Response of neuroblastoma cells to RF currents as a function of the signal frequency

The neuroblastoma cell line NB69 (lot No. 03I019/2008, item No. 99072802) was purchased from the European Collection of Authenticated Cell Cultures (ECACC, Salisbury, UK). The cells were periodically tested for mycoplasma contamination (PCR) and response to chemical and physical treatments, including cytostatic agents or hyperthermia.

Cells were plated in 75 cm² culture flasks containing D-MEM medium (Biowhittaker, Lonza, Verviers, Belgium) supplemented with 10% (v/v) foetal bovine serum, 1% L-glutamine and 1% penicillin-fungizone (Gibco, Invitrogen, Camarillo, CA, USA). Cells were grown in an incubator (Forma Scientific, Thermo Fisher, Waltham, MA, USA) with a 37 °C, 5% CO₂, humidified atmosphere. Every seventh day, when confluent, the culture was passed by detaching the cells with 0.05% trypsin + 0.02% EDTA (Sigma, Saint Louis, Missouri, USA) in HBSS and seeding them in a new flask. The remaining cells were plated in 60 mm-diameter Petri dishes (Nunc, Roskilde, Denmark), at a density of 8160 cells/cm². For immunofluorescence analysis, the cells were seeded on glass coverslips placed inside the Petri dishes.

The electric treatment was applied at day 4th after seeding. The exposure system has been described in detail in previous papers [13, 18, 19]. Briefly, electric current was delivered by pairs of sterile, stainless steel electrodes, designed ad hoc for in vitro stimulation, that were fitted inside the Petri dishes. For CRET exposure, the electrode pairs inserted in the experimental dishes were connected in series to a Multifrequency signal generator (INDIBA®, Barcelona, Spain) custom-engineered to supply RF electric currents within a 350 kHz – 650 kHz range. For sham-exposure, the electrode pairs placed inside all control dishes were also connected to the generator, though they remained unenergized. The stimulation pattern consisted of 5-min pulses of 350 kHz, 448 kHz, 570 kHz or 650 kHz, sine wave current administered at a subthermal density of 50 μA/mm². Each pulse was followed by a non-stimulation lapse of either 0 min or 25 min (short-term treatments), or 3 h and 55 min (other treatments). Except for the short-term treatment, the described pulse-interpulse cycle was repeated along total intervals of 4, 12 or 24 h. The cultures were grown in two separate, identical CO₂ incubators (Thermo Fisher Scientific, Waltham, MA, USA). The stimulation parameters and the atmosphere inside the incubators (temperature: 37 °C, relative humidity: 90% and CO₂: 5%) were constantly monitored. The electromagnetic environment inside the incubators was also monitored [18].

The effect on cell viability of a 24-h treatment with each of the selected frequencies was assessed through Trypan Blue exclusion assay. The cells were detached from the plates using 0.05% trypsin + 0.02% EDTA, stained with 0.4% Trypan Blue (Sigma, Steinheim, Germany) diluted in PBS 1:4, and counted in a Neubauer chamber. At least three experimental replicates per frequency were conducted.

The applied standard gating procedure has been described in detail elsewhere [16, 17, 19]. Briefly, at the end of a 24-h treatment with the 448 kHz signal the cells were trypsinized, harvested and stained with propidium iodide for 1 h at room temperature. In each experimental repeat three cell suspensions were processed per experimental condition. The relative fractions of subG0/G1 (indicative of cell death) G1, S and G2/M subpopulations were determined through DNA content quantification (FACScalibur, Becton Dickinson, Franklin Lakes, NJ, USA). CellQuest 3.2 software was used for data acquisition (20,000 events per sample) and analysis. Suitable gating strategies were applied to exclude debris and aggregates.

