Fibroblast growth factor receptor 3 inhibition by short hairpin RNAs leads to apoptosis in multiple myeloma

Introduction

Multiple myeloma is an aggressive neoplastic disease characterized by the increased proliferation and extended life span of monoclonal plasma cells. The t(4;14)(p16.3;q32.3) translocation occurs in ∼10% to 20% of multiple myeloma, resulting in ectopic expression of functional fibroblast growth factor receptor 3 (FGFR3; ref. 1). FGFR3 is a membrane-spanning tyrosine kinase receptor that has a high affinity for fibroblast growth factors (2). FGFR3 contains three glycosylated extracellular immunoglobulin-like domains, a transmembrane domain, and a split intracellular tyrosine kinase domain. Under ligand stimulation, FGFR3 undergoes dimerization and tyrosine autophosphorylation, resulting in cell proliferation or differentiation. Depending on the cellular context, this effect is achieved through different signal transduction pathways (3, 4). FGFR3 is not expressed or has a very low level of expression in B-cell lineage, but is overexpressed as a consequence of the translocation (1, 5, 6). Interestingly, FGFR3 gene mutations associated with human skeletal dysplasia have also been identified in some multiple myeloma with t(4;14). Activating mutations lead to a transforming event, suggesting that overexpression of activating mutations of FGFR3 may play an oncogenic role in multiple myeloma (5, 7–12). Thus, FGFR3 could become a specific therapeutic target in patients suffering from myeloma characterized by the presence of 4;14 translocation.

The possibility of RNA interference is one of the most exciting new discoveries in functional genomics of the past decade. The potential for the inhibitory effect on the expression of specific genes for experimental and therapeutic purposes is currently under investigation (13). RNA interference is a posttranscriptional process triggered by the introduction of double-stranded RNA which leads to gene silencing in a sequence-specific manner. In mammalian systems, the RNA interference effect has been observed by expression of 21 to 23 base transcripts capable of forming duplexes, or via vector-based expression of short hairpin RNAs (shRNA). Although chemically synthesized siRNA is a functional molecule by itself, direct application of siRNA is accompanied by several disadvantages, including an immediate disappearance of the knockout effect due to the lack of siRNA amplification mechanisms in mammalian cells and the inconvenience and high expense associated with its use. On the contrary, shRNA-expressing vectors, which work as platforms to produce a large amount of shRNA for a relatively longer period, can potentially circumvent these problems.

In this study, we designed several shRNAs and transfected them into three multiple myeloma cell lines: KMS-11, OPM-2, and NCI-H929. These cell lines are characterized by the presence of t(4;14), FGFR3 overexpression, and mutation—Y373C and K650E in KMS-11 and OPM-2 cell lines, respectively. We compared gene expression profiles of shRNA-treated sample with those of vector control sample using microarray analysis. Understanding how global gene expression patterns change after inhibition of FGFR3 and refining the functional role of FGFR3 may lead to targeted shRNA-based therapeutics in appropriate patients with multiple myeloma.

Materials and Methods

Construction of Short Hairpin RNA Expression Plasmids

Using siRNA design software and avoiding the hot mutation areas of FGFR3, six shRNA sequences were designed to encode two complementary sequences of 21 nucleotides homologous to a segment of FGFR3 gene separated by a seven-nucleotide space. A Basic Local Alignment Search Tool (BLAST) search ensured the sequences were not homologous to any other genes. We synthesized the sequences (City of Hope, DNA Synthesis Core Facility) and dissolved the oligonucleotides at a final concentration of 25 μmol/L. For the formation of duplexes, the forward and reverse oligonucleotides were incubated at 80°C for 2 minutes in annealing buffer (100 mmol/L NaCl) and kept in a water bath until the temperature reached 35°C. Double-stranded DNA oligonucleotides encoding shRNA were ligated into the BbsI site of the psiRNA-hH1GFPzeoG2 vector, which has a green fluorescent protein marker and resistance to the antibiotic Zeocin (InvivoGen, San Diego, CA), and then transformed into GT116 E. coli competent cells (InvivoGen). We confirmed the presence of the shRNA insert on an agarose gel after enzyme digestion. Sequencing (ABI PRISM 377 DNA Sequencer) was used to verify the positive clone.

