Antitumor abscopal effects in mice induced by normal tissue irradiation using pulsed streamer discharge plasma

An antitumor abscopal effect is occasionally observed in radiotherapy and plasma treatment. It is a remote antitumor effect induced by tumor irradiation that delays the growth of other distant, nonirradiated tumors. In this study, it was demonstrated that the plasma irradiation of normal tissues (not tumors) also leads to an abscopal effect. When a pulsed streamer discharge was irradiated to the left flanks of mice where no tumor existed, the growth of murine colorectal carcinoma CT26 tumors in their right limbs was delayed. This abscopal effect was significant for mice with small tumors before plasma irradiation, whereas it was not significant for those with large tumors before plasma irradiation. The abscopal effect induced by normal tissue irradiation was compared to the antitumor effect induced by direct tumor irradiation. Contrary to our expectation, normal tissue irradiation delayed the tumor growth equally or more than the direct tumor irradiation under the present experimental conditions.


Introduction
Atmospheric pressure nonthermal plasma is an emerging technology used for cancer treatment. The plasma irradiation of tumors can reduce their growth in various mouse models [1][2][3][4]. Plasma is a partially ionized gas generated by applying a high voltage to the gas. It produces reactive oxygen and nitrogen species (RONS) via collisions of gaseous molecules/atoms with electrons accelerated by the high electric field in the plasma. The plasma-generated RONS acting on cancer cells are considered to be responsible for the antitumor effects [4,5].
Recent studies suggested that plasma irradiation also activates antitumor immune responses [6][7][8][9][10]. Our previous study [11] suggested that the plasma irradiation of murine melanoma B16F10 tumors in mice may activate an adaptive immune response specific to B16F10. This was supported by the high secretion level of inflammatory cytokine interferon-γ from the splenocytes of plasma-irradiated mice when co-cultured with B16F10 cells. Our following study [12] showed that plasma irradiation induces a long-term, systemic antitumor effects on mice. In that experiment, plasma-irradiated B16F10 tumors were resected from mice; then 2 weeks later, fresh B16F10 cells were reinoculated at sites distant from the resected sites. The plasma irradiation of primary tumors delayed the growth of reinoculated tumors. The increased level of cytotoxic T cells (CD8 + ) observed in the reinoculated tumors of plasma-irradiated mice suggested the activation of adaptive immune responses. Other researchers also reported the possibility of plasma-induced antitumor immunity by vaccinating mice with plasma-treated cancer cells [13], irradiating tumors with plasma [14], and applying plasma-treated saline to tumors [15]. In [12], it was also suggested that the plasma-induced immunity may inhibit recurrence after tumor resection.
Our previous study [11] also showed a plasma-induced abscopal effect, which is an antitumor effect on distant nonirradiated tumors. The plasma irradiation of one tumor in mice with two distant B16F10 tumors reduced the growth of another nonirradiated tumor. The plasma-induced abscopal effect was also observed for murine breast cancer 4T1 in another group [14]. The abscopal effect has been occasionally observed in RT [16,17], but its mechanism is not clear. Immune system activation is considered to be responsible for this effect.
In the experiments of plasma-induced abscopal effects described above [11,14], the tumors were irradiated with plasma. However, our recent experiments showed that the abscopal effect is also observed when the plasma is irradiated to normal tissues of tumor-bearing mice. In this study, the abscopal effect induced by normal tissue irradiation is reported using a murine colorectal carcinoma CT26 tumor mouse model. For the irradiation, a pulsed streamer discharge was used. The streamer discharge is a branching, filamentary shaped nonthermal plasma typically generated in atmosphericpressure air or oxygen-nitrogen mixtures [21]. It produces many types of RONS which may be effective for the cancer treatment [22].

Plasma device
The plasma device for generating the streamer discharge consisted of a high-voltage rod electrode with a hemispherical tip and a grounded plate electrode to put a mouse on [11,12]. A streamer discharge was generated between the hemispherical tip and mouse skin. The plate electrode was heated to 37 • C to prevent the mouse from losing heat. The rod 3 mm in diameter was concentrically placed inside a glass tube with 4 mm inner diameter. The hemispherical tip was protruded 1 mm from the glass tube end. Oxygen gas humidified with a water bubbler was flowed through a glass tube at a rate of 0.5 L min −1 (relative humidity ⩾ 90%). Humid oxygen was used because a large amount of reactive oxygen species production was expected. A 5 mm width grounded metal tape was wrapped around the glass tube at 20 mm from the tube end to generate a dielectric barrier discharge inside it, which stabilized the streamer generation. More details of the plasma device and streamer discharge characteristics are given in section S1 of supplementary materials (available online at stacks.iop.org/JPhysD/55/17LT01/mmedia).

Cells and animals
CT26 cells purchased from RIKEN BioResource Center (Ibaraki, Japan) were cultured in RPMI media supplemented with 1% penicillin-streptomycin and 10% fetal bovine serum at 37 • C with 5% CO 2 . The experiments were performed in two different laboratories at Hongo and Komaba. In Hongo experiments, BALB/c mice (female, 7 weeks old) were purchased from Charles River Laboratories Japan, Inc. (Kanagawa, Japan), while in Komaba experiments, the mice (female, 6 weeks old) were purchased from CLEA Japan, Inc. (Tokyo, Japan). The mice were maintained in animal facilities for 1 week before starting the experiments. The hair at tumor inoculation and plasma irradiation sites was removed by applying depilatory cream. The right limbs of the mice were inoculated subcutaneously with 2 × 10 5 cells in a 100 µL phosphate-buffered saline (figure 1(a)). In Hongo experiments, 10% matrigel (BD, NJ, USA) was mixed with saline. The mice were anesthetized using isoflurane during plasma irradiation. All animal experiments were approved by and conducted in accordance with the guidelines of the Animal Ethics Committee of the University of Tokyo.

Plasma irradiation
In the present experiments, the abscopal effect was not stably observed. To obtain a significant result, six experiments were  respectively. The increase in pulse width increased the discharge intensity. To generate discharges with approximately equivalent intensities but with different pulse widths, the gap between the hemispherical tip and mouse skin was set to 3 mm larger than the threshold gap distance at which a spark discharge occurred. The threshold gap distance was approximately 5 mm for the 20 ns pulse and 10 mm for the 90 ns pulse. Thus, discharges of not the same but with similar intensities were obtained with different pulse widths. Six days after cell injection (day 6), the mice were randomized into three groups: (1) normal tissue irradiation, (2) tumor irradiation, and (3) control groups. There were 4-8 mice in each group, depending on the experiment. The left flank of mice in the normal tissue irradiation group was irradiated with plasma (figure 1), whereas it was directly irradiated to tumors in the tumor irradiation group. The distance between the plasma irradiation point and cell injection point was 2-3 cm. Plasma irradiation was done for 10 min/day and was kept from days 6 to 10. Tumor sizes were measured using calipers and tumor volumes were calculated as (4/3)π × (length/2) × (width/2) 2 .
Skin surface temperature of mice was measured using infrared thermography (Avio, H2640, NEC). The increase in skin temperature during plasma irradiation was approximately 1 • C ( figure S2 in supplementary materials). Additionally, plasma irradiation did not cause any visible damage to the mouse skin. Therefore, it can be assumed that the streamer discharge did not have considerable thermal effects on the treatment of mice.
In Komaba experiments, immune cells in tumors were measured using flow cytometry, but the results are not shown because it was a preliminary analysis. Results of flow cytometry analysis will be presented in future works.

Effect of plasma irradiation on tumor growth
The effect of plasma irradiation on tumor growth suppression was significant in some experiments, but it was marginal in others. The tumor growth curves in figure 2 show that in experiment #2, (A) exhibits a significant effect, whereas in experiment #3, (B) only exhibits a marginal effect. The tumor growth curves in all experiments (#1-#6) are shown in figure S3 in supplementary materials. The antitumor effect caused by normal tissue irradiation is an abscopal effect, whereas that caused by tumor irradiation is not, since the plasma was directly irradiated to the tumors. Although both antitumor effects may The tumor volumes on the final day, V f , were used to examine the efficacy of plasma treatment. For example, the tumor volumes on day 27 were used for experiment #1, and those on day 21 were used for experiment #2 (table 1). To compare the V f in experiments #1-#6 with different final days, V f was normalized by dividing it by the average tumor volume of the control mice on the final day, V f,ctrl , in each experiment. Figure 3 shows the cumulative frequency of normalized tumor volumes on the final day, V f,n (= V f /V f,ctrl ), including all mice in experiments #1-#6. It appears that normal tissue irradiation delayed the tumor growth. Contrary to our expectations, the tumor irradiation caused less antitumor effects than the normal tissue irradiation.

