Targeting the cancer cell cycle by cold atmospheric plasma

Abstract

Cold atmospheric plasma (CAP), a technology based on quasi-neutral ionized gas at low temperatures, is currently being evaluated as a new highly selective alternative addition to existing cancer therapies. Here, we present a first attempt to identify the mechanism of CAP action. CAP induced a robust ~2-fold G2/M increase in two different types of cancer cells with different degrees of tumorigenicity. We hypothesize that the increased sensitivity of cancer cells to CAP treatment is caused by differences in the distribution of cancer cells and normal cells within the cell cycle. The expression of γH2A.X (pSer139), an oxidative stress reporter indicating S-phase damage, is enhanced specifically within CAP treated cells in the S phase of the cell cycle. Together with a significant decrease in EdU-incorporation after CAP, these data suggest that tumorigenic cancer cells are more susceptible to CAP treatment.

Figure 1. CAP effect on the cell migration and velocity distribution among the chosen cell population.

The dependence of the cells migration rates on the length of the CAP treatment from 0 to 180 seconds are shown for WTK, 308 and PAM 212 cells (Figure 1A–1C, left panel). SEM (standard error of mean) is shown with an error bars. For each time point ~50–60 cells were analyzed. The difference between groups were considered statistically not significant for p-value > 0.05 (no mark) and for p-value < 0.05 (single star), p-value < 0.01 (double star) and p-value < 0.001 (triple star) are considered significant. The probability distribution functions (PDF) of the cell migration rates (distribution of the cell velocities) for untreated cells (control, blue) and cells treated with cold atmospheric plasma for 60 seconds (plasma treated, red) are shown for WTK, 308 and PAM 212 cells (Figure 1A–1C, right panel). The experimental data (~250 cells of each cell type were analyzed for each condition) is fitted with inverse Gaussian distributions.

Figure 2. Cell viability studies in two threshold regimes.

MTT assay was performed on: (A) wild type keratinocytes (WTK), (B) 308 cells (papilloma) and (C) PAM212 cells (carcinoma). Control (not treated) is shown in blue, CAP treated for 60 seconds – in red and CAP treated for 120 seconds – in grey cells at 0, 4 and 24 hour time points. SEM is shown with an error bar. The experiment was repeated 3 times for each cell type and condition. Three stars denote extremely significant statistical difference (p-value < 0.001), one star – statistically significant difference (p-value < 0.05), data was considered statistically not different for p-value > 0.05.

Figure 3. Identification of the cell cycle change in G2/M-phase.

Figure 3A–3C Bright-field images of wild type keratinocytes (WTK), epidermal papilloma (308 cells) and epidermal carcinoma (PAM212 cells) cells are shown with a magnification 10x. Figure 3D–3L show cell cycle studies: Propidium Iodide content (horizontal axis) and normalized number of cells (vertical axis) are shown. The controls are shown in blue and cells CAP treated for 60 seconds are in red. The ratio of the number of cells (treated to untreated) in G2/M-phase with coefficients of variation (CV, in percents) is shown in the right top corner of each figure. Figure 3D – 3F shows cell cycle measurements in ~4 hours; Figure 3G–3I in ~24 hours; and Figure 3J–3L in ~48 hours after the CAP treatment for WTK, 308 and PAM 212 cells respectively. The data is shown for ~25000–50000 cells for each experimental condition. The measurements were repeated 2–3 times.

Figure 4. Studies of the cell population's distribution during the cell cycle after CAP treatment.

The detailed studies of the cell cycles for control and CAP treated for 60 seconds cells are shown. The correlation between DNA content (Propidium Iodide) and DNA-replicating cells (EdU-component) is shown for: Figure 4A, 4B – normal cells (wtk); Figure 4C, 4D–papilloma cells (308 cells); and Figure 4E, 4F–carcinoma cells (PAM212 cells) in ~24 hours after CAP treatment. Coefficient of variation (CV) was used to characterize the Propidium Iodide data (linear scale) and standard deviation (StDev) was used for EdU-data (log scale) in each phase. Fractions of the number of cells in each cell cycle phase are shown in percents. The data is shown for ~25000 cells for each experimental condition. The experiments were repeated 2–3 times.

Figure 5. Cell synchronization for G2/M increase studies.

308 cells were synchronized with nocodazole and analysis of the cell cycle was carried out for various time points. Figure 5A shows the time evolution of the 308 cells: the control cells (not treated cells) is shown in blue, the cells pre-treated with nocodazole is in orange and nocodazole pre-treated cells which were CAP treated for 60 seconds – in red. The cell cycle evolution after nocodazole removal is shown at ~0, 4 and 24 hour points. The cells were treated with CAP at time point ~2 hours after nocodazole removal (the state of the cell cycle for this time point is not shown). The final state of the systems control (not-treated cells), cells pre-treated with nocodazole only and cells pre-treated with nocodazole and treated with CAP is shown for time point ~24 hours. The change in ratio between G2/M fractions of the cells at each experimental condition is marked with a red square. Figure 5B shows detailed cell cycle studies of 308 cells for chosen time point ~24 hours .The DNA content (Propidium Iodide, linear scale) and DNA-replicating cells (EdU-component, log scale) characterized by CV (in percents) and StDev (arb. units) values, respectively. The data is shown for ~25000 cells. The measurements were repeated 2–3 times.

Figure 6. CAP targets the cell cycle phase.

The time sensitive studies of the 308 cells cell cycles are shown for the time points ~0 hour, ~4 hour and ~24 hours points after 60 seconds of CAP treatment. The control (not treated cells) is shown in blue, CAP treated cells is in red. Figure 6A shows the time evolution of the cell cycles (DNA content – vs. – DNA replicating cells) of control and CAP treated cell. The changes in EdU and γH2A.X signals are characterized by K-S maximum difference D_max and D_cr values at the selected time point. Figure 6B shows the correlation between DNA content and γH2A.X reporter for G0/G1 cell phase; Figure 6C shows correlation between DNA replicating cells (EdU) and γH2A.X for 308 cells in S-phase; and Figure 6D shows correlation between DNA content and γH2A.X for the 308 for G2/M-phase for ~0, ~4 and ~24 hours time points. ~25000 cells are shown for each experimental condition. Changes in the fraction of cells between the control (not treated) and CAP treated cells are shown for each cell phase: cell number decreases if the ratio is less than 1. The statistical description of the signals shown in the Figure 6A is a maximum differences between two distributions in the S-phase with confidence interval 99.9%; in the Figures 6B–6D D_cr – value is shown: D_cr < 1.3581 (p > 0.05) is considered not statistically significant and D_cr > 1.9526 (p < 0.001) is extremely statistically different.

Figure 7. The schematic model of the CAP impact on the cell cycle.

Figure 7A is the schematic representation of the cells distribution during the cell cycle phase for analyzed types of cells: WTK, 308 and PAM212. Figure 7B is schematic of the CAP affect on the cell cycle with consequent cell response.