The protocol has been described in detail in previously published articles [17‐19]. Briefly, the expression of p-ERK1/2, p-p38, p53, Bax and caspase-3 in samples exposed on coverslips to the 448 kHz current during 30 min, 12- and/or 24-h, was immunofluorescence determined. The cells were fixed with 4% paraformaldehyde and incubated overnight at 4 °C with rabbit polyclonal anti-p-ERK1/2 (1:500; cat n: 44-680G, Thermo Fisher, Bengaluru, India), mouse monoclonal anti-p-p38 (1:500; cat. n: #9216), mouse monoclonal anti-p53 (1:200; cat. n: #2524), rabbit polyclonal anti-caspase-3 (1:400; cat. n: #9664), the three of them from Cell Signalling (Danvers, MA, USA) and rabbit polyclonal anti-Bax (1:100; cat. n: sc-6236) from Santa Cruz Technologies (Texas, USA). Afterwards, the samples were fluorescence stained for 1 h at room temperature with Alexa Fluor® 488 goat anti-rabbit IgG (1:500; cat n: A-11034) or with Alexa Fluor® 568 goat anti-mouse (1:500; cat. n: A-11031), both from Molecular Probes (Oregon, USA) and the cell nuclei were counterstained with bisBenzimide H33258 (Sigma). In each experimental repeat 3 coverslips per experimental group were photomicrographed and analyzed. Twenty microscope fields (800 cells per field) were randomly selected per coverslip and the percents of immunoreactive cells were calculated over the total cell number, revealed by Hoechst counterstaining of the nuclei using fluorescence microscopy (Nikon Eclipse TE300; Melville, USA) and Computer-Assisted Image Analysis (Analy-SIS, GMBH, Munich, Germany). All analyses were performed in duplicate and repeated at least 3 times for each of the analyzed proteins.

The protocols for electrophoresis and Western blotting have been described in detail elsewhere [16]. Briefly, p-ERK1/2, were analyzed at the end of the initial 5-min stimulation with the 448-kHz signal, and at 30 min, 4 h, 12 h or 24 h from the first exposure onset. ERK1/2 and p-EGFR were analyzed at the end of the initial 5-min stimulation and 30 min afterwards. p-JNK and p-p38 expressions were analyzed at 30 min from the initial 5-min stimulation, and cyclin D1, p53, p27 and Bax were analyzed at 12 and 24 h. After CRET- or sham-stimulation the cells were harvested in hypotonic lysis buffer (10 mM Tris-HCl, 10 mM KCl, 1 mM dithiothreitol, 1 mM EDTA, 1 mM PMFS, 10 μg/ml leupeptin, 5 μg/ml pepstatin, 100 mM NaF, 20 mM β-glycerophosphate, 20 mM sodium molybdate, 0.5% Triton X-100 and 0.1% SDS). Afterwards, the proteins were separated, transferred to nitrocellulose membranes and immunostained overnight at 4 °C for rabbit polyclonal anti-p-EGFR (1:1000; cat. n: #3777), mouse monoclonal anti-p-p38 (1:1000), anti-ERK1/2 (1:1000; cat. n: #9102), anti-p-JNK (1:1000; cat. n: #4668) and anti-p53 (1:1000), them all from Cell Signalling (Danvers, MA, USA), as well as for mouse monoclonal anti-p27 (1:300; cat. n: AHZ0452; Invitrogen, Carlsbad, CA, USA), mouse monoclonal anti-Cyclin D1 (1:500; cat. n: NCL-CYCLIN D1-GM; Novocastra, Newcastle, UK), rabbit polyclonal anti-Bax (1:500; Santa Cruz Technologies) and rabbit polyclonal anti-p-ERK1/2 (1:1000; Thermo Fisher). Mouse monoclonal anti-β-actin (1:5000; cat. n: A5441; Sigma, Rehovot, Israel) was used as loading control. The membranes were incubated for 1 h at room temperature with anti-rabbit IgG conjugated to IRdye 800 CW (1:10000; cat. n: 926–32,211; LI-COR Biosciences, Nebraska, USA) and with anti-mouse IgG conjugated to IRdye 680 LT (1:15000; cat. n: 926–68,020; LI-COR Biosciences). Then, the membranes were scanned with a LI-COR Odyssey scanner (LI-COR Biosciences). When ECL-chemiluminescence was needed, the membranes were incubated with ECL-anti-mouse IgG horseradish peroxidase-linked antibody (1: 3000; cat. n: NA931; GE Healthcare, Little Chalfont, Buckinghamshire, UK) or with ECL-anti-rabbit IgG horseradish peroxidase-linked antibody (1: 3000; cat. n: NA934; GE Healthcare). The obtained bands were densitometry analyzed (PDI Quantity One 4.5.2 software, BioRad, Munich, Germany). At least three experimental replicates were conducted per time interval and protein. The data were normalized over the load controls and sham-exposed controls.

