Low Intensity and Frequency Pulsed Electromagnetic Fields Selectively Impair Breast Cancer Cell Viability

Introduction A common drawback of many anticancer therapies is non-specificity in action of killing. We investigated the potential of ultra-low intensity and frequency pulsed electromagnetic fields (PEMFs) to kill breast cancer cells. Our criteria to accept this technology as a potentially valid therapeutic approach were: 1) cytotoxicity to breast cancer cells and; 2) that the designed fields proved innocuous to healthy cell classes that would be exposed to the PEMFs during clinical treatment. Methods MCF7 breast cancer cells and their normal counterparts, MCF10 cells, were exposed to PEMFs and cytotoxic indices measured in order to design PEMF paradigms that best kill breast cancer cells. The PEMF parameters tested were: 1) frequencies ranging from 20 to 50 Hz; 2) intensities ranging from 2 mT to 5 mT and; 3) exposure durations ranging from 30 to 90 minutes per day for up to three days to determine the optimum parameters for selective cancer cell killing. Results We observed a discrete window of vulnerability of MCF7 cells to PEMFs of 20 Hz frequency, 3 mT magnitude and exposure duration of 60 minutes per day. The cell damage accrued in response to PEMFs increased with time and gained significance after three days of consecutive daily exposure. By contrast, the PEMFs parameters determined to be most cytotoxic to breast cancer MCF-7 cells were not damaging to normal MCF-10 cells. Conclusion Based on our data it appears that PEMF-based anticancer strategies may represent a new therapeutic approach to treat breast cancer without affecting normal tissues in a manner that is non-invasive and can be potentially combined with existing anti-cancer treatments.


Introduction
There is a growing interest in the use of electromagnetic fields as an anticancer treatment [1][2][3][4][5]. The search for new therapeutic strategies is particularly active in the field of oncology where standard antineoplastic treatments, based on chemotherapeutic drugs and/or radiotherapy, possess potentially detrimental secondary effects and on their own often fall short of providing a complete and resilient recovery. Fueling this recent interest is the fact that extremely low-frequency and low-intensity pulsed electromagnetic fields (PEMFs) have been shown to be innocuous, possibly even beneficial [4,[6][7], to normal cell types. On the other hand, certain malignant cell classes have been shown to be particularly vulnerable to their effects [5,[8][9][10]. A potential value of extremely low frequency PEMFs hence lies in their use as an adjuvant treatment to more traditional chemo-and radiotherapies with the aim of reducing their dosage, mitigating any harmful secondary side effects and enhancing patient prognosis. Despite recent successes, however, the types of signals applied and cancer classes tested varied widely, producing a wide range of killing efficiencies and succeeding in forestalling concurrence in this area of research [1,[3][4][5]. A clear determination of the types of cancer most susceptible to PEMFs and their subsequent optimization for targeted killing will be needed before they can be used to selectively remove cancer cells from a heterogeneous population of malignant and healthy cells.
Here we show that the ability of ultra-low intensity and frequency PEMFs to selectively kill breast cancer cells depends exquisitely on field parameters. MCF-7 breast cancer cells are selectively vulnerable to PEMFs within a discrete window of PEMF signal parameters and times of exposure with resolutions of mTeslas and tens of minutes, respectively. Using five independent means of monitoring cancer cell death we obtained identical findings; selective killing of MCF7 cells was best achieved with PEMFs of 3 mT peak-to-peak magnitude, at a pulse frequency of 20 Hz and duration of exposure of only 60 minutes per day. By stark contrast, this same pulsing paradigm (cytotoxic to MCF-7s) was innocuous to normal MCF-10 breast cells. PEMF-based therapeutic strategies might thus provide a manner to control certain classes of cancer while minimally implicating healthy tissues. . The cells were maintained at 37uC in a standard tissue culture incubator (Vitaris AG, Baar, Switzerland) in an atmosphere of 95% humidity and 5% CO 2 .

PEMFs exposure system
The PEMF exposure setup, described in Text S1 and illustrated in Figure S1 A-C, was housed inside a standard cell culture incubator (Vitaris AG, Baar, Switzerland) providing a humidified environment at 37uC, but lacking CO 2 regulation. The cells were exposed to an asymmetric pulsed magnetic field while continuously monitoring the field strength and temperature. The nonexposed (control) cells were placed within the same incubator for identical periods, but shielded from the magnetic fields by a mu metal enclosure surrounding the coils. Thus, all cells were exposed to the same climate and temperature.