Cell Culture

Three human myeloma cell lines, all characterized by the presence of t(4;14) and FGFR3 expression, were used. KMS-11 cell line was kindly provided by Dr. P. Lief Bergsagel (Department of Hematology and Oncology, Mayo Clinic, MN). OPM-2 cell line was purchased from Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (German Collection of Microorganisms and Cell Cultures, Brunswick, Germany). NCI-H929 cell line was obtained from the American Tissue Culture Collection (Manassas, VA). Cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum, 1% (v/v) penicillin, and streptomycin (100 μg/mL), and maintained at 37°C in a 5% CO₂ atmosphere.

Transfection of Multiple Myeloma Cell Lines

DNA was purified with the Plasmid Maxi Kit (Qiagen, Valencia, CA). Prior to electroporation, cells were washed with serum-free RPMI 1640 and resuspended with hypoosmolar buffer (Eppendorf, Germany) at a concentration of 2.0 × 10⁶ viable cells/mL. The cell suspension was mixed with 30 μg/mL plasmid in a 4 mm gap electroporation cuvette (Eppendorf). The cells were electroporated using a multiporator (Eppendorf) at 700 V with a time constant of 80 μs. The electroporated cells were left in the cuvette for 5 to 10 minutes at room temperature and then transferred to 10 mm dishes containing warm (37°C) complete medium and placed at 37°C in 5% CO₂ atmosphere. Forty-eight hours after electroporation, 100 μg/mL Zeocin (Invitrogen, Carlsbad, CA) was added to select the transfectants.

Sequential Immunophenotype/Fluorescence In situ Hybridization Analysis

Approximately 1 × 10⁶ cells were processed for sequential immunophenotype/fluorescence in situ hybridization (FISH) analysis. Cytospin slides were made from each cell line according to standard immunocytochemistry procedures. A pre-B leukemia cell line (ALL-1) was used as a negative control. Immunocytochemistry was done using the DAKO EnVision System/horse radish peroxidase with 3-amino 9-ethylcarbazole (DAKO Corporation, Carpinteria, CA). A 1:50 optimal FGFR3 antibody dilution was established.

Slides from each cell line were scanned for the presence or absence of FGFR3 antigen expression (immunophenotype) and representative bright-field images were captured and photographed according to their slide location by the BioView Duet Imaging System (Bioview, Rehovet, Israel). Next, the slides were fixed in 3:1 methanol/acetic acid for 1 hour and allowed to air dry. FISH was done according to the BioBlue protocol of the manufacturer (Bioview). The LSI IGH/FGFR3 dual color, dual fusion probe (Vysis, Inc., Downer's Grove, IL) was used to detect t(4;14)(p16.3;q32.3). After hybridization, the slides were rescanned on the imaging system for the presence or absence of the t(4;14) fluorescent signal pattern in the representative cells photographed for FGFR3 antigen expression for each cell line. The 900 kb 4p16/FGFR3 probe was labeled with SpectrumOrange (red signal) whereas the 1.5 Mb 14q32/IGH DNA probe was labeled with SpectrumGreen. In a normal cell, two red and two green signals reflect two intact copies of each gene. In a cell containing t(4;14), one red (FGFR3), one green (IGH), and two fusion signals [one fusion on each derivative chromosome, der(4) and der(14)] will be observed.

Western Blot Analysis

Cells were harvested, washed with ice-cold PBS, and lysed in radioimmunoprecipitation assay buffer (1× PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS) with freshly added inhibitors (100 μg/mL phenylmethylsulfonyl fluoride, 1 mmol/L sodium orthovanadate, and 30 μL/mL aprotinin). The samples were separated on 4% to 12% Tris-glycine gel (Invitrogen) and then transferred onto a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech, Piscataway, NJ). The membranes were immunoblotted with antibodies against FGFR3, B-cell chronic lymphocytic leukemia/lymphoma 2 (BCL2), and myeloid cell leukemia sequence 1 (MCL1; Santa Cruz Biotechnology, Santa Cruz,CA). The immunoblots were detected by the Western-Light system (Applied Biosystems, Foster City, CA). Values of the relative amounts of protein were determined by the densities of the bands by ImageQuant version 5.2.