Three plates with NB69 cultures were used per each of 3 experimental groups: untreated controls (C), samples treated with the p-ERK1/2 inhibitor: 20 μM PD98059 (PD), and samples exposed to CRET in the presence of the inhibitor (PD + CRET). The cultures were seeded at a density of 8160 cells/cm² and incubated for 4 days. Next, PD and PD + CRET samples received the inhibitor and PD + CRET samples were exposed to the electric treatment for 12 or 24 h. After electrical stimulation, Bax expression (immunoblot at 12 h) as well as cell cycle and apoptosis rate (flow cytometry at 24 h) were analyzed following the protocols described above.

The effects on cell death and viability after 24 h of intermittent exposure at the selected signal frequencies are summarized in Fig. 1. Compared to their respective control samples, those exposed to 350 or 448 kHz currents showed statistically significant decreases (p = 0.0008 and p = 0.0360, respectively) in the average number of living cells. By contrast, at higher frequencies the average number of living cells remained unchanged with respect to controls (p = 0.507 at 570 kHz) or it was slightly but significantly increased (p = 0.0474 at 650 kHz). Regarding cell death, the average incidence of necrosis and/or late apoptosis in the control samples was 6% of the total cell count. This death rate was significantly increased over controls in the samples exposed to 448 kHz (37%; p = 0.0167) and 570 kHz and (18%; p = 0.0109) but it did not differ from that of controls in cultures exposed to higher or lower frequency signals. Taken together, these data support the hypothesis that the cell response to electrical stimulation at sub-thermal doses is a nonlinear function of the signal frequency, being 448 kHz the most effective one in simultaneously inducing cytostatic and cytotoxic effects in NB69. On the basis of this, the 448 kHz signal was chosen as suitable to investigating the processes underlying the cytostatic/cytotoxic action of CRET sub-thermal electrostimulation.

As shown in Fig. 2, flow cytometry revealed a statistically significant increase (40% over controls; p = 0.0136) in the percent of apoptotic nuclei, represented by the subG0/G1 peak. The electric treatment also induced slight decreases below controls in the fraction of cells in phase G1 (8%; p = 0.0294: statistically significant) and S (4%; p = 0.4075: non-significant). The possibility that these effects were mediated by sequential changes in proteins involved in apoptosis or cell cycle progression was investigated.

The tumour suppressor protein p53 promotes growth arrest, apoptosis and cellular senescence through transcriptional activation or repression of target genes such as Bax, Bak, Bcl or caspases [22]. The expression of p53, Bax and caspase-3 was analyzed in samples exposed intermittently to the 448 kHz signal for 12 h (4 exposure pulses of 5 min) or 24 h (8 pulses of 5 min). The results in Fig. 3 show that compared to their respective controls, the treated samples showed significantly increased expression of protein p53, both at 12 h (p = 0.0340) or 24 h (p < 0.0001) from the stimulation onset. This effect was confirmed by immunofluorescence analysis, although this test only could identify statistically significant (p = 0.065) differences with respect to controls after 24 h of exposure (Fig. 4). As for Bax expression, it experienced a significant (p = 0.0076), transient increase after 12 h of intermittent exposure and returned to levels equivalent of those in controls after 12 additional hours of treatment (Fig. 3). On the other hand, the immunofluorescence analysis did not provide relevant information on potential effects on Bax expression (Fig. 4). Regarding caspase-3 expression, the NB69 line is known to have very low rates of caspase-3-positive cells [23, 24]. Indeed, that rate was less than 3% in our control samples. Under these conditions the immunoblot analysis was inefficient to assess caspase-3 expression both in the exposed and in controls samples. However, the immunofluorescence assay did reveal significant increases in the amount of caspase-3+ cells, both at 12 h (68% over controls; p < 0.0001) and 24 h of treatment (52%; p = 0.0165; Fig. 4). It must be noted that the analysis of the images at 12 h of treatment revealed a large number of cells with cytoplasmic caspase labeling, while at 24 h the number of caspase+ cells was low, but their labeling had a nuclear location (Fig. 4a).