PEMFs treatment
MCF7 and MCF10 cells were seeded in T25 flasks (SPL Life Sciences, Korea) at concentrations of 6.5610 5 cells/ml and 6.7610 5 cells/ml, respectively. After 24 hours of being plated the cells were washed with PBS (Life Technologies Corporation, Gibco, Paisley, United Kingdom), given fresh medium and exposed to PEMFs for the first of three daily trials; media was not changed from this point onward. An asymmetric pulsed magnetic field of 6 ms interval at a repetition rate of 20 and 50 Hz were applied at flux densities of 2.0, 3.0 and 5.0 mT (peak-to-peak) for 1 hour/day for three days. Whereas exposure to PEMFs at a repetition rate of 20 Hz caused a significant increase in cancer cells death that was dependent on PEMF amplitude, PEMFs applied at a repetition rate of 50 Hz did not produce any noticeable effects over cell viability and were not dealt with further in this manuscript ( Figure S2 A-B). To test for effects of different exposure durations, cells were exposed to PEMFs of 3 mT magnitude and at a repetition rate of 20 Hz for 30, 60 or 90 minutes per days for one, two or three days. Cells were collected and analyzed on the first, second or third day for analysis, depending on the assay being conducted.

Trypan blue assay
After a given PEMF exposure protocol, cells were detached, spun down at 1200 rcf for 5 min, resuspended in 1 ml of PBS and incubated in trypan blue at 1:1 (Sigma-Aldrich, St. Louis, MO, USA). A homogeneous suspension of cells was then deposited into a Burker chamber (BRAND GMBH + CO KG, Wertheim Germany), viewed microscopically and counted. The percentage of dead cells was obtained by calculating the ratio of trypan blue positive cells in treated and untreated samples. In some cases cells were allowed to recover for up to 48 hours after their last PEMF exposure. These cells were then detached, stained with trypan blue (Sigma-Aldrich, St. Louis, MO, USA) and the number of dead cells calculated relative to control.

Apoptosis determination by DNA strand break detection
Apoptosis was measured by means of an Apo-direct kit (BD biosciences, Allschwil, Switzerland) that labels DNA strand breaks using FITC-dUTP. After each treatment 5610 5 cells were collected and then fixed and stained accordingly to the manufacturer's instructions. The assay was run on a FACS Calibur (BD Biosciences, Allschwil, Switzerland) flow cytometer using the positive and negative controls provided in the kit as well as an additional positive (death) control given by exposing MCF7 or MCF10 cells to 1 mM H 2 O 2 overnight. H 2 O 2 applied in this manner resulted in 87% 6 2% (+/-SD, n = 4) and 82% 6 3% (+/-SD, n = 4) lethality in MCF7 and MCF10 cells, respectively. The FITC fluorescence (520 nm) was detected in the FL1 channel and quantifies the amount of DNA strand breaks. For each measurement, 20,000 cells were acquired and analyzed by Flow Jo software (vers. 7.6.5) (Tree Star Inc. ON, USA).

Analysis of cellular electrical properties by means of Impedance microflow cytometer
Impedance flow cytometry (IFC) was conducted on a prototype provided by Amphasys AG (Root Lä ngenbold (LU), Switzerland). Concisely, the apparatus consists of a microfluidic chip, outfitted with a pair of microelectrodes that measure changes of electrical impedance as cells pass through dual interrogation points in response to an alternating current at four user-defined frequencies in the mid frequency (MF) and high frequency (HF) bands [11][12][13][14][15]. The obtained data (amplitude, phase and cell velocity) were automatically converted into a standard FCS3 format and analyzed with Flow Jo (vers. 7.6.5) (Tree Star Inc. ON, USA).
After treatment cells were collected, resuspended in PBS at a concentration of 4-5610 6 cells/ml and pumped through the chip at a maximum velocity of 1 cm per second, 500-1000 cells per second. For each measurement, 20,000 cells were analyzed at a frequency of 0.5 MHz to monitor apoptosis [11][12][13]15] or 9 MHz to determine metabolic status [11][12][13][14][16][17]. Each sample was run in parallel with polystyrene beads (8 mm) (Sigma-Aldrich, St. Louis, MO, USA) to obtain a standard signal response over the entire frequency spectrum, establishing a set point.