Apoptosis Analysis Using Flow Cytometry

Apoptosis was assayed with the Annexin V apoptosis detection kit (BD Biosciences, San Jose, CA) following the instructions of the manufacturer. Cells were harvested and washed twice with cold PBS, resuspended in binding buffer at a concentration of 1 × 10⁶ cells/mL, and then 5 μL of Annexin V and 5 μL of propidium iodide were added to 100 μL cells. After incubation for 15 minutes at room temperature in the dark, 400 μL of binding buffer were added to the cells. Stained cells were analyzed by flow cytometry (City of Hope, Core facility) within 1 hour.

Reverse Transcriptase-PCR and Quantitative Real-time PCR

Total RNA was extracted from the cells using RNeasy Mini kits (Qiagen) according to the instructions of the manufacturer. Total RNA was reverse transcribed and amplified using Superscript one-step reverse transcriptase-PCR (RT-PCR) system with a platinum Taq polymerase (Invitrogen, Life Technologies). The FGFR3-specific primers (forward primer 5′-TGCTGAATGCCTCCCACG-3′, reverse primer 5′-CGTCTTCGTCATCTCCCGAG-3′) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH)–specific primers (forward primer 5′-CCACATCGCTCAGACACCAT-3′, reverse primer 5′-CCAGGCGCCCAATACG-3′) were employed for amplification of FGFR3 and GAPDH. To avoid the background of products amplified from genomic DNAs, primers for detecting expressions of FGFR3 and GAPDH were designed to exist on two different exons. FGFR3 and GAPDH were amplified as follows: 1 cycle of 50°C for 30 minutes, 94°C for 2 minutes; 25 cycles of denaturing at 94°C for 30 seconds, annealing at 53°C for 45 seconds, and extending at 72°C for 1 minute, followed by a final elongation step at 72°C for 10 minutes.

To confirm and validate the results of mRNA expression, quantitative real-time PCR was employed to measure FGFR3 expression. The forward and reverse primers were the same as described as above. The Pre-Developed TaqMan Assay Reagents control kit (Perkin-Elmer Applied Biosystems, Boston, MA), which included forward and reverse primers and a VIK-probe of GAPDH, was used as an internal control. As a positive control, a pcDNA3.1-FGFR3 plasmid and a GAPDH-pT7T3D-PAC plasmid, which contain the full-length FGFR3 and GAPDH structure genes, respectively, were used to generate the standard curve. The PCR conditions were as follows: 1 cycle of 50°C for 2 minutes, 95°C for 10 minutes; 40 cycles of 95°C for 15 seconds, 60°C for 60 seconds. Each data point was done in quadruplicate. Quantification of gene expression was done using the ABI PRISM 7700 Sequence Detection System (Perkin-Elmer Applied Biosystems). The mRNA amounts of FGFR3 and GAPDH were calculated according to the standard curve, and the ratio of FGFR3 to GAPDH was calculated to determine the normalized amount of FGFR3 mRNA.

Microarray Analysis

Human Genome U133 Plus 2.0 array (Affymetrix, Santa Clara, CA) was used for mRNA expression profiling. This array comprised more than 54,000 probe sets and 1,300,000 distinct oligonucleotide features. It can be used to analyze expression levels of more than 47,000 transcripts and variants, in addition to 38,500 human genes, at the same time.

