To achieve successful silencing of FGF2 and NUDT6, we initially compared the effects of three Silencer^TM Select Pre-designed siRNA variants on FGF2 and NUDT6 mRNA expression. In each case, all three tested sequences significantly decreased the targeted mRNA (p < 0.05, data not shown). FGF2 siRNA variant s132467 and NUDT6 siRNA variant s220427 had the greatest effect on their respective target mRNAs, and were therefore chosen for use in all subsequent experiments. The mRNA knockdown was dose- and time-dependent (Figure 1 and Figure 2). Significant knockdown of either target was evident within 48 hours following treatment with as little as 50 nM siRNA. As shown in Figure 1a, FGF2 knockdown with 100 or 200 nM siRNA (but not lower doses) resulted in a significant (3-fold) increase in NUDT6 mRNA. In contrast, even the lowest dose (50 nM) of NUDT6 siRNA significantly increased FGF2 mRNA expression, in conjunction with NUDT6 knockdown (Figure 1b).

The effects of siRNA knockdown of the target genes were detectable as early as 24 hours and persisted for up to 96 hours (Figure 2). The up-regulation of FGF2 following NUDT6 knockdown was evident within 24 hours, and remained elevated for the 96-hour duration of the experiment (Figure 2a). NUDT6 mRNA was also up-regulated following FGF2 silencing, although the response was slower in onset, not being detectable until 48 hours after siRNA treatment. Interestingly, NUDT6 mRNA remained elevated at 72 and 96 hours following FGF2 siRNA treatment, even after FGF2 mRNA levels had returned to normal (Figure 2b). Taken together these data strongly support the hypothesis that FGF2 and NUDT6 mRNAs are mutually regulatory. We next examined the effect of the siRNA knockdowns on the levels of the encoded proteins.

Alternative translation initiation results in synthesis of five isoforms of FGF2 ranging from 18 to 34 kDa in size, with distinct intracellular localizations and actions [21,22,23]. The low molecular weight (LMW) 18kDa FGF2 isoform is secreted by a non-classical pathway [24] and binds to high affinity cell surface receptors [25]. The high molecular weight (HMW) FGF2 isoforms contain nuclear and nucleolar localization signals and have cell survival, proliferation and immortalization activities [26,27]. In order to determine if NUDT6 knockdown differentially affected the expression of the various FGF2 isoforms, Western blot analysis was used to determine the relative abundance of LMW and HMW FGF2 following siRNA knockdown of either FGF2 or NUDT6. As shown in Figure 3a, C6 glioma cells predominantly express three FGF2 isoforms; the LMW 18 kDa isoform and two HMW isoforms of 21 and 23 kDa, with the 23 KDa isoform being most abundant.

FGF2 knockdown significantly attenuated all three isoforms, whereas NUDT6 knockdown resulted in up-regulation of all FGF2 isoforms. Densitometric analysis did not detect any significant differential effect on individual isoforms (not shown). As depicted in Figure 3b, total FGF2 immunoreactivity was significantly up-regulated more than 3-fold following NUDT6 knockdown, resulting from comparable up-regulation of each of the three major FGF2 isoforms.

Confocal laser scanning microscopy (CLSM) analysis confirmed that immunoreactive FGF2 protein was present in the cytoplasmic, perinuclear and nuclear compartments in unsynchronized cells transfected with scrambled control siRNA (Figure 4a, top panels). In contrast, NUDT6 immunoreactivity showed significant nuclear staining, as well as punctate extra-nuclear staining consistent with the mitochondrial localization previously described [28] for rat NUDT6 (Figure 4b, top panel). Following 72 hours of FGF2 siRNA transfection, FGF2 immunostaining in both nucleus and cytoplasm was markedly attenuated, whereas it was markedly up-regulated in both compartments following NUDT6 knockdown (Figure 4a, middle and bottom panels). Conversely, NUDT6 immunostaining was dramatically up-regulated following FGF2 siRNA treatment, and greatly attenuated by NUDT6 knockdown (Figure 4b, middle and bottom panels). These data indicate that the FGF2 and NUDT6 protein levels are mutually regulated by their antisense RNA partner in C6 glioma cells.

