Since response to platinum-based therapy in non–small-cell lung cancer (NSCLC) is poor, the present study was designed to rationally identify novel drug combinations in cell models including the A549 cell line and the cisplatin-resistant subline A549/Pt, characterized by reduced sensitivity to cisplatin-induced apoptosis and by upregulation of efflux transporters of the ATP binding cassette (ABC) superfamily. Given the molecular features of these cells, we focused on compounds triggering apoptosis through different mechanisms, such as the mitochondria-targeting drug arsenic trioxide and the phenanthridine analog sanguinarine, which induce apoptosis through the extrinsic pathway. Sanguinarine, not recognized by ABC transporters, could overcome cisplatin resistance and, when used in combination with arsenic trioxide, was synergistic in A549 and A549/Pt cells. The arsenic trioxide/sanguinarine cotreatment upregulated genes implicated in apoptosis activation through the extrinsic pathway. Drug combination experiments indicated that tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) treatment improved arsenic trioxide/sanguinarine efficacy, a feature associated with a striking apoptosis induction, particularly in the cisplatin-resistant variant. Thus, a synergistic interaction between sanguinarine and arsenic trioxide could be obtained independent of relative cell sensitivity to arsenic trioxide, and an enhanced apoptosis induction could be achieved in combination with TRAIL through modulation of the extrinsic apoptotic pathway. Antitumor activity studies supported the interest of drug combinations including TRAIL in NSCLC, indicating that drug-resistant NSCLC cells can efficiently be killed by the combination of proapoptotic agents. Our results suggest that the molecular changes occurring in treated cells may be exploited to rationally hit surviving cells.
Lung cancer is the leading cause of cancer-related death throughout the world, and non–small-cell lung cancer (NSCLC) account for 75% of lung cancer cases (Jemal et al., 2011). Although recent advances in the management of NSCLC have improved its prognosis, more than half of patients with NSCLC exhibit advanced-stage disease at diagnosis, and no curative therapy currently exists for such patients (Pastorino, 2010). Platinum-based combination therapies in advanced NSCLC patients with good performance status have shown a significant improvement in overall survival and quality of life (Seve and Dumontet, 2005). However, limited response to platinum-based therapy is frequent; thus, the identification of molecular changes occurring in tumor cells after treatment could be helpful in improving the understanding of the mechanisms of drug resistance, in an attempt to define personalized chemotherapy strategies. In addition, difficulty of treatment with single or combination treatments with cisplatin has been associated with high toxicity and limited impact on survival. Thus, research toward novel drugs with improved efficacy against NSCLC is mandatory. In this regard, arsenic trioxide (As2O3; ATO), used as mitochondria-targeting drug in patients with acute promyelocytic leukemia (Zhang et al., 2001), is active in vitro in several solid tumor cell lines (Miller et al., 2002; Smith et al., 2010). ATO has multiple mechanisms of action, and it exhibits different effects on differentiation and apoptosis, depending on the concentration (Chen et al., 1996). At low, clinically achievable doses (1–2 μM), ATO displays potent activity against acute promyelocytic leukemia, but little toxicity (Niu et al., 1999). Preclinical studies have demonstrated that ATO can induce apoptosis and inhibit tumor cell growth in a wide variety of solid tumors (Miller et al., 2002), including A549 human NSCLC cells, with a poorly understood mechanism (Jin et al., 2006; Walker et al., 2010). Because of the many pathways relevant to the effects of ATO, the potential exists for synergistic interactions providing enhanced therapeutic benefits. Moreover, combination therapy may allow the administration of lower doses of ATO, minimizing toxicity and potential drug antagonism.
Sanguinarine [13-methyl[1,3]benzodioxolo[5,6-c]-1,3-dioxolo[4,5-i]phenanthridinium], a benzophenanthridine alkaloid found in Papaveraceae, such as the roots of Sanguinaria canadensis and the seeds of the Argemone mexicana (Tandon et al., 1975; Mahady and Beecher, 1994), endowed with antibacterial and anti-inflammatory properties (Lenfeld et al., 1981; Godowski 1989), is a promising compound with proapoptotic effects in human NSCLC A549 cells (Jang et al., 2009). In fact, sanguinarine induces apoptosis in a variety of cancer cells after cell cycle arrest (Adhami et al., 2004), caspase activation (Kim et al., 2008), depletion of cellular GSH (Debiton et al., 2003), modulation of Bcl-2 family members (Ahsan et al., 2007; Lee et al., 2012; Xu, et al., 2012), and upregulation of death receptor 5 [DR5; tumor necrosis factor–related apoptosis-inducing ligand R2 (TRAIL-R2)] (Hussain et al., 2007). Of note, it has been observed that sanguinarine sensitizes human gastric adenocarcinoma cells to TRAIL-mediated apoptosis (Choi et al., 2009) and that it induces apoptosis of human osteosarcoma cells through the extrinsic pathway (Park et al., 2010). Because of its preferential ability to induce apoptosis in cancer cells over normal cells (Ichikawa et al., 2001), TRAIL can be regarded as a promising targeted drug (Takeda et al., 2007) which induces apoptosis in a variety of cancer cells by interacting with its death receptors DR4 (TRAIL-R1) and DR5 (TRAIL-R2) on the target cell surface (Johnstone et al., 2008). It has been reported that TRAIL induces apoptosis in NSCLC cell lines and inhibits the growth of NSCLC xenografts (Bhojani et al., 2003; Hao et al., 2004).
