In Situ OH Generation from O2 − and H2O2 Plays a Critical Role in Plasma-Induced Cell Death

Reactive oxygen and nitrogen species produced by cold atmospheric plasma (CAP) are considered to be the most important species for biomedical applications, including cancer treatment. However, it is not known which species exert the greatest biological effects, and the nature of their interactions with tumor cells remains ill-defined. These questions were addressed in the present study by exposing human mesenchymal stromal and LP-1 cells to reactive oxygen and nitrogen species produced by CAP and evaluating cell viability. Superoxide anion (O2 −) and hydrogen peroxide (H2O2) were the two major species present in plasma, but their respective concentrations were not sufficient to cause cell death when used in isolation; however, in the presence of iron, both species enhanced the cell death-inducing effects of plasma. We propose that iron containing proteins in cells catalyze O2 − and H2O2 into the highly reactive OH radical that can induce cell death. The results demonstrate how reactive species are transferred to liquid and converted into the OH radical to mediate cytotoxicity and provide mechanistic insight into the molecular mechanisms underlying tumor cell death by plasma treatment.


Introduction
Cold atmospheric plasma (CAP), an ionized gas, has many biological applications including wound healing, surgical procedures, disinfection, and even cancer treatment [1][2][3][4][5]. Dielectric barrier discharge (DBD) and the plasma jet are two methods for producing CAP; both generate various kinds of reactive oxygen and nitrogen species (ROS and RNS, respectively), including hydroxyl radical (OH), hydrogen peroxide (H 2 O 2 ), ozone (O 3 ), atomic oxygen (O), superoxide anion (O 2 − ), nitric oxide (NO), and peroxynitrite anion (ONOO − ) [6], which are considered as the most biologically relevant components of plasma. Reactive species composition in CAP can be altered by regulating the voltage, frequency, working and feeding gases, and humidity.
While many studies have shown that CAP is an efficient disinfectant and can also kill normal as well as tumor cells [7][8][9], it remains unclear which reactive species are chiefly responsible for these biological effects.
Since tissues and cells are immersed in liquid, studies have mostly focused on the interaction of the plasma with a liquid medium. ROS and RNS undergo conversion into different types of reactive species when transferred from gas to liquid phase. Our previous work showed that O 2 − and H 2 O 2 can permeate in distilled water to a greater extent than other species and may interact with cellular components [10]. Computer simulations have shown that OH, HO 2 , and H 2 O 2 can travel deep into a liquid layer to reach biomolecules [11], and that O 2 − , ONOO − , NO 3 − , O 3 , H 2 O 2 , and HNO X are the predominant species generated after treatment of a 50-400 μM thick water layer with a DBD plasma device [12]. Another study that measured O 2 − and OH using the spin trapping compound CYPMPO and detected the signal by electron spin resonance (ESR) spectrometry showed that O 2 − and OH density varied according to plasma jet settings, although the range of concentrations was not reported [13]. OH radicals in 3 ml of aqueous solution produced by atmospheric-pressure He plasma jet measured using terephthalic acid (TA) as a spin trapping compound were present at a concentration of 3.3 uM [14]. The concentration of OH and O 2 − in various ionic solutions was about 1-10 μM after a 3 min Ar plasma treatment, as measured by ESR [15]. However, there is no information about whether these species can interact with biomolecules in liquid medium during plasma treatment.
The present study investigated this question in human mesenchymal stromal cells (MSCs) and LP-1 myeloma tumor cells exposed to CAP generated by a plasma jet. The results demonstrate that O 2 − and H 2 O 2 are the two major reactive species in liquid but are present at concentrations that are insufficient to cause cell death; this was ultimately induced by the OH radical generated in situ upon exposure of cells to O 2 − and H 2 O 2 in the plasma. These findings provide insight into the molecular mechanisms underlying plasma-induced tumor cell death, and may also provide a basis for generating a more powerful plasma enriched with particular reactive species for biological applications such as cancer treatment.