In order to confirm that the changes in FGF2 protein expression detected by CLSM reflected actual alterations in cellular protein content, we performed a fluorescence-activated cell sorting (FACS) analysis of FGF2 content in the C6 cells following siRNA mediated knockdown of FGF2 or NUDT6. As shown in Figure 5a, FGF2 siRNA significantly reduced the total cellular FGF2 content (as indicated by the mean staining intensity) 72 hours following siRNA transfection, whereas transfection with the NUDT6 siRNA significantly increased the FGF2 levels. Furthermore, as shown in Figure 5b, the up-regulation of cellular FGF2 immunoreactivity following NUDT6 knockdown was detectable within 48 hours, and the effect persisted for the duration of the experiment. Similarly, FGF2 siRNA transfection significantly suppressed FGF2 within 48 hours following transfection, and the inhibition persisted for up to 4 days following transfection (Figure 5b).

We have previously reported that endogenous FGF2 is required for C6 glioma cell proliferation and that FGF2 antisense oligonucleotides inhibit C6 cell proliferation and colony formation in soft agar [29]. However, the physiological significance of the NUDT6 gene for FGF regulation of cell proliferation has not been fully clarified. We therefore examined the effect of FGF2 and NUDT6 knockdown on C6 glioma cell growth. As shown in Figure 6a, knockdown of FGF2 significantly inhibited cell proliferation, detectable within 48 hours post-transfection. This inhibition persisted for the duration of the experiment, becoming highly significant at 96 hours post-transfection (0.72 ± 0.08 × 10⁵ vs. 2.07 ± 0.18 × 10⁵ for the negative control transfected cells). Remarkably, knockdown of NUDT6 also resulted in inhibition of cell proliferation, although this did not reach significance until 4 days post-transfection, and achieved only 31% inhibition vs. control, compared to 65% inhibition seen following FGF2 knockdown.

In order to determine if the inhibition of cell proliferation was a direct result of perturbation of FGF2 expression, we examined the effect of exogenous FGF2 addition to siRNA transfected cells. As shown in Figure 6b, recombinant FGF2 caused a modest but non-significant increase in proliferation in cells transfected with a negative control siRNA. In contrast, FGF2 siRNA transfection caused a highly significantly 64% inhibition of cell proliferation, and this effect was completely reversed by addition of 10 ng/mL FGF2 to the culture medium. In contrast, addition of exogenous FGF2 had no significant effect on proliferation of cells transfected with NUDT6 siRNA.

Over-expression of recombinant NUDT6 results in perturbation of cell cycle progression [5,19]. We used siRNA knockdown to examine the effect of endogenous NUDT6 and FGF2 on cell cycle progression. Cells were transfected with control or gene-specific siRNAs, then arrested and synchronized in G0/G1 by serum starvation (0.01% FBS for 48 hours) as previously described [19]. As shown in Figure 7a, cells transfected with a negative control siRNA rapidly re-entered the cell cycle following re-addition of 10% FBS, with 16.77 ± 1.3% of control cells in S-phase within 3 hours following serum replacement. The number of control cells in S-phase increased further to more than 20% by 6 hours, and remained at this level for the duration of the experiment. Knock-down of FGF2 resulted in a highly significant decrease in the number of cells entering S-phase, an effect which was detectable as early as 3 hours after serum addition (5.93 ± 0.8% vs. 16.77 ± 1.3% in controls), and which persisted for up to 24 hours. Consistent with the cell proliferation results observed in Figure 6, knock-down of NUDT6 also caused a delay in S-phase re-entry, although the effect was not as profound as that seen with FGF2 knockdown (Figure 7a). At 3 hours after serum readdition, the number of cells in S phase was 10.63 ± 1.3% in NUDT6 knockdowns vs. 5.93 ± 0.8% in FGF2 knockdowns. The effect of NUDT6 knockdown on S phase reentry persisted for up to 12 hours, but by 24 hours there was no significant difference from controls.