Based on this background, the aim of the present study was to identify novel treatments effective in NSCLC cells exploiting cell response to design effective combinations. Specifically, two proapoptotic agents endowed with different mechanisms of action, i.e., ATO and sanguinarine, were combined in NSCLC cells, including the parental A549 cell line and the cisplatin-resistant variant A549/Pt. Because the analysis of molecular changes occurring in tumor cells after ATO/sanguinarine cotreatment evidenced an involvement of the extrinsic apoptotic pathway, TRAIL was used in an attempt to improve the synergistic interaction. When the efficacy of TRAIL cotreatment was assessed, an enhanced apoptotic response was observed in drug-resistant cells. TRAIL-based combinations were also effective when tested in xenografts, thereby supporting that drug-resistant NSCLC cells can efficiently be killed by the combination of well tolerated proapoptotic agents.
Materials and Methods
This research protocol was approved by the Ethics Committee for Animal Experimentation of the Fondazione IRCCS Istituto Nazionale Tumori of Milan, in accordance with institutional guidelines.
Cell Culture and Drugs.
The NSCLC A549 and H460 cell lines, the ovarian carcinoma IGROV-1 cell line, and the cisplatin-resistant sublines A549/Pt, H460/Pt, and IGROV-1/Pt1 were maintained in RPMI 1640 medium (BioWhittaker Lonza, Italy) supplemented with 10% fetal bovine serum (Life Technologies, Monza, Italy). The A549/Pt, H460/Pt, and IGROV-1/Pt1 variants were generated as previously described (Perego et al., 1998; Bertolini et al., 2009). Resistance was stable for at least 6 months when cells were grown in the absence of selecting agent. The human osteosarcoma U2-OS cell line was grown in McCoy’s 5A medium (BioWhittaker Lonza) and supplemented with 10% fetal bovine serum (Life Technologies). U2-OS transfected cells (Beretta et al., 2010) were cultured in the presence of G418 (400 μg/ml; Life Technologies). The human colon carcinoma HT29 cell line and the mitoxantrone-resistant HT29/MIT subline (Perego et al., 2001) were maintained in RPMI 1640 medium (BioWhittaker Lonza) and supplemented with 10% fetal bovine serum (Life Technologies). The human colon carcinoma LoVo cell line and the doxorubicin-resistant LoVo/DX subline (Grandi et al., 1986) were maintained in F12 medium (BioWhittaker Lonza) supplemented with 10% fetal bovine serum (Life Technologies). Cell cultures were routinely checked to ensure they were mycoplasma-free, and experiments were carried out using cell lines at similar passages following thawing from a frozen stock. All experiments were carried out in medium plus 10% fetal bovine serum (Life Technologies), with the exception of the transfections (described later). Cisplatin (Teva Italia S.r.l., Milan, Italy) was diluted in saline. Sanguinarine (provided by Indena, Milan, Italy) was dissolved and diluted in water. ATO (Sigma-Aldrich GmbH, Munich, Germany) was dissolved in 1 N NaOH and diluted in saline. Soluble TRAIL was obtained from Vinci Biochem S.r.l. (Firenze, Italy) and diluted in saline. Topotecan (Hycamtin; GlaxoSmithKline S.p.A., Verona, Italy) was dissolved and diluted in water. Paclitaxel (Indena, Milan, Italy) was dissolved in dimethylsulfoxide and diluted in water. Gemcitabine (SUN Pharmaceuticals Italia, Milan, Italy) was dissolved and diluted in saline.
Cell Growth Inhibition Assays and Drug Interaction Analyses.
Cell sensitivity was assessed by growth-inhibition assays (Perego et al., 1998). Exponentially growing cells were seeded in duplicate in six-well plates at 19,000–25,500 cells/cm2. After 24 hours, cells were exposed to different concentrations of drugs alone or in combination for 72 hours, and cells were counted at the end of treatment. Before counting, culture medium was removed and adherent cells were harvested using trypsin and counted with a cell counter (Beckman Coulter, S.p.A., Milan, Italy). IC50 is defined as the concentration causing a 50% inhibition of cell growth compared with control. The resistance index is the ratio between the IC50 of resistant and sensitive cells. Drug interaction was evaluated using a standard approach that allows a value to be assigned to a drug combination [synergism index (SI)] that indicates synergistic drug interaction. The SI is equivalent to the ratio between expected cell growth and observed cell growth (synergism: S I >1) (Zanchi et al., 2005).