Total RNA was extracted from the cells using RNeasy Mini kits (Qiagen). An agarose gel was run and absorbance was then checked at 260 and 280 nm to determine the concentration and purity of the RNA, ensuring that the highest quality RNA was hybridized to the gene expression arrays. Double-stranded cDNA was synthesized from total RNA. An in vitro transcription reaction was then done to produce biotin-labeled cRNA from the cDNA. Labeled cRNA was hybridized in oven to Human Genome U133 Plus 2.0 array. Immediately following hybridization, the probe array underwent an automated washing and staining protocol on the fluidics station. Once the probe array was hybridized, washed, and stained, it was scanned. Digitized image data were processed using Affymetrix GeneChip Analysis Suite software version 5.0. Comparison analysis was used to compare expression profiles from the vector-only control (designated as the baseline) and the shRNA-transfected sample (designated as the experiment). To be precise, the Affymetrix microarrays contained 11 to 20 pairs of oligonucleotide probes for each target RNA. The analysis compared the different values of each probe pair in the baseline array to its matching probe pair on the experiment array. Signal values, which represent abundance of the transcript, were calculated using the one-step Tukey's biweight estimate. Wilcoxon's signed rank test was used to generate the detection P value and change P value. All analyses were done by the University of California, Irvine DNA microarray core facility.

Results

Short Hairpin RNA Inhibition of FGFR3 Protein Expression

ShRNA sequences were cloned into a corresponding psiRNA-hH1GFPzeoG2 vector. Using electroporation, the recombinants were transfected into human multiple myeloma KMS-11, OPM-2, and NCI-H929 cell lines which possess the t(4;14) translocation with FGFR3 overexpression. Seven days after Zeocin selection, cells were harvested and the inhibitory effect on FGFR3 protein expression was determined from cell lysates via Western blot. Three shRNAs that showed the most reduction in FGFR3 expression were selected. The sequences and shRNA structure are shown in Fig. 1A. To enhance the inhibition efficiency, we then tested whether cotransfection of cells with a combination of these three constructed plasmids, which will produce three shRNAs (called FRS) targeting three different sites of mRNA, could increase the inhibitory effects. Plasmid DNA (30 μg/mL, 10 μg each) in electroporation buffer was cotransfected into the same three cell lines. For comparison, vector control cells were set at 100% FGFR3 expression in their respective cell lines. The combination of three plasmids was more effective at inhibiting FGFR3 expression in all three cell lines in comparison with that of any single shRNA (data not shown). In KMS-11 cells, FRS reduced FGFR3 expression to 21.21%. In OPM-2 cells, FRS resulted in a reduction of FGFR3 expression to 16.22%. In NCI-929 cells, FRS reduced FGFR3 expression to 25.02%. FRS showed significant reduction of FGFR3 protein expression in all three cell lines (P < 0.01; Fig. 1B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

A, the sequences of two complementary oligonucleotides necessary to create shRNA and the inhibition of FGFR3 protein expression by shRNAs. Both oligonucleotides are designed such that the first four bases create a 5′ overhang compatible with BbsI. In the sense strand (top), the 5′ overhang is followed by an A (transcription initiation point of the human H1 promoter), then the target sequence of 21 nucleotide, 7 bases (TCAAGAG) for the spacer region and the inverted 21 nucleotide sequence. The sense strand ends with TT to reconstitute the T5 terminator sequence. Each shRNA consists of a 21 bp double-stranded region corresponding to the target sequence and a small loop formed by the spacer region. Each shRNA targets a site within coding sequences for human FGFR3 transcript variant 1 (GenBank accession number NM_000142) and transcript variant 2 (GenBank accession number NM_022965). B, for measurement of FGFR3 expression, KMS-11, OPM-2, and NCI-H929 were harvested 7 d after Zeocin selection. Total cell lysates were probed with an anti-FGFR3 antibody. Coomassie brilliant blue (CBB) was used as loading control. Data were normalized to FGFR3 expression of vector only control being 100%. Columns, mean of triplicate experiments; bars, SD. **, P < 0.01, compared with control (t test).

