Synergistic Effect of H2O2 and NO2 in Cell Death Induced by Cold Atmospheric He Plasma

Abstract

Cold atmospheric pressure plasmas (CAPPs) have emerged over the last decade as a new promising therapy to fight cancer. CAPPs' antitumor activity is primarily due to the delivery of reactive oxygen and nitrogen species (RONS), but the precise determination of the constituents linked to this anticancer process remains to be done. In the present study, using a micro-plasma jet produced in helium (He), we demonstrate that the concentration of H₂ O₂ , NO₂^- and NO₃^- can fully account for the majority of RONS produced in plasma-activated buffer. The role of these species on the viability of normal and tumour cell lines was investigated. Although the degree of sensitivity to H₂ O₂ is cell-type dependent, we show that H₂ O₂ alone cannot account for the toxicity of He plasma. Indeed, NO₂^- , but not NO₃^- , acts in synergy with H₂ O₂ to enhance cell death in normal and tumour cell lines to a level similar to that observed after plasma treatment. Our findings suggest that the efficiency of plasma treatment strongly depends on the combination of H₂ O₂ and NO₂^- in determined concentrations. We also show that the interaction of the He plasma jet with the ambient air is required to generate NO₂^- and NO₃^- in solution.

O and NO

R

A

P

Girard, Arbabian Fleury, Grard Bauville, Vincent , Dutreix,

O, NO and NO

O

O , acts in synergy O

O and NO

and NO

Cancer is a leading cause of death worldwide and its incidence rate increases with the age of the population, the exposure to carcinogens and the modern lifestyle of the population. About two thirds of patients defeat their disease, and the combined action of surgery, radiotherapy and chemotherapy accounts for most cured cases1.

Alongside with these classical therapies, new therapies have emerged, such as anti-angiogenic therapy and immunotherapy1. However, therapy resistance has been observed with every type of therapy that is available today, including poly-chemotherapy, radiotherapy, immunotherapy, and molecular targeted therapy2. Importantly, sequencing of primary tumors has revealed that therapy-resistant clones already exist prior to targeted therapy, demonstrating that tumor heterogeneity in primary tumors confers a mechanism for inherent therapy resistance2.

Therefore, there is still the need of a new therapy that can overcome this problem.

There are numerous publications showing that cold atmospheric pressure plasmas (CAPPs) are eective against tumour cells both in vitro and in vivo (ref. 3 and references therein). CAPPs are partially ionised gases containing a complex and reactive environment consisting of ions, electrons, free radicals, strong localised electric eld, UV radiation, and neutral molecules. CAPPs devices are classied in three categories: direct plasma sources that use the target as a counter electrode [e.g. oating electrode dielectric barrier discharge (FE-DBD)]; indirect plasma sources that do not use the target as a counter electrode (e.g. plasma jets); and hybrid plasma sources that combine the benets of direct and indirect plasma sources410. Dierent gases can be used to produce CAPPs such as Helium (He), Argon (Ar), Nitrogen (N2), ambient air, or a mixture of gases6,7. All the plasma sources developed for biomedical applications have in common that the major reactive molecules produced in CAPPs emerge when the components of the partially ionized gas (atoms, molecules, ions and electrons) interact with the molecules of the surrounding air, i.e. O2, N2 and H2O, and with the biological sample which is usually a wet surface (e.g. cells in medium)1114. Consequently, the plasma composition and the subsequent eects on cells can vary enormously depending on the plasma source, the plasma settings, the ambient conditions and the biological target12,15.

Despite this large variability in the plasma composition, it is now widely accepted that the principal mode of plasma-cell interaction is the delivery of reactive oxygen species (ROS) and reactive nitrogen species (RNS) that

Universit Paris- LPGP, CNRS, Universit Paris-Sud,

SCIENTIFIC REPORTS

1

www.nature.com/scientificreports/

Figure 1. Scheme of the plasma device. (A) Photograph of the home-made plasma jet system. (B) Schematic illustration of the plasma jet used in this study. (C) Photograph showing the interaction between the plasma jet and a solution of PBS(Ca2+/Mg2+) poured in a 12 well plate.

can be generated in or transferred into the liquid phase surrounding the biological target16,17. Both short-lived (O, OH, O2, 1O2, NO, NO2) and long-lived (H2O2, NO2, NO3, O3) species have been detected in the CAPPs but also in the plasma-treated liquids17. However, several groups have shown that the anti-cancer activity of CAPPs was as eective in vitro whether the cells, in cell culture medium or in buer solution, were directly exposed to plasma treatment, or the cell culture medium or buer solution was rst exposed to plasma treatment (so-called plasma-activated medium or plasma-stimulated medium or conditioned medium) and then added to medium-free cells14,1824. This implies that long-lived species play a major role in in vitro anti-cancer capacity of CAPPs. Indeed, several publications have shown that H2O2 (hydrogen peroxide), NO2 (nitrite) and NO3

(nitrate) are formed at concentrations ranging from M to mM in CAPP-treated solutions12,14,22,2426, H2O2 being

a central player in the cytotoxicity of CAPPs21,2729. The aim of this study was to identify the main long-lived reactive species generated in a simple buered solution by a He plasma jet operating in ambient air at low gas ow, and their contribution to the plasma-induced cell death in normal and cancer cell lines.

Normal human skin broblasts (NHSF) were kindly provided by Dr Meng-Er Huang (Institut Curie, Orsay, France) and were used at passages below 12. MRC5Vi is a SV40-transformed and immortalized cell line derived from the normal human lung broblasts MRC530. HCT116 are human colon cancer cells and Lu1205 are human melanoma cell lines. Dulbeccos modified Eagle Medium (DMEM) with 4.5 g/l glucose, L-glutamine (L-gln) 100X, penicillin-streptomycin 100 (10000 U/ml) and fetal calf serum (FCS) were from Eurobio (France). The cells were grown in DMEM containing 10% FCS, P/S 1X and L-gln 1X at 37C, 5% CO2 in an humidied atmosphere. The cells were regularly checked for mycoplasma contamination using VenorGeM

Advance Mycoplasma Detection Kit (Biovalley, France).

The plasma source used in this study is a nanosecond pulsed atmospheric pressure cold plasma micro-jet. It consists of a stainless steel needle, inserted inside a dielectric tube made of quartz. The needle is connected to a homemade high voltage generator, while the ground electrode, made of copper, is wrapped around the dielectric tube. Refer to Fig.1A,B for more details regarding the relative position of the dierent constituents and their dimensions. The plasma is created by a dielectric barrier discharge (DBD) with axial symmetry by applying high voltage pulses (amplitude of 8 kV, rise time of 280 ns and full width at half maximum of 540 ns) at a repetition rate of 20 kHz to the internal electrode31. Pure helium (Alphagaz 2 He type S11, Air Liquide, France) is injected through the needle at a ow rate of 50 or 400 sccm (cm3/min), regulated by a owmeter (GF40-SA46, Brooks instrument, ServInstrumentation, France). In these experimental conditions, the plasma jet propagates for about a 1cm through ambient air outside the quartz tube.

SCIENTIFIC REPORTS

2

www.nature.com/scientificreports/

The micro-plasma jet was set up vertically with the gas owing downwards for interaction with buered solutions covering the various cellular models adhered to the bottom of plate wells. The plasma propagated through a capillary tube, and either the plasma or its gaseous effluent entered the buer solutions with little admixture of the surrounding air (Fig.1C).

Another plasma reactor allowing the shielding of the plasma jet with a gas of pure O2 (Air Liquide, France), instead of ambient air, was used to evaluate the contribution of the gaseous environment to the toxicity of the He plasma jet. The plasma jet structure, albeit dierent, is very similar to the other one. Refer to Figure S1 for more details regarding the relative position of the dierent constituents and their dimensions. The oxygen ow was set to 5 slm and the He ow to 100 or 400sccm. The plasma is created by applying high voltage pulses (amplitude of 5.5kV, rise time of 110ns and full width at half maximum of 260ns) at a repetition rate of 20kHz to the internal electrode.