We next examined the effect of siRNA treatment on cell cycle progression. As shown in Figure 7b, both FGF2 and NUDT6 knockdown resulted in significantly increased accumulation of cells in the G0/G1 phases of the cell cycle, and significant decreases in the population of S-phase cells, consistent with the results observed in Figure 7a. However, NUDT6 knockdown resulted in a significant reduction in the percentage of cells in G2/M (1.33 ± 0.3% vs. 3.5 ± 0.4% in controls) whereas FGF2 knockdown resulted in a significant increase in the percentage of cells in G2M (6 ± 0.8%). These data are consistent with our previous report that FGF2 promotes progression from G1 to S phase [19], and further indicate that coordinated expression of both FGF2 and NUDT6 is required for normal cell cycle progression, although at different stages of the cycle.

In order to examine the functional implications of FGF2 and NUDT6 expression on C6 glioma cell behaviour, we assessed the effect of knockdown on features of cell behaviour that characterize the neoplastic phenotype, specifically: anchorage-independent growth, adhesion to extracellular matrix proteins, cell migration, invasiveness, and wound healing. Knockdown of either FGF2 or NUDT6 significantly reduced cell adhesion to wells coated with collagen I, collagen IV, and laminin I, but not to fibronectin or fibrinogen (data not shown). Knockdown of either target also significantly inhibited anchorage-independent growth in soft agar by more than 60% compared to controls (Figure 8a). Similarly, knockdown of either FGF2 or NUDT6 caused a highly significantly delay in wound healing over the 5 day duration of the experiment (Figure 8b). There was no significant effect of either knockdown on cell migration or on cell invasiveness as measured using commercially available kits (data not shown).

Taken together, these findings support the hypothesis that the complementary RNAs encoding FGF2 and NUDT6 are mutually regulatory, and that silencing of either gene results in up-regulation of the corresponding antisense partner, with consequent disruption of cell cycle progression and cell proliferation.

Culture medium, fetal bovine serum (FBS) and all other culture reagents were from Gibco BRL (Burlington, Ont., Canada). Gels, membranes and other materials used in Western blotting were, unless otherwise noted, from Bio-Rad Laboratories, Hercules, CA. Anti-FGF2 IgG was obtained from Santa Cruz Biotechnology, Santa Cruz, CA; and anti-NUDT6 and anti-actin were from ProteinTech Group, Chicago, IL, (all rabbit polyclonal). Alexa Fluor^TM 488 (green) goat anti-rabbit fluorescent IgG conjugate was from Invitrogen, Carlsbad, CA. All other chemicals were from Sigma-Aldrich Chemical Company, St. Louis, MO, unless otherwise specified.

Rat C6 glioblastoma cells, obtained from the American Type Culture Collection (Manassas, VA, USA), were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, penicillin (100 units/mL) and streptomycin (0.1 mg/mL), and kept at 37 °C in a humidified, 5% CO₂ chamber. Transfection with the specific Silencer^TM siRNAs (Applied Biosystems, Austin, TX) was according to the manufacturer’s instructions, with Lipofectamine^TM RNAiMAX (Life Technologies Corp., Carlsbad, CA, USA) used as the transfection reagent.

Total RNA was isolated using the RNeasy^TM kit from Qiagen Inc., Mississauga, Ont., Canada, as previously described [17], and reverse-transcribed with Omniscript^TM reverse transcriptase (Qiagen). Amplification of the transcripts used 35 cycles of 45 sec at 95 °C, 60 sec at 57 °C and 60 sec at 72 °C with the primers listed in Table 1 (all obtained from Invitrogen), and the HotStar Taq^TM master mix (Qiagen). DNA levels were subsequently normalized against the β-actin PCR product amplified from the same RT reaction.

C6 cells were tranfected as required and incubated for 72 hours. The cells were trypsinized, pelleted at 400 × g for 10 min, washed with phosphate-buffered saline (PBS), and resuspended in RIPA buffer (50 mM Tris, 150 mM NaCl, 50 mM Na₂HP04, 1 mM Na₃V0₄, 1 mM NaF, 5 mM EDTA, 5 mM EGTA, 0.25% Na deoxycholate, 0.1% NP40) supplemented with a protease inhibitor cocktail (Sigma-Aldrich Canada Ltd., Oakville, ON, Canada). Protein was extracted following 30 minutes on ice with frequent vortexing and quantified with a Bradford assay using coomassie reagent (Pierce, Rockford, IL). Equal amounts of protein (30 μg) from each lysate were separated with SDS-PAGE using 12% NuPAGE® Novex® Bis-Tris mini gels in a Xcell SureLock^TM Novex Mini-Cell (Life Technologies Corp., Carlsbad, CA, USA) and transferred to a nitrocellulose membrane for western analysis as previously described [16]. Immunodetection of proteins was conducted with fluorescent conjugated secondary antibodies anti-mouse 750 (1:2500) and anti-rabbit 680 (1:5000) (Life Technologies Corp., Carlsbad, CA, USA) against mouse monoclonal FGF-2 (1:200, sc-135905) and rabbit polyclonal Actin (1:1000, sc-1616) (Santa Cruz Biotechnology, Inc., CA, USA). Immunoreactive bands were detected using the LI-COR Biosciences Odyssey^TM Imager and quantified using ImageJ.

