The effect of local non‐thermal plasma therapy on the cancer‐immunity cycle in a melanoma mouse model

Abstract Melanoma remains a deadly cancer despite significant advances in immune checkpoint blockade and targeted therapies. The incidence of melanoma is also growing worldwide, which highlights the need for novel treatment options and strategic combination of therapies. Here, we investigate non‐thermal plasma (NTP), an ionized gas, as a promising, therapeutic option. In a melanoma mouse model, direct treatment of tumors with NTP results in reduced tumor burden and prolonged survival. Physical characterization of NTP treatment in situ reveals the deposited NTP energy and temperature associated with therapy response, and whole transcriptome analysis of the tumor identified several modulated pathways. NTP treatment also enhances the cancer‐immunity cycle, as immune cells in both the tumor and tumor‐draining lymph nodes appear more stimulated to perform their anti‐cancer functions. Thus, our data suggest that local NTP therapy stimulates systemic, anti‐cancer immunity. We discuss, in detail, how these fundamental insights will help direct the translation of NTP technology into the clinic and inform rational combination strategies to address the challenges in melanoma therapy.


| INTRODUCTION
Skin cancer is the most common cancer worldwide, and melanoma is responsible for 65% of skin cancer-related deaths. 1 The incidence of melanoma is growing globally at a rate faster than other malignancies, and distressingly, the patient demographic includes younger individuals compared to other cancer types. 1,2 Early-stage melanomas are usually treated with surgical excision, and in the majority of cases, treatment is curative. 3 However, certain patients experience relapse and further dissemination of the disease. For more advanced stages, surgical resection is no longer an option and systemic treatment is required for managing metastasis. 4,5 While significant advancements in metastatic melanoma have been made in the last decade with the introduction of immune checkpoint blockades and targeted therapies, challenges still remain, including significant associated adverse effects, limited efficacious populations, and development of resistance. 6 Therefore, novel therapeutic strategies and combinations are still required for both early-and late-stage melanoma treatment.
Non-thermal plasma (NTP) technology has been emerging onto the oncology scene, and medical NTP devices operate by ionizing a gas (e.g., argon, helium, air) at atmospheric pressure and room temperature. NTP systems for biomedical applications have been thoroughly characterized for in vitro systems, which include component analysis (e.g., pulsed-electric fields, UV radiation), 7,8 gas-phase measurements of excited and reactive species, 9 and liquid chemistry quantification. 10,11 Preclinical research has reported the ability of NTP treatment to induce immunogenic cell death (ICD) in several cancer types including melanoma and even indicate abscopal effects in vivo. [12][13][14][15] Treatment of cancerous cells in vitro increased tumor immunogenicity through the emission of multiple danger signals such as surfaceexposed calreticulin (CRT), secreted adenosine triphosphate (ATP), and released high mobility group box protein 1 (HMGB1). [15][16][17][18] Furthermore, co-culturing NTP-treated cancer cells with ex vivo immune cells, particularly immature dendritic cells (DCs), increased DC maturation, cytokine release, and phagocytosis. 15,18,19 Using the gold standard vaccination assay to test ICD-inducing agents, our laboratory has also demonstrated that the short-lived reactive oxygen species (ROS) generated by NTP, particularly •OH, •NO, O/O 3 , are the main effectors of NTP-induced ICD. 13 Thus, NTP is similar to other localized ROS therapies, such as photodynamic therapy 20,21 and electrochemical therapy, 22,23 in which harmful side effects can be modulated, while local anti-cancer responses and systemic, immune responses can still be induced. Therefore, NTP is an attractive therapy for superficial skin cancers, which are easily accessible for treatment with the device.
While clinical pilot studies with NTP have only recently started, first results on cutaneous skin lesions indicate a promising treatment modality. [24][25][26][27] Metelmann et al. applied NTP to six patients with locally advanced squamous cell carcinoma of the oropharynx as part of their palliative treatment. Treatment not only reduced the need for pain medication and reduced odor in infected ulcers, but partial remission was observed in two patients for at least 9 months. [24][25][26] Friedman et al. treated five patients diagnosed with actinic keratosis, which are pre-cancerous skin lesions with the potential to develop into squamous cell carcinoma. After a single NTP treatment, 9 of the 17 treated lesions were fully resolved and 3 showed significant improvement in a 1-month follow-up. 27 Most importantly, in both these clinical studies, no serious adverse effects were reported, which indicates the high tolerability of NTP therapy.
As promising clinical reports with NTP treatment are becoming available and more preclinical reports are focusing on the immunological implications of treatment, it is crucial and urgent to gain deeper insight into NTP treatment effects. Of note, a current gap in understanding is on how NTP therapy affects the cancer-immunity cycle: a series of events directed by the patient's immune system, which leads to effective killing of cancerous cells. 28 This understanding would provide application insight into how NTP can be used in the clinic and strategically combined with existing therapies to further enhance patient anti-cancer immunity for greater clinical benefit.
In this study, we investigated the anti-cancer effects of NTP