Given the dimensions of both plasma sources and the ows used, Reynolds numbers (Re) between 7 and 55 can be determined. The ows used in this study were, thus, laminar, with very similar Re for both plasma sources (e.g. for 400sccm, Re= 55 vs Re= 48 in the rst (cf. Fig.1) and second (cf. Figure S1) plasma setup, respectively).

In in vitro experiments, 1105 to 4105 cells (depending on the cell type) were seeded per well in 12-well plates and incubated for 24 to 72h so that the cells are between 50 to 70% conuent at the time of plasma treatment. For direct plasma treatment, cell culture medium was removed, the cells washed 2 times with phosphate buered saline containing 0.9 mM CaCl2 and 0.49 mM MgCl2 [called PBS(Ca2+/Mg2+) in this manuscript], and 500l of PBS(Ca2+/Mg2+) were added to the cells. The cells were then exposed to He plasma in open air for dierent times, as shown in Fig.1C. At the end of the plasma treatment, the plates were le at room temperature for 1h protected from light. For indirect plasma treatment, 500l of PBS(Ca2+/Mg2+) were added to each well of a 12-well plate and treated with He plasma, resulting in plasma-activated PBS(Ca2+/Mg2+). Parallel to that, the cell culture medium was removed from wells where cells had been incubated, the cells washed 2 times with PBS(Ca2+/Mg2+) and then exposed for 1 h to the plasma-activated PBS(Ca2+/Mg2+). In both cases (direct and indirect treatment), 2.5ml of DMEM containing 10% FCS, P/S 1X and L-gln 1X were added to the cells aer-wards, and the plates were incubated at 37C and 5% CO2 in a humidied atmosphere for 24h. No dierence in cell behaviour between cultures exposed to the gas ow and unexposed cultures has been observed. For shielding experiments, wells of 12 well plates were lled with 3ml of PBS(Ca2+/Mg2+) so that the buered solution reaches the top of the wells.

In order to determine the presence of air impurities (N2, O2, H2O) in the plasma jet outside the quartz tube, optical emission spectroscopy was performed. The light emitted by the plasma jet was collected by a 10cm focal length optical lens and its intensity detected with a 75cm focal length spectrometer (Acton SP2750 with a 1800 grooves per mm grating blazed at 500nm) coupled with a 1340 pixel detector (Pixis from Roper Scientic). The emission spectra of the molecular bands of OH (at around 309nm), N2(C) (Second Positive System at around 337 nm) and N2+(First Negative System at around 391 nm) and the atomic lines of He (at around 706 nm) and O (at around 777 nm) were recorded and normalized to the time of acquisition.

O Cells at 50 to 70% conuence in 12-well plates were washed 2 times with PBS(Ca2+/Mg2+) and then exposed to 500 l of PBS(Ca2+/Mg2+) containing increasing concentration of H2O2.

The plates were le at room temperature for 1 h protected from light. Thereaer, 2.5 ml of DMEM containing 10% FCS, P/S 1X and L-gln 1X were added to the cells, and the plates were incubated at 37C and 5% CO2 in a humidied atmosphere for 24h.

To assess for the cell viability, the cell culture medium was removed from the plates, the cells washed once with DMEM without phenol red and covered with 500l of DMEM without phenol red containing 0.5mg/ml thiazolyl blue tetrazolium bromide (MTT) (Sigma-Aldrich). The cells were incubated 23h at 37C until purple precipitate was visible. The resulting intracellular purple formazan was then solubilized in the dark for 2h in isopropanol 95%/0.4 N HCl. Spectrophotometric quantication was performed at 470nm.

O + + VO The concentration of H2O2 in untreated and plasma-treated PBS(Ca2+/Mg2+) was determined using two methods. In the rst method, H2O2 reacts with sodium orthovanadate to produce pervanadate, which is colourless32. In the second method, H2O2 reacts with titanium oxysulfate to produce pertitanic acid, which is yellow33,34. The formation of each product is detected spectrophotometrically. For the establishment of H2O2 standard curves by Na3VO4-based assay (method 1), serial dilutions of H2O2 were prepared in 500l of PBS(Ca2+/Mg2+), and Na3VO4 was added to a nal concentration of 1mM. For the establishment of H2O2 standard curves by TiOSO4-based assay (method 2), serial dilutions of H2O2 were prepared in 400 l of PBS(Ca2+/Mg2+), 15 l of 200 mM NaN3 were added and then 200 l of 2% TiOSO4 diluted in 3 M H2SO4. NaN3 is used to scavenge nitrites and other ROS that can interfere with TiOSO4. For the determination of H2O2 concentration in plasma-treated PBS(Ca2+/Mg2+) by the method 1, 500l of PBS(Ca2+/ Mg2+) containing (direct treatment) or not (indirect treatment) 1 mM Na3VO4 were exposed to He plasma for various times. For indirect treatment, Na3VO4 was added post treatment to plasma-treated PBS(Ca2+/Mg2+) from a stock solution at 200 mM. For the determination of H2O2 concentration in plasma-treated PBS(Ca2+/ Mg2+) by the method 2, 500l of PBS(Ca2+/Mg2+) were exposed to He plasma for various times. Thereaer, 400 l of plasma-treated PBS were mixed to 15l of 200mM NaN3 and 200l of 2% TiOSO4 diluted in 3M H2SO4. The samples were incubated protected from light for 30min at room temperature to allow the reaction to occur, and

SCIENTIFIC REPORTS

3

www.nature.com/scientificreports/

the absorbance was measured at 260 and 270 nm (method 1) or at 407 nm (method 2). All reagents (Na3VO4, TiOSO4, H2O2 and NaN3) were from Sigma-Aldrich.

+ + The quantication of nitrite and nitrate was performed using the nitrate/nitrite colorimetric assay kit (Cayman) according to the suppliers instructions.

+ + The pH of untreated and treated buered solutions was taken using a SevenEasy pH meter S20 tted with a InLab Micro electrode (Mettler Toledo).

All optical densities were recorded at room temperature in a double beam spectrophotometer (UVIKON XS, SECOMAM, Servilab, France) using quartz cuvettes

with a light path of 10 mm (Hellma). Known concentrations of H2O2, NaNO2 and NaNO3 were prepared in PBS(Ca2+/Mg2+). NaNO2 and NaNO3 were from Sigma-Aldrich.

Results were plotted using a Microso Excel soware as mean standard deviation. Student t-test was used to check the statistical signicance (*p<0.05, **p<0.01, ***p<0.001).

O + +

There are compelling evidences in the literature that plasma-induced liquid H2O2 plays a major role in the cellular toxicity of plasma-treated aqueous solutions14,2428. Therefore, we wanted to precisely determine the concentration of H2O2 induced by our He plasma jet in a very simple buer, phosphate buered saline (PBS) containing

Ca2+ and Mg2+, named PBS(Ca2+/Mg2+) hereaer. As the cells are exposed for one hour to PBS, we add the cations Ca2+ and Mg2+ as they contribute to maintain cell adhesion35. It has been shown that H2O2 can react in solution with Na3VO4 to yield pervanadate32. Therefore, we based our assay on a change of the absorbance of Na3VO4 upon its oxidation by H2O2. At rst, we recorded the absorbance of dierent concentrations of Na3VO4 in PBS(Ca2+/Mg2+) between 200 and 400nm, and found a concentration dependent increase of the optical density (O.D.) (Figure S2). For a concentration of 1 mM Na3VO4, the O.D. below 250 nm were higher than 3, closed to the maximum of the measurement range of the spectrophotometer (3.5). Then, we prepared solutions of 1mM Na3VO4 in 500 l of PBS(Ca2+/Mg2+) and added increasing concentrations of H2O2. We observed a concentration-dependent decrease of the absorbance of Na3VO4 in the spectral range 250320nm and a slight increase in the range 320400nm (Fig.2A). Based on these results, we focused on the change of O.D. at 260 and 270nm. By repeating the measurements several times, we obtained a linear correlation between the change of O.D. at both wavelengths and the H2O2 concentration (Fig.2B). Note that these correlations are true for concentrations of

