Introduction

Osteosarcoma is the most common primary bone malignancy and is characterized by histologic heterogeneity and high genetic instability (Luetke et al. 2014). Despite the drastic improvement in survival rates introduced by optimized chemotherapy programs, over the past decades, few advances have been made for patients with poor response to the conventional therapies (Anderson 2016). Typically, osteosarcoma chemotherapy involves agents such as doxorubicin, cisplatin, cyclophosphamide, methotrexate, and etoposide, which are commonly used alone or in combination. Nonetheless, chemotherapy can be toxic to healthy tissues or organs, and induce secondary malignancies. Moreover, chemoresistance is still a limitation to therapy (Luetke et al. 2014).

A diet rich in plant polyphenols has been documented to protect against several diseases, e.g., cardiovascular, neurodegenerative and gastrointestinal disorders, and cancer (Gomes et al. 2008; González-Gallego et al. 2010; Pandey and Rizvi 2009). Bone health maintenance is the consequence of a dynamic equilibrium between bone formation and degradation. Regarding this, polyphenols have been documented to show positive effects on the interaction between osteoblast and osteoclast activities, as reviewed (Horcajada and Offord 2012; Shen et al. 2012). In a collagen-induced arthritis model, the flavonoid fisetin (3,3′,4′,7-tetrahydroxyflavone) was found to decrease the incidence and severity of arthritis (Lee et al. 2009). Moreover, fisetin was reported to modulate bone mineral density, micro-architecture parameters, and bone markers, by repressing RANKL-induced osteoclast differentiation/activity and by stimulating osteoblast differentiation (Choi et al. 2012; Kim et al. 2014; Léotoing et al. 2013, 2014; Sakai et al. 2013). Regarding its anti-cancer properties, fisetin has been described as an anti-inflammatory agent inhibitor of NF-κB activation (Liao et al. 2009; Sung et al. 2007) and as an apoptosis inducer in multiple human cancer cell lines, e.g., (Khan et al. 2008; Murtaza et al. 2009; Suh et al. 2009; Ying et al. 2012). In addition, apoptosis induction by fisetin has been linked to p53-dependent pathways, e.g., (Chen et al. 2002; Li et al. 2011; Lim and Park 2009). Concerning osteosarcoma, fisetin was previously reported as inducer of apoptosis in the wild-type p53 U2OS cell line (Li et al. 2015).

In response to cell damage, p53 can induce cell cycle arrest and cell death commitment. Previously, fisetin treatment resulted in decreased proliferation and apoptosis in the wt p53 and wt Rb U2OS cell line (Li et al. 2015; Ying et al. 2012). This cell line exhibits epithelial morphology and is described as near hypertriploid, with most chromosomes showing abnormalities. The MG-63 and Saos-2 cell lines are both hypotriploid, and whereas the MG-63 cell line grows with fibroblastic shape, the Saos-2 cell line displays epithelial morphology. In addition, the MG-63 cell line showed TP53 sequence rearrangements and promoter hypermethylation, while Saos-2 was found a p53-null mutant (Di Fiore et al. 2014; Masuda et al. 1987).

In this study, the cytotoxic action and cell cycle effects of fisetin on osteosarcoma p53-deficient cell lines were investigated. For this, cells were exposed to fisetin, the chemotherapeutic agent etoposide, and fisetin:etoposide combinations. Cell cycle effects were analyzed and markers of cell cycle deregulation were identified. The study of differential regulation by fisetin:etoposide combinations allowed the identification of relevant biomarkers for the cytotoxic and cytostatic actions of fisetin on osteosarcoma.

Materials and methods

Cell culture media and reagents

Dulbecco’s modified Eagle’s medium, fetal bovine serum, l-glutamine, antibiotics, amphotericin B, and trypsin–EDTA (0.25% Trypsin, 1 mM EDTA) were purchased from Life Technologies (Carlsbad, CA, USA). Dimethyl sulfoxide (DMSO), sodium dodecyl sulfate (SDS), propidium iodide, and ribonuclease were purchased from Sigma-Aldrich (St. Louis, MO, USA). Fisetin, etoposide, and crystal violet (CV) were purchased from Merck Millipore (Billerica, MA, USA).

