Necroptosis Triggered by ROS Accumulation and Ca2+ Overload, Partly Explains the Inammatory Responses and Anti-Cancer Effects Associated With 1Hz, 100 mT ELF-MF in vivo

Background: Focus on application of non-ionizing, extremely low frequency magnetic elds (ELF-EMF) as an alternative approach for treating cancer is rapidly rising nowadays. Nevertheless, little is known about the underlying anti-tumoral mechanism of action of them. Methods: In the present study, for the rst time, we reported that along with apoptosis, 2 h/day exposure to 100 Hz, 1 mT ELF-EMF for a 5-day period, can induce necroptosis, a specic type of programed necrotic cell death, by promoting RIPK1/RIPK3/MLKL pathway which may also be responsible for observed pro-inammatory responses in vivo, evident from an increase in plasma levels of proinammatory cytokines including TNF-α, IL-1β, IL-2, IL-6, IL-17A and IFN-γ. Alongside, 30-day exposure to this system could also signicantly suppress tumor growth and expression of markers of tumor cell proliferation, angiogenesis, and metastasis, namely Ki-67, CD31, VEGFR2 and MMP-9. Results: The number of tumor inltrating lymphocytes (TILs), especially CD8+ T h cells were signicantly increased following exposure to ELF-EMF. Interestingly, pretreating cancer cells with N-acetyl cysteine, a free-radical scavenger, or verapamil, an L-type calcium channel blocker in vitro, could diminish observed necroptotic and apoptotic responses while pretreating with calcium chloride, could aggravate responses. Conclusions: Overall, results of present study demonstrated that along with apoptosis, necroptosis is also a prominent form of cell death induced by exposure to ELF-EMF which is also dependent on elevated intracellular levels of ROS and calcium.

a prominent form of cell death induced by exposure to ELF-EMF which is also dependent on elevated intracellular levels of ROS and calcium.

Background
Although modulatory effects of electromagnetic elds (EMFs) on living cells and tissues has long been documented in vitro and in vivo, our understanding from mechanisms of these interactions still remains in its infancy and far from being completely understood [1,2]. In this context, recent studies on extremely low frequency electromagnetic elds (ELF-EMFs), produced by several household devices and industrial networks, have shown that these non-ionizing waves can affect multiple biological processes in living matter including expression of genes and proteins, as well as modulating physiological pathways governing cellular proliferation, differentiation or integrity either in direct or indirect ways [3][4][5][6][7].
Alongside, the therapeutic potential of ELF-EMF in treatment of cancer has long been documented in vitro and in vivo [8][9][10][11]. Regardless of few exceptions, ELF-EMF exposure promotes programmed cell death of cancer cells in vitro. Similarly, a tumor retardation response has been demonstrated in tumor bearing mice following exposure to ELF-EMF in vivo [8,[12][13][14][15]. More importantly, the advantage of this approach is its low damaging effects toward adjacent normal tissues. Therefore, many scientists have underscored the potential of ELF-EMF to be applied as an adjunctive therapy or even a primary therapeutic modality in treatment of cancer. Unfortunately, diversities in applied magnetic elds strengths, in addition to the lack of a generally accepted mechanism of action, has made it di cult to compare the results of exposure to ELF-EMF in different studies and come up with a unit treatment protocol.
Previously reported by Tatarov et al., 6 h/day exposure to a 100 mT, 1 Hz magnetic eld for a total of 4week period could successfully suppress breast tumor growth in nude mice [16]. Thus, in present study, was applied the same electromagnetic eld to identify types of cancer cell's responses to ELF-EMF and studying underlying molecular and cellular pathways. Currently, induction of "apoptosis" is the main proposed explanation for anti-tumorigenic effects observed following exposure to ELF-EMF [17]. Apoptosis is a "caspase" dependent type of programmed cell death which is typically together with antiin ammatory responses [18][19][20]. Very oddly however, we noticed a vigorous pro-in ammatory response in vivo with a very signi cant increase in serum TNF-α concentration following 6 h/day ELF-EMF exposure at the end of 40-day period which could not be simply explained by classical apoptotic pathway's characteristics. Similar results were also recorded when daily exposure time was reduced to 2 h/day though with less intensities. As these responses are typical characteristics of another type of programmed cell death, namely "necroptosis", we further investigated the possibility of necroptosis occurrence in response to ELF-EMF exposure. Finally, as ELF-EMF has shown to affect reactive oxygen species (ROS) formation and perturb intracellular calcium levels, we further investigated how modulation of these two factors can affect induction of necroptosis using their speci c modulators.