Plasma generation and characterization
CAP was generated by a plasma jet system consisting of a 1 mm powered electrode enclosed in a quartz tube, with a grounded outer electrode wrapped around a 6.0 mm diameter dielectric tube (Fig 1). The system also included a gas flow controller, high-voltage power supply, oscilloscope, and plasma jet. A gas flow of 2 slm for He/Ar was used at voltages of 10 kHz/8 kV for He and 10 kHz/10 kV for Ar. The detailed experimental setup is shown in S1 Fig.

Cell culture
Human MSCs and LP-1 cells [16] were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, 50 µg/ml streptomycin, and 2 mM L-glutamine (all from Hyclone, Logan, UT, USA), in an atmosphere of 5% CO 2 at 37°C. MSCs were trypsinized and refreshed weekly and only cells from nine or fewer passages were used in the experiments.

Plasma treatment
Normally, cells were cultured in a 24 well plate in 300 ul RPMI1640 medium at a concentration of 2 × 10 5 cells. For cell viability assay, MSC cells were cultured in a 96 well plate in 100 ul RPMI1640 medium at a concentration of 1 × 10 5 cells. Luminescence was measured 24 h after plasma treatment by directly adding 100 ul of Cell-Titer-Glo reagent into the cells, which could avoid trypsinization of MSC cells and losing some un-adherent cells after plasma treatment.

Cell viability assay
MSCs were seeded in RPMI1640 medium in 96 well optical plates at a concentration of 10 5 cells/100 ul well. After plasma treatment for 30 s, or 1 or 2 min, cells were grown for further 24 h. Cell viability was assessed using the CellTiter-Glo assay (Promega, Madison, WI, USA) according to the manufacturer's instructions. CellTiter-Glo measures luminescence to quantify the level of ATP, which is positively correlated with cell viability. Briefly, 100 ul reagent were added to 100 ul cells and the mixture was lysed by placing on an orbital shaker for 2 min, followed by a 10 min incubation at room temperature. Luminescence was measured with a microplate reader (Varioskan Flash; Thermo Scientific, Waltham, MA, USA).

Observation of morphological changes by microscopy
Cell morphological changes after plasma treatment were examined and imaged using an IX51 inverted phase contrast light microscope (Olympus, Tokyo, Japan). MSCs were normally adherent, with a polygonal shape; dying cells became shrunken and rounded, and detached from the plate.

Western blotting
Cell pellets were lysed in lysis buffer containing 50 mM Tris, 150 mM NaCl, 1% Nonidet P40, and 0.25% sodium deoxycholate. Cell debris was removed by centrifugation for 5 min at 14,000 rpm before sample buffer was added. After boiling, samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (Bio-Rad, Hercules, CA, USA), which were blocked with PBS containing 5% low-fat milk and 0.1% Tween 20. Membranes were probed with antibodies against human ferritin heavy chain (FTH1) (1:1000) and β-actin

Coomassie Blue staining
Holo-and apo-transferrin were dissolved at concentrations of 0.1 mg/ml in 300 μl Ca 2+ /Mg 2+ -free DPBS and treated with Ar plasma for 2 or 8 min. Buffer was added and the samples were boiled, then separated by 12% SDS-PAGE and stained with Coomassie Blue for 20 min, followed by destaining for 1 h.

Statistical analysis
All experimental conditions were prepared in triplicate and experiments were repeated at least three times. Data are presented as mean ± SD. Differences between groups were evaluated using the Mann-Whitney U test. P < 0.05 was considered statistically significant.