FRS Effectively Inhibits FGFR3 Protein Expression with t(4;14)

Our research group has previously conducted much work on KMS-11 cells and NCI-H929 cells (14). Comparatively, K650E mutation, found in OPM-2 cells, showed more transforming activity in vitro and in vivo. As a result, we focused our study on OPM-2 cells. Figure 2 shows typical cells of t(4;14) and FGFR3 expression levels before and after FRS transfection using the BioView Duet Imaging System in OPM-2 cells. The Bioview imaging system enables the sequential analysis of single cells by both immunocytochemistry and FISH, thus allowing us to test for FGFR3 expression and t(4;14) concomitantly. Figure 2A shows FGFR3 protein expression for OPM-2 cells by immunocytochemistry. Figure 2B shows the FISH results for the same cells. The 1.5 Mb 14q32/IGH probe was labeled with SpectrumGreen and the 900 kb 4p16.3/FGFR3 probe was labeled with SpectrumOrange. Normal cells without t(4;14) should have two red dots and two green dots, whereas a cell carrying the translocation will display one red (FGFR3), one green (IGH), and two red/green fusion (yellow) dots on der(4) and der(14). Figure 2B shows an OPM-2 cell with two red/green fusion signals, one on derivative 4 (red) and one on derivative 14 (green), indicating that this cell is positive for t(4;14). The cell line also contains multiple copies of chromosome 4 and chromosome 14, accounting for the extra red and green signals seen in the figure. Three days after FRS transfection, FGFR3 protein expression was undetectable by antibodies (Fig. 2C), although t(4;14) was detected in the cells by FISH analysis (Fig. 2D). Moreover, the entire cell (Fig. 2D) is shown with green fluorescence, confirming that transfection occurred with FRS, due to the presence of a green fluorescent protein marker in the psiRNA-hH1GFPzeoG2 vector. These results confirm that overexpression of FGFR3 due to t(4;14) can be specifically and effectively inhibited by FRS.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

T(4;14) translocation and FGFR3 expression before and after FRS transfection. A, positive FGFR3 staining in the OPM-2 cell line. B, FISH analysis for the same cell. This cell is positive for t(4;14), indicated by presence of two red/green fusions, one fusion on derivative 4 and one on derivative 14. C, OPM-2 cell at day 3 posttransfection. FGFR3 expression was not detected by immunocytochemistry. D, FISH analysis on the same cell indicated in C. The t(4;14) is present (two red/green fusions), however, FGFR3 expression is at background level. The whole cell in D shows green fluorescence to show that this cell was transfected with FRS-psiRNA-hH1GFPzeoG2.

Time-Dependent Down-Regulation of FGFR3 Protein by FRS

Because vector-based shRNA can produce a longer period of inhibition, we initially tried to establish a stable transfected cell line without FGFR3 expression. Unexpectedly, almost all of the cells died after about 2 weeks under Zeocin selection. We harvested the cells every 24 hours to test for FGFR3 protein expression using Western blot. There were no FGFR3 protein expression changes 24 hours after transfection, but FGFR3 expression continued to decrease with time during the 2- to 5-day period, from 80.39% expression on day 2 to 12.18% expression on day 5 (Fig. 3). The inhibitory effect did not show significant changes after 5 days (data not shown). So we chose day 5 cells for the subsequent experiments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Time-dependent down-regulation of FGFR3 by FRS. Cells were harvested at indicated times and total cell lysates were probed with an anti-FGFR3 antibody. The densities of each band were determined by ImageQuant software. Coomassie brilliant blue was used as loading control in each lane. Data were normalized by FGFR3 expression of the parental cell control, defined as 100%. Columns, means of three independent experiments; bars, SD. *, P < 0.05; **, P < 0.01, compared with control (t test).

Inhibition of FGFR3 Induces Cell Apoptosis

After transfection, fewer viable cells were always observed under the microscope in the FRS-transfected cells compared with control. We used Annexin V staining and flow cytometry to determine if cells had undergone apoptosis after FRS transfection. Both electroporation and Zeocin inhibition caused cell death. Although dead cells were difficult to completely eliminate from suspension cells, an obvious increase of Annexin V–positive cells could still be observed after FRS transfection (Fig. 4, left), indicating an increase of apoptosis. Because BCL2 and MCL1 were most intensively studied in multiple myeloma, we measured whether BCL2 and/or MCL1 decreased after FRS transfection using Western blot. As shown in Fig. 4 (right), both BCL2 and MCL1 decreased. Flow cytometry and Western blot results indicated inhibition of FGFR3-induced apoptosis in OPM-2 cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Confirmation of apoptosis by flow cytometry and Western blot after FRS transfection. Five days after FRS transfection, cells were harvested and stained with Annexin V and analyzed by flow cytometry. Left, top and middle, typical results of flow cytometry. Left, bottom, percentage of Annexin V–positive cells. Columns, mean of triplicate experiments; bars, SD. *, P < 0.05, compared with control (t test). Right, total cell lysates were probed with anti-BCL2 and anti-MCL1 antibodies. Coomassie brilliant blue was used as loading control.