Cell culture and exposure conditions

Human osteosarcoma cell lines MG-63, Saos-2, and U2OS were purchased from American Type Culture Collection (ATCC, Rockville, MD, USA). The cell lines were maintained and subcultured in complete medium, i.e., Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, 2 mM l-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2.5 µg/ml amphotericin B. Cells were incubated in a humidified atmosphere of 5% CO2 at 37 °C. Subconfluent cells were trypsinized with trypsin–EDTA (0.25% trypsin, 1 mM EDTA) and subcultured at a split ratio of 1:6. For exposure experiments, concentrated stock solutions of fisetin (200 mM) or etoposide (60 mM) were prepared in DMSO. Final dilutions for exposure were prepared in complete medium. Final DMSO concentration was 0.1% for all conditions, including control. Each experiment was conducted with freshly thawed aliquots of these authenticated cell lines.

Clonogenic assay

MG-63, Saos-2, or U2OS cells were seeded in 6-well plates (100 cells/well) and allowed to adhere overnight at 37 °C. After adhesion, cells were washed with phosphate-buffered saline (PBS) and incubated in medium containing fisetin for 48 h. After exposure, cells were washed with PBS and incubated in complete Dulbecco’s modified Eagle’s medium for 7 days. During the 7-day recovery period, the medium was replaced with fresh medium every other day. After this period, the medium was carefully removed, and cells were fixed and stained for 30 min with crystal violet (CV) solution (0.2% CV and 10% ethanol). The culture plates were washed with distilled water and allowed to dry overnight. Colony image was acquired in a G:BOX CHEMI HR-16 BioImaging system (Syngene, Frederick, MD, USA) and clonogenic assay parameters colony number and area were determined with ImageJ (Abràmoff et al. 2004).

Cell growth and proliferation

Cell growth and proliferation was assessed by crystal violet (CV) assay. The CV assay allows quantification of cell growth based on the quantification of intracellularly retained CV. MG-63 and Saos-2 cells were seeded at a density of 5 × 104 cells/well in 24-well plates. After overnight adhesion, cells were exposed to fisetin, etoposide, or combinations for 48 h. Subsequently, cells were carefully washed with PBS and fixed/stained with CV solution (0.2% CV and 10% ethanol) for 30 min. After staining, cells were carefully washed with distilled water and, after drying, intracellular CV was dissolved in 900 μl of 1% SDS. The culture plates were shaken for 1 h and absorbance was measured at 570 nm using a Synergy HT Multi-mode Microplate Reader (BioTek Instruments, Winooski, VT, USA).

Combination index (CI) analysis

The CI analysis was performed according to the method of Chou-Talalay (Chou 2006) in MG-63, and Saos-2 cells. Briefly, data from CV assay dose–response curves for fisetin and etoposide were considered, together with CV assay data for the different mixtures fisetin:etoposide. The CI values for positive, independent, and negative interaction effects were determined using the CompuSyn software (Chou and Martin 2005). The CI is a quantitative measure of the degree of interaction between two or more drugs. For these studies, a CI = 1 denotes an independent effect of fisetin and etoposide, > 1 a negative interaction between the two drugs, and < 1 a positive interaction between the two drugs for the inhibition of cell growth and proliferation.