Animals
Adult inbred Balb/C mice (~20-25 g) were obtained from Pasteur Institute of Iran (Tehran, Iran) and were housed in groups of ve to six under regular 12 h light/dark cycles. Mice were allowed to access water and food freely and whole experiment was carried out between 10.00 A.M to 13 P.M each day using regular room light and temperature. All animal experiments were performed according to the ARRIVE guidelines and National Institutes of Health guide for the care and use of Laboratory animals (NIH Publications No. 8023, revised 1978). At the end of experiment, all mice were sacri ced by cervical dislocation. The Whole experiment, design of the study and included techniques were approved by the ethics committee of the Motamed Cancer Institute, ACECR, Iran.
EMF exposure systems ELF-EMF exposure system applied herein was identical to the one previously described by our group [22].
The exposure device consisted of two couples of square coils in a U-shape solenoid like con guration and an ironic core which guaranties required homogeneity and maximizes B-eld levels. Each mentioned coil has a side length equal to 2cm that is covered with 1250 single-strand copper wires. Coils were positioned in a way that the shortest distance among their ends was equal to 15 cm. In order to obtain a tunable magnetic eld within the range of 1-100 mT in intensity and 1-60 Hz in frequency, a wave form generator (GW Instek SFG-1000 Series, South Korea) was coupled to the coils. To minimize possible effects of ELF-EMF on control group, the incubator of ELF-EMF non-treated group (control group) was located in a place which received EMFs similar to those of background. Using a scope coupled parallel to a resistor in series with coils, waveforms, frequency and magnetic eld intensity B were carefully monitored. The inhomogeneity close to the ironic plates edge was about 10%. Using a Lakeshore ® Gauss meter, point by point calibration of the net-like plastic plate located at the bottom of exposure zone was performed and an 8 × 8 cm square at central part was determined as exposure zone where eld alterations were less than 2%.

In vitro studies
Cell proliferation assay Throughout the study, cells in passage number 3-4 were applied to minimize variations induced by different cellular characteristics. Cells were detached from asks with the use of 0.025% Trypsin-EDTA after reaching an 80% con uency, and were then neutralized using 10% FBS and centrifugated. Following dispersion of pellets, Cells were seeded in 6-well plates with an initial density of 5×10 4 cells/well. Cells were exposed to 100 mT, 1 Hz magnetic elds daily for 2 h. Exposures continued for 5 consecutive days, the time equal to 2-3 population doubling time, and at the end of the fth day, cells were detached using 0.025% Trypsin-EDTA and cell viability was measured according to trypan blue staining protocol. The percent of cell inhibitory effect following ELF-EMF treatment was calculated by dividing the differences between counted cells of sham and ELF-EMF exposed groups by total number of cells grown in sham group (formula 1):

Percent of cell growth inhibition = [(Sham Group Number of cells -ELF-EMF Group Number of cells )/ Sham Group Number of cells ] × 100
In some performed studies, in order to evaluate the role of intracellular ROS in ELF-EMF induced antiproliferative effects, cells were pretreated with 2.5 mM N-acetyl cysteine (NAC), a potent scavenger and inhibitor of intracellular ROS accumulation 30 min prior to daily ELF-EMF exposures. In addition, to evaluate the role of Ca 2+ concentrations on ELF-EMF induced effects, cells were pretreated with 5 mM calcium chloride, a precursor of calcium, or 10 µM verapamil, a potent inhibitor of L-type voltage gated channels 30 min prior to daily ELF-EMF exposure. Dosages applied in present study were chosen according to a preliminary screening test and were those which could induce desired biological effects without presenting signi cant cytotoxicity.

Flowcytometry analysis of programmed cell death
In order to evaluate probable induction of programmed cell death following exposure to ELF-EMF, MC4-L2 cell lines were seeded at a density of 5 × 10 5 cells/well in 6-well plates and exposed to 100 mT, 1Hz ELF-EMFs 2h daily for 5 days. At the fth day, cells were collected immediately after exposure and stained with annexin V/ propidium iodide kit prior to analysis on a Beckman Coulter FC500 cytometer. Similarly, for analyzing the role of ROS or Ca 2+ on ELF-EMF mediated effects, cells were pretreated with reagents mentioned in previous section 30 min prior to daily exposures.

Intracellular reactive oxygen species detection
To evaluate changes in intracellular levels of reactive oxygen species (ROS) following exposure to ELF-EMF, cells were stained with 2′, 7′-dichloro uorescin diacetate (DCFH-DA, Sigma, US) uorescent dye and then analyzed by owcytometry. In brief, following ELF-EMF exposure, cell media were replaced with prewarmed serum-free media containing 5 μM DCFH-DA and incubation continued for 30 min. Cells were then washed three times with PBS to remove extra DCFH-DA and detached using 0.25% trypsin-EDTA for uorescent intensity measurement by owcytometry. For each experiment, 10 4 cells were counted and read under ow cytometer. To obtain reliable results, each experiment was repeated at least three times.
In some experiments, effect of NAC, calcium chloride and verapamil pretreatment on cellular ROS levels following exposure to ELF-EMF was also studied.