Apoptosis determination by Annexin V staining
An Annexin V/Propidium iodide (BD biosciences, Allschwil, Switzerland) assay was used to monitor the progression of apoptosis at distinct stages. Monitoring the dual staining pattern of Annexin V (FITC-conjugated) and propidium iodide (PI) allowed for the identification of early (Annexin V + and PI -) and late apoptosis as well as cells having undergone necrosis (dead cells, Annexin V and PI +). After each treatment, 3610 5 cells were collected and stained as specified by the manufacturer's instructions. Staining was assayed on a FACS Calibur (BD Biosciences, Allschwil, Switzerland), recording 20,000 cells for each measurement. Fluorescence was detected in the FL1 and FL2 channels for FITC (Annexin V) and PI, respectively. Data were acquired and analyzed by Flow Jo software (vers. 7.6.5) (Tree Star Inc. ON, USA).

Statistical analyses
All histogram data were presented as mean 6 SD (standard deviation) of at least 3 independent experimental runs (range = 3 to 5) consisting of between 1 to 3 replicates for each biological parameter analyzed. The exact number of measurements in each presented data point is reported in the figure legend and is indicated in brackets (n). Statistics were performed using the Wilcoxon Rank-Sum Test (two-tailed) comparing each treated sample to relative control (sham-exposed sample) for all the cell lines used. A p-value ,0.05 was considered statistically significant (*) and a p-value , 0.005 was considered highly significant (**).

PEMFs increase breast cancer cell death as detected by Trypan Blue inclusion
Our objective was to devise a set of treatment protocols that could potentially translate into the clinical arena to slow cancer growth, while proving harmless to healthy tissues. We focused on a breast cancer cell model given our previous success using PEMFs to slow their growth [8]. To ascertain the sensitivity of normal and cancer cells to PEMFs we exposed MCF7 breast cancer cells and their normal breast epithelial counterparts, MCF10s, to PEMFs of magnitudes between 2 mT and 5 mT and at a repetition rate of 20 Hz for 1h per day for three days. Following the last exposure (day 3) all samples were harvested and stained with trypan blue to quantify cell death and compared to otherwise identically treated control (non-exposed) cultures. A highly significant reduction in the percentage of surviving MCF7 cells was observed in response to exposure to 3 mT PEMFs. By contrast, exposure of identical MCF-7 cultures to PEMFs of either 2 mT or 5 mT amplitudes resulted in less significant levels of cell death (Fig 1A). On the other hand, exposure to 3 mT PEMFs, which proved the most cytotoxic to MCF-7 cancer cells, was innocuous to ''wild type'' MCF10 cells (as were 2 and 5 mT PEMFs) and moreover, appeared to have even accentuated their survival (mitigating resting levels of apoptosis) relative to unexposed cells (also see Figure S5). We next sought to determine the best exposure interval to 3 mT PEMFs to kill breast cancer cells. Figure 1B depicts cell death as a function of duration of exposure to 3 mT PEMFs (20 Hz). Cells were exposed to 3 mT PEMFs for either 30, 60 or 90 minutes per day for 3 days before assaying for cell death. MCF7 cells were most susceptible to PEMF exposures of 60 minutes duration, whereas exposure times 50% shorter (30 minutes) or 50% longer (90 minutes) than this resulted in significantly less amounts of cell killing ( Fig 1B). Once again, MCF10 cell viability was not compromised by PEMF exposure of any duration. Indeed, PEMFs appeared to make MCF10 cells more resistant to undergoing apoptosis, particularly in response to the 60-minute exposure regimen that proved most cytotoxic to MCF7 cells ( Figure S5). The data thus reveals a discrete set of PEMF parameters (magnitude, frequency and duration of exposure) that are most cytotoxic to breast cancer cells, whereas the identical set of PEMFs parameters were apparently harmless to non-malignant cell types (also see Figures S3 and S4).
To ascertain whether the PEMFs-induced cytotoxicity reported here is a cumulative response or requires a threshold level of cellular insult to become evident, we treated cells with 3 mT PEMFs for either 60 or 90 minutes per day for 1, 2, or 3 days and next quantified the total number of dead and living cells. Whereas in the unexposed cultures total cell number steadily increased throughout the three days of trial, exposure to 60 or 90 minutes of PEMFs per day either totally abrogated or slowed the increase in cell number, respectively (Fig 2). On the other hand, the absolute number of dead (trypan blue positive) cells did not scale down in proportion to the decrease in total cell number as might be expected if cell proliferation was simply being slowed, but instead, increased. Notably, on the third day, in response to 60 minutes of daily exposure to PEMFs (3 mT), the total number of cells in the culture decreased, whereas the total number of dead cells increased, by -40% (+/-6% (SD); n = 12) ((total cells in control sample -total cell in treated sample)/total cells in control sample)) and +20% (+/-13% (SD); n = 12) ((dead cells in control sampledead cell in treated sample)/dead cells in control sample)), respectively, indicating heightened cytotoxicity in response to PEMFs. Figure 3 shows that the increase in cell loss with time is greatest in cultures treated for 60 minutes per day, rather than 90 minutes per day.