For immunofluorescence studies, knockdown transfections were performed in dishes containing microscope coverslips. Following 72 hours incubation, the cells were fixed with ice-cold 2% paraformaldehyde (with 9 mg/mL disodium hydrogen orthophosphate and 6 mg/mL L-lysine) for 15 min, permeabilized with ice-cold 0.1% Triton X-100 in PBS for 15 min and then blocked with 3% bovine serum albumin (BSA) in PBS for one hour (13,18). Staining was by sequential exposure to primary antibodies and their fluorescently tagged secondary antibodies, each for an hour at room temperature, followed by staining for DNA using propidium iodide (PI, 50 ng/mL in PBS) for 10 min. All antibodies were diluted in 0.1% BSA-PBS as follows: anti-FGF2, 2 μg/mL; anti-NUDT6, 3 μg/mL; and Alexa Fluor^TM 488 (green) anti-rabbit fluorescent IgG conjugate, 40 μg/mL. Control staining to eliminate antibody non-specificity was performed by application of secondary antibodies without prior exposure of the cells to the primaries, or following incubation with the primary antibody’s preimmunizing peptide. The coverslips were upended onto slides with a drop of glycerol-PBS (Citifluor^TM, Marivac, Halifax, NS, Canada), and image analysis used the standard operating software on the Zeiss LSM 510 microscope.

Cells were plated at a density of 2 × 10⁵ cells/cm² on culture dishes in normal (10% FBS) DMEM and allowed to attach overnight. The cells were then washed with PBS, transfected as appropriate, and incubated for 48 hours in 0.01% FBS DMEM (serum deprivation) to synchronize them in G0. The media were then replaced with normal DMEM and the cells harvested at indicated intervals for DNA content analysis. In brief, the trypsinized cells were centrifuged (400 × g, 10 min at 4 °C), fixed overnight in 70% ethanol at 4 °C and incubated for 30 min at room temperature in PBS with 50 ng/mL propidium iodide and 100 units/mL ribonuclease A [19]. To assess FGF2 content, transfected cells were harvested on successive days, fixed in ethanol, and incubated with an anti-FGF2 antibody−Alexa Fluor secondary antibody combination. Fluorescence intensities were determined in a Becton Dickinson FACSCalibur ^TM flow cytometer, the proportion of cells in different phases of the cycle estimated using the ModFit program. Cellular FGF2 immunofluorescence following siRNA treatment was quantified using FCS Express V3 as previously described [19]. Minimum and maximum fluorescence cutoffs were set following background subtraction using cells stained with secondary antibody only. FGF2 immunoflourescence values were determined from the modal fluorescence intensity readings for each treatment. siRNA transfections were performed in triplicate, and three samples from each transfection were used for immunofluorescence readings. Each reading was repeated three times, and the average fluorescence recorded for each sample.

In the proliferation assay, the cells were grown in low serum (1% FBS) DMEM (control) or 1% FBS-DMEM with 10 ng/mL recombinant human FGF2 (EMD Biosciences), then harvested on successive days and counted through a Coulter counter.

The assessment of cell transformation and tumorigenic tendency was conducted following knockdown transfections and 72 hours incubation of the cells, and all assays were performed using CytoSelect ^TM Assay kits (Cell Biolabs, Inc., San Diego, CA, USA) according to the manufacturer’s instructions for each kit:

All data were compared using the ANOVA with Student-Newman-Keuls procedure, and are presented as mean ± standard error of the mean (SEM). Differences of p < 0.05 were considered significant.