| Physical characterization of the NTP device
A microsecond-pulsed dielectric barrier discharge (DBD) plasma system was used for all experiments. While NTP treatment with a DBD plasma system has been performed in the past, 12,29,30 direct comparisons with our system cannot be made due to the different electrical parameters of the systems used. Therefore, we first performed a pilot study on C57BL/6J mice bearing subcutaneous B16F10 melanoma tumors (n = 4-5) to determine the ideal, therapeutic NTP treatment intensity ( Figure S1). From the study, the optimal NTP treatment parameters were determined (Table 1) and the thermal and electrical profile of this NTP regime was further investigated.
Since NTP discharge characteristics are highly influenced by their treatment target, 31 all measurements were performed in situ, on the surface of a mouse, analogous to treatment of tumors (details in Section 5). Using a high-sensitivity, cooled thermal imaging camera, the temperature of the discharge and changes to the mouse skin were monitored over time, to determine whether treatment was thermally well-tolerated. Selected thermal images are shown ( Figure 1a) and the full image collection process is also provided (Video S1). At baseline, the temperature of the mouse skin was around 22 C, increased up to 31 C within 10 s of NTP treatment (NTP on), and returned to baseline within 40 s after treatment was completed (NTP off) (Figure 1b).
Images of the skin taken immediately after NTP treatment also indicated no visible necrosis or thermal damage to the skin ( Figure S2).
Further tests showed that NTP treatment up to 1 min did not further increase the temperature of the skin ( Figure S3 and Video S2). At F I G U R E 1 Temperature and energy characterization of non-thermal plasma (NTP) treatment. (a) The development of the NTP discharge was observed, and the temperature of the treated mouse skin was monitored over time (t, seconds). (b) At the end of the 10 s treatment, the mouse skin reached 31 C but returned to baseline (22 C) within 40 s. (c) The spatial temperature profile was also measured across the skin under the NTP device (represented by the red line). The temperature was highest directly under the NTP device, while the skin 6.5 mm away (represented by the dotted lines) was unaffected. (d) The schematic setup used to measure the (e) voltage and current of a single NTP discharge on the surface of the mouse is shown. (f) The NTP applicator was adjusted over a range of distances and the (g) temperature and (h) energy per pulse was measured. Data are represented as mean ± SEM (n = 5-14). The yellow and white arrows in (a) and (f) indicate the position of the dielectric barrier discharge (DBD) electrode and skin surface, respectively approximately 6.5 mm away from the center of treatment, the temperature of the surrounding tissue was unaffected ( Figure 1c). Altogether, the rise in temperature associated with NTP therapy is minimal. Therefore, these data suggest that NTP temperature is not associated with its therapeutic effect.
In our previous reports, we have indicated the importance of the delivered NTP treatment energy in dictating biological response in vitro. 32 Therefore, we determined the NTP energy per pulse on the mouse during treatment (Figure 1d). This was calculated based on voltage and current measurements (Figure 1e) of a single pulse and after accounting for displacement current (details in Section 5). At a 1 mm application distance, the energy per pulse (ε pulse ) on the mouse was 0.9 ± 0.3 mJ. Since the delivered energy of the plasma is highly subjected to the application distance, both the temperature and energy profiles were also studied for a range of distances. As the application distance increased, the temperature rise on the skin surface decreased, up to 5 mm where no changes were observed   Of particular interest is the IL-6/JAK/STAT3 pathway, which is known to drive tumor proliferation, survival, and even metastasis. 35 This pathway is hyperactive in many types of cancers, including melanoma, and hyperactivation of STAT3 in tumor-infiltrating immune cells is associated with an immunosuppressive effect. 36,37 Here, detailed list of the core enrichment genes and analysis scores and the genes from the differential gene expression (DESeq2) analysis are also provided (Tables S1 and S2, respectively).
Taken together, our data suggest that NTP could reduce tumor burden and prolong mouse survival via modulating the tumor profile and microenvironment and stimulating a dynamic anti-cancer immune response. Therefore, in the next sections, we investigated how NTP treatment affects tumor immunogenicity and stimulates the anticancer immunity cycle over time (Days 7, 10, and 14).

| NTP affects tumor immunogenicity
Given that NTP has been reported to induce ICD, and an increase in the UPR pathway was measured in the transcriptome analysis ( Figure 2f), we evaluated how NTP affects tumor immunogenicity.
Tumors were resected, sectioned, and stained individually for specific cancer ligands. Quantification was performed with our developed software, which identifies cell nuclei within the tumor and analyzes the signal around it (Figure 3a).
The expression of calreticulin (CRT) and CD47 in the tumor was evaluated, as they are key signals for the innate immune response.
CRT is a hallmark of ICD and a potent pro-phagocytic signal that is counterbalanced by CD47, an anti-phagocytic signal. 39  Taken together, the data indicate that NTP treatment modulates the expression of cellular proteins in the tumor, and the durability of the NTP effect seems transient for some proteins while it is persistent for others.