The exposure effects on the kinetics of the expression of cyclin D1, p27, p-p38, p-ERK and the membrane receptor p-EGFR, which regulate cell proliferation and cell cycle, were investigated. Immunoblot analysis of the samples treated during 12 h revealed significant subexpression of the proteins p27 (23.6% below controls; p = 0.0019) and cyclin D1 (18.17% below controls; p = 0.0161), both involved in the control of G1 phase progression (Fig. 5). After 12 additional hours of intermittent exposure, such transient subexpression of proteins was followed by overexpression of p27 (21.8% over controls; p = 0.0093) and by return of the expression of cyclin D1 to levels that did not differ significantly (p = 0.4624) from those in controls.

It has been reported that cyclin D1 expression and activation in the early G1 phase is stimulated, through transcriptional and post-transcriptional mechanisms, by mitogenic signaling pathways such as Ras-ERK [25]. Thus, the possibility was investigated that protein p-ERK1/2, from the Ras-ERK pathway, mediates the electroinduced fluctuations in cyclin D1 expression observed throughout the CRET treatment (Fig. 5). The immunoblot data showed no significant changes in the expression of total ERK1/2 at 5 (p = 0.5405) and 30 min (p = 0.2797) of treatment, when compared to the corresponding controls. No changes were detected in the expression of the active form of the protein, p-ERK1/2 at the end of the initial 5-min exposure, either (p = 0.1453). However, significant overexpression of p-ERK1/2 (28.2% over controls; p = 0.0214) was observed at 30 min of treatment (the initial 5-min exposure plus 25 min without exposure), followed by underexpression (11.7% below controls; p = 0.143) at 4 h (after a second exposure pulse) and 12 h (21.4% below controls, after a fourth pulse; p = 0.0267). After 24 h of treatment (8 exposure pulses) the p-ERK1/2 expression levels did not differ significantly (p = 0.4489) from those in controls. In contrast to this early response of p-ERK1/2, other proteins such as MAPK p38, JNK or the membrane protein EGFR involved in reception of extracellular signals and in Ras-ERK pathway activation [26], did not experience significant changes in the expression of their active forms, p-p38, p-JNK and p-EGFR, after the initial 5 or 30 min of treatment (Fig. 5).

The CRET effects on the early expression of p-ERK and p-p38 were also analyzed by immunofluorescence. After 30 min of exposure, a significant increase (10%; p = 0.0261) was observed in the percent of p-ERK1/2+ cells, which in control samples averaged 60% (Fig. 6). In contrast, the average rate of p-p38+ cells at 30 min (around 50% in controls) was not changed (p = 0.2299) by the CRET treatment. These results were consistent with those obtained by immunoblot analysis (Fig. 5).

The possibility that the observed effects of CRET on p-ERK1/2 could influence cell cycle regulation and/or apoptosis was also investigated. The expression of Bax, which was significantly increased (30% over controls; p = 0.0340) after 12 h of electric treatment (Fig. 3), remained unaffected both in samples treated with the p-ERK1/2 inhibitor only and in those stimulated electrically in the presence of the inhibitor (Fig. 7a and b). By contrast, the rate of apoptosis was significantly increased (p = 0.0235) after 24 h of CRET exposure in the presence of the inhibitor. As for cell cycle, while after 24 h of exposure to CRET there was a modest decrease in the rate of cells in S phase (4% below controls, Fig. 2), when applied in the presence of the inhibitor the electric treatment induced a 17% increase (p = 0.0574) in the proportion of cells in said cycle phase and a significant decrease (p = 0.0213) in G0/G1 phase (Fig. 7c). None of the treatments, either applied together or separately, induced significant effects in other cycle phases.