H2O2 1mM. We then exposed 1mM Na3VO4 in PBS(Ca2+/Mg2+) to either a ow of He or a He plasma for 2 and 4min, and we recorded the absorbance of the solutions between 250 and 400nm. While the absorption spectrum of an untreated solution of 1mM Na3VO4 was identical to those of the solutions only exposed to the He gas, we observed a time-dependent change of the absorbance of plasma-treated solutions (Fig.2C). Interestingly, the absorption spectra obtained aer 2 and 4min of plasma treatment resemble those obtained aer 800 and 2000M of H2O2 treatment, respectively (Fig.2D). Because plasma treatment also leads to the formation of nitrite (NO2)

and nitrate (NO3) in solution16,17, we checked that there was no change in the absorbance at 260 and 270 nm of 1mM Na3VO4 incubated in the presence of either NaNO2 or NaNO3 for concentrations up to 3mM (data not shown). Collectively, these results strongly support H2O2 as the major plasma-induced ROS that interact in solution with Na3VO4.

To conrm this hypothesis, 500 l of PBS(Ca2+/Mg2+) containing (direct treatment) or not (indirect treatment) 1 mM Na3VO4 were exposed to He plasma at a gas ow of 50 sccm for dierent times of treatment up to 4 min. For the indirect treatment, Na3VO4 was added post treatment. We recorded the absorbance at 260 and 270 nm of the treated solutions, and used the equations shown in Fig.2B to determine the concentration of the plasma-induced H2O2. We noticed that the concentration of H2O2 at a given time was identical in both conditions (Fig.2E), suggesting that short-lived RONS produced in solution do not play a role in the reaction with Na3VO4. These results conrm that H2O2 is the major plasma-induced ROS that interact with Na3VO4. We also observed that the concentration of H2O2 increases almost linearly with the time of treatment to inect at 4 min (Fig.2E). This inection is likely due to the non-linearity of the response for H2O2 concentration >1 mM (see Fig.2B), and it was not observed if the plasma-treated solutions of PBS(Ca2+/Mg2+) were diluted before adding Na3VO4 (insert of Fig.2E). From the data presented in Fig.2E, we determined that about 400 M of H2O2 are produced per minute of He plasma treatment at a gas ow of 50sccm.

We also measured H2O2 concentration using titanium oxysulfate solution (TiOSO4)33,34. By this method, we found that about 300M of H2O2 are produced per minute of He plasma treatment at a gas ow of 50sccm (Figure S3). Together, these results demonstrate that the concentration of H2O2 produced in PBS(Ca2+/Mg2+) by our He plasma device can range from a few hundred micromolar to a few millimolar, according to the time of treatment.

To quantify NO2 and NO3 produced in the buer solution by He plasma, 500l of PBS(Ca2+/Mg2+) were exposed to He plasma for 1, 2, 3 and 4min and the amount of NO2 and NO3 was quantied using a colorimetric assay kit, as described in the Material and

Methods section. We found a time-dependent increase of the concentration of each compound, NO2 concentration being higher than the NO3 concentration (Fig.3). From these experiments, we determined that about 400M of NO2 and 100M of NO3 are produced in PBS(Ca2+/Mg2+) per minute of He plasma treatment at a gas ow of 50sccm.

SCIENTIFIC REPORTS

4

www.nature.com/scientificreports/

Figure 2. Quantication of H2O2 produced in PBS(Ca2+/Mg2+) by He plasma. (A) Absorption spectra between 250 and 400nm of 1mM Na3VO4 solutions in PBS(Ca2+/Mg2+) and incubated with increasing concentration of H2O2. (B) Correlation between the change in the optical density at 260nm and 270nmof 1mM Na3VO4 solutions and the concentration of H2O2. The data are the mean sd of 12 independent experiments. The equations and correlation coefficients in black and red were derived from the linear regression of the values at 260 and 270nm, respectively. (C) Absorption spectra between 250 and 400nm of 1mM Na3VO4 solutions in PBS(Ca2+/Mg2+) exposed to He gas or He plasma for 2 and 4min. (D) Comparison between the absorption spectra of H2O2 solutions at 800M and 2mM and the absorption spectra of plasma-activated

PBS(Ca2+/Mg2+) aer 2 and 4min of treatment. The spectra are the average SD of 3 independent experiments. (E) Solutions of PBS(Ca2+/Mg2+) containing (direct treatment) or not (indirect treatment) 1mM Na3VO4 were exposed to He plasma for 1, 2, 3, or 4min. For indirect treatment, Na3VO4 was then added to plasma-treated PBS. The optical density of each solution was recorded at 260 and 270nm, and the concentration of H2O2

determined using equations shown in panel B. The data are the mean SD of 12 independent experiments (t-test **p<0.01). Insert: Solutions of PBS(Ca2+/Mg2+) were exposed to He plasma and were diluted 2x, 4x and 8x before adding Na3VO4. The concentration of H2O2 in each solution was determined as mentioned above by taking into account the dilution factors. The data are the mean SD of 9 independent experiments. For the experiments described in the panels C, D, E and F, the He ow was set to 50sccm and the output voltage to 8kV.

SCIENTIFIC REPORTS

5

www.nature.com/scientificreports/

Figure 3. Quantication of NO2 and NO3 produced by He plasma. Solutions of PBS(Ca2+/Mg2+) were exposed to He plasma for 1, 2, 3, or 4min and the concentration of NO2 and NO3 determined as described in Material and Methods. Note that NO2 and NO3 were not detectable in untreated (0min) PBS(Ca2+/Mg2+). The data are the mean SD of 3 independent experiments. The He ow was set to 50sccm and the output voltage to 8kV.

O, NO and NO + + To evaluate if the chemical modications in PBS(Ca2+/Mg2+) can be attributed essentially to the formation of H2O2, NO2 and NO3 following plasma treatment, we performed UV spectrum analysis36,37. At rst we recorded the absorption spectra in PBS(Ca2+/Mg2+) of each of these compounds at known concentrations. For NO2 and NO3, we used stock solutions of NaNO2 and NaNO3, respectively. We found that H2O2 poorly absorbs between 200 and 300nm, with a maximum absorbance around 204nm (Fig.4A). Indeed, a 10mM solution of H2O2 has an absorbance at 204 nm of 1.8 0.1. In marked contrast, both NaNO2 and NaNO3 solutions strongly absorb between 200 and 250 nm, but not between 250 and 300 nm, with a maximum of absorbance at 210 nm (A210nm) (Fig.4B,C). For example, A210nm = 2.48 0.08 for a solution of NaNO2 at 0.5 mM, and

A210nm=1.56 0.04 for a solution of NaNO3 at 0.2mM (Fig.4B,C).

As we previously demonstrated that approximately 400 M of H2O2, 400 M of NO2 and 100 M of NO3 are generated per minute of He plasma treatment at a gas ow of 50 sccm (see above), we then looked at the absorbance of a mixed solution of 800 M of H2O2, 800 M of NO2 and 200 M of NO3 (Fig.4D). As these concentrations are expected to be produced in PBS(Ca2+/Mg2+) by He plasma aer 2 min of treatment, we also recorded the absorbance of plasma-activated PBS(Ca2+/Mg2+) at such conditions (Fig.4E). Because 800 M of NO2 alone gives rise to a A210nm above the limits of the measurement range of the spectrophotometer, serial dilutions (dil 2x, 4x and 8x) were performed. As shown in Fig.4D,E, the absorption spectra of plasma-activated PBS(Ca2+/Mg2+) were very similar to the absorption spectra of a mixed solution of 800 M of H2O2, 800 M of

NO2 and 200M of NO3. To conrm these results, we superimposed the absorption spectra (dil x4 and x8) of plasma-activated PBS(Ca2+/Mg2+) and the absorption spectra (dil x4 and x8) of a mixed solution of 800 M of H2O2, 800M of NO2 and 200M of NO3 (Fig.4F). Indeed, for each dilution, the two absorption spectra were very similar suggesting that the three main long-lived species generated in PBS(Ca2+/Mg2+) by He plasma are H2O2, NO2 and NO3.