Fluorescence microscopic detection of calcium in ux
The effects of ELF-EMF exposure on intracellular calcium levels was studied using FURA4-AM uorescent die. In brief, on the fth day of treatment, cells were incubated with FURA4-AM at a nal concentration of 1 μM, 15 min prior to exposure to ELF-EMF. In some experiments, cells were also treated with NAC, calcium chloride or verapamil following incubation with FURA4-AM for 30 min prior to ELF-EMF exposure.
Before ELF-EMF exposure, the uorescence background associated with cells were recorded under an LSM510 inverted uorescence microscope. At the end of the 2h exposure period, cells were again imaged and uorescent intensities were determined using ImageJ software. Data were reported as mean± SD of arbitrary units of three independent experiments performed on different days.

Acridine orange staining
To evaluate the increase in number of AVOs, a hallmark of induction of autophagy, following exposure to ELF-EMF, MC-4L2 cells were stained with acridine orange uorescent dye. In some experiments, cells were also treated with NAC, calcium chloride or verapamil 30 min prior to ELF-EMF exposure. In brief, following ELF-EMF exposure, cells were incubated with acridine orange with a nal concentration of 5 μg/ml for 15 min and then green and red uorescence were visualized utilizing an LSM510 inverted uorescence microscope. Intensity of red uorescence, as a response to AO accumulation in AVOs, was measured using ImageJ software and reported as mean ± SD of intensity of 3 independent experiments.

In vivo studies
In vivo EMF exposure experiments in mice 14 mice were randomly divided in to two groups, each containing 7 mice, one receiving 100 mT, 1 Hz ELF-EMF 2h daily for a period of 28-day and the other receiving no treatment, serving as control group. For development of MC-4L2 tumors, 10 6 cells were subcutaneously injected to right ank of one mouse and allowed to growth for 3 weeks. Next, established tumor was dissected and minced in to several 1-2 mm 3 pieces under sterile condition. Each section was then subcutaneously grafted in to one mouse and allowed to grow up to a mean size of 150 mm 3 . Following this point, exposures were initiated and continued up to 28 days. Mice were routinely monitored for possible cytotoxicity during the 28-day period and weighed every day using a digital balance. Tumor sizes were also measured every day using a digital caliper and tumor sizes were calculated using following formula: At the end of the 28-day treatment period, mice were sacri ced, their tumors were dissected, weighed and xed in 4% paraformaldehyde for further analysis. The graph of tumor size growth as the function of time was drawn for both groups and compared with each other. Weight of tumors at the end of 28-day treatment period was also measure in both groups and compared with each other.

H&E staining and Immunohistochemistry
Tumors were dissected from mice bearing MC4-L2 tumors and xed in 4% paraformaldehyde overnight. Fixed tumors were then dehydrated in an alcohol gradient, followed by para n-embedment and sectioning in to slices with a thickness of 4 µm. Following depara nization in xylene and rehydration, Hematoxylin and eosin (H&E) staining and immunohistochemistry (IHC) were performed based on a established protocol [23]. Primary antibodies utilized herein included anti-VEGFR2 (#2479L, CST), anti-MMP9 (#13667T, CST), anti-Ki67 (#9449T, CST), anti-CD31+ (#3528S, CST), anti-CD8+ (#85336S, CST) and anti-CD4+ (sc-13573 L, Santa Cruz Biotechnology). Following an overnight incubation with primary antibodies, sections were carefully washed and incubated for about an hour with secondary antibody at room temperature. Sections were then visualized through staining with diaminobenzidine and nucleus counterstaining with hematoxylin. Images were taken from sections using an Olympus light microscope. Tumor in ltrating lymphocytes (TILs) number in H&E stained sections was also counted by an expert pathologist.

TUNEL assay
Following embedment in para n, tumor tissues were sectioned in to 4 µm slices and subjected to TUNEL assay using a commercially available detection kit (Biovision, Canada) according to the manufacturer's provided guideline. Following this step, each section was stained with diaminobenzidine and counterstained with hematoxylin. Images were taken using an LSM510 uorescence microscope.

Multi-Analyte ELISArray for analyzing serum levels of in ammatory cytokine
At the end of the treatment period, mice blood samples were collected by heart puncture. Serum samples were immediately separated by centrifugation and stored at -80 °C until further analysis. A MEM-004A Multi-Analyte ELISArray Kit (Qiagen, US) was used for analyzing 12 in ammatory cytokines including IL-1α, IL-1β, IL-2, IL-4, IL-6, IL-10, IL-12, IL-17A, TNF-α, IFNγ, G-CSF and GM-CSF. Procedure was performed according to the manufacturer's provided instructions and data were represented as fold changes compared to non-treated control group.