Effect of tumor volumes before plasma irradiation
To examine the efficacy of plasma irradiation, the plasmairradiated mice were classified into three groups: (a) Efficacious group, in which the tumor volume on the final day was much smaller than the average of the control group, satisfying V f ⩽ V f,ctrl − 0.5σ f,ctrl , where σ f,ctrl is the standard deviation in the tumor volume of the control mice on the final day of each experiment.   The analysis showed that the mice in the efficacious group had smaller tumor volumes on day 6, which was the starting day of plasma irradiation, than those in the inefficacious group, as shown in figure 4. The mice with small tumor volumes on day 6 had a high efficacy both in the normal tissue irradiation and tumor irradiation groups.
To further investigate the influence of tumor volume on day 6, V 6 , on plasma irradiation efficacy, the V f,n of all experiments #1-#6 were plotted together as a function of V 6 . Figure 5(a) shows the plots for the normal tissue irradiation and control groups. The V f,n of control mice was distributed higher and lower than 1.0 (= normalized average volume of control mice for each experiment), irrespective of V 6 . However, the V f,n of normal tissue-irradiated mice tends to be distributed lower than 1.0 for small V 6 (e.g. V 6 ⩽ 70 mm 3 ), whereas it was distributed higher and lower than 1.0 for large V 6 (e.g. V 6 > 70 mm 3 ). This suggests that normal tissue irradiation is effective for mice with small V 6 . Tumor irradiation also showed a similar tendency, as demonstrated in figure 5(b).
To statistically analyze this hypothesis, the plots in figures 5(a) and (b) were divided into two groups based on the value of V 6 : V 6 ⩽ 70 mm 3 and V 6 > 70 mm 3 . Figure 6(a) shows the average and standard deviation of V f,n for all mice satisfying V 6 ⩽ 70 mm 3 . Both normal tissue and tumor irradiation showed significant effects in reducing tumor growth. In contrast, for mice satisfying V 6 > 70 mm 3 , almost no effects were observed ( figure 6(b)). Thus, normal tissue and tumor irradiation significantly reduced the tumor growth in mice with a small V 6 , whereas no significant effect was observed in mice with a large V 6 . The threshold value of V 6 seems to be at around 70-90 mm 3 in this series of experiments.
In experiment #1, four of the eight mice in the normal tissue irradiation group had tumors on left limbs, which was the opposite side of other mice in experiments #1-#6, and the plasma was irradiated to the right flanks. All of the four right flank-irradiated mice were in the efficacious group. This suggests that normal tissue irradiation to either the left or right flank was effective, but the number of mice for right flank irradiation (N = 4) was not large enough to obtain a significant result.
In this study, high-voltage pulse widths of 20, 60, and 90 ns were used. However, figure 4(a) suggests that the pulse width does not have significant effects on the abscopal effect.

Discussion
In RT, there are few studies on the abscopal effect induced by normal tissue irradiation. Camphausen et al [23] observed a dose-dependent abscopal effect induced by normal tissue irradiation using Lewis lung carcinoma and T241 fibrosarcoma tumor mouse models. Irradiation with 10 Gy × 5 and 2 Gy × 12 showed abscopal effects, but the effect of the former irradiation was more marked. The abscopal effect was found to be mediated by p53, as shown in using p53 knockout mice and p53 inhibitor. It was hypothesized that local inflammation caused by normal tissue irradiation makes p53 to produce cytokines and other factors, which may exert some effects on distant tumors. The p53-mediated abscopal effect was also reported in tumor-irradiated mice [24].
Contrary, there are also some studies in RT reporting that normal tissue irradiation has no positive abscopal effects. Shiraishi et al [25] reported that normal tissue irradiation (6 Gy) in mice with CT26 tumors has no abscopal effects. Yasuda et al [26] reported that normal tissue irradiation (2 Gy × 10) in mice with subcutaneous and liver metastasis CT26 tumors enhanced the growth of metastatic tumors.
Thus, this phenomenon is not well-understood. It is unclear whether the effect is positive or negative. This study showed the positive abscopal effect of normal tissue irradiation. Here, tumor volume before plasma irradiation (day 6) was considered as a factor that determines the efficacy of this treatment.
The even or less efficacy of tumor irradiation compared to normal tissue irradiation was contrary to our expectation.
Tumor irradiation was expected to be more effective because it may produce DAMPs that trigger the immune response. However, this study implies that normal tissue irradiation might be more efficacious at least under the present conditions. For clinical use, normal tissue irradiation is preferable because it is easier than irradiating tumors. This difference between normal tissue and tumor irradiation should be clarified in future studies.
Some questions that need to be solved in future studies arise from our results.
(a) The mechanism of this phenomenon should be examined.
It might be the p53-mediated local inflammation, as suggested in RT [23], or other possible mechanisms. (b) It should be examined why the abscopal effects were preferably observed in mice with small tumor volumes before plasma irradiation. (c) The abscopal effect was also caused by tumor irradiation using plasma [11,14]. It should be examined whether the mechanisms of the abscopal effects caused by normal tissue and tumor irradiation are the same or not. (d) Adaptive immune activation may occur simultaneously with the abscopal effects caused by tumor irradiation [11]. It should be determined if the abscopal effect caused by normal tissue irradiation also leads to adaptive immune activation.
This study supports the existence of abscopal effects caused by normal tissue irradiation with significant difference, which was only reported in few studies. The efficacy of this treatment was shown to depend on the tumor volume before plasma irradiation. The elucidation of the mechanism indicated in (a) is currently ongoing. So far, immune cells in tumors have been analyzed using flow cytometry as a preliminary analysis. The mechanism will be discussed in future studies.

Conclusions
The abscopal effect caused by normal tissue irradiation was demonstrated using pulsed streamer discharge in CT26 tumor-bearing mice. It supports the existence of abscopal effects caused by normal tissue irradiation, while few studies have reported this phenomenon for RT. The abscopal effect was observed with significant difference in mice with small tumor volumes on day 6, whereas it was not observed in those with large tumor volumes on day 6. The threshold tumor volume on day 6 seemed to be at around 70-90 mm 3 in this series of experiments. Direct tumor irradiation was also performed to compare it with normal tissue irradiation. Contrary to our expectations, tumor irradiation caused less antitumor effects than normal tissue irradiation.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).