Confirmed Down-Regulation of FGFR3 mRNA Expression by Quantitative Real-time PCR following Reverse Transcription-PCR

RT-PCR and quantitative real-time PCR were done to examine changes of FGFR3 mRNA expression levels under the influence of FRS. RT-PCR showed that the mRNA level of FGFR3 was significantly lower (down-regulated) in FRS-transfected cells than in both parental cells and vector control cells on day 5 (Fig. 5A). To confirm the down-regulation of FGFR3 mRNA, we used quantitative real-time PCR to test FGFR3 mRNA expression. The standard curve formulas, Y = 44.475 − 3.661X (r² = 0.998) for FGFR3 and Y = 43.826 − 3.567X (r² = 0.999) for GAPDH, were derived from serial dilution of positive control. The mRNA amount of FGFR3 was calculated as described in Materials and Methods. The mean ratio of FGFR3 mRNA to GAPDH mRNA was 19.88% for FRS samples, whereas the mean ratio for the control was set as 100%. The average mRNA amount in FRS samples was 5.03-fold less than that of the vector control (P < 0.001). Results of a typical analysis are shown in Fig. 5B.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Down-regulation of FGFR3 mRNA expression after FRS transfection. Expression of FGFR3 was examined by RT-PCR (A) and quantitative real-time PCR (B). RT-PCR and real-time PCR were done using primers as described in Material and Methods. GAPDH was used as an internal control. A, RT-PCR products were determined on a 2% agarose gel (left) and analyzed by ImageQuant Software. The values of the ratio of FGFR3 to GAPDH were calculated and normalized by parental cells when the value of parental cell was defined as 100% (right). B, typical results of the amplification plots of FGFR3 and GAPDH by real-time PCR (left). Standard curves for FGFR3 and GAPDH were generated using serial dilution of control plasmids. FGFR3 values were normalized to GAPDH, and expressed as a percentage of control. Each sample was assayed in replicates of four. Columns, mean; bars, SD. *, P < 0.05; **, P < 0.01, compared with control (t test).

Gene Expression Profiling

To better understand changes in global gene expression after inhibition of FGFR3, Human Genome U133 plus 2.0 array was used to analyze expression levels. As shown in Table 1 (panel A), there were almost no changes seen in the expression level of the housekeeping genes GAPD and β-actin in vector control and FRS samples. Of the two probe sets (204379_s_at and 204380_s_at) that target FGFR3 in U133 plus 2.0 array, 204380_s_at targeted the CDS sequence and 204379_s_at targeted 3′ downstream of CDS. FRS effectively inhibited the coding sequence of FGFR3 (Table 1, panel B). No changes were observed for FGFR1, FGFR2, and FGFR4, although the FGFR family is known to be highly homologous.

Although it is impossible to elucidate all the sequential genetic changes that may occur after the inhibition of FGFR3, we found that most of the changes were associated with apoptosis, immune response, and translational activity. Examination of apoptosis-related genes revealed lower expression of the antiapoptotic BCL2 and MCL1, whereas expression of the proapoptotic BID increased. CYCS, known to be involved in apoptosis and regulated by BCL2, increased, validating our previous observations (Fig. 4, right). Aside from the BCL2 gene family, expression of members of the caspase family, CASP2 and CASP6, increased whereas expression of CASP8- and FADD-like apoptosis regulator CFLAR decreased. Expression of the inhibitor of apoptosis proteins (IAP) genes, BIRC2, BIRC3, and BIRC5, also decreased. In addition, several members of the NF-κB gene family, which are known to be important in transcription of antiapoptotic factors, such as RELB, NF-κB2, REL, NF-κB1A, and BCL3, were found to have decreased expression (Table 1, panel C).