Statistical analysis
For evaluation of statistical differences between two group, unpaired Student's t-test and for comparison of more than two groups one-way analysis of variances (ANOVA) with tukey's post hoc test was applied. Throughout the study, data were expressed as the mean ± the standard deviation of the mean (SD). Differences were considered signi cant when P < 0.05.

Results
Exposure to ELF-EMF induced a pro-in ammatory response in vivo resembling those of necroptosis As depicted in Fig. 1A, both 2 h and 6 h/day exposure to ELF-EMF could effectively suppress tumor growth rate and nal weight compared to non-exposed mice. Nevertheless, no signi cant differences in rate of tumor growth was observable between 2 h-and 6 h/day exposed groups at the end of 40-day treatment period which was contrary to the previous report by Tatarov et al. [16] Moreover, represented in Fig. 1B, TUNEL positive cells in ELF-exposed group were signi cantly higher in number compared to the non-treated group. Again, TUNEL positive cells in 2-and 6 h/day exposed group were not signi cantly different in number with each other. Strangely, in addition to classical highly dense bright blue spots, detectible in nuclei of apoptotic cells (representative of condensed DNA), an atypical smeared form of DNA with no changes in the nuclear gure or size and density of blue color was also observable in TUNEL positive cells. Whereas positive cells in TUNEL assay are usually considered as apoptotic cells due to the presence of fragmented DNA, different studies have demonstrated that necrotic cells, either in its programmed form or the other, also possess fragmented DNA which can interact with TUNEL in vivo, and thus, can be considered as another marker for induction of a necroptotic response in vivo [24]. Also, examining H&E stained sections of ELF treated and non-treated groups, a signi cantly raised number of tumor in ltrated cells (TILs) was observable in both 2 h-and 6 h-/day exposed group compared to the control (Fig. 1C).
Identifying possible changes in pro le of 12 pro-/anti-in ammatory cytokines in collected blood samples at the end of the treatment period, a vigorous pro-in ammatory response, evident from a very noticeable increase in TNF-α, IL-6 and IL-1β concentrations ( ~ × 3, × 2.7 and × 3 fold increase respectively) was detected following both 2 h and 6 h/day exposures to 100 mT, 1 Hz ELF-EMF ( Fig. 2A). Similarly, concentrations of other pro-in ammatory cytokines including IL-2, IL-12 and IL-17A were also signi cantly raised. Notably, concentration of IFN-γ, an important cytokine with tumor suppressive effects, was also signi cantly raised while the concentration of pleiotropic anti-in ammatory cytokine IL-4 remained almost unchanged. This pro-in ammatory response was highly representative of those observed during necroptosis in vivo [25]. Based on previous reports, increased TNF-α concentration induces expression of several pro-in ammatory cytokines including IL-6, IL-1β, IL-17, etc. during a necroptotic response.
In parallel, secretion of speci c molecules containing DAMPs from cell undergoing necroptosis and production of other cytokines including IL-1α react with neighboring cells and promotes antigen presenting cells (APC) recruitment and activation [26]. This further promotes initiation of an effective adaptive immune response. Consistent with these ndings, as depicted in Fig. 2B, the number of tumor in ltrated lymphocytes was signi cantly increased compared to non-treated cells and the proportion of CD8 + T cells was signi cantly increased compared to the non-treated group which can be attributed to the speci c pattern of secreted cytokines. These evidences, further potentiate the possibility of induction of a necroptotic response following ELF-EMF exposure.
Interestingly, despite of the signi cant increase in TNF-α and several other pro-angiogenic cytokines levels, the effect of treatment with ELF-EMF on MMP-9 and VEGFR2 expression, as well as the number of CD31 + cell in tumor site was dwindling, proposing a net anti-angiogenic effect on exposed mice (Fig. 2B).
Similarly, previous reports have also demonstrated that Ripk3 −/− mice undergo similar changes and the process of cutaneous wound healing becomes signi cantly impaired in them. What can be possibly proposed herein, is that very high concentration of TNF-α at site of action may have resulted in downregulation of these receptors and had led to following consequences.