Flow Cytometry Analysis of Caspase-3 Cleavage and TRAIL Receptor Expression.
For caspase-3 cleavage fluorescence-activated cell sorter (FACS) analyses, both floating and adherent cells were harvested 72 hours after 1 hour of treatment with 100 μM cisplatin, washed in phosphate-buffered saline (PBS), and fixed by exposing, for 15 minutes at room temperature, to 3% paraformaldehyde (Sigma-Aldrich) dissolved in PBS. After washing with PBS, cells were resuspended and permeabilized in ice-cold methanol (−20°C) for 1 hour. After washing with PBS, cells were resuspended and incubated for 45 minutes in a PBS–bovine serum albumin (PBS-BSA) solution (PBS containing 1% BSA and 0.1% Tween 20). Cells were then resuspended and incubated for 1 hour in 100 μl of PBS-BSA solution containing 1:50 diluted anticleaved caspase-3 antibody. Cells were then washed with PBS-Tween 20, and then resuspended and incubated for 30 minutes in the dark in 100 μl of PBS-BSA solution containing 1:500 diluted AlexaFluor 594 (or 488)–conjugated goat anti-rabbit antibody (Molecular Probes, Life Technologies). Samples were then washed and resuspended in 1 ml of PBS for final reading with a FACScan flow cytometer (Becton Dickinson, Milan, Italy).
The expression of TRAIL receptors was measured by flow cytometry as described previously (Perego et al., 2006). Cells were seeded in 24-well plates, and 24 hours later, they were exposed to the drug for 24 hours. Cells were then harvested and incubated for 30 minutes at 4°C with 1 μg of biotinylated antihuman TRAIL-R1 or -R2 antibody (R&D Systems, Milan, Italy), washed twice with PBS, and incubated in PBS containing 2 μl of streptavidin phycoerythrin or fluorescein isothiocyanate (Becton Dickinson). Cells were then washed, and samples were immediately used for flow cytometric analysis (FACScan; Becton Dickinson). Data were analyzed using Cell Quest (Becton-Dickinson).
Alkaline Comet Assay.
The alkaline comet assay (Trevigen Inc., Bologna, Italy) was performed, according to the manufacturer’s instructions, 24 hours after seeding in control cells and in cells exposed to sanguinarine (3 μM, for 3 or 24 hours) or to ATO (80 μM, 3 hours). Topotecan was used as a positive control. In brief, the cells were suspended in low melting agarose and layered onto slides. Cells were then lysed to remove cellular proteins and to release the DNA. Electrophoresis was carried out under alkaline conditions, which allow detection of single-strand breaks. After neutralization, DNA was stained with SYBER Green. The slides were scored using a fluorescence microscope equipped with a video camera, and quantitative assessment of DNA damage was obtained using the Comet Assay IV software (Perceptive Instruments, Suffolk, UK). One hundred cells were monitored for each sample in at least two independent experiments. The tail moment is defined as the product of the percentage of DNA in the comet tail and the tail length expressed in micrometers.
RNA Extraction and Expression Analysis with TaqMan Arrays.
Harvesting of cells, RNA extraction, and DNase digestion were carried out with the RNAqueous-4PCR Kit (Ambion Europe Ltd., Huntingdon, UK), according to the manufacturer’s instructions, as previously described (Bertolini et al., 2009). The TaqMan Arrays (Human ABC Transporter Panel, Human Apoptosis Panel; Life Technologies) were placed in the Micro Fluidic Card Sample Block of an ABI Prism 7900 HT Sequence Detection System (Life Technologies). The calculation of the threshold cycle values was performed using the SDS 2.2 software (Life Technologies), after automatically setting the baseline and the threshold.
Exponentially growing cells were seeded in 25-cm2 flasks and, 24 hours later, were exposed to different concentrations of sanguinarine, ATO, TRAIL, or a combination of the drugs for 24, 48, or 72 hours. At the end of treatment, floating and adherent cells were harvested for detection of apoptotic cells. Apoptosis was evaluated by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay (Roche, Mannheim, Germany). After harvesting, cells were fixed in paraformaldehyde, permeabilized in a solution of 0.1% Triton X-100 in 0.1% sodium citrate and then incubated in the TUNEL reaction for 1 hour. After washing, samples were analyzed by flow cytometry using Cell Quest software using markers to quantify TUNEL-positive (FL1 > 101) cells (Becton-Dickinson).
Caspase-3 and caspase-8 proteolytic processing was analyzed by Western immunoblotting as described previously (Perego et al., 1999). In brief, samples were fractionated by SDS-PAGE and blotted on nitrocellulose membranes. Blots were preblocked in PBS containing 5% (w/v) dried nonfat milk and then incubated overnight at 4°C with antibodies to cleaved caspase-3 (Asp175; Cell Signaling Technology, Danvers, MA) and cleaved caspase-8 (Asp391-18C8; Cell Signaling Technology). An antiactin antibody (Sigma-Aldrich) was used as control for loading. Antibody binding to blots was detected by chemiluminescence (GE Healthcare, Milan, Italy). Three independent experiments were performed. Band intensities were quantified by NIH ImageJ software.