Assessment of PEMF-induced apoptosis by detecting DNA strand breaks
Our Flow Cytometric (FCM) determination of apoptosis was assayed with identical PEMF parameters (days of consecutive exposure, durations of exposure, field amplitudes and frequency) as those utilized for trypan blue assessment of killing efficiency with identical results. Figure 4A shows an overlay of MCF7 cells exposed to PEMFs of three distinct intensities (2, 3 or 5 mT) for 60 minutes per day. A shift to the right (greater FL1-H values) of a cell population reflects greater DNA damage. As previously demonstrated, MCF7 cancer cells are particularly vulnerable to 3 mT PEMFs. Figure 4B shows the extent of 3 mT PEMF-induced DNA strand breaks following 30, 60 or 90 minutes exposures per day. Once again, 60 minutes of 3 mT PEMFs for three consecutive days gave the greatest DNA damage in MCF7 cancer cells. And, once again, stronger fields (5 mT) or longer exposures (90 minutes per day) were less cytotoxic to MCF7 cells (Fig 4A-D). Further paralleling our trypan blue results, MCF10 normal breast epithelial cells were not harmed by any of the PEMF paradigms tested, particularly those observed to be especially cytotoxic to MCF7 cells. Indeed, a slight protective effect (a leftward shift to lower FL1-H values) was again discerned in MCF10 cells in response to the PEMF parameters that were most cytotoxic to MCF7 breast cancer cells (Fig 4E; see also Figure S5). To investigate if the previously described increase in DNA fragmentation observed in MCF7 cells after 3 days of PEMF treatment was cumulative with time, we stained cells after 1, 2 or 3 consecutive days of exposure to either 60 or 90 minute of 3 mT PEMFs. Although PEMF-induced DNA damage increased with time, it only really obtained significance from control levels after the third day and was particularly pronounced in response to 60-minute daily exposures (figure 5 A-D). Our FCM analysis thus corroborates and strengthens our trypan blue results, indicating that treatment with 3 mT PEMFs for 60 minutes per day were most effective at killing MCF7 breast cancer cells while leaving healthy cell classes (MCF10) unharmed.

Determination of PEMF-induced apoptosis by Impedance Flow Cytometry
Impedance Flow Cytometry (IFC) assesses real-time cell viability by monitoring cellular electrical properties in behaving cells [11][12][13]15]. In the dot plot generated from monitoring the entire cell population's electrical characteristics at a scan frequency of 0.5 MHz dead cells reside in the far lower left quadrant (low impedance phase and magnitude values). PEMFs produced a shift in MCF7 cells to the lower left quadrant, particularly in response to 3 mT PEMFs, which gave the greatest separation between living (right) and dying (left) cells (Fig 6A). Figure 6B shows the results of MCF7 cells exposed to 3 mT PEMFs for either 30, 60 or 90 minutes per day for three days. In agreement with our previous trypan blue and FCM assessment of apoptosis, cells exposed to 60 minutes of 3 mT PEMFs per day exhibited the greatest percentage of dead cells as detected by IFC (Fig 6 C-D). In stark contrast, yet in further confirmation of our previous results, MCF10 cells were slightly benefitted by these same PEMF parameters (Fig 6E, see also Figure S5).

Assessment of cell metabolic status after PEMF treatment with IFC
At higher scan frequencies the IFC discerns metabolic status [11][12][13][14]16]. At a scan frequency of 9 MHz the IFC detects two populations of cells, the right-most population (higher phase values) reflects cells experiencing the initial stages of metabolic stress [11][12][13][14][16][17]. After three days of exposing MCF7 cells to PEMFs the magnitude of right-most population augmented, the greatest right-shift coinciding exactly with those parameters (3 mT, 20 Hz, for 60 min/day for 3 days) producing the greatest cell death in response to PEMFs (Fig 7 A-D). And, once again, MCF10 normal breast cells were apparently benefitted by PEMFs as determined by IFC analysis at 9 MHz (Fig 7 D, see also Figure  S5). Due to the relatively broad scope of the phenotype (metabolic Values represent the averages of at least 4 independent experiments (n = 4, 12, 4 for 2, 3 and 5 mT, respectively; n = 5, 12, 8 for 30, 60 and 90 minutes, respectively) for MCF7 cells (average 6 SD). A total of 5 independent experiments (average 6 SD) is provided for MCF-10 cells for all conditions. MCF10 were unresponsive to PEMFs (3 mT, 60 minutes per day for three days) (also see Figure S3). 50 Hz PEMFs (3 mT for 60 minutes a day for three days) was less effective at killing MCF-7 cells (see Figure S2). The potential recovery of MCF-7 cancer cells following PEMF treatment is addressed in Figure S6. doi:10.1371/journal.pone.0072944.g001 stress) the effect is the largest we have measured in response to PEMFs (see next, see also Figure S5).
To independently validate that IFC effectively detects apoptosis and metabolic status in our cell system we treated MCF7 cancer and MCF10 normal cells with 1 mM H 2 O 2 to evoke cell death to an extent of 87% 6 2% (+/-SD, n = 4) and 82% 6 3% (+/-SD, n = 4), respectively. When analyzed by IFC at a scan frequency of 0.5 MHz cells treated with H 2 O 2 were displaced to the far lower left quadrant (Fig 8A; cf Fig 6A-D). Also, confirming that a cell population undergoing the initial stages of metabolic stress is indeed shifted to the right (in IFC scans at 9 MHz) we obtained an analogous right-shift in MCF7 cells after overnight exposure to 1 mM H 2 O 2 (Fig 8B; cf Fig 7A-D). Hence, IFC does appear to be a viable method to monitor cancer cell viability.

Assessment of PEMF-induced apoptosis by Annexin V staining
To further corroborate our trypan blue, FCM and IFC data demonstrating the induction of apoptosis in MCF7 cancer cells in response to PEMF exposure, we performed Annexin V/PI assays, discriminating cells in early apoptosis (Annexin V+/PI-) from dead and damaged cells (propidium iodide +). MCF7 (cancer) and MCF10 (normal) cells were directly exposed to the PEMFs paradigms we previously found to be most cytotoxic to MCF7 cells, 3 mT for 60 minutes per day. Figure 9A shows that PEMF treatment resulted in a 13% increase in Annexin V+ MCF7 cells relative to control, quantitatively agreeing with our other PEMFinduced cytotoxic assessments assayed with trypan blue (treatedcontrol: 11% dead cells), FCM (treated -control: 14% dead cells), IFC at scan frequency of 0.5 MHz (treated -control: 16% dead cells) and IFC at scan frequency of 9 MHz (treated -control:  The most significant degree of cell impairment was seen after 3 days (4 independent experiments with 3 replicates/experiment (n = 12)) (p-values, left to right: 0.3246, 0.02032, 0.00004) (also see Table 1 for the mean, high value, low value and average absolute deviation from median). (B) MCF7 cancer cells treated with 3 mT PEMFs for 90 min/day for 1, 2 or 3 days. Overall, 90 min/day of exposure produced less cytotoxicity than 60 min/day. Data were generated from 4 independent experiments with 2 replicates/experiment (n = 8) (p-values, left to right: 0.2595, 0.02953, 0.00015) (also see table 2 for the mean, high value, low value and average absolute deviation from median). doi:10.1371/journal.pone.0072944.g003 25%). As previously demonstrated with all the other apoptosis assays we performed, MCF10 cells were not adversely affected by these same PEMF parameters (Fig 9B) (also see Figure S5).

Discussion
Motivated by studies demonstrating the safety of very low frequency and intensity PEMFs [4,6] and extending from our previous work [8], demonstrating that MCF7 cancer cells are selectively vulnerable to 20 Hz pulsed electromagnetic fields, we investigated the effects of PEMFs on human breast epithelial cells of malignant (MCF7) and non-malignant (MCF10) phenotypes. Cytotoxic sensitivity to certain PEMFs parameters was entirely restricted to the malignant phenotype and exhibited a clear dependency on the duration, frequency and intensity of the PEMFs employed. Specifically, breast cancer cells of the MCF7 lineage were most vulnerable to PEMFs of 3 mT magnitude, at a repetition rate of 20 Hz and for an exposure interval of 60 minutes per day (Fig 1 A-C). These same PEMF parameters, although cytotoxic to MCF7 cells, were slightly protective to non-malignant breast epithelial cells of an identical host lineage, MCF10 (see Figure S5). For these experiments we limited our analysis to within three days of exposure to remain within the realm of a clinically feasible therapeutic strategy. Three days was also chosen as an appropriate end point to our analysis as it avoided the overgrowth of control cells. In a tissue culture paradigm such as ours, staying below cell confluence would minimize the potential contributions of cell density/contact-induced changes in biochemical status or nutrient deprivation to our measured differences. The possibility hence remains, that increasing the number of days of exposure to PEMFs may enhance the specificity and efficiency of cancer cell killing. The design of longer time course experiments will be the focus of our future studies. Nonetheless, our results, although relatively modest are sufficiently provocative (in terms of their reproducibility and selectivity) to merit future studies aimed at further evolving this approach and yet, are consistent with previous studies demonstrating that sensitivity to electromagnetic fields depends on the signal parameters used as well as the type of cells exposed to the fields [5,7,9,[18][19].
For this study we focused our attention on PEMF parameters that: 1) could practically translate into the clinical arena with reference to duration of exposure and 2) were innocuous to healthy cell classes collaterally exposed to PEMFs during clinical treatment. Our results are notable given that: 1) our most effective exposure time to induce cancer cell (MCF7) death was only one hour per exposure rather than 3-72 hours as previously reported [5,[20][21] and; 2) the field paradigms we designed were apparently innocuous to normal cells (MCF10). As of yet, we have not achieved complete ''selective'' killing with PEMFs. Although this objective might be achieved with further fine-tuning of the PEMF parameters (exposure magnitude, duration, signal shape, number of days of treatment) we cannot then exclude the possibility that other tissues type might then be implicated in the death pool. Quite notable, however, were the diametrically opposed responses of MCF7 (cancer) and MCF10 (normal) cells to PEMFs, widening the cytotoxic gap between exposed cancer and exposed normal cells. Potentially, PEMFs might prove useful as a non-invasive adjuvant treatment to be combined with other common anticancer therapies.
The selective killing of cancer cells with PEMFs was corroborated by four independent methodologies using five different analytical paradigms, covering the full gambit of stages leading to ultimate cell death. Firstly, our trypan blue results gave the number of cells in a late stage of cell dying known as ''postapoptotic necrosis'' or ''secondary necrosis'' (Fig 1 A-B, 2 A-D and 3 A-B) [18,[22][23]. Secondly, our FCM analysis detected DNA breaks prior to cell death [17,24] and occurring downstream of calcium-stimulated caspase activation (Fig 4 A-E and 5 A-D) [25]. Thirdly, we investigated the progression of apoptosis using Impedance Flow Cytometry (IFC) that detects changes in the electrical properties of cells reflecting physiological status [11][12][13][14][15][16][17]24,[26][27] at two frequencies: 1) 0.5 MHz, to ascertain the number of cells having undergone apoptosis (Fig 6 A-E) [11][12][13]15] and 2) 9 MHz, to monitor changes that coincide with the onset of cellular stress (Fig 7 A-E) [11][12][13][14][16][17]. Several recent publications have supported the value of IFC to gauge cell viability [11][12][13][14][15][16][17]27]. Finally, we employed an Annexin V/PI assay to distinguish early apoptotic cells from damaged or already dead cells (Fig 9 A-B) [28][29]. In all five assays of cell viability identical PEMF parameters produced the greatest degree of cell damage to MCF-7 breast cancer cells, 3 mT intensity for 60 minutes a day, demonstrating a clear and discrete window of vulnerability of MCF7 cells to PEMFs of given characteristics. Stronger fields, longer exposures, or higher frequencies to these empirically determined values (3 mT, 20 Hz, 60 minutes exposures per day) were less cytotoxic to MCF7 cells, clearly demonstrating the importance of field optimization for the eventual killing of malignant cell classes with PEMFs.