FGFR3 signaling is in part involved in the RAS/mitogen-activated protein kinase (MAPK) pathway (3, 15) leading to cell proliferation. We investigated genes associated with the RAS/MAPK signal transduction pathway. Down-regulation of RAS, RAF, MAPK14, MAP2K5, and MAP3K1 was observed. The expression of IL-6R and VEGF genes, which are upstream from the RAS/MAPK cascade and function to promote multiple myeloma cell proliferation, were decreased (Table 1, panel D).

Microarray results also revealed down-regulation of MYC and JUN, as well as changes in several genes not commonly associated with multiple myeloma (i.e., SAS, GADD45A, TLR4, and KLF4; Table 1, panel E).

Discussion

Both synthetic siRNA and vector-based shRNA have been tested for their ability to down-regulate protein expression in mammalian cells. Currently, vector-based shRNA is more popular than chemically synthesized siRNA for knocking out a target gene. In this study, we found that not all of the shRNA we generated can effectively inhibit FGFR3. Our results are consistent with results of other groups, demonstrating that siRNA targeting distinct mRNA usually results in different silencing efficiency (16, 17). Because shRNA incompletely inhibits expression of the target gene and evokes different inhibition efficiency, it is important to find ways to increase inhibition in different cell lines. Ji et al. (18) reported that cotransfection of cells with two or more siRNA duplexes, targeting different sites on the same mRNA, resulted in enhanced gene silencing when compared with transfection with each single siRNA. Similarly, we cotransfected three different cell lines with three shRNAs, and we have found that cotransfection is more efficient for the inhibition of FGFR3 expression in all of the cell lines.

After FRS transfection, we observed fewer viable cells in FGFR3-inhibited cells than in cells containing the vector only. On the basis of this observation and previous reports (10), we hypothesized that inhibition of FGFR3 will lead to decreased multiple myeloma cell proliferation and increased apoptosis. Whether a cell lives or dies is largely determined by the ratio of proapoptotic to antiapoptotic proteins, such as the BCL2 family of proteins and caspases (19, 20). The BCL2 family consists of both antiapoptotic, such as BCL2 and MCL1, and proapoptotic molecules. BCL2, which includes a transmembrane domain that anchors to the outer mitochondrial membrane, can block the release of cytochrome c from mitochondria. In contrast, BID can promote cytochrome c release from mitochondria, activating the caspase cascade. Not surprisingly, down-regulation of BCL2 and up-regulation of BID can both make mitochondria release cytochrome c and direct the cells to undergo apoptosis, as our microarray results indicated (CYCS was 1.4 in signal log ratio change). For antiapoptotic molecules, BCL2 and MCL1 have been intensively studied in multiple myeloma. However, there is a great deal of discrepancy. Some studies reported that BCL2 and/or MCL1 increased in multiple myeloma patients, whereas others have shown no change at all (21–23). Our studies showed that both BCL2 and MCL1 were down-regulated. Other interesting apoptosis-related gene expression alterations are found in CFLAR and NF-κB. Munshi et al. (24) compared gene expression profiles from the bone marrow of a multiple myeloma patient with that of her identical twin, and revealed that CFLAR increased 29.9-fold whereas FGFR3 increased 9.0-fold. Our study found that after inhibition of FGFR3, CFLAR has decreased 6.06-fold (equal to −2.6 signal log ratio). This finding suggests that FGFR3 and CFLAR may have a causative relationship in the development of multiple myeloma. Our microarray results also showed down-regulation of members of the NF-κB gene family. Nuclear factor κB (NF-κB) is a regulatory protein that activates transcription of a number of genes, including antiapoptotic factors (25). Increased NF-κB activity may protect tumors from apoptosis (26–28). Inhibition of NF-κB by PS-341 showed enhanced sensitivity of myeloma cells to chemotherapeutic agents (29). Our findings, revealing a pattern of changes in the expression of apoptosis-related genes, may explain how cells undergo apoptosis after inhibition of FGFR3.