The catalytic activity of caspase-3 and caspase-8 was measured as the ability to cleave the specific substrates N-acetyl-Asp-Glu-Val-Asp-pNA (DEVD-pNA) and N-acetyl-Ile-Glu-Thr-Asp-AMC (IETD-AMC) by means of the APOPCYTO/caspase-3 and APOPCYTO/caspase-8 kits (MBL International, Eppendorf srl, Milan, Italy), respectively, according to the manufacturer’s instructions. The hydrolysis of the specific substrates was monitored by spectrofluorometry with 380-nm excitation and 460-nm emission filters. Results were expressed as relative fluorescence units normalized to untreated cells.
A pool of synthetic small-interfering RNAs (siRNAs) (Silencer Select Pre-designed and Validated siRNA; Ambion, Life Technologies) was used to knock down TRAIL-R2 gene expression. Preliminary experiments were performed to define optimal transfection conditions, and transfection efficiency was monitored using a fluorescent double-strand RNA (dsRNA) oligomer (BLOCK-iT; Life Technologies). The efficiency of downregulation of target expression was assayed by real-time polymerase chain reaction (TaqMan assay; Life Technologies) and by flow cytometry as described earlier. Exponentially growing cells were seeded in six-well plates, and when subconfluent, they were transfected with 100 nM TRAIL-R2 siRNA or control siRNA (Silencer Select Negative Control siRNA; Ambion, Life Technologies) using Lipofectamine RNAiMax (7.5 μl/well; Life Technologies). For each well, 100 μl of vehicle-siRNA mix prepared in Opti-MEM I (Life Technologies) was added to 0.5 ml of Opti-MEM I for 18 hours, and then the medium was replaced with fresh medium. Twenty-four hours later, cells were harvested and seeded for cell sensitivity assays. Parallel cultures were used to assess the efficiency of knock down of the target gene by quantitative real-time and FACS analysis, as described earlier. Cell sensitivity was determined as described earlier.
In Vivo Experiments.
In vivo experiments were performed on 6–8-week-old, female CD-1 nude mice (Charles River, Calco, Italy). Mice were maintained under specific pathogen-free conditions, keeping temperature and humidity constant. Mice were given sterile food and water ad libitum. The research protocol was approved by the Ethics Committee for Animal Experimentation of the Fondazione IRCCS Istituto Nazionale Tumori of Milan, according to institutional guidelines. At day 0, 5 × 106 exponentially growing cells were subcutaneously injected into the right flank of mice. Each group included —six to eight animals. Tumor growth was followed by biweekly measurements of tumor diameters with a Vernier caliper, and tumor volume (TV) was calculated according to the following formula: TV (mm3) = d2 × D × 0.5, where d and D are the width and the length, respectively. ATO was dissolved in 1 M NaOH (66 mg/ml), diluted in saline, and administered by intraperitoneal injection (0.5 mg/kg) for 5 consecutive days/week, with treatment repeated for 3 weeks. Sanguinarine was administered by oral gavage (20 mg/kg) for 5 consecutive days/week, with treatment repeated for 3 weeks. Because this model is known to be particularly resistant, treatments with ATO and sanguinarine were initiated 4 days after cell injection. TRAIL was diluted in saline and subcutaneously administered (30 mg/kg) 4 times every 4 days, with treatment starting 7 days after cell injection. Efficacy of drugs and combinations was assessed as the mean percentage of tumor volume inhibition (TVI) in drug-treated versus control mice as follows: TVI% = 100 – [mean tumor volume of treated mice/mean tumor volume of control mice × 100]. Mice were killed by cervical dislocation 32 days after cell injection.
Statistical analyses were carried out using GraphPad Prism (version 5.02; GraphPad Software Inc., La Jolla, CA) as detailed throughout the article.
Phenotype of A549 and A549/Pt Cells.