Cell cycle analysis

Cell cycle analysis was performed in MG-63 and Saos-2 cells as described in a previous work, with a few modifications (Ferreira de Oliveira et al. 2014). For this, cells on 6-well plates at the density of 3 × 105 cells/well were allowed to adhere at 37 °C overnight. After adhesion, cells were washed with PBS and treated with fisetin, etoposide, and combinations as indicated and additionally exposed for 48 h. Subsequently, cells were detached by trypsinization and centrifuged at 700 g for 5 min. The cells were resuspended in PBS, the cell suspension was recentrifuged as previously (700 g, 5 min) and the cell pellets were resuspended in 1 ml of 85% ethanol at 4 °C after which they were stored at − 20 °C until analysis. For cell cycle analysis, samples were centrifuged at 700 g for 5 min, and the cell pellets were resuspended in 0.8 ml of PBS. Cell suspensions were then filtered through a nylon mesh (35-µm pore), and ribonuclease (final 50 µg/ml) and propidium iodide (final 50 µg/ml) were added to the filtered suspension. For measurements of relative fluorescence intensity of PI-stained nuclei, a Coulter Epics XL Flow Cytometer (Beckman Coulter, Hialeah, FL, USA) equipped with an argon laser (15 mW, 488 nm) was used. Acquisitions were performed using SYSTEM II software v. 3.0 (Beckman Coulter, Hialeah, FL, USA). The number of events analyzed for each sample was ~ 5000. Histogram outputs were used for analysis and frequency results were determined with the software FlowJo (Tree Star, Ashland, OR, USA).

RNA extraction and qPCR

The oligonucleotide primers used were described previously (Coutinho et al. 2017). Primer3 platform (Rozen and Skaletsky 2000) was used for oligonucleotide primer design and primer specificity was confirmed by the UCSC In-Silico PCR tool (http://genome.ucsc.edu/cgi-bin/hgPcr?command=start). The TRIzol method was used to extract total RNA. For this, MG-63 and Saos-2 cells were lyzed in TRI reagent (Life Technologies, Carlsbad, CA, USA). Organic phase separation was achieved in Phase Lock Gel Heavy tubes (5 PRIME Inc., Boulder, CO, USA). The aqueous phase was mixed with 70% ethanol (1:1), and RNA was incubated with DNase I and purified using NZY Total RNA Isolation kit (NZYTech, Lisbon, Portugal). For cDNA synthesis, 2 μg total RNA was reverse-transcribed with 2.5 μM Oligo (dT)18, using the NZY First-Strand cDNA Synthesis Kit (NZYTech, Lisbon, Portugal). The cDNA samples were prediluted in ultrapure water (1:10). The final individual qPCR reactions contained iTaq Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA), 150 nM each gene-specific primer, and 1:4 (v/v) prediluted cDNA (final 1:40). The qPCR program included denaturation for 1 min at 95 °C, and 60 cycles of amplification including denaturation for 5 s at 94 °C, annealing for 15 s at 58 °C, and extension for 15 s at 72 °C. After qPCR, a melting temperature program was performed. At least three qPCR technical replicates were performed per sample from each of three independent biological assays. Cycle thresholds and average PCR efficiencies were estimated from the baseline-corrected fluorescence data using Real-Time PCR Miner (Zhao and Fernald 2005). Gene expression of exposed cells was calculated relative to control cells and normalized with the GAPDH reference gene, following the Pfaffl method (Pfaffl 2001).

Statistical analysis

All data were derived from at least three independent experiments. The inhibitory concentrations (ICs) were determined from the growth assay with the R program for Windows version 3.1.1 (core team, VIE, Austria) from statistical packages drm and drc. Results are mean ± standard deviation (SD), except for gene expression data, where mean ± standard error of the mean (SE) are used. Results were analyzed by one-way ANOVA and compared with the Holm–Sidak statistical test. When normality could not be reached for at least one sample group, Kruskal–Wallis one-way ANOVA on ranks was performed, with post-hoc Tukey’s test analysis. For one-way ANOVA, the software package Sigmaplot for windows, version 11.0 (Systat Software Inc, San Jose, CA, USA) was used. For all comparisons, differences were considered statistically significant at p < 0.05.

Results

Inhibition of colony formation and cell recovery

Cell recovery was assessed by the number colonies relative to control and area occupied by colonies relative to total well area and relative to control (Fig. 1). After 48 h exposure to fisetin, U2OS and Saos-2 cell lines had decreased recovery in the exposures ranging from 10 to 20 µM in a dose-dependent manner (p < 0.05), as assessed by the two parameters. In contrast, MG-63 cell recovery was not significantly affected by fisetin for the same exposure period and concentrations. MG-63 and Saos-2 cells were then selected to test cytotoxic effects as well as cell cycle effects of fisetin used as single agent or combined with etoposide.