| NTP affects dendritic cells and antigen presentation
Following an increase in tumor immunogenicity, it is crucial for antigen-presenting cells to be recruited and activated for the development of specific anti-cancer immunity. 28 Dendritic cells (DCs), which are key antigen-presenting cells, must capture tumor antigens from dead and dying cancer cells for processing and antigen presentation. 40 Therefore, we evaluated the presence of DCs (MHC-II + /CD11c + ) in the tumors using flow cytometry analysis. The detailed gating strategy is provided in supplementary information ( Figure S4). Since tumors were too small for analysis immediately at the end of NTP treatment , and 14 also showed higher amount of CD11c + antigen-presenting cells, with localization occurring predominantly in the periphery of the tumor (black arrows) compared to the center (red arrows). c) CD103 + DCs, a subtype responsible for cross-presentation of antigens to T cells, were also measured in the tumor. Flow cytometry analysis of the spleen (d, e) and TDLN (f, g) also showed a higher frequency of DCs and CD103 + DCs following NTP treatment. Data are represented as mean ± SEM and each biologically independent sample is shown (n = 3-5). Statistical significance was determined using the generalized linear mixed model and the NTP treatment condition was compared to that of the control for each day. The fold change (FC) compared to the untreated control is also reported for statistically significant comparisons. *p ≤ 0.05 (detailed p values are reported in the text) with NTP showed a higher density of CD11c + cells compared to the untreated, and localization occurred predominantly in the periphery ( Figure 4b).

T A B L E 2 Histological scoring of immune cell populations in the tumor
CD103 + DCs, a subtype with high capacity for antigen crosspresentation to T cells, 41 were also quantified in the tumor with flow cytometry to inform on the antigen presentation stage of the cancerimmunity cycle. A slight increase was observed on Day 10 in the NTP treatment group (11.2 ± 1.5%) compared to untreated ( Taken together, the data suggest that NTP treatment can enhance anti-cancer immunity by increasing antigen-presenting cells to the tumor (primarily in the periphery) and enhancing their migration to the surrounding lymph organs for antigen presentation.

| NTP affects T cells in the tumor-draining lymph node
Presentation of captured tumor antigens from DCs must result in the priming and activation of T cells in order for the anti-cancer immunity cycle to progress. 28 Therefore, we evaluated the cytotoxic (CD8 + ) and helper (CD4 + ) T-cell populations in both the spleen and lymph node after NTP treatment. Using FoxP3 and CD25 lineage markers, CD4 + T cells were further discriminated between T regulatory cells (Tregs; FOXP3 + /CD25 + ) and non-Tregs (FOXP3 À /CD25 À ). Therefore, it appears that following NTP treatment, adaptive immune cells are also stimulated to have a more anti-cancer profile.
Taken together, it appears that NTP treatment can transiently potentiate the next steps of the cancer-immunity cycle in the TDLN as more CD8 + and non-regulatory CD4 + T-cell activation was measured along with less exhaustion. This parallels the measured increase in IFN-γ. Moreover, it has been reported that an increase in ICOS expression in both CD4 and CD8 T cells parallels an increase in the CD8 + /Treg ratio. 45 Indeed, while we did not observe statistically significant changes to the T-cell subtype populations, the CD8 + /Treg ratio was significantly higher in NTP-treated mice (Figure 5c), which is advantageous for developing robust anti-cancer immunity.  Based on RNA sequencing analysis of the tumor, we observed the signaling pathway most affected by NTP was the IL-6/JAK/STAT3 pathway, which was downregulated after treatment. This is highly favorable, as hyperactivation of this pathway has been associated with tumor progression, immunosuppression, and poor clinical prognosis and therefore, warrants independent follow-up studies to validate these findings. In-line with previous in vitro and in vivo reports, [12][13][14][15]18 our NTP treatment also induced acute UPR activation and CRT emission, which are characteristic of ICD in the tumor. 48,49 This was followed by higher infiltration of DCs into the tumor and more migration to the spleen and TDLN. Moreover, a higher frequency of CD103 + DCs was observed in both the tumor and TDLN of mice receiving NTP treatment, which indicates more DCs are primed for antigen presentation to T cells. 41  Altogether, these fundamental insights can help direct the translation of NTP technology into the clinic and inform potential combination strategies for enhanced patient outcome.

| NTP affects T cells in the tumor and cytotoxic immune responses
In fact, in recent years, there has been an increased endeavor to find optimal combination strategies against tumors