Sensitivity of the A549 NSCLC cell line and the cisplatin-resistant variant A549/Pt to drugs of different classes, with particular reference to those used in the clinical setting, was examined using growth-inhibition assays after 72-hour drug exposure (Table 1). The A549/Pt variant, which displayed an ∼8-fold degree of resistance to cisplatin (Fig. 1A), was characterized by reduced sensitivity to topotecan, gemcitabine, and paclitaxel (P < 0.05, unpaired t test of IC50 values of parental vs. resistant cells; Table 1). A549/Pt cells exhibited reduced susceptibility to drug-induced apoptosis as evidenced by flow cytometry analyses of caspase-3 cleavage, in comparison with A549 cells, when exposed to equimolar cisplatin concentrations (Table 2). Cross-resistance to the proapoptotic agent ATO was also found (Fig. 1B; Table 1). The A549/Pt variant exhibited a multidrug-resistant phenotype that was related to increased levels of ABC transporters (Fig. 2). Thus, in our search for non–cross-resistant drugs and novel drug combinations, we selected sanguinarine, which is not a substrate for well known ABC transporters (Supplemental Fig. 1). Cell sensitivity assays after 72-hour exposure to sanguinarine indicated that the agent was capable of almost completely overcoming cisplatin resistance in NSCLC cells because it inhibited proliferation of drug-resistant cells with an IC50 only slightly different (i.e., P < 0.05, unpaired t test of IC50 values of parental vs. resistant cells, but resistance index of 1.5) from that observed in parental cells (Fig. 1C; Table 1). When additional cisplatin-resistant cell lines were examined for sensitivity to sanguinarine (72-hour exposure), no cross-resistance was observed. Indeed, the IC50 value of the platinum-resistant NSCLC cell variant H460/Pt (1.43 ± 0.2 μM) was similar to that of the H460 cell line (1.42 ± 0.1 μM). Similar IC50 values were also obtained in the ovarian carcinoma IGROV-1 cell line (1.39 ± 0.4 μM) and in the platinum-resistant IGROV-1/Pt1 variant (1.17 ± 0.1 μM).
DNA Damage Induced by Sanguinarine or ATO on A549 and A549/Pt Cells.
To investigate the specific cellular effects of sanguinarine on A549 and A549/Pt cells, we analyzed the possible induction of DNA damage using the alkaline comet assay in cells exposed to 3 μM drug concentrations (Table 3). The DNA topoisomerase I inhibitor topotecan was used as a positive control. We found that sanguinarine was capable of inducing DNA breaks in parental and resistant cells, as evidenced by the mean tail moment values. Specifically, DNA damage in A549 cells appeared more pronounced 3 hours after drug exposure as compared with 24 hours, whereas the 24-hour exposure resulted in particularly effective DNA damage induction in the cisplatin-resistant cell line. Moreover, under conditions similar to those used for sanguinarine (i.e., short-term exposure to an equimolar concentration around IC80 of resistant cells, that is 80 μM), ATO could induce DNA damage in A549 cells. A reduced extent of damage was observed in the A549/Pt variant (Supplemental Fig. 2).
Drug Combination Studies in A549 and A549/Pt Cells.
Because sanguinarine appeared to overcome cisplatin resistance of NSCLC cells, we assessed sensitivity of A549 and A549/Pt cells to a simultaneous combination treatment with sanguinarine and different concentrations of cisplatin, topotecan, and ATO by growth-inhibition assay following 72-hour drug exposure. Although the combined treatments of sanguinarine with cisplatin or topotecan did not prove a marked synergistic effect (Supplemental Tables 1 and 2), when we evaluated the effect of the simultaneous treatment between sanguinarine and the apoptosis inducer ATO, we observed a synergistic interaction which was particularly evident in A549/Pt cells, as indicated by the SI values (Fig. 3).
Analysis of the Expression of Genes Belonging to the Apoptotic Pathway.
To define the determinants of cellular response of cisplatin-sensitive and -resistant NSCLC cells to the synergistic ATO/sanguinarine treatment, with particular reference to the induction of apoptosis-related genes, we analyzed the mRNA levels of genes belonging to the apoptotic pathway by using TaqMan Arrays. Gene-expression analysis was performed on cells exposed to equitoxic drug concentrations of sanguinarine (1 and 2 μM for A549 and A549/Pt, respectively), ATO (10 and 30 μM for A549 and A549/Pt, respectively), or a combination of the two for 24 hours. In A549 and A549/Pt cells, both single and combined treatment induced the expression of different proapoptotic factors such as caspase-1, which displayed the greatest increase in A549 cells (Supplemental Table 3). Overall, a prevalence of induction of genes involved in the extrinsic apoptotic pathway was observed (Table 4).
Analysis of the Expression of TRAIL Receptors.
Since the pattern of gene expression modulation suggested an involvement of the extrinsic apoptotic pathway in the synergistic interaction between sanguinarine and ATO, we evaluated the expression levels of both proapoptotic TRAIL receptors in A549 and A549/Pt cells following exposure to sanguinarine and ATO alone or in combination (24-hour drug exposure; Table 5). After 24-hour single-agent exposure, no substantial modulation of TRAIL-R1 (DR4) was observed in both cell lines (Table 5). Under the same conditions, the combined treatment with sanguinarine and ATO produced an increase in TRAIL-R2 (DR5) levels in the cell lines, whereas the induction of TRAIL-R2 was documented in A549/Pt cells also after exposure to ATO alone (Table 5).
Effects of Combined Treatment with ATO, Sanguinarine, and TRAIL.
Based on the previously mentioned results (Tables 4 and 5), we examined the effect of a simultaneous combined treatment including sanguinarine, ATO, and exogenous TRAIL. When A549 and A549/Pt cells were exposed for 72 hours to different concentrations of ATO and subtoxic concentrations of sanguinarine and TRAIL, we observed a strong increase in the synergism described between ATO and sanguinarine, as documented by the SI values (Fig. 4). Interestingly, the increase in the synergistic interaction observed when adding TRAIL was particularly evident in cisplatin-resistant A549/Pt cells.