O + + To assess the toxicity of He plasma at a gas ow of 50sccm, we used normal primary skin broblasts (NHSF), normal transformed lung broblasts (MRC5Vi), human colon cancer cells (HCT116), and human melanoma cells (Lu1205). The cells were exposed directly or indirectly to He plasma for dierent times of treatment, and the cell viability was measured 24hours post treatment. We observed for the four types of cells, a decrease in the % of cell viability as a function of the treatment time (Fig.5A). Moreover, and as previously reported14,19,22, we did not observe a dierence between the two modes of treatment (i.e. direct versus indirect) suggesting that the cytotoxicity of He plasma is essentially due to plasma-induced long-lived species in solution (Fig.5A). The two tumour cell lines tested in this study (HCT116 and Lu1205) were slightly more resistant to the toxic eect of He plasma than the two normal cell lines (NHSF and MRC5Vi) especially for short (<4 min) treatment times (Fig.5A).

As plasma-induced liquid H2O2 is a key ROS involved in the toxicity of several cold atmospheric plasmas14,22,28, we then measured the cytotoxicity of known concentrations of H2O2 with respect to the four cell types. Although we observed a concentration-dependent cell death for all cell types, the two cancer cell lines (HCT116 and Lu1205) were more resistant to H2O2-induced cell death than the two normal cells (NHSF and MRC5Vi)

(Fig.5B). This behaviour resembles to that observed aer plasma treatment (Fig.5A,B), suggesting that H2O2 plays a central role in the cellular toxicity of He plasma. However, if we consider a concentration of H2O2 of 800 M, which is induced in PBS(Ca2+/Mg2+) aer 2 min of He plasma (see Fig.2), the % of viable cells, for the four cell

SCIENTIFIC REPORTS

6

www.nature.com/scientificreports/

Figure 4. The absorption spectra of a mixture of H2O2, NO2 and NO3 match the absorption spectra of plasma-activated PBS(Ca2+/Mg2+). Absorption spectra between 200 and 300nm of dierent concentrations of H2O2 (A), NO2 (B) and NO3 (C) prepared in PBS(Ca2+/Mg2+). The data are the mean SD of 5 (H2O2)

and 4 (NO3 and NO2) independent experiments. (D) Absorption spectra of a mixture of 800M H2O2, 800M NO2 and 200M NO3 prepared in PBS(Ca2+/Mg2+). The data are the mean SD of 4 independent experiments. (E) 500l of PBS(Ca2+/Mg2+) were exposed to He plasma at a ow rate of 50sccm for 2min, and the absorption spectra were recorded between 200 and 300nm. The data are the mean SD of 3 independent experiments. (F) Comparison between the absorption spectra of a mixture of 800 M H2O2, 800M NO2 and 200 M NO3 (curves shown in panel D) and the absorption spectra of plasma-activated PBS(Ca2+/Mg2+) (curves shown in panel E). As indicated in the panels D, E, and F, the solutions containing H2O2, NO2 and NO3 or plasma-activated PBS(Ca2+/Mg2+) were diluted 2x, 4x, or 8x in PBS(Ca2+/Mg2+) before the spectroscopic measurements.

types, is higher aer a H2O2 treatment compared to a He plasma treatment (Fig.5A,B). Indeed, the % of viable cells in response to 800M of H2O2 compared to 2min of He plasma treatment was about 4% compared to 45% for NHSF, 15% compared to 55% for MRC5Vi, 45% compared to 70% for HCT116, and 30% compare to 70%

SCIENTIFIC REPORTS

7

www.nature.com/scientificreports/

Figure 5. The concentration of plasma-induced H2O2 in PBS(Ca2+/Mg2+) does not fully explain the sensitivity of cells to He plasma. (A) Normal (MRC5Vi and NHSF) and tumour cells (HCT116 and Lu1205) were exposed in PBS(Ca2+/Mg2+) to He plasma (direct treatment) or to plasma-activated PBS(Ca2+/Mg2+) (indirect treatment) for 1, 2, 4 and 8min, as described in Material and Methods. The He ow was set to 50sccm and the output voltage to 8kV. (B) The same cells were exposed to increasing concentration of H2O2 in

PBS(Ca2+/Mg2+). The cell viability assay was performed 24h post treatment. The data are the average SD of 3 to 4 independent experiments (He plasma treatment, t-test p> 0.05) and 5 to 8 independent experiments (H2O2

treatment). The continuous and dashed red lines indicate the percentage of viable cells aer 2min of He plasma treatment and aer 800M of H2O2, respectively.

SCIENTIFIC REPORTS

8

www.nature.com/scientificreports/

for Lu1205. At longer times of treatment by He plasma (i.e. 4min), the % of viable NHSF and MRC5Vi cells is almost identical to those obtained at the equivalent H2O2 concentration (i.e. 1.6mM) (Fig.5). In contrast, the % of viable HCT116 and Lu1205 cells is always higher in response to a treatment of H2O2 than to a He plasma treatment, in the concentration and time range considered in this study (Fig.5). These data strongly suggest that other RONS than H2O2 also contribute to the toxicity of the He plasma.

O, NO and NO

All the experiments described above were performed at a He gas ow of 50sccm. To investigate the eect of the gas ow on the RONS induced in PBS(Ca2+/Mg2+), we determined the concentration of H2O2, NO2 and NO3

in the buer solution aer a He plasma treatment at a gas ow of 400sccm. We found that the concentration of H2O2 (Fig.6A) and of NO2 and NO3 (Fig.6B) is lower at 400sccm compared to 50sccm. Indeed, aer 2min of

He plasma treatment at 400 sccm, the concentration of H2O2 was about 300M (instead of 800 M at 50 sccm), while the concentrations of NO2 and NO3 were about 500M and 150M, respectively (instead of 800M and 200 M at 50 sccm). In order to verify if the absorption spectrum of a mixture of these RONS at these concentrations could fully reproduce the absorption spectrum of a solution of PBS(Ca2+/Mg2+) treated for 2 min with a He plasma at a gas ow of 400 sccm, we recorded and compared the absorption spectra of PBS(Ca2+/Mg2+) solution containing 300 M of H2O2, 500 M of NO2 and 150 M of NO3 (Fig.6C), and of plasma-activated

PBS(Ca2+/Mg2+) aer 2min of treatment (Fig.6D). The overlay of the spectra for each of the two conditions, at the same dilution factor (Fig.6E), suggests that the mixture of these three species at the measured concentrations can adequately reproduce the chemistry generated in the buer solution aer 2min of He plasma at a gas ow of 400sccm.

O + + We used MRC5Vi cells, as control of normal cells, and HCT116, as control of tumour cells to assess the role of H2O2 in the toxicity of the He plasma at a gas ow of 400 sccm. As reported above for a He plasma operating at a gas ow of 50 sccm (see Fig.5A), the indirect treatment is as efficient as the direct treatment in inducing cell death also at a gas ow of 400sccm (Fig.7). Moreover, HCT116 cells were also found more resistant than MRC5Vi to the plasma treatment at a gas ow of 400sccm (Fig.7), thus conrming the results obtained at 50sccm (see Fig.5A). We showed that at a gas ow of 400sccm, the He plasma generates about 300M H2O2 per min of treatment (see Fig.6A). From the sensitivity of each cell line to H2O2 (see Fig.5B), if the toxicity arises only from plasma-induced liquid H2O2, then the % of viable cells should range between 60% (for MRC5Vi) to 95% (for HCT116) aer 2min of He plasma treatment, and between 30% (for MRC5Vi) to 60% (for HCT116) aer 4 min of He plasma treatment. We found that aer 2 min of He plasma treatment, the % of viable cells was about 20% for MRC5Vi and 60% for HCT116, while aer 4min of treatment, these values dropped to 4% and 40%, respectively (Fig.7). Therefore, the % of viability obtained aer plasma treatment is lower than those determined aer H2O2 treatment alone, demonstrating that H2O2 alone cannot account for the toxicity of plasma-activated PBS(Ca2+/Mg2+) at a gas ow of 400 sccm, as it was also observed at 50sccm.