ELF-EMF exposure induces necroptosis and apoptosis programmed cell deaths in vitro
To further con rm the hypothesis of induction of necroptotic cell death by ELF-EMF, different in vitro cellular studies were performed. Mimicking chronically exposed condition in vivo, cells were exposed to the same ELF-EMF for 5 days, equivalent to about 3 population duplicating time of cells in vitro [27]. It is almost accepted that ELF-EMF mediated biological effects are linked to calcium overload and ROS overproduction [28]. Consequently, we also examined whether these two factors could be involved in necroptotic like responses mediated by ELF-EMF exposure. To this end, in some experiments, cells were also treated with H 2 O 2 , a source for overproduction of ROS; n-acetyl cysteine (NAC), a potent ROS scavenger; verapamil, a potent L-type calcium (Ca 2+ ) channel blocker; or calcium chloride, a source for raised extracellular Ca 2+ concurrent to the ELF-EMF exposure.
Initially, inhibitory effects of ELF-EMF exposure on MC4-L2 cell proliferation was evaluated by trypan blue staining. As depicted in Fig. 6A, exposure to ELF-EMF could signi cantly reduce the number of live cells compared to non-treated group. Interestingly, these inhibitory effects were completely abolished upon cotreatment of cells with NAC or verapamil while excess extracellular Ca 2+ could potentiate anti-proliferative effects of ELF-EMF. In parallel, cells were also stained with annexin V and PI, and analyzed with owcytometry. Generally, in necroptotic cell death, phosphatidyl serine is externalized prior to cell membrane penetration and secretion of necroptotic bodies [29]. Furthermore, similar to necrosis, PI can also internalize through disintegrated membrane of necroptotic cell and stain DNA. Thus, full necroptotic response can be understood from summing up PI (+)/AV (-) population, representing cells undergoing necrosis and PI (+)/AV (+) population, representing cells undergoing late necroptosis. However, the later population can also account for cells undergoing late apoptosis. Thus, AV/PI staining is not a highly selective test for differentiating apoptosis from necrosis. Despite this, during apoptosis, an increase in both early and late apoptotic subpopulations is expected and if late apoptosis is suddenly raised while early apoptosis remained intact means that other forms of programmed cell death including necroptosis may also be involved. As depicted in Fig. 6A, ELF-EMF treatment could signi cantly increase double AV/ PI positive and AV positive PI negative subpopulation of cells which is indicative of induction of apoptosis, necroptosis or both. Despite this, only a trivial increase in number of PI-/AV + cells was observable following exposure to ELF which is highly representative of necroptotic cell death rather than apoptotic one. Similarly, co-treatment with NAC and verapamil could signi cantly reduce number of cell deaths while excess Ca 2+ could aggravate the results.
To further illustrate the nature of the programmed cell death, western blot analysis was performed for identi cation of changes induced in protein levels of pro-caspase 3, cleaved caspase 3, procaspase 9, cleaved caspase 9 and procaspase 8, cleaved caspase 8, following exposure to ELF-EMF. As illustrated in Fig. 3, concentrations of cleaved caspase − 9, -3 and − 8 were signi cantly increased following ELF-EMF exposure, highly proposing that ELF-EMF induces apoptosis through both intrinsic and extrinsic pathways. Interestingly, co-treating cells with verapamil could signi cantly reduce amounts of cleaved caspase − 3, -8 -9 upon exposure to ELF while co-treating cells with NAC could only signi cantly suppress cleavage of caspase − 9 while amounts of cleaved caspase − 3 and − 9 remained almost unchanged. With this data in mind, one can hypothesize that ROS overproduction upon ELF exposure is part of the scenario behind ELF-EMF exposure induced apoptosis and rather, calcium overload, as antagonized upon co-treating with verapamil may be the prevalent event in induction of apoptosis. Oddly however, upon introducing excess calcium into the extracellular environment and exposing to ELF-EMF, no signi cant enhancement in cleaved caspases − 3, -9 and − 8 in comparison to ELF-EMF exposure alone was observed while owcytometry analysis demonstrated a signi cant increase in number of late apoptotic population upon co-treating with calcium chloride compared to ELF-EMF alone. Thus, calcium overload may have increased number of AV+/PI + population with a cell death pattern independent from apoptosis ( Fig. 6A). As necroptosis can also increase the number of cells in this subpopulation and considering data obtained from in vivo studies, we hypothesized that ELF-EMF exposure may induce calcium overload with in turn, can trigger cell death by necroptosis. This hypothesis become stronger considering previous reports in literature, proposing necroptosis as an event strongly dependent on calcium overload together with manifestation of ROS overproduction. Recently it has been shown that ROS overproduction may result in auto phosphorylation of RIPK-1, the initiator of necroptosis cascade [30]. It seems that exposure to ELF-EMF could result in similar cascade of events.
To con rm occurrence of necroptosis as another form of programmed cell death following ELF-EMF exposure, fold changes in protein levels of RIPK1, p-RIPK1, RIPK3, p-RIPK3, MLKL and p-MLKL was examined following exposure to ELF-EMF and co-treatment with calcium chloride, NAC and verapamil. As was expected, levels of p-RIPK1, p-RIPK3 and p-MLKL were signi cantly raised following ELF-EMF exposure alone and calcium chloride co-treated group (Fig. 4) while no signi cant changes were observable in other treatment groups. Based on these data, we concluded that calcium overload following ELF-EMF exposure can induce necroptotic cell death in addition to apoptosis.