To determine whether the drug interaction between sanguinarine, ATO, and TRAIL resulted in an enhancement of apoptotic cell death, we performed TUNEL assays 48 and 72 hours after drug exposure. In A549 cells, the 48-hour combined treatment of 1 μM sanguinarine, 10 μM ATO, and 10 ng/ml TRAIL did not produce any evident apoptosis increase with respect to the sum of the effects of single-agent treatment (Table 6), whereas in the Pt-resistant cells, the increase following the 48-hour combined exposure (2 μM sanguinarine, 30 μM ATO, and 10 ng/ml TRAIL) was remarkably higher (+26.86% ± 15%; Table 6). The analysis of apoptosis performed after 72-hour drug exposure indicated that the combined treatment with sanguinarine, ATO, and TRAIL was effective in apoptosis induction in A549 cells and noticeably in A549/Pt cells (+44.1% and +88.8%, respectively; data not shown).
We then assessed whether the increased apoptotic response induced by combined treatments was sustained by caspase activation in both cell lines. Immunoblotting results indicated that the 48-hour simultaneous combined treatment with equitoxic concentrations of ATO, sanguinarine, and TRAIL determined an evident increase of the cleaved forms of caspase-3 in both cell lines (Fig. 5A). However, the extent of release of the active fragments of caspase-3 was remarkably higher in cisplatin-resistant than in parental cells with the ATO/sanguinarine combined treatment and with the triple combination which included TRAIL. Likewise, a 48-hour combined treatment with the three agents was effective in releasing caspase-8 cleavage products (43-, 41-, and 18-kDa fragments) in both cell lines (Fig. 5B). A noticeable cleavage was already observed in parental cells exposed to ATO alone, but not for the other compounds that were used at subtoxic concentrations.
Consistent with the proteolytic processing of caspase-3 and caspase-8, an increased catalytic activity of both enzymes was observed by the in vitro hydrolysis of the fluorogenic substrates (Fig. 6). A similar and negligible catalytic activity of caspase-3 and caspase-8 (<0.1 relative fluorescence units) was measured in untreated A549 and A549/Pt cells (data not shown), whereas a marginal increase in the activation of both enzymes was found in cells after 48-hour exposure to ATO, sanguinarine, or TRAIL, respectively (Fig. 6). By contrast, a 1.8-fold to 3-fold higher activation of both caspases was observed after exposure to combined treatments ATO/sanguinarine, ATO/TRAIL, and sanguinarine/TRAIL. When ATO, sanguinarine, and TRAIL were added to A549 and A549/Pt cells in combination, a significant increase in the activation of both caspases was detected (Fig. 6). Such an effect was significantly higher in A549/Pt cells with respect to parental cells. Of note, 72 hours after the different combined treatments, it was not possible to determine caspase activity because the high levels of dead cells made the measurement unreliable.
Additionally, we examined the modulation of mRNA levels of caspases-3 and -8 under all treatment conditions, and we confirmed the lack of significant up-modulation of the studied caspases observed using TaqMan microfluidic cards after treatment with ATO, sanguinarine, and their combination. Interestingly, caspase-3 mRNA levels displayed a 2-fold increase only in A549/Pt cells after treatment with ATO in combination with TRAIL and sanguinarine (Supplemental Fig. 3).
Analysis of the Role of TRAIL-R2 in the Synergistic Interaction between ATO, Sanguinarine, and TRAIL in A549/Pt Cells.
To establish the relevance of the extrinsic apoptotic pathway in the observed synergistic interaction between ATO, sanguinarine (S), and TRAIL, using a suitable siRNA, we knocked down TRAIL-R2, whose expression was modulated by ATO/S treatment (Fig. 7). Since the synergistic interaction was particularly evident in A549/Pt2 cells, we evaluated the effect of TRAIL-R2 knockdown prior to and after the combined treatment with ATO, S, and TRAIL in resistant cells. The efficiency of siRNA transfection was confirmed through a fluorescein-labeled dsRNA oligomer (Fig. 7A), whereas the efficiency of knockdown of the target gene was confirmed by quantitative real-time (Fig. 7B) and FACS analysis (Fig. 7C). When TRAIL-R2–knockdown cells were exposed to the simultaneous ATO/S combined treatment or to the triple combination including TRAIL, we found that the increase in the synergistic interaction triggered by TRAIL addition could not be achieved differently from cells transfected with negative control siRNA. This evidence was documented by the relative SI values (Fig. 7D).
Antitumor Activity of the Drug Combinations.