Down-regulation of RAS and RAF genes was also observed. Both of these genes are important factors in the RAS/MAPK pathway, which promotes cell growth. Our results are consistent with other reports, which found that FGFR3-induced transformation could be inhibited by cotransfection with dominant-negative forms of RAS or RAF. Although there is no evidence that the MAPK14 (also known as p38) pathway directly promotes the growth of multiple myeloma cells, studies indicated that p38 promotes multiple myeloma cell growth via paracrine secretion of interleukin 6 (IL-6) and vascular endothelial growth factor (VEGF; ref. 30). Microarray results also showed that both MAPK14 and VEGF were decreased after inhibition of FGFR3. IL-6, which is generated mainly by marrow stromal cells, is a major growth and survival factor for multiple myeloma cells (31–33). IL-6 binding to IL-6 receptor triggers homodimerization of gp130, which is believed to activate the RAS/MAPK and Janus-activated kinase/STAT pathways (34). Overexpression of FGFR3 can substitute for IL-6 receptor signaling in IL-6–dependent murine B9 myeloma cells (10). Our microarray results showed that despite lack of changes seen in IL-6 expression levels, inhibiting FGFR3 might also knock out IL-6 receptor expression. For MAPK and MAPK kinase, the important functional regulations are phosphorylation or dephosphorylation. However, because microarray results reveal only mRNA expression levels, no conclusions can be made regarding any regulating effects on MAPK and MAPK kinase.

Interestingly, expression of Kruppel-like factor 4 (KLF4), which has not yet been linked to multiple myeloma, was 32-fold (equal to 5 signal log ratio) higher in FRS-transfected OPM-2 cells than in the vector only controls. KLF4 is a key transcriptional regulator of cell differentiation and proliferation, which can reduce tumorigenicity of colon and bladder cancer cells (35, 36). Despite uncertainty about which signaling pathways are involved, the microarray results imply that KLF4 may play a role in multiple myeloma disease progression. Another interesting finding is that inhibition of FGFR3 down-regulates the oncogene MYC. MYC expression is tightly controlled by mitogen availability in normal cells, but it is usually expressed in a deregulated or elevated manner in tumor cells, including multiple myeloma (37–39).

Recently, preclinical studies targeting FGFR3 in human myeloma cells lines with SU5402, 3-[(3-(2-carboxyethyl)-4-methylpyrrol-2-yl)methylene]-2-indolinone, and PD173074, 1-tert-butyl-3-[6-(3,5-dimenthoxy-phenyl)-2-(4-diethylaminobutylamino)-pyrido[2,3-d]pyrimidin-7-yl]-urea, found that FGFR3 inhibition was associated with decreased viability and increased apoptosis (40–42). SU5402 and PD173074, known inhibitors of FGFR1 (43), can inhibit FGFR3 phosphorylation. SU5402 and PD173074 can significantly reduce viability of KMS-11 cells, but only a modest reduction in viability was observed for the OPM-2 and NCI-H929 cell lines. One possible explanation for this difference is that KMS-11 cells express the Y373C mutation that activates FGFR3 by promoting constitutive dimerization in the absence of the ligand. This specific targeting may limit the clinical usage of these two compounds because FGFR3 is expressed in ∼70% of t(4;14) patients and only a small fraction of these patients acquire an activating mutation of FGFR3 (11, 44). In addition, SU5402 and PD173074 do not specifically target FGFR3. FRS, however, targets and cleaves mRNA. According to the microarray results, expression of none of any other FGFR family member genes changed except FGFR3 after FRS transfection. FRS can specifically inhibit FGFR3 regardless of FGFR3 phosphorylation. Thus, the clinical benefit of FRS could be broad.

In conclusion, our study proposes a new method of inhibiting FGFR3 expression, leading to genetic changes related to apoptosis. Whereas our results need further confirmation, shRNA-induced inhibition of FGFR3 in myeloma cells transfected with FRS is worthwhile to test as a novel therapeutic tool for patients with FGFR3 overexpression by t(4;14) in multiple myeloma. Testing of the therapeutic efficacy of shRNAs in the xenograft model of FGFR3-induced multiple myeloma and primary myeloma cells from t (4;14) patients, using inducible shRNA expression vector, is ongoing in our lab.