Fig. 1
figure 1

Colony formation ability after 48 h exposure to fisetin. After exposure, cells were allowed to recover in complete medium for 7 days after which colonies were stained with crystal violet and clonogenic assay parameters were analyzed with the ImageJ software. a Recovering colonies after exposure to fisetin; b number of colonies relative to control (%); c area occupied by colonies relative to total well area (%). The data are mean ± SD (n = 3). Statistical analysis was performed using the one-way ANOVA test (p < 0.05). Significant differences (asterisks) are shown relative to the control group

Full size image

Inhibition of cell proliferation

MG-63 and Saos-2 cell growth decreased in a dose-dependent manner from 10 to 150 µM fisetin (Fig. 2). Compared to etoposide, fisetin was less cytotoxic. Considering only the tested concentrations, fisetin was significantly cytotoxic starting from 40 μM for MG-63 cell line and 20 μM for Saos-2 cell line (p < 0.05, Fig. 2b, c).

Fig. 2
figure 2

MG-63 and Saos-2 cell growth for single-agent exposures. Cells were exposed to fisetin or etoposide for 48 h and growth was evaluated by CV assay. a Cell morphology and confluence was assessed by inverted microscopy; scale bar: 100 μm; b MG-63 cell growth and proliferation; c Saos-2 cell growth and proliferation. The data are mean ± SD (n = 3). Statistical analysis was performed using the one-way ANOVA test (p < 0.05). Significant differences (asterisks) are shown relative to the control group. Fisetin concentrations: 0, 5, 10, 20, 40, 100, and 150 μM; etoposide concentrations: 0, 0.5, 1, 1.5, 2, 5, and 10 μM. The dose–response models presented are the Brain–Cousens log–logistic model for fisetin exposure and the Cedergreen–Ritz–Streibig model for etoposide exposure

Full size image

From the dose–response curves shown in Fig. 2, inhibitory concentrations IC30 and IC50 were estimated (Table 1). Cells were then exposed, as indicated in Fig. 3, to the combinations of fisetin and etoposide near IC30 concentrations.

Table 1 Inhibitory concentrations of fisetin and etoposide
Full size table
Fig. 3
figure 3

MG-63 and Saos-2 cell growth for double-agent combination exposures. Cells were exposed to fisetin and etoposide for 48 h, and growth was evaluated by CV assay. a Proliferation of MG-63 cells exposed to combination treatment; b proliferation of Saos-2 cells exposed to combination treatment. The data are mean ± SD (n = 3). Statistical analysis was performed using the one-way ANOVA test (p < 0.05) for MG-63 data and one-way ANOVA on ranks for Saos-2 data (p < 0.05). Significant differences between control and each experimental group are marked with asterisks. Significant differences between paired experimental groups are marked with triangles. CI combination index

Full size image

A range of combination indices were observed, depending on the combination, between 1.51 (negative interaction fisetin:etoposide) and 0.59 (positive interaction fisetin:etoposide). For comparison, the effect of etoposide and fisetin at the largest concentrations is shown in Fig. 3, as isolated agents are presented in Table 1.

Cell cycle analysis

Exposure to the most anti-proliferative combinations investigated, i.e., 36 μM fisetin:0.5 μM etoposide for MG-63 cells and 45.5 μM fisetin:1 μM etoposide for Saos-2 cells, resulted in ~ 40% inhibition of cell growth/proliferation, as shown by CV assay (Fig. 3a, b). Exposure to fisetin or etoposide as single agents at ~ 15 and ~ 40% inhibitory concentrations was associated with decreased percentage of cells in G0/G1-phase and increased percentage of cells in G2/M-phase, for both cell lines (p < 0.05; Fig. 4a, b). For cells exposed to fisetin and etoposide as single agents, etoposide induced a larger difference in % G0/G1 and G2/M cells. When considering double-agent exposure, a decrease in % cells in G0/G1-phase and an increase in % cells in G2/M-phase were also observed in cells exposed to combinations of fisetin and etoposide, relative to control (p < 0.05; Fig. 4c, d). However, for both cell lines, despite the increase in % G2/M-phase cells compared to control, a significant decrease was observed in % G2/M-phase cells from the least to the most anti-proliferative combinations (p < 0.05; Fig. 4c, d). In addition to this, in Saos-2 cell line, despite the decrease in % S-phase cells compared to control, a significant increase was observed in % S-phase cells from the least to the most anti-proliferative combination (p < 0.05; Fig. 4d).