The effect of the combination of ATO, sanguinarine, and TRAIL on the in vivo growth of A549/Pt cells xenografted in nude mice was examined (Fig. 8). In this model, single treatments with ATO, sanguinarine, or TRAIL could induce a slight inhibition of tumor growth, the TVI being 44%, 28%, and 36%, respectively. The effects of the combinations of ATO plus sanguinarine (TVI = 38%) or sanguinarine plus TRAIL were not superior to single agents, with an inhibition of 38% and 18%, respectively. An increased inhibition of tumor growth (TVI = 64%) was obtained with the combined treatment of ATO and TRAIL. Similarly, an improved efficacy was observed in mice treated with the three agents in combination, in which the TVI was 65%. No treatment-related toxicities were recorded (Supplemental Fig. 4).
In the present study, we used the A549-A549/Pt model pair, in which acquired resistance to cisplatin was associated with reduced susceptibility to drug-induced apoptosis, for investigating novel approaches to modulating drug response through interference with apoptosis. It was our goal to identify rational strategies for drug combinations designed on the basis of mechanistic assumptions of drug action and on cell response to treatment. Drug-resistant cells exhibited a markedly reduced sensitivity to the antiproliferative effect of a variety of conventional antitumor agents used in the treatment of NSCLC. The observed multidrug-resistant phenotype was associated with increased expression of selected transporters of the ABC superfamily, including P-glycoprotein (P-gp). Increased P-gp expression has been reported to account for reduced sensitivity to paclitaxel and topotecan (Gatti et al., 2009). The role of P-gp and other ABC transporters in gemcitabine efflux and resistance is more controversial, due to the capability of drug transporters to efflux the deaminated metabolite (Rudin et al., 2011), although our results suggest a possible role. An interesting finding of our study was the activity of sanguinarine in colon carcinoma cells displaying overexpression of P-gp/ABCB1 or breast cancer resistance protein (BCRP)/ABCG2 and in cisplatin-resistant A549/Pt cells with increased mRNA levels of ABCB1 and ABCG2. Of note, sanguinarine was also effective in additional platinum-resistant cell lines, the NSCLC H460/Pt variant, and the ovarian carcinoma IGROV-1/Pt1. Reduced sensitivity of A549/Pt cells to ATO was not unexpected because it has been previously shown that ATO is effluxed by ABC transporters such as P-gp, multidrug resistance-associated proteins 1 (MRP1), and 3 (MRP3) typically overexpressed by drug-resistant cells (Smith et al., 2010; Zhao et al., 2010; Ong et al., 2012). Thus, ATO could not overcome cisplatin resistance, but when combined with sanguinarine, produced a synergistic effect in both sensitive and drug-resistant cells. Indeed, our study shows that a synergistic interaction can be obtained when combining two agents endowed with a proapoptotic effect (Jin et al., 2006; Jang et al., 2009) in different tumor cell lines including NSCLC cells. The proapoptotic effect of sanguinarine could be ascribed to its capability to damage DNA as shown by comet assay performed after short-term exposure to the compound. Of note, ATO was also endowed with DNA damaging activity as expected based on its capability to trigger reactive oxygen species generation.
Our in vitro data also demonstrate that it is possible to effectively exploit response of tumor cells to drugs to tailor further treatments. In fact, consistent with such a strategy, we tested the possible improvement of cell sensitivity to ATO/sanguinarine cotreatment as a consequence of TRAIL-R2 upregulation, and found that a striking synergism could be achieved with the triple combination including TRAIL. Indeed, a marked cell death could be triggered by the addition of soluble TRAIL, which was per se completely inactive, in NSCLC cells. Interestingly, cells resistant to cisplatin were particularly sensitive to the effect of the triple combination. Under our experimental conditions, although a remarkable antiproliferative effect could be achieved when combining ATO with sanguinarine, the optimal effect was obtained with the concomitant treatment including TRAIL. Our findings are consistent with the reports showing that sanguinarine sensitizes human tumor cells to TRAIL-mediated apoptosis (Choi et al., 2009) and induces apoptosis of human tumor cells through the activation of the extrinsic pathway (Park et al., 2010).
The marked synergistic interaction further increased by TRAIL was related to an efficient activation of caspase-3 and caspase-8 in both cell lines. The activation of cleavage of caspase-8 by ATO itself in parental cells may be related to their increased propensity to undergo apoptosis as compared with resistant cells in which the most striking apoptotic response was found after exposure to the triple combination. Of note, in resistant cells, such an efficient activation of caspases resulted in improved apoptotic cell death, which massively occurred already at 48 hours in cells exposed to the triple combination. The increased activation of caspase-3 and caspase-8 could only in part be predicted by the analysis of the pattern of mRNA expression in which caspase-3 mRNA displayed a 2-fold increase exclusively after the triple combination in resistant cells, suggesting the multiple layers of regulation acting between RNA and protein levels/activities. In keeping with such a concept, although a marked upregulation of the mRNA levels of caspase-1 was observed in A549 cells after the combined treatment ATO/S, this finding was not validated at the protein level (data not shown).