NO O So far, our results demonstrated that the three main species generated in PBS(Ca2+/Mg2+) by He plasma are H2O2, NO2 and

NO3, and that H2O2 is essential, but not sufficient, to account for the toxicity of He plasma. These observations prompted us to investigate the role of NO2 and NO3 in the toxicity of plasma-activated PBS(Ca2+/Mg2+). To do so, NHSF, MRC5Vi, HCT116 and Lu1205 cells were exposed to dierent solutions of PBS(Ca2+/Mg2+) containing H2O2 and/or NO2 and/or NO3 at the concentrations obtained aer 2min of He plasma treatment at a gas ow of 50 sccm (i.e. 800 M H2O2, 800 M NO2 and 200 NO3). We show that in the range of concentrations used in this study, NO2 and/or NO3 are not toxic to the cells (Fig.8 and Figure S4A), and that the sensitivity of each cell line to H2O2 treatment is not enhance by the addition of NO3 (t-test p> 0.05) (Fig.8). In contrast, a mixture of H2O2 and NO2 triggered more cell death than H2O2 alone, and again the addition of NO3 to H2O2/NO2 mixture did not change the % of viable cells (t-test p > 0.05) suggesting that NO3 does not contribute to cell death (Fig.8). Finally, and most importantly, we found that the % of viable cells in response to a mixture of H2O2/

NO2 (or H2O2/NO2/NO3) is not statistically dierent to the % of viable cells in response to plasma-activated PBS(Ca2+/Mg2+) (t-test p> 0.05) (Fig.8). Because H2O2 can react with NO2 in weakly acid to acid aqueous solutions to form peroxinitric acid38, we also monitored the pH of plasma-activated PBS(Ca2+/Mg2+) as a function of treatment time at a gas ow of 50sccm. For comparison, we also checked the pH of PBS(Ca2+/Mg2+) containing a mixture of H2O2/NO2/NO3 corresponding to the concentrations expected aer plasma treatment. We found a treatment time-dependent decrease of the pH of plasma-activated PBS(Ca2+/Mg2+) but not of reconstituted buered solutions (Figure S4B). Indeed, a drop of the pH from 7.2 to 6 was observed aer 8 min of treatment. Collectively, our results strongly suggest that plasma-induced-H2O2 and -NO2 in PBS(Ca2+/Mg2+) act in synergy, possibly in part via the formation of peroxynitrite, to induce cell death.

O, NO and NO + + To assess for the role of atmospheric ambient air in the formation of plasma-induced RONS, the He plasma jet was shielded from the atmosphere (ambient air) by a gas of pure O2. As the experimental setup used for this specic study was slightly dierent to the one used so far (see Material and Method), at rst we decided to measure the concentration of H2O2, NO2 and NO3 produced in these new experimental conditions. Using a He gas ow at 100 and 400sccm, we found that the production of H2O2 was 63 and 35 M per min, respectively (Figure S5A), while the production of total NOx (NO2+NO3)

was 33 and 12M per min, respectively (Figure S5B). These values are lower than those measured with the other

SCIENTIFIC REPORTS

9

www.nature.com/scientificreports/

Figure 6. Increasing the He ow decreases the concentration of plasma-induced H2O2, NO2 and NO3.

500l of PBS(Ca2+/Mg2+) per well in 12 well plate were exposed to 50 or 400sccm of He plasma for 1, 2 and 4min. The concentration of H2O2 (A) and NO2/NO3 (B) were determined using the Na3VO4-based method and the nitrate/nitrite colorimetric assay kit, respectively. The data are the mean SD of 5 (H2O2) and 4 (NO3 and NO2) independent experiments. (C) Absorption spectra of a mixture of 300M H2O2, 500M NO2 and 150M NO3 prepared in PBS(Ca2+/Mg2+). The data are the mean SD of 4 independent experiments. (D) 500l of PBS(Ca2+/ Mg2+) were exposed to He plasma at a ow rate of 400sccm for 2min, and the absorption spectra were recorded between 200 and 300nm. The data are the mean SD of 3 independent experiments. (E) Comparison between the absorption spectra of a mixture of 300M H2O2, 500M NO2 and 150M NO3 (curves shown in panel C) and the absorption spectra of plasma-activated PBS(Ca2+/Mg2+) (curves shown in panel D). As indicated in the panels C, D, and E, the solutions containing H2O2, NO2 and NO3 or plasma-activated PBS(Ca2+/Mg2+) were diluted 2x, 4x, or 8x in PBS(Ca2+/Mg2+) before the spectroscopic measurements.

plasma jet (see Figs2,3 and 6), but can be explained at least by the larger volume of treated PBS(Ca2+/Mg2+) used here (3ml instead of 0.5ml) and the lower output voltage (5.5kV instead of 8kV). Nevertheless, we found again that increasing the gas ow leads to lower the concentration of these RONS in the plasma-treated solution. Using a shielding gas of pure O2 surrounding the plasma jet, we found that aer 4min of treatment, the concentration of

H2O2 drops by 36% (322M with ambient air to 206M with pure O2) (Fig.9A) while the concentration of total

SCIENTIFIC REPORTS

10

www.nature.com/scientificreports/

Figure 7. Percentage of viable cells aer a He plasma treatment at a gas ow of 400sccm. MRC5Vi and HCT116 cells were exposed in PBS(Ca2+/Mg2+) to He plasma (direct treatment) or to plasma-activated PBS(Ca2+/Mg2+) (indirect treatment) for 1, 2, 4 and 8min. The He ow was set to 400sccm and the output voltage to 8kV. The cell viability assay was performed 24h post treatment. The red lines indicate the percentage of viable cells aer 2min of He plasma treatment. The data are the average SD of 4 independent experiments (t-test p>0.05)

NOx (NO2+NO3) drops by 96% (226M with ambient air to 9M with pure O2) (Fig.9B). These reductions of the production of H2O2 and NO2/NO3 follow the observed decrease of the light intensity emitted in the plasma jet by OH and N2(C)/N2+ of about 1 and 3 orders of magnitude when a shielding of pure O2 is applied (Fig.10 and Table1). Concomitantly, the shielding gas of pure O2 also prevented the acidication of the plasma-treated PBS(Ca2+/Mg2+) (Figure S6).

The application of cold atmospheric pressure plasmas (CAPPs) in cancer treatment is one of the main active elds of research in Plasma Medicine. The proof-of-concept has been largely demonstrated in vitro and to a lesser extent in vivo (for a recent review see ref. 3). Although the dierent groups working in this eld used dierent plasma devices with dierent plasma chemistries and cell lines derived from dierent tumours3, all the dierent types of CAPPs were eective, indicating that the eects of plasma seem to be uniform and are not restricted to a particular type of tumour. One fundamental insight arising from all these studies is that plasma-induced changes

SCIENTIFIC REPORTS

11

www.nature.com/scientificreports/

Figure 8. Synergic eect of H2O2 and NOs species on cell viability. NHSF (A), MRC5Vi (B), Lu1205 (C), and HCT116 (D) cells were untreated (Ctr) or exposed to a mixture of NO2 and NO3, H2O2 alone, a mixtureof H2O2 and NO2 or NO3, or a mixture of H2O2, NO2 and NO3. The concentrations of H2O2, NO2 and NO3 were those determined aer a He plasma treatment of 2min at a gas ow of 50sccm (see Fig.3). The cell viability assay was performed 24h post treatment. The data are the mean SD of 6 independent experiments. The cell viabilities aer an indirect He plasma treatment at a gas ow of 50sccm are also indicated.

in the liquid environment of the cells play a key role in plasma-cell interactions, and thus to the cell fate. As mentioned by D. Graves in 2012: The successful development of plasma biomedicine applications will hinge in signicant measure on controlling the actions of the RONS created in the plasma by generating only the species that are needed and delivering them to the right place at the right time in the right concentration16. To date, it is unanimously recognized that RONS, among them H2O2, NO2 and NO3, are the central players in the antitumor activities of CAPPs16,17. The aim of this study was to precisely determine the concentration of each of these species in solution aer a He plasma treatment and to address the following question: is the production of one or more of these species in solution sufficient to explain the cellular toxicity of the He plasma device?