Fig. 4
figure 4

Cell cycle distribution of cells exposed to fisetin and etoposide. Cells were stained with propidium iodide and the cell cycle was analyzed by flow cytometry. Single-agent exposure in a MG-63 cells and b Saos-2 cells. Double-agent exposure in c MG-63 cells and d Saos-2 cells. The data are mean ± SD (n = 3). Statistical analysis was performed using the one-way ANOVA test (p < 0.05). Significant differences (asterisk) are shown relative to the control group

Full size image

The gene expression of cell cycle regulators was analyzed after exposure to fisetin, etoposide, and combinations. For single-agent exposures, in the MG-63 cell line, etoposide increased cyclin E1 RNA levels, while in Saos-2 cells, the flavonoid fisetin decreased cyclin B1 RNA levels (p < 0.05, Fig. 5). For double-agent exposures (combinations), significant differences were found only in MG-63 cells, viz., decrease in cyclin B1, Cdk1, and Cdk2 RNA levels. Compared to the chemotherapeutic agent etoposide, the combinations fisetin:etoposide significantly decreased cyclin E1, Cdk1, and Cdk2 gene expression in MG-63 cells. In Saos-2 cells, the average cyclin E1 gene expression was lower when cells were exposed to fisetin:etoposide combination compared to etoposide as single agent, although not reaching statistical significance (p < 0.05, Fig. 5b).

Fig. 5
figure 5

Gene expression of cell cycle regulators. Cells were exposed for 48 h to fisetin (single agent), etoposide (single agent), or combinations (double agent). Exposures in a MG-63 cells and b Saos-2 cells. The data are mean ± SD (n = 3). Statistical analysis was performed using the one-way ANOVA test (p < 0.05). Significant differences are shown relative to the control group (asterisks) or to other experimental groups (triangles)

Full size image

Discussion

Targeted osteosarcoma therapy is still hindered by the genetic instability and complexity of the disease (Kansara et al. 2014; Yang and Zhang 2013). Combined therapy regimens could, therefore, prove beneficial, particularly for poor responders.

Herein, the main goal was to test cellular effects of fisetin in cell lines with deficient or absent wt p53 expression. In the wt p53 U2OS cells, Li et al. (2015) reported IC50 for fisetin near 70 micromolar for 48 h exposure, using MTS assay, while Ying et al. (2012) reported IC50 for fisetin near 130 micromolar (48 h exposure, MTT assay). In this study, a biphasic behavior was observed for the inhibition of cell growth by fisetin. Previously, at low concentrations, fisetin was found to enhance preosteoblast proliferation (Léotoing et al. 2014), and this effect could have played a role on the enhanced osteosarcoma proliferation up to 10 μmol observed in this study, resulting in a biphasic effect. Moreover, for some cell lines described in the literature, exposure to low fisetin concentrations did not result in significant inhibition of proliferation or viability (e.g., Kim et al. 2014; Touil et al. 2011a, b; Ying et al. 2012). Despite the increase in cell proliferation upon exposure to low-dose fisetin observed in this work, Saos-2 cells suffered inhibition of colony formation for 10 and 20 μmol concentrations. This observation shows that, even at subtoxic doses, fisetin decreased colony formation, suggesting a cytostatic effect in cells recovered from low-dose fisetin exposure. Depending on cell type and exposure times, in the previous studies, micromolar doses of fisetin showed various effects on the cell cycle, ranging from increase in % cells in G0/G1-phase and reduction in % cells in S-phase, to emergence of sub-G1 population with decreased % G0/G1 population, and increased % cells in S-phase and G2/M-phase, e.g., (Adan and Baran 2015; Chen et al. 2002; Haddad et al. 2006; Khan et al. 2008; Li et al. 2011; Lu et al. 2005; Pal et al. 2013; Ying et al. 2012).