To assess the impact of the extrinsic pathway on the described synergistic interaction, we conducted functional studies. Since knock down of TRAIL-R2 in resistant cells resulted in the attenuation of the increase in synergism obtained upon TRAIL addition evident in untransfected and in negative control transfected cells, it is possible to state that the upregulation of TRAIL-R2 triggered by the ATO/S combination is needed for an efficient effect of the TRAIL-based triple combination. Taken together, the obtained cellular results demonstrate that it is possible to prime platinum drug-resistant cells to undergo apoptotic cell death using TRAIL by modulating the expression of genes belonging to the extrinsic pathway. Of note, sensitization to TRAIL-induced apoptosis has been reported as a result of exposure to other conventional cytotoxic agents, including cisplatin (Baritaki et al., 2007).
With regard to how the proposed mechanism of action of each drug may relate to TRAIL-mediated apoptosis, our data indicate that sanguinarine and ATO may act by increasing the mRNA of TRAIL, an effect that was observed in parental cells after exposure to sanguinarine and ATO and to their combination, and that has been reported to occur in myeloma cells (Wu et al., 2010). Interestingly, in these cells, ATO was found to activate both the intrinsic and extrinsic apoptosis pathways as shown by activation of caspase-9 and caspase-8, respectively (Wu et al., 2010). However, the proapoptotic effect of ATO may, under other circumstances, rely on the intrinsic pathway because cell death can occur via an intrinsic Bcl-2-regulated route in acute promyelocitic leukemia cells (Scholz et al., 2005). In the present article, the observation of caspase-8 activation by ATO supports an involvement of the extrinsic pathway. Moreover, it was found that stimulation of the intrinsic apoptotic pathway by ATO, but not by sanguinarine, was supported by induction of a loss of the mitochondrial membrane potential and by p53 upregulation in A549/Pt cells (data not shown). Of note, the knock down of TRAIL-R2 indicated that the mechanism of synergism is to preferentially trigger the extrinsic apoptotic pathway.
Based on the promising in vitro data, the effect of the drug combination was tested in vivo using the A549/Pt cells xenografted in nude mice, after optimization of single drug treatments. An improved antitumor activity was observed with two different TRAIL-based combinations, i.e., ATO/TRAIL and ATO/sanguinarine/TRAIL. An interesting but not obvious observation of in vivo studies was the evidence of a modest activity of soluble TRAIL in A549/Pt tumors. Since TRAIL has been shown to engage survival pathways in TRAIL-resistant cancer (Fulda, 2013), such evidence suggests that TRAIL does not trigger such pathways in our drug-resistant model. Both ATO and sanguinarine, when administered as a single agent, produced a marginal tumor growth inhibition which could not be increased by their combination.
When TRAIL was combined with other agents, a marked in vivo antitumor efficacy was already observed for the combination of ATO and TRAIL. Sanguinarine, endowed with a slight antitumor activity per se, could not improve the effect of the combination of ATO with TRAIL. These results support the relevance of in vivo preclinical testing to optimize novel drug combinations, and suggest that pharmacokinetics issues may be crucial in achievement of stronger inhibition of tumor growth. In fact, TRAIL is characterized by a short half-life. The possible activation of counterproductive effects by TRAIL per se further strengthens the validity of our combination approach that may thereby prevent the survival of cells resistant to TRAIL treatment. The promising feature of TRAIL-based treatments has recently emerged in other tumor types (Piggott et al., 2011), a phenomenon that further strengthens that cell response to treatment can be exploited to kill drug-resistant cells.
Taken together, our results support that, despite the acquisition of reduced susceptibility to apoptosis by the selecting agent (i.e., cisplatin), NSCLC cells can efficiently be killed by the combination of proapoptotic agents. A peculiar finding of the present study was the particularly favorable drug interaction observed in platinum-resistant cells despite resistance to ATO.
The authors thank Carmelo Carlostella for kindly providing TRAIL for in vivo studies.
Participated in research design: Gatti, Cossa, Tinelli, Carenini, Arrighetti, Pennati, Cominetti, De Cesare, Zunino, Zaffaroni, Perego.
Conducted experiments: Gatti, Cossa, Tinelli, Carenini, Pennati, Cominetti, De Cesare.
Performed data analysis: Cossa, Arrighetti.
Wrote or contributed to the writing of the manuscript: Gatti, Perego.
- Received September 26, 2013.
- Accepted December 16, 2013.
This work was supported in part by the European Community Integrated Project CHEMORES [Grant 037665], the AIRC Special program Molecular Clinical Oncology–5 per mille Project [Grant 2010-9998], the Fondazione CARIPLO [Grant 2011-0490], and the Fondazione Guido Berlucchi.
- ATP binding cassette
- arsenic trioxide
- breast cancer resistance protein
- bovine serum albumin
- death receptor
- double-strand RNA
- fluorescence-activated cell sorter
- non–small-cell lung cancer
- phosphate-buffered saline
- synergism index
- small-interfering RNA
- tumor necrosis factor–related apoptosis-inducing ligand
- tumor volume
- tumor volume inhibition
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics