutant protein to the addition of the aFGF ligand. On ligand stimulation of IL-6–starved cells, B9-WT proliferation was greater than that seen for parental B9 and B9-MINV cells, and the already increased proliferation of B9-TD cells was further enhanced (Figure4). Some B9-WT clones expressing the highest levels of FGFR3 (eg, clone #3 and clone #6) also exhibited IL-6 independence. These data together demonstrate an enhanced responsiveness of FGFR3-expressing cells to IL-6 stimulation and the acquisition of IL-6 independence proportional to the degree of activation of FGFR3.

The mechanism of enhanced survival of B9-TD cells in the absence of IL-6 and ligand was next examined. There was little difference in the cell-cycle status of B9, B9-WT, and B9-TD in the presence or absence of IL-6. It was noted, however, that B9-TD cells had fewer sub-G₀ apoptotic cells in the absence of IL-6 and ligand than did either controls or B9-WT. Indeed, with the use of a Tunel assay, the vast majority of parental B9 cells were apoptotic 5 days after withdrawal of IL-6, whereas only 53% of pooled B9-TD cells were undergoing apoptosis (Table 1). B9-MINV and pooled B9-WT had between 85% and 88% apoptotic cells in the absence of IL-6 and ligand. These data demonstrate that activated FGFR3 prevents apoptosis of IL-6–dependent cells in the absence of IL-6.

Because the STATs are a key mediator of cytokine signaling, Western blot analysis was performed with the use of antibodies that bind to tyrosine phosphorylated STAT1, STAT3, or STAT5. STAT5 was phosphorylated in all cell lines independent of the presence or absence of IL-6 and was not further activated by FGFR3. STAT1 and STAT3 were phosphorylated in all of the cell lines after IL-6 stimulation. However, STAT3 was also constitutively phosphorylated (in the absence of IL-6) in both B9-WT clones expressing high levels of FGFR3 and in B9-TD cells (Figure 5). To confirm that STAT3 was constitutively activated in cells expressing high levels of wild-type FGFR3 or mutant FGFR3, EMSAs were performed. In the absence of IL-6 stimulation, both B9-WT and B9-TD cells displayed a protein-DNA complex that was only present in parental B9 cells in the presence of IL-6 (Figure 6). This complex could be competed with an excess of unlabeled IRF-1 but was not affected by a molar excess of nonspecific oligonucleotide. It was demonstrated that the complex contained STAT3 because a peptide-specific STAT3 antibody could supershift the complex in B9-TD cells.

Because we have previously shown that the IL-6 survival signal in myeloma is mediated by up-regulation of bcl-x_L,8 we next examined the expression of this protein by Western blots and demonstrated that the baseline level of bcl-x_L is higher in B9-TD cells expressing constitutively activated FGFR3 than in B9 controls (Figure7A). With the withdrawal of IL-6 from B9 or B9-MINV control cells, there is a decrease in the amount of bcl-x_L protein over 48 hours. In contrast, the expression of bcl-x_L in B9-TD clones remained relatively constant after the withdrawal of IL-6 from the culture medium. Some B9-WT clones expressing high levels of FGFR3 also exhibit a marked increase in baseline expression of bcl-x_L (Figure 7B). Of interest, the addition of ligand actually decreases bcl-x_L expression in B9-WT clones, perhaps because of down-regulation of the receptor.

Western blots were also used to examine the expression levels of bcl-x_S, bax, and bcl-2. There was no change in the expression of bax because of the presence of FGFR3 either in the presence or absence of IL-6. Levels remained similar to those expressed by parental B9 cells. Expression of bcl-2 and bcl-x_S was not detected in B9 cells or in the FGFR3-expressing derivatives (data not shown).

Although overexpression of bcl-x_L may explain the enhanced survival of B9 cells overexpressing FGFR3, it does not explain the enhanced proliferative rate of the transduced cell lines in response to IL-6. Because cytokines not only activate the JAK-STAT signaling pathway but also function to phosphorylate members of the MAPK family, we next examined ERK-, p38-, and SAPK-induced activation. Although p38 was induced by IL-6, no change in SAPK, ERK1, ERK2, or p38 activity was noted in the FGFR3-expressing cell lines (not shown). Thus, the mechanism of enhanced IL-6 responsiveness in FGFR3-expressing cells remains unknown.

The t(4;14) translocation occurs in approximately 25% of MM cell lines and patient samples and results in the ectopic expression ofFGFR3 from the der1414 and IgH-MMSET hybrid messenger RNA transcripts from the der4.18 Furthermore, activating mutations of the translocated FGFR3 have been identified in some myeloma samples, indicating an important role for the FGFR3 signaling pathway in these cases.14 One such activating mutant of FGFR3 is known to cause TD II, a severe form of human dwarfism, when the mutation arises in the germline.21,22 Prior investigation of the role of FGFR3in dwarfism has demonstrated that such mutations result in constitutive activation of the receptor independent of ligand.23,24 This activation leads to phosphorylation and translocation of STAT1 to the nucleus followed by up-regulation of the cell cycle regulator p21, resulting in growth arrest in chondrocytes.25 In a related TD I mutation ofFGFR3, STAT1 is also activated, along with the MAPK pathway, with a resulting increase in apoptosis secondary to increased levels of bax and decreased bcl-2.26 The net effect of these mutations in human chondrocytes is, therefore, growth arrest and cell apoptosis, explaining the chondrocyte growth arrest viewed clinically. Clearly, such changes would not explain why activation of FGFR3 promotes the growth or development of MM cells.

Clues to the latter come from the observation that signaling through gp130 by LIF in cardiac myocytes up-regulates bcl-x_L via STAT1.27 Because the myeloma-promoting cytokine IL-6 also signals through gp130 to up-regulate bcl-x_L in myeloma cells8 and because FGFR3 phosphorylates STAT1,25,26 we hypothesized that FGFR3 could enable cells to bypass gp130 signaling by constitutively up-regulating STATs and thus bcl-x_L in myeloma cells. To examine the consequences of FGFR3 overexpression, IL-6–dependent murine B9 myeloma cells were engineered to express either wild-type human FGFR3 or FGFR3-TD containing the K650E-activating mutation. We observed an increased level of responsiveness to IL-6 in the resultant cell lines that correlated with the expression level of FGFR3. Indeed, some cells expressing the constitutively active FGFR3-TD or high levels of wild-type FGFR3 exhibited complete independence from IL-6. Consistent with these results, others have shown that BaF3 cells overexpressing FGFR3 also exhibit sustained proliferation in the absence of ligand,23 suggesting that, in cytokine-dependent hematopoietic cells, the net effect of FGFR3 overexpression is to promote rather than to inhibit cell growth.

The concept that FGFR3 can produce seemingly opposing signals in chondrocytes and cytokine-dependent BaF3 cells (or as shown here in IL-6–dependent B9 cells) has precedence, as cytokine signaling via gp130 may also simultaneously induce contradictory intracellular signaling pathways.36 The central orchestrator of this gp130-signaling dichotomy appears to be STAT3 in that phosphorylation of STAT3 in response to IL-6 up-regulates cyclins and down-regulates p21, resulting in cell cycle transition.36 In contrast, when STAT3 is suppressed, signaling via gp130 can induce up-regulation of p21 and growth arrest, as viewed in human dwarfism.25 

We initially hypothesized that the effect of FGFR3 in conferring IL-6 independence occurred via enhanced phosphorylation of STAT1 as previously observed in chondrocytes; however, we found that expression of FGFR3 in B9 cells did not alter the phosphorylation of STAT1 but did constitutively phosphorylate STAT3. Because STAT3 is implicated in gp130 signaling6 and because gp130 signaling also up-regulates bcl-x_L,8,37 38 we next examined bcl-x_L expression. Indeed, B9-TD clones and clones expressing high levels of wild-type FGFR3 also express a relatively constant amount of bcl-x_L protein regardless of whether IL-6 is present or absent, and this baseline level is higher than that expressed by parental B9 cells.

It has previously been reported that the activating TD II mutation of FGFR3 results in a level of kinase activity of the receptor higher than can be achieved by maximally stimulating wild-type FGFR3 with ligand.24 From this observation, we theorize that the wild-type receptor would activate the same pathways as FGFR3-TD, but the magnitude of activation is expected to be less. This theory would appear to be true with our cell lines because there is a greater proliferative effect of expressing mutant FGFR3 than when the wild-type receptor is maximally stimulated with aFGF. Indeed, a dose effect of FGFR3 was observed for STAT3 phosphorylation and for IL-6 independence in high- and low-expressing FGFR3 clones.

Whereas phosphorylation of STAT3 and resulting overexpression of bcl-x_L likely explains the reduction in apoptosis of FGFR3-expressing cells after IL-6 withdrawal, it does not account for the enhanced proliferative response to IL-6 that was observed. In this regard, activated FGFR3-expressing cells exhibit an enhanced proliferative response to IL-6 at all concentrations of cytokine in comparison with controls, and cells expressing the wild-type protein demonstrate an enhanced proliferative response when ligand is present. Cells expressing high levels of wild-type FGFR3 may also show growth in the complete absence of IL-6. Although ERKs are thought to be involved in the enhanced proliferative response of IL-6–stimulated myeloma cells6,7 and FGFR3 has been shown to be capable of phosphorylating ERK1 and ERK2 in other systems,26,39 we were unable to demonstrate enhanced ERK phosphorylation secondary to FGFR3 overexpression. Similarly, activity of the other MAP kinase family members SAPK9 and p3840 was not altered by FGFR3 overexpression. The mechanism of cytokine-enhanced proliferation induced by FGFR3, therefore, remains unexplained.

Around 25% of myelomas translocate the FGFR3 gene. This fact alone may confer a survival advantage to B cells, as cells overexpressing FGFR3 proliferate at a higher rate than do controls in the presence of ligand. It can be further hypothesized that selective pressures subsequently result in the acquisition of activating FGFR3 mutations. A report by Fracchiolla et al41suggests that FGFR3 mutations occur rarely in MM. However, the authors chose to analyze DNA rather than cDNA by SSCP, which is likely to increase the background and decrease the sensitivity of this assay. Furthermore, DNA analysis for detection of mutations would be more sensitive if tumor cells were selected or if only samples containing at least 30% tumor cells were analyzed. Because serial dilutions of DNA from a cell line known to have an FGFR3 mutation, such as KMS11, and of DNA from a cell line with nonmutated FGFR3 were not performed to determine the minimal concentration of tumor DNA required to detected the abnormal migrating band by SSCP, the true incidence of FGFR3 mutation remains unknown.

In conclusion, FGFR3-overexpressing myeloma cells proliferate in the absence of IL-6, exhibit an enhanced proliferative response in the presence of IL-6, and exhibit decreased apoptosis on IL-6 withdrawal. Cells overexpressing FGFR3 phosphorylate STAT3 and up-regulate bcl-x_L expression, thus inhibiting apoptosis following cytokine withdrawal. The mechanism of FGFR3-enhanced proliferation is not yet elucidated.

We speculate that an initial translocation of FGFR3 into the IgH chain switch region in a germinal center B cell likely provides the affected B cells with enhanced growth potential, with signaling occurring through both the IL-6 receptor and FGFR3. Further, activating mutations of FGFR3 would enable the cells to entirely escape the need for ligand stimulation, whether by IL-6 or aFGF, and would result in enhanced proliferation and a competitive advantage over other B cells. These findings provide a plausible, albeit partial, explanation for the myeloma-promoting properties of FGFR3 and provide a basis for the development of therapies targeting this pathway.

We are grateful to Dr Brent Zanke and Dr Ian Dubé for advice, as well as to Nathan Faliconi, Fawzy Saad, Darrin Cappe, Meenakshi Bali, and Christine Dodgson for technical assistance.