Etoposide is frequently administered in patients with poor response to therapy. Despite its anti-cancer role as topoisomerase II inhibitor, a large body of evidence indicates that it could lead to secondary malignancies (Azarova et al. 2007). In this study, fisetin was shown less cytotoxic to osteosarcoma cells, compared to etoposide. Combination index values were calculated for different mixtures to determine positive or negative interactions between fisetin and etoposide. The two mixtures with highest fisetin concentration and lowest etoposide concentration showed more positive interaction effects for the inhibition of cell growth, pointing to fisetin usefulness in osteosarcoma combined therapy.

DNA damage triggers a set of cellular responses, including cell cycle arrest and commitment to cell death. It is well known that cell cycle arrest can increase the sensitivity of cancer cells to etoposide (Walker and Nitiss 2002). Comparing only cells exposed to fisetin:etoposide mixtures, an association was found with increased accumulation of cells in G1-phase and S-phase and decreased % of cells in G2-phase. CCNE1 (cyclin E1 gene) and CDK2 (cyclin-dependent kinase 2 gene) govern G1/S-phase transition and S-phase were previously found amplified and overexpressed in osteosarcoma (Lockwood et al. 2011; Vella et al. 2016). Moreover, silencing of CCNE1 or CDK2 was previously shown to suppress osteosarcoma proliferation (Vella et al. 2016; Wang et al. 2014). In general, when considering only fisetin:etoposide combinations, an increase was observed in % cells in S-phase for the mixtures with highest fisetin concentration for both cell lines. Combinations with lower CI values, i.e., more positive interactions between fisetin and etoposide, were associated with increased % of cells in S-phase, in both cell lines. In MG-63 cells, exposure to 38.6 μM fisetin alone (Fig. 4a) or to mixture fisetin:etoposide—36 μM:0.5 μM (Fig. 4c) originated very similar cell cycle profiles. Notwithstanding, when considering gene expression, there is no significant difference between control and cells exposed to fisetin for the genes analyzed in the MG-63 cell line. For this reason, it cannot be ruled out that other regulators of the cell cycle might be involved in accumulation at the S-phase, even though a decrease was observed in cyclin E1, Cdk1, and Cdk2 gene expression in cells exposed to fisetin:etoposide combination, and such decrease would be expected to impact cell cycle progression. Notwithstanding, in other cancer cell lines reported in the literature, fisetin was shown to decrease the protein levels of cyclin E, cyclin A, cyclin D1, Cdk1, Cdk2, and Cdk4, while increasing p21 levels (Khan et al. 2008; Li et al. 2011; Lu et al. 2005; Sung et al. 2007; Syed et al. 2011). The cell cycle effects of etoposide rely mostly on its action as topoisomerase II inhibitor, and previously, fisetin was described as a moderate topoisomerase II inhibitor in vitro (Bandele and Osheroff 2007; Strick et al. 2000). Therefore, it would be relevant in the future to investigate whether it shows similar effects on osteosarcoma cell lines. Concerning the possibility of induction of secondary malignancies by fisetin, at physiological concentrations, this flavonol induces the differentiation of pre-osteoblasts to mature osteoblasts, a feature opposite of that found in bone malignancy (Sung et al. 2007). More importantly, in a pharmacokinetics study in mice, fisetin was found to be accumulated at ~ 84 μg/(g tissue) in kidneys, corresponding to ~ 24 mmol/g, without negative effects being reported (Touil et al. 2011a). Moreover, in another work from this team, this fisetin dose was continuously administered in mice for 2 weeks and resulted in decreased Lewis lung tumor development, while showing no toxic effects (Touil et al. 2011b). These observations suggest that fisetin can inhibit tumorigenesis and would not induce secondary malignancies at the concentrations tested in this study.

In conclusion, fisetin combined with etoposide induced cell cycle arrest, which was found associated with decreased expression of cell cycle regulators commonly overexpressed in many osteosarcoma samples. For this reason, the use of fisetin combined with chemotherapeutic drugs warrants further studies.