Results and Discussion
A 2 slm gas flow was used to produce CAP by plasma jet. Two types of plasma-He + H 2 O and Ar-were tested. He + H 2 O plasma was produced at 10 kHz/8kV with 1% H 2 O in He gas (1.5 slm dry He gas with 0.5 slm humid He gas), while Ar plasma was produced at 10 kHz/10 kV. The distance between the plasma jet and liquid was fixed at 1.5 cm. The 2 slm of plasma gas flow resulted in slight evaporation of the 300 μl volume of medium in the 24 well plate, which was reduced by about 5 μl after plasma treatment for 2 min (data not shown). Next, MSCs cultured in 100 ul RPMI1640 medium in a 96 well plate were treated with plasma for various times (10, 30, 60, and 120 s) and cell viability was measured after 24 h. He + H 2 O and Ar plasma induced MSC apoptosis in a time-dependent manner (Fig 2A); this was accompanied by morphological changes after 120 s of plasma treatment, with cells gradually shrinking and becoming rounded (Fig 2B). A time series detection by flow cytometry of annexin-V and PIstained cells showed that the number of annexin-V + cells gradually increased over 24 h following a 60 s Ar plasma treatment (Fig 2C).
To evaluate the interaction between plasma and liquid, a computer simulation was used to calculate the distribution of different radicals in liquid water with the following set of equations.
The boundary conditions and parameters of this model were described in our previous work [22]. Among the variety of radicals in the plasma gas phase, certain species were transferred further than others: while most disappeared or were converted into other species at a depth of 1 mm, O 2 − and H 2 O 2 reached depths of up to 2 mm (Fig 3A).
When scavengers for particular species were added in order to exclude the effect of the particular ROS and RNS generated during the process of plasma treatment, we first confirmed that these scavengers are non-toxic to cells at the working concentrations used in the assays (data not shown). The results showed that depletion of O 2 − or H 2 O 2 , but not of other species such as OH, NO, ROO•, O 3 , ONOO − , or 1 O 2 , annihilated the effects of plasma treatment, as seen by the changes in cell viability ( Fig 3B) and cell morphology (Fig 3C). These results indicate that O 2 − and H 2 O 2 were the predominant species for induction of apoptosis generated by plasma treatment. Moreover, when 500-600 μl of culture medium were added to each well (yielding a depth of about 3 mm in a 24 well plate), the plasma treatment induced little or no cell death (data not shown).    The basic equations were the same as those previously described. Given the dielectric layer at the bottom, the charge conservation equation was included to calculate the effect of charge accumulation. The Poisson equation for the dielectric area was also applied.
The simulation showed that the penetration depth of O 2 − was increased by applying a +20 V bias voltage, while that of H 2 O 2 also increased slightly (Fig 6A).  Applying a bias voltage increased the penetration depth of O 2 − , which may have been partly converted to H 2 O 2 , thereby increasing its concentration. The experimental setup for testing the above simulation is shown in Fig 6B. Several sizes for the cover (A) and bias voltage node (B) were tested, and diameters of 1.5 and 5 mm, respectively, were ultimately used. Based on the simulation, the penetration of O 2 − was predicted to be deeper and cover a broader area. Indeed, after applying a bias voltage of +20 V, the area of cell death cell was increased significantly by a 60 s He + H 2 O or Ar plasma treatment (Fig 6C). Furthermore, to exclude the effect of H 2 O 2 , a scavenger was added to the medium before the bias voltage was applied. Although O 2 − penetration was still deeper and covered a larger area, in the absence of H 2 O 2 , the area of cell death was unchanged (data not shown). These results demonstrate that H 2 O 2 and O 2 − are the major reactive species that induce cell death. However, each species alone was insufficient to produce this effect, suggesting that a synergistic interaction occurs.
The suggestion of OH radical production being catalyzed by iron [23] has been experimentally validated by the discovery of several Fenton-like reactions [24][25][26]. In the iron-catalyzed Haber-Weiss reaction, OH is produced from H 2 O 2 and O 2 − radicals [26].
It is now widely acknowledged that the Haber-Weiss reaction does not occur in the absence of metal catalysts [27]; the reaction has been directly observed in the gas phase in the presence of iron [28].
We set out to determine whether the Haber-Weiss reaction occurs in a plasma-treated cell system. Findings from previous studies suggested that plasma treatment could provide both H 2 O 2 and O 2 − species at μM concentrations. Cells typically express several iron proteins such as ferroportin [29,30], lactoferrin receptor [31], transferrin [32,33], divalent metal transporter 1 [34,35], and ferritin [36,37], among others. These proteins could conceivably catalyze H 2 O 2 and O 2 − radicals into a highly reactive OH radical that has more potent biological effects.
To test this hypothesis, transferrin-which is known as holo-or apo-transferrin depending on whether it is or is not bound to iron, respectively-was examined. Both forms of the protein (0.1 mg/ml in 300 μl PBS solution) were treated with Ar plasma and protein degradation was assessed by gel electrophoresis. After an 8 min plasma treatment, holo-but not apo-transferrin showed significant degradation (Fig 7A), suggesting that the presence of iron resulted in the catalysis of the Haber-Weiss reaction, thereby enhancing the effects of the plasma. It has been reported that the superoxide anion may also induces the release of ferrous iron from transferrin [38], further facilitating the Haber-Weiss reaction and consequent OH formation. Ferritin expression was then knocked down with miR-200b [39] and cells were examined for their sensitivity to plasma treatment; the sequences of miR-200b and the complementary site in the FTH1 gene are shown in Fig 7B. Cell viability was decreased by knockdown of FTH1 gene expression (Fig 7C). After 24 h of transfection, cells were treated with Ar or He + H 2 O plasma (data not shown) for 20 s followed by a 24 h incubation. The downregulation of ferritin expression by miR-200b as compared to the control transfection was confirmed by western blotting (Fig 7D), and was associated with a decrease in cell sensitivity to Ar plasma treatment (Fig 7E). These data suggest that iron proteins act as catalysts that transform H 2 O 2 and O 2 − into the highly reactive OH radical, which then induces cell death. Iron was added in the form of Fe(III)-EDTA to the plasma-treated cell system to enhance OH production and consequently cell death. To circumvent the possibility that the penetration of iron-catalyzed OH would not be sufficiently deep to affect MSCs given their adherence, an LP-1 cell suspension [16] was used in this experiment. To increase the probability of interaction between OH and LP-1 cells, the cells were treated with He + H 2 O or Ar plasma for 20 s at a relatively high concentration (5 × 10 6 /ml). Fe(III)-EDTA (1 μM) had no effect on cell viability by itself, but enhanced plasma-induced cell death (Fig 8A). This was confirmed by flow cytometry in which apoptotic cells were visualized by annexin V-FITC and PI staining ( Fig 8B).  CAP can produce various reactive species such as the OH radical, which is the most highly reactive and can damage most macromolecules including DNA, proteins, lipids, or polysaccharides. OH is 100 fold more potent than H 2 O 2 and O 2 − and can affect molecules located a few nanometers from its site of generation [40]. However, given its high reactivity, OH has an extremely short half-life (in the nanosecond range). Thus, although CAP can generate OH radicals at μM concentrations, these species are unlikely to reach cells and cause damage to biomolecules; this can only be achieved by long-lived H 2 O 2 species [41,42]. As such, it is debatable whether plasma medicine constitutes nothing more than H 2 O 2 treatment. The present study showed for the first time that two major reactive species-H 2 O 2 and O 2 − -can penetrate to a depth that is sufficient to reach cells; we also propose a novel model of in situ generation of OH, which acts as the final effector causing cellular damage (Fig 9). iron-catalyzed OH production thereby greatly influencing cell viability. These findings provide a basis for achieving optimal biological effects by plasma treatment using this in situ mechanism for OH generation.

Conclusion
In conclusion, the present study demonstrated for the first time by biological experiments that H 2 O 2 and O 2 − are the two major reactive species that are produced by plasma treatment. Given that the concentration of either species is insufficient to induce cell death, we propose that the OH radical generated in situ of the cells by the Haber-Weiss reaction ultimately causes cell damage and death. Supporting Information S1 Fig. Experimental setup for plasma treatment. The apparatus consisted of a gas flow controller, high-voltage power supply, oscilloscope, and plasma jet. (TIF)