At rst, we would like to draw attention to the fact that it is difficult to strictly compare the measured concentration of each species obtained in one study (including ours) to other published studies insofar as dierent types of CAPPs and biological targets are used. Any parameters of the experimental setup (e.g. the nature of the gas, the gas ow, the applied voltage, the distance between the plasma and the solution, the composition of the solution, the volume of the solution )15,21 play a role in the amount of plasma-induced RONS in solution. Hereaer

SCIENTIFIC REPORTS

12

www.nature.com/scientificreports/

Figure 9. The concentrations of H2O2, NO2 and NO3 in plasma-treated PBS(Ca2+/Mg2+) are decreased in the presence of a shielding gas of pure oxygen. Three milliliters of PBS(Ca2+/Mg2+) were set per well in12 well plates, and exposed to He plasma at a gas ow of 100sccm for the indicated period of times, and atan output voltage of 5.5kV. The treatments were performed in the presence (with O2) or absence (no O2) ofa shielding gas of oxygen at a gas ow of 5 slm. (A) The concentration of plasma-induced H2O2 for each time point was determined by using the Na3VO4 assay and the TiOSO4 assay. (B) The concentrations of NO2 and NO3 aer 4min of plasma treatment were determined by the Griess assay.

are some selected examples regarding the concentration of plasma-induced H2O2 in different conditions: [H2O2] = 30 M in 500 l of MEM medium aer 1 min of He plasma at a gas ow of 5 L/min24. [H2O2] = 6 M in 300 l of phenol-free RPMI 1640 medium aer 1 min of He +0.25% O2 plasma at a gas ow of 8 L/min22; [H2O2] = 60 M in 1 ml of phenol-free RPMI 1640 medium aer 1 min of Ar plasma at a gas ow of 3 L/min28; [H2O2] = 32 M in 5 ml of phenol-free RPMI 1640 medium aer 1 min of Ar plasma at a gas ow of 3 L/min39; [H2O2]=190M in 3ml of PBS(Ca2+/Mg2+/Glucose) aer 1min of Ar plasma at a gas ow of 1.5L/min.

Using two dierent assays (one based on Na3VO4 and the other on TiOSO4), we showed that in our standard experimental conditions500 l of PBS(Ca2+/Mg2+) exposed to a He plasma jet operated at 8 kV and at a gas ow of 50 sccm - around 400 M of H2O2 are produced per min, a value which is quite high compare to those cited above. Yang et al. reported that the concentration of ROS measured aer plasma treatment decreases with increasing the complexity of the targeted solution22. We carried out our experiments in a simple buered solution [PBS(Ca2+/Mg2+)], which is devoid of amino acids, vitamins and other compounds, such as glucose or serum found in all cell culture media. The presence of some of these components in the cell culture media during plasma exposure might interfere with the formation of H2O2, or react with H2O221. Furthermore, we used a very small

He ow rate (50sccm), when compared to most published data for which He ow rates of few liters per min were used15,21,23,24,4042, and, as further discussed below in the text, we found that the concentration of H2O2 in solution is higher as the gas ow is lower.

We also found that the rate production of NO2 and NO3 is 400 M and 100 M per min, respectively, at a gas ow of 50 sccm. As for H2O2, the concentration of these RNS in solution is also highly dependent on the

SCIENTIFIC REPORTS

13

www.nature.com/scientificreports/

Figure 10. A shielding gas of pure oxygen strongly inhibits the presence of OH and N2 molecules in the gas phase. The emission spectra of the molecular bands of (A) OH (at around 309nm), (B) N2(C) (Second Positive System at around 337nm) and (C) N2+ (First Negative System at around 391nm) and the atomic lines of (D)

He (at around 706nm) and (E) O (at around 777nm) were recorded in the absence (no shielding) or presence (pure O2 shielding) of a shielding of pure O2. The gas ow of He was set to 100sccm and the plasma is created by applying high voltage pulses with amplitude of 5.5kV.

Gas species Ratio (no shielding/O2 shielding)

OH (309nm) 33 N2(C) (337nm) 2155

N2+ (391nm) 1475 He (706nm) 0.72 O (777nm) 0.22

Table 1. Ratio of the relative intensities of the light emission of the molecular bands of OH, N2(C) andN2+ and of the atomic lines of He and O from the plasma jet in the absence or presence of a shielding gas of pure O2. The data are derived from the experimental values measured from the emission spectra of each species shown in Fig.10. Note that the emission intensity of the molecular bands of OH, N2(C) and N2+ drop drastically in the presence of the shielding gas of pure O2.

SCIENTIFIC REPORTS

14

www.nature.com/scientificreports/

experimental setup [this study, see also14,25]. By looking at the absorption in the UV range (200300nm) of solutions of PBS(Ca2+/Mg2+) exposed to a He plasma jet and solutions of PBS(Ca2+/Mg2+) containing a mixture of H2O2, NO2 and NO3, we have been able to demonstrate that these species account for the main long-lived

RONS induced by our plasma in the buer solution. Aiming the understanding of how these species accumulate in solution during the He plasma treatment, we used a shielding gas of pure O2 isolating the plasma jet from the ambient air. We found that the surrounding atmosphere has a greater impact on the formation of NO2 and NO3 than on the formation of H2O2 in solution. These results are in good agreement with those published by Tresp et al. who used argon as the feeding gas12. In buer solution, the dissociation of water molecules by energetic particles from the plasma can generate hydroxyl radicals that recombine to form H2O212,27,43,44. The water molecules can already be present in the feeding gas (e.g. using a humidied feeding gas)14,27,45 or arise from the humidity in the ambient air, through which the plasma propagates, and from the water vapor evaporated from the water outer layer of the solution targeted with the plasma44. In our experimental conditions, we used as feeding gas Helium Alphagaz 2 that contains only traces of H2O2 (<0.5ppm). Therefore, water molecules mainly arise from the ambient air, and at the gas/liquid interface. Using a shielding gas of pure O2, we showed that both pathways contribute almost equally to the liquid H2O2 induced in the buer solution by the plasma. Concomitantly, we showed that the concentration of NO2 and NO3 drops drastically in the presence of the shielding gas of pure O2. This was expected as nitrites and nitrates are formed in plasma-treated buer solutions through the dissolution of nitrogen oxides produced by gas-phase reactions of dissociated N2 and O246. By increasing the He gas ow, we found a decrease of the concentration of the three species in the buer solution. This should result from the fact that less air is admixed to the plasma jet channel when the gas ow is higher and, thus, less RONS are produced in the gas phase.

The eects of theses plasma-generated species on mammalian cells were investigated on four dierent cell types: 2 normal cell types (NHSF and MRC5Vi) and 2 cancer cell lines (HCT116 and Lu1205). We conrmed several published data showing that plasma-activated medium is as efficient as the direct treatment of cells in triggering cell death14,1824. Although we did not investigate further the route leading to cell death, it is well documented that apoptosis (a programmed cell death) and necrosis (a non physiological process) are the two main cell death pathways that have been described aer CAPP treatments23,26,29,4749, and are likely involved in our study. Several investigators have shown that cancer cells are more susceptible to plasma-induced cell death than normal or healthy cells20,41,50,51. It was proposed that the distribution of the cells within the cell cycle may account for the higher susceptibility of cancer cells to CAPP treatment52. In a recent review, Yan et al. proposed that cancer cells tend to express more aquaporins on their cytoplasmic membranes, which may cause the H2O2 uptake speed in cancer cells to be faster than in normal cells53. However, these observations contrast with other published data26,49,54 and our data (this study) showing that the cancer cells are more resistant to CAPP treatment than the normal or healthy cell types. To explain such discrepancies, further investigations are required but it is necessary to consider several parameters such as the plasma device used in each experimental setup, the concentration of RONS produced in solution, the nature of the treated solution (e.g. PBS versus cell culture medium) and the type of targeted cells. Regarding this last point, an eective comparison between the responses of normal and cancer cells to CAPP treatment should be performed between cell lines derived from the same tissue53.

Our observations also conrm that, at least regarding cell death, long lifetime species such as H2O2 and NO2 fully account for the toxicity of CAPPs. We have demonstrated that the sensitivity of the four cell lines to the He plasma treatment parallels their individual sensitivity to H2O2, thus pointing to H2O2 as a central player in plasma-induced oxidative stress24,28, and that the concomitant production of NO2 exacerbates H2O2 toxicity. In weakly acid to acid solutions, peroxynitric acid can be formed by the interaction of NaNO2 and H2O238,45,55. As the He plasma treatment led to acidication of PBS(Ca2+/Mg2+), it is reasonable to think that peroxynitrite is formed, especially at long treatment times. Peroxynitrite can induce both cellular apoptosis and necrosis depending on the production rates, endogenous antioxidant levels and exposure time56, and therefore could contribute to plasma-induced cell death. We propose that the ability of the cells to cope with these two RONS (H2O2/NO2)

is the major signal that triggers the cell fate in response to our He plasma device. Nonetheless, we cannot not exclude that others plasma-induced RONS, such as nitric oxide (NO), radical hydroxyl (HO.), superoxide anion (O2)29,57,58 could contribute, to a less extent, to plasma toxicity. At the level of the cellular response, the control of the intracellular redox homeostasis59, the activation of MAPK pathways25,60, the down regulation of survival signal transduction pathway18, the epigenetic and cellular changes that are induced by CAPP in a cell type-specic manner61, the distribution of the cells within the cell cycle52, the expression of aquaporins53 are all endpoints to take into account to evaluate the eectiveness of CAPPs as a new antitumor strategy.

1. Urruticoechea, A. et al. Recent Advances in Cancer Therapy: An Overview. Current Pharmaceutical Design 16(1), 3 (2010).2. Kim, S. New and emerging factors in tumorigenesis: an overview. Cancer Management and Research 7, 225 (2015).3. Schlegel, J., Kritzer, J. & Boxhammer, V. Plasma in cancer treatment. Clinical Plasma Medicine 1(2), 2 (2013).4. Schutze, A. et al. The atmospheric-pressure plasma jet: a review and comparison to other plasma sources. Plasma Science, IEEE Transactions on 26(6), 1685 (1998).

5. Laroussi, M. & Akan, T. Arc-Free Atmospheric Pressure Cold Plasma Jets: A Review. Plasma Processes and Polymers 4(9), 777 (2007).

6. Fridman, G. et al. Applied Plasma Medicine. Plasma Processes and Polymers 5(6), 503 (2008).7. Lu, X., Laroussi, M. & Puech, V. On atmospheric-pressure non-equilibrium plasma jets and plasma bullets. Plasma Sources Science and Technology 21(3), 034005 (2012).

8. Hoffmann, C., Berganza, C. & Zhang, J. Cold Atmospheric Plasma: methods of production and application in dentistry and oncology. Medical Gas Research 3, 21 (2013).

9. Isbary, G. et al. Cold atmospheric plasma devices for medical issues. Expert Review of Medical Devices 10(3), 367 (2013).10. Keidar, M. et al. Cold atmospheric plasma in cancer therapy. Physics of Plasmas 20(5), 057101 (2013).

SCIENTIFIC REPORTS

15

www.nature.com/scientificreports/

11. Mericam-Bourdet, N., Laroussi, M., Begum, A. & Karakas, E. Experimental investigations of plasma bullets. Journal of Physics D: Applied Physics 42(5), 055207 (2009).

12. Tresp, H., Hammer, M. U., Weltmann, K.-D. & Reuter, S. Eects of Atmosphere Composition and Liquid Type on Plasma-Generated Reactive Species in Biologically Relevant Solutions. Plasma Medecine 3(12), 45 (2013).

13. Yusupov, M. et al. Reactive molecular dynamics simulations of oxygen species in a liquid water layer of interest for plasma medicine. Journal of Physics D: Applied Physics 47(2), 025205 (2014).

14. Wende, K. et al. Identication of the biologically active liquid chemistry induced by a nonthermal atmospheric pressure plasma jet. Biointerphases 10(2), 029518 (2015).

15. Cheng, X. et al. The eect of tuning cold plasma composition on glioblastoma cell viability. PLoS One 9(5), e98652 (2014).16. Graves, D. B. The emerging role of reactive oxygen and nitrogen species in redox biology and some implications for plasma applications to medicine and biology. Journal of Physics D: Applied Physics 45(26), 263001 (2012).

17. Jablonowski, H. & von Woedtke, T. Research on plasma medicine-relevant plasmaliquid interaction: What happened in the past ve years? Clinical Plasma Medicine 3(2), 42 (2015).

18. Tanaka, H. et al. Plasma-Activated Medium Selectively Kills Glioblastoma Brain Tumor Cells by Down-Regulating a Survival Signaling Molecule, AKT Kinase. Plasma Medecine 1(34), 265 (2011).

19. Strassenburg, S. et al. Comparison of Biological Eects on Human Keratinocytes Using Dierent Plasma Treatment Regimes. Plasma Medecine 3(12), 57 (2013).

20. Utsumi, F. et al. Selective cytotoxicity of indirect nonequilibrium atmospheric pressure plasma against ovarian clear-cell carcinoma. Springerplus 3, 398 (2014).

21. Yan, D. et al. Principles of using Cold Atmospheric Plasma Stimulated Media for Cancer Treatment. Sci Rep 5, 18339 (2015).22. Yang, H. et al. Eects of atmospheric pressure cold plasma on human hepatocarcinoma cell and its 5-uorouracil resistant cell line. Physics of Plasmas 22(12), 122006 (2015).

23. Virard, F. et al. Cold Atmospheric Plasma Induces a Predominantly Necrotic Cell Death via the Microenvironment. PLoS One 10(8), e0133120 (2015).

24. Mohades, S. et al. Evaluation of the eects of a plasma activated medium on cancer cells. Physics of Plasmas 22(12), 122001 (2015).25. Ahn, H. J. et al. Targeting Cancer Cells with Reactive Oxygen and Nitrogen Species Generated by Atmospheric-Pressure Air Plasma. PLoS One 9(1), e86173 (2014).

26. Hirst, A. M. et al. Low-temperature plasma treatment induces DNA damage leading to necrotic cell death in primary prostate epithelial cells. Br J Cancer 112(9), 1536 (2015).

27. Winter, J. et al. Tracking plasma generated H2O2 from gas into liquid phase and revealing its dominant impact on human skin cells. Journal of Physics D: Applied Physics 47(28), 285401 (2014).

28. Bekeschus, S. et al. Hydrogen peroxide: A central player in physical plasma-induced oxidative stress in human blood cells. Free Radic Res 48(5), 542 (2014).

29. Xu, D. et al. In Situ OH Generation from O2- and H2O2 Plays a Critical Role in Plasma-Induced Cell Death. PLoS One 10(6), e0128205 (2015).

30. Huschtscha, L. I. & Holliday, R. Limited and unlimited growth of SV40-transformed cells from human diploid MRC-5 broblasts. J Cell Sci 63, 77 (1983).

31. Douat, C. et al. Dynamics of colliding microplasma jets. Plasma Sources Science and Technology 21(3), 034010 (2012).32. Bennett, P. A., Dixon, R. J. & Kellie, S. The phosphotyrosine phosphatase inhibitor vanadyl hydroperoxide induces morphological alterations, cytoskeletal rearrangements and increased adhesiveness in rat neutrophil leucocytes. J Cell Sci 106 (Pt 3), 891 (1993).

33. Eisenberg, G. Colorimetric Determination of Hydrogen Peroxide. Industrial & Engineering Chemistry Analytical Edition 15(5), 327 (1943).

34. Sattereld, C. N. & Bonnell, A. H. Interferences in Titanium Sulfate Method for Hydrogen Peroxide. Analytical Chemistry 27(7), 1174 (1955).

35. Hornby, J. E. Measurements of cell adhesion. II. Quantitative study of the eect of divalent ions on cell adhesion. J Embryol Exp Morphol 30(2), 511 (1973).

36. Panngom, K. et al. Non-thermal plasma treatment diminishes fungal viability and up-regulates resistance genes in a plant host. PLoS One 9(6), e99300 (2014).

37. Ercan, U. K. et al. Chemical Changes in Nonthermal Plasma-Treated N-Acetylcysteine (NAC) Solution and Their Contribution to Bacterial Inactivation. Sci Rep 6, 20365 (2016).

38. Starodubtseva, M. N., Cherenkevich, S. N. & Semenkova, G. N. Investigation of the interaction of sodium nitrite with hydrogen peroxide in aqueous solutions by the chemiluminescence method. Journal of Applied Spectroscopy 66(3), 473 (1999).

39. Schmidt, A. et al. Non-thermal plasma activates human keratinocytes by stimulation of antioxidant and phase II pathways. J Biol Chem 290(11), 6731 (2015).

40. Brun, P. et al. Helium generated cold plasma nely regulates activation of human broblast-like primary cells. PLoS One 9(8), e104397 (2014).

41. Guerrero-Preston, R. et al. Cold atmospheric plasma treatment selectively targets head and neck squamous cell carcinoma cells. Int J Mol Med 34(4), 941 (2014).

42. Ishaq, M., Bazaka, K. & Ostrikov, K. Intracellular eects of atmospheric-pressure plasmas on melanoma cancer cells. Physics of Plasmas 22(12), 122003 (2015).

43. Bruggeman, P. & Schram, D. C. On OH production in water containing atmospheric pressure plasmas. Plasma Sources Science and Technology 19(4), 045025 (2010).

44. Norberg, S. A., Tian, W., Johnsen, E. & Kushner, M. J. Atmospheric pressure plasma jets interacting with liquid covered tissue: touching and not-touching the liquid. Journal of Physics D: Applied Physics 47(47), 475203 (2014).

45. Brisset, J.-L. & Pawlat, J. Chemical Eects of Air Plasma Species on Aqueous Solutes in Direct and Delayed Exposure Modes: Discharge, Post-discharge and Plasma Activated Water. Plasma Chemistry and Plasma Processing, 1 (2015).

46. Lu, X., Naidis, G. V., Laroussi, M. & Ostrikov, K. Guided ionization waves: Theory and experiments. Physics Reports 540(3), 123 (2014).

47. Ahn, H. J. et al. Atmospheric-pressure plasma jet induces apoptosis involving mitochondria via generation of free radicals. PLoS One 6(11), e28154 (2011).

48. Kumar, N. et al. Induced apoptosis in melanocytes cancer cell and oxidation in biomolecules through deuterium oxide generated from atmospheric pressure non-thermal plasma jet. Sci Rep 4, 7589 (2014).

49. Lunov, O. et al. Cell death induced by ozone and various non-thermal plasmas: therapeutic perspectives and limitations. Sci Rep 4, 7129 (2014).

50. Kalghatgi, S., Kelly, C., Cerchar, E. & Azizkhan-Cliord, J. Selectivity of Non-Thermal Atmospheric-Pressure Microsecond-Pulsed Dielectric Barrier Discharge Plasma Induced Apoptosis in Tumor Cells over Healthy Cells. Plasma Medecine 1(34), 249 (2011).51. Siu, A. et al. Dierential Eects of Cold Atmospheric Plasma in the Treatment of Malignant Glioma. PLoS One 10(6), e0126313 (2015).

52. Volotskova, O., Hawley, T. S., Stepp, M. A. & Keidar, M. Targeting the cancer cell cycle by cold atmospheric plasma. Sci Rep 2, 636 (2012).

SCIENTIFIC REPORTS

16

www.nature.com/scientificreports/

53. Yan, D. et al. Toward understanding the selective anticancer capacity of cold atmospheric plasmaa model based on aquaporins (Review). Biointerphases 10(4), 040801 (2015).

54. Kaushik, N. K., Kaushik, N., Park, D. & Choi, E. H. Altered antioxidant system stimulates dielectric barrier discharge plasma-induced cell death for solid tumor cell treatment. PLoS One 9(7), e103349 (2014).

55. Oury, T. D., Tatro, L., Ghio, A. J. & Piantadosi, C. A. Nitration of tyrosine by hydrogen peroxide and nitrite. Free Radic Res 23(6), 537 (1995).

56. Szabo, C., Ischiropoulos, H. & Radi, R. Peroxynitrite: biochemistry, pathophysiology and development of therapeutics. Nat Rev Drug Discov 6(8), 662 (2007).

57. Kaushik, N. et al. Responses of Solid Tumor Cells in DMEM to Reactive Oxygen Species Generated by Non-Thermal Plasma and Chemically Induced ROS Systems. Scientic Reports 5, 8587 (2015).

58. Kim, S. J. & Chung, T. H. Cold atmospheric plasma jet-generated RONS and their selective eects on normal and carcinoma cells. Scientic Reports 6, 20332 (2016).

59. Weiss, M. et al. Cold Atmospheric Plasma Treatment Induces Anti-Proliferative Eects in Prostate Cancer Cells by Redox and Apoptotic Signaling Pathways. PLoS One 10(7), e0130350 (2015).

60. Ishaq, M. et al. Atmospheric gas plasma-induced ROS production activates TNF-ASK1 pathway for the induction of melanoma cancer cell apoptosis. Mol Biol Cell 25(9), 1523 (2014).

61. Park, S. B. et al. Dierential Epigenetic Eects of Atmospheric Cold Plasma on MCF-7 and MDA-MB-231 Breast Cancer Cells. PLoS One 10(6), e0129931 (2015).

This work was partially supported by the LabEx LaSIPS through the project TCP-NAT, the CNRS PEPS PlasmaMed through the project Plasma-Tox, the CNRS Fderation LuMat through the project Traitement des Cancers par Plasmas, and Institut Curie (UMR3347 budget). PMG also wish to thank its former colleagues at Institut Curie CNRS UMR3348, and especially Evelyne Sage and Mounira Amor-Guret for their continuous support.

P.-M.G. wrote the main manuscript text. P.-M.G., A.A., J.S.S. and M.D. designed the biological experiments. J.S.S., M.F., G.B. and V.P. designed the plasma devices. P.-M.G., A.A. and J.S.S. performed the experiments. P.-M.G. and J.S.S. prepared the gures.

Supplementary information accompanies this paper at http://www.nature.com/srep

Competing nancial interests: The authors declare no competing nancial interests.

How to cite this article: Girard, P.-M. et al. Synergistic Eect of H2O2 and NO2 in Cell Death Induced by Cold Atmospheric He Plasma. Sci. Rep. 6, 29098; doi: 10.1038/srep29098 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS

17