A clear window of vulnerability of cancer cells to PEMFs exists; more is not necessarily better. That weaker fields, or less exposure to them, are less lethal, upon first impression, might seem somewhat intuitive. However, the fact that stronger, or longer, exposure to fields is less efficient at killing, implies some specifically of biological action, rather than a straightforward dose-dependent accumulation of generalized damage over a susceptible cell status. The validity of the described window effect is implicitly substantiated within the context of our data presented herein, the fact that five independent assays (four distinct methodologies) of measuring cell viability gave the identical result and produced similar magnitudes of cell death (also see Figure S5). The cytotoxic-dependency on exposure duration was so robust that it was also apparent when examining the time course in the development of cytotoxicity during three days of consecutive PEMF exposure. That is, 60-minute daily exposures to PEMFs gave greater ratios of cell death (figure 3) and greater amounts of DNA fragmentation (figure 5) than 90 minutes of daily exposure. Moreover, the PEMF parameters that were most cytotoxic to MCF7 breast cancer cells proved most beneficial to MCF10 normal breast cells. Similar window effects have been reported in the field of electromagnetics and have been openly discussed in the literature, yet there are no accepted models to explain their existence [19,[30][31]. Within the Protection Guidelines Report of the International Commission on Non-Ionizing Radiation [30] it is stated, ''Interpretation of several observed biological effects of AM (amplitude modulated) electromagnetic fields is further complicated by the apparent existence of ''windows'' of response in both the power density and frequency domains. There are no accepted models that adequately explain this phenomenon, which challenges the traditional concept of a monotonic relationship between the field intensity and the severity of the resulting biological effects.'' At this juncture, however, the relative contributions of an actual slowing of cell proliferation and the induction of cell death to the overall effect of PEMFs is unclear (cf figure 2), as is the rate and extent of absorption of dead cells by the culture after their demise. Therefore, although cell cycle withdrawal possibly resulting from PEMFs may contribute to observations reported here, the most directly measurable effect is that of induced apoptosis. Nonetheless, the capacity of PEMFs to slow the proliferation of a cancer cell class also would be positive clinical outcome and of relevance in advancing PEMF-based anti-cancer therapies.
The molecular mechanisms whereby cancerous (MCF7) cells are compromised yet, healthy (MCF10) cells are not fully understood and yet, of utmost importance for the ultimate development of PEMF-based strategies to combat cancer and will be the focus of our future investigations. We speculate that the window effect observed in this study results from changes in intracellular calcium handling in response to PEMF exposure. Calcium signaling is renowned for its multimodal effects relying on intracellular calcium increments that: 1) result from both calcium influx across the cell surface membrane and release from intracellular membrane-delimited compartments; 2) are simultaneously coded in space, time and holding level; 3) exhibit negativeand positive-feedback regulatory mechanisms and; 4) are coordi- Figure 7. MCF7 and MCF10 cell metabolic status analyzed by IFC at 9 MHz. (A) Dot plots generated from MCF7 cells after exposure to PEMFs of 2, 3 or 5 mTs and in control (non-exposed) samples and analyzed at a scan frequency of 9 MHz. Exposed samples exhibited a larger rightside population, particularly after exposure to 3 mT PEMFs. (B) Dot plots of MCF7 cells after exposure to 30, 60 or 90 minutes of PEMFs (3 mT, 20 Hz) per day for 3 days; the right-side population was preferentially enhanced in response to 60 minutes exposures. (C) Histograms depicting the percentage increase in the size of the right population normalized to controls after exposure to 2, 3 or 5 mT PEMFs for 60 minutes. Each value is the average of 4 independent experiments (1 replicate/experiment, n = 4) (6 SD). P-values, left to right: 0.00879, 