Elf-emf Exposure Induces Necroptotic-like Morphological Alterations In Mc4-l2 Cells
To further illustrate the speci c mode of cell death in MC4-L2 cell lines following ELF-EMF exposure, changes in cellular morphology was investigated by transmission electron microscopy (TEM). TEM images obtained from normal MC4-L2 cells demonstrated well-preserved cytoplasmic organelles, as well as smoothly shaped nuclei encompassing heterochromatin shaped chromatin (Fig. 5). Contrarily, H 2 O 2 treated cells were used as apoptosis control group. This group of cells were mostly characterized by a marginated chromatin and formation of budding vesicles in the absence of swollen organelles. Upon examining ELF-EMF treated cells, in addition to the cells with typical apoptotic morphology, a large part of the cells demonstrated a translucent cytoplasm (yellow arrows), associated with disintegrated membrane (Blue arrow), organelle swelling (Red arrow) and condensed chromatin (Green arrow); all of which are typical characteristics of a necroptotic like cell death [31]. The most prevalent form of morphological changes in other groups was either apoptotic form (NAC co-treated) or normal form (NAC control, verapamil control and verapamil co-treated groups).
ROS overproduction and Ca 2+ overload interplay following ELF-EMF exposure is responsible for necroptotic events in MC4-L2 cell lines in vitro DCHF-DA and FURA-4AM staining were used to evaluate possible enrolment of ROS and calcium in induction of ELF-EMF mediated necroptosis respectively. The mean uorescent intensity of each group in DCHF-DA experiment was measured by owcytometry and for FURA-4AM experiment was obtained by ImageJ software. Based on results, ELF-EMF exposure could signi cantly increase both ROS and intracellular Ca 2+ levels in treated cells (Fig. 6B and D). Interestingly, while NAC co-treatment could only suppress ELF-EMF mediated ROS overproduction, calcium chloride co-treatment could enhance both intracellular overproduction of ROS and calcium overload simultaneously. These ndings were extremely helpful in understanding the interconnection of ROS and calcium in development of ELF-EMF mediated necroptosis as blockade of ROS overproduction in the presence of intracellular calcium overload could suppress development of necroptosis while induction of ROS overproduction and calcium overload by calcium chloride co-treatment could induce necroptosis without having net effect on apoptosis.
Furthermore, blockade of both calcium and ROS overload by verapamil could completely diminish both necroptosis and apoptosis induced by ELF-EMF. This strongly suggests that increase in intracellular calcium levels may happen following elevation of ROS overproduction.

ELF-EMF exposure induces autophagic response which is dependent on elevation of intracellular ROS concentration
To evaluate autophagic responses following ELF-EMF exposure, cells were stained with AO/PI uorescent dyes. As depicted in Fig. 6C, exposure to ELF-EMF resulted in a signi cant increase in number of acidic vesicular organelles including lysosomes and autophagosomes which are depicted by orange dots in cytoplasmic region of cells. To evaluate the putative role of ROS and Ca2 + in autophagic response, cells were again pretreated with NAC, calcium chloride and verapamil. Interestingly, only administration of NAC could signi cantly reduce the number of autophagosomes and no signi cant differences were observed in autophagosome content of cells pretreated with calcium chloride and verapamil following exposure to ELF-EMF in comparison to cells which were only treated with ELF-EMF. Thus, it can be concluded that elevated intracellular ROS play a more important role in induction of autophagy in response to exposure to ELF-EMF.

ELF-EMF exposure induced caspase-independent program cell death is not in association with autophagic response
Comparing data obtained from autophagy responses with those related to caspase-independent program cell death didn't demonstrate a direct relation between the intensity of autophagic response and the extension of caspase independent programmed cell death observed. The important point noticed herein was that although administration of verapamil couldn't signi cantly decrease observed autophagic response, it could completely reverse caspase independent programmed cell death. Despite of a clear relation between increased autophagy and induction of necroptosis, this pattern didn't happen in our case.

Discussion
To our best knowledge, this is the rst study providing evidence for ELF-EMF mediated necroptotic cell death, which is different from currently accepted apoptosis hypothesis worldwide. Necroptotic cell death was initially suspected from vigorous in ammatory responses and appraisal of TNF-α concentration in vivo; beyond secretion of DAMPs as a result of cytoplasmic membrane disintegration, necroptosis have shown to promote pro-in ammatory responses through activation of expression of a vast number of genes. In this context, large part of up-regulated genes, including pro-in ammatory cytokines, are similar to those activated by TNFα pathway but at higher levels. Overexpression of pro-in ammatory cytokines seems to be a common characteristic of necroptosis as necroptosis stimulators other than TNFα including TRAIL can also signi cantly increase their expression. Finally, it has been suggested that overproduction of cytokines in TNFα-induced necroptosis takes place in a manner independent from DAMPs release and is highly dependent on NF-κB and p38 activation [32].
Necroptosis hypothesis was further con rmed by observation of a signi cant increase in phosphorylated levels of RIPK1, RIPK3 and MLKL proteins in parallel to a negligible change in cleaved caspase-3 level following exposure to ELF-EMF. This was concurrently associated with a signi cant increase in formation of autophagosomes and induction of autophagy, evident from an increase in the intensity of orange color emitted by acridine orange upon accumulation in acidic vacuoles. Interestingly, no signi cant changes in autophagic responses were observable following pretreatment with calcium chloride and verapamil in exposed cells while NAC could signi cantly attenuate autophagy both at the baseline and following ELF- necroptotic cell death and activation of caspases (presumably caspase 8) can suppress autophagic cell death [33]. Other studies have shown that T-cell activation results in formation of a DISC-like complex formed from assembly of FADD caspase-8 and RIPK1 on autophagosomic membranes [34]. This model strongly suggests that a certain level of autophagy is necessary for physiological proliferation of T-cells and manipulation of each of these components can affect proliferation. For instance, deletion of FADD and caspase-8 has shown to enhance autophagy and inhibit proliferation of T cells while inhibition of RIPK1 could restore proliferation of T-cells lacking normal FADD functionality. Recently it has been discovered that knocking out RIPK3 can also restore T cell proliferation [35]. Thus, it is possible that unregulated necroptosis and autophagy may induce T-cell death during clonal expansion and regulation of caspase 8 may have an important role in modulation of T-cell proliferation.
Monitoring intracellular calcium concentrations in different exposed groups, we demonstrated that cytoplasmic Ca 2+ overload per se is not effective in induction of necroptosis. Consistent with numerous previously existing reports, pointing out enhancing effects of ELF-EMF on intracellular Ca 2+ concentration, here we observed an increase in Ca 2+ concentration following ELF-EMF exposure. Among co-treatment strategies for understanding the main players of ELF-EMF mediated necroptosis, only calcium chloride could enhance necroptosis which upon molecular examination, could signi cantly rise both intracellular ROS and calcium levels. Since NAC co-treatment could signi cantly reduce necroptosis event, evident from reduced p-MLKL compared to ELF-EMF group in western blotting and also owcytometric analysis of AV+/PI + and AV-/PI + subpopulations, one can conclude that existence of ROS overproduction is essential in parallel to calcium overload for induction of necroptosis.
Different studies have pointed out that the role of ROS overproduction in induction of necroptosis may be cell dependent. For instance, it has been shown that in mouse embryonic broblast cells, TNF-α and BV6 induced necroptosis mainly takes place by RIP1-induced accumulation of ROS [36]. Similarly, based on Vanlangenaker et al. the main mechanism of Inhibitor of Apoptosis (cIAP)1induced anti-necroptotic event is suppression of RIP1-/RIP3-induced ROS overproduction [37]. Also, in an animal model of tuberculosis infection, TNF-α could accelerate mitochondrial ROS overproduction in mycobacteria infected macrophages by activating RIPK1/ RIPK3 pathway [38]. Contrarily however, in HT29 colon carcinoma cells or human monocyte-derived cell line THP-1, application of ROS scavengers couldn't reverse TNF-α induced necroptosis [39,40]. Based on our results, necroptosis in MC4-L2 cell line may be activated upon accumulation of overproduced ROS in response to ELF-EMF exposure. Furthermore, overproduction of TNF-α in vivo upon exposure to ELF-EMF and accumulation of intracellular ROS may have been the underlying mechanism of induction of necroptosis. However, further studies for clari cation of the underlying mechanism is highly recommended.
Calcium mediated necroptosis has also been discovered in Caenorhabditis elegans. In this context, hyperactivating MEC-4 ion channel by mutation, results in in ux and accumulation of calcium up to a toxic level in C. elegans which can in turn, induce necroptotic neurodegeneration. Similarly, treating L929 cell lines with TNF-α results in appraisal of cytoplasmic calcium and induction of necroptosis [41,42]. In addition, application of calcium chelators in L929 cells can completely diminish necroptosis, proposing an executive role for calcium in necroptosis [43]. Studies on HT29 cells have also demonstrated that following induction of necroptosis by TNF-α treatment, translocation of MLKL to the cytoplasmic membrane induces a vigorous calcium in ux through the TPRM7, a non-voltage-sensitive ion channel ion channel, that is an early characteristic of necroptosis [44]. Notably, this ion channel functions downstream of MLKL and its knocking down doesn't have any effect on MLKL phosphorylation and trimerization. Consistently, here we demonstrated that application of verapamil, an antagonist of voltage gated calcium channels could diminish necroptosis while co-treating with calcium chloride aggravated it. Possible effect of verapamil may be the lessening of net calcium in ux and calcium chloride effect may be potentiation of calcium in ux through TPRM7 ion channels. However, further studies are highly recommended for proving correctness of this statement.

Conclusion
Overall, herein we con rmed that ELF-EMF exposure can induce tumor suppressing effects in vivo which is characterized by a signi cant increase in pro-in ammatory cytokines, most importantly TNF-α. In vitro studies also con rmed that ELF-EMF treatment induced cell death is not merely mediated by apoptosis.
Necroptosis is another form of programmed cell death which is activated by ELF-EMF exposure and may be the underlying reason for the signi cant pro-in ammatory responses observed in vivo. Furthermore, in the context of MC4-L2 cell lines, both ROS accumulation and calcium overload are critical for induction of ELF-EMF mediated necroptosis. Finally, as recent studies have shown that necroptosis may be a good solution for treatment of cell lines resistant to chemotherapies, application of ELF-EMF in treatment of chemo-resistant cancers may become of great interest.

Availability of data and material
Not applicable.

Competing interests
The authors declare no competing interests.  2h-and 6h/day exposure to ELF-EMF for a 30-day period could effectively reduce tumor growth rate. No differences however were not observable in tumor growth rate between two treated groups. Right panel: tumor weights at the end of treatment period was signi cantly lower in ELF exposed group compared to the non-exposed control ones. (B) Tunnel assay, demonstrating a signi cantly higher number of cells with fragmented DNA (green uorescent) for both ELF-EMF exposed groups compared to their non-treated counterparts. Note the differences between the shapes of nuclei in ELF-EMF treated groups (low brightness, not condensation and presence in smeared form) and their controls (highly condensed blue color). (C) H&E staining of tumor sections at the end of treatment period, demonstrating a signi cantly increased number of TILs in tumor site compared to nonexposed specimen.

Figure 2
Exposure to ELF-EMF induces a pro-in ammatory response in vivo. (A) examining pro le of 12 pro/antiin ammatory cytokines using ELISA technique demonstrated a signi cant increase in majority of in ammatory cytokines, including IFN-γ (5× folds), TNF-α (3×folds), IL-6 (×3 fold) and IL-1β (×3 folds) in ELF-EMF exposed groups. Note that 6h exposure could more effectively induce in ammatory responses compared to the 2h/day treatment regimen. (B) IHC staining of tumor sections also demonstrated a signi cant decline in expression of angiogenesis (VEGFR2, MMP9 and CD31), proliferation markers (Ki-67). Contrarily the number of CD8+ cells signi cantly raised following ELF-EMF exposure to ELF-EMF treatment. Also, exposure to ELF-EMF treatment reduced ER and Her2 expression levels at the end of treatment, while results of 2h and 6h/day exposure was con icting on PR expression. These results highly suggest the importance of patients monitoring during the time of ELF-EMF adjuvant therapy. ELF-EMF exposure can induce apoptotic cell death in vitro. ELF-EMF exposure signi cantly increased concentrations of cleaved caspase -9, -3 and -8, proposing that ELF-EMF can induce apoptosis through both intrinsic and extrinsic pathways. Meanwhile, co-treating cells with verapamil could signi cantly reduce amounts of cleaved caspase -3, -8 -9 upon exposure to ELF while co-treating cells with NAC could only suppress cleavage of caspase -9. Finally, introducing excess calcium into the extracellular Page 24/27 environment of cells and exposing them to ELF-EMF, did not signi cantly enhance amounts of cleaved caspases -3, -9 and -8 in comparison to ELF-EMF exposure alone. ELF-EMF exposure can induce necroptosis in vitro. Levels of p-RIPK1, p-RIPK3 and p-MLKL were only signi cantly raised following ELF-EMF exposed and calcium chloride co-treated groups while no signi cant changes were observable in other treatment groups. Co-treatment with verapamil and NAC could completely reverse appraisal of these proteins phosphorylation upon exposure to ELF-EMF. These data highly suggests that ELF-EMF can induce necroptosis which can become aggravated by calcium overload.