0.0017 and 0.07033. (D) Size of right population as a function of exposure duration and normalized to each respective control (unexposed) sample; the right-side population was preferentially enhanced in response to 60 minutes exposures. Each value is the average of 4 independent experiments (1 replicate/experiment, n = 4) (6 SD). P-values, left to right: 0.6786, 0.0017 and 1. (E) Dot plots generated from MCF10 cells exposed to 3 mT PEMFs (20 Hz) for 60 minutes/day for three days and in control (unexposed) samples, revealing essentially no change in response to treatment. The dot plots shown were generated from cells of the same experimental date and are representative of cells responses observed in all of the independent experiments with identical conditions. Also see Figure S8, for the spread of individual measurements. doi:10.1371/journal.pone.0072944.g007 nated by dynamic changes in membrane organization [32][33]. As a commonly reported consequence of PEMF exposure is elevations of intracellular calcium level [34] one possibility is that PEMFs mediate their effects via influencing intracellular calcium signaling pathways. In the context of this report 3 mT PEMFs at a frequency of 20 Hz for 60 minutes per day would create the ''correct'' combination of calcium signals that would most effectively result in cell death. Indeed, it has been previously shown that chelating or augmenting intracellular calcium accordingly spares or compromises MCF7 survival, respectively [35][36][37]. The shift to the right observed at 9 MHz in IFC (Fig 4 A-D) likely reflects changes in membrane complexity and cytoplasmic reorganization (change in whole-cell capacitance) [11][12][13][14][16][17] that coincide with the establishment of cytomorphological features that reflect the modulation of biochemical pathways that, in turn, regulate the delicate balance between cell proliferation and apoptosis including, modifications in mitochondrial metabolism downstream of changes in intracellular calcium levels [16][17]33,38]. Future studies of ours will focus on the effects of PEMFs over cytosolic calcium increments.
Non-malignant MCF10 cells were unaffected, or even fortified, by the PEMF paradigms producing the greatest damage in MCF7 cells, revealing another level of specificity of action and supporting the possibility that it may be ultimately feasible to selectively remove cancer cells from an organism without implicating normal tissues in the death pool using PEMF-based technologies (Figs 1 A-B, 4E, 6E, 7E, 9B and Figure S5). The immunity of MCF10 cells to PEMFs might suggest that their endogenous calcium homeostatic mechanisms are capable of buffering, or even exploiting, small increments in intracellular calcium concentrations, whereas MCF7 cells are not able to withstand even modest perturbations in cytosolic calcium levels, a supposition that is supported by recently amplitude histograms (left (above, inset) and right (vertically expanded)) generated from MCF7 cells exposed to PEMFs or H 2 O 2 at 0,5 MHz. PEMFs produce similar population displacements to the lower left quadrant of the dot plot (lower phase and magnitude values) as in H 2 O 2 treated samples: untreated cells (gray), cells exposed to PEMFs (3 mT, 20 Hz for 60 min/day for 3 days; 25% dead cells: black) and cells incubated overnight with H 2 O 2 (1 mM; 87% dead cells: light blue). (B) Amplitude histograms correspondent to dot plots generated from MCF7 cells and analyzed by IFC at 9 MHz. H 2 O 2 treatment (1 mM; producing 87% cell death) caused the displacement of the entire cell population to the right; horizontal bars indicate inclusion gates. The shift to the right upon death induction is clearly shown in the overlay of controls (untreated; gray), PEMF-exposed cells (black) and H 2 O 2 treated samples (light blue) in panel C. The dot plots were generated from cells of the same experimental date and are representative of cells responses observed in all of the independent experiments with identical conditions. Trypan blue inclusion was used to quantify the percentage cell death in the H 2 O 2 treated samples. doi:10.1371/journal.pone.0072944.g008 published studies [36][37]. In further support for such a calciumdependent mechanism of preferential killing of malignant cells it has been shown that Panaxydol, a derivative of Panax ginseng that induces sustained elevations in cytosolic calcium, preferentially induces apoptosis in cancer cells (including MCF7s) but not normal cells [39]. Such a selective calcium-dependent mechanism of cancer cell killings may eventually help in the refining of PEMFbased technologies to better execute the preferential killing of breast cancer cells in clinical settings.

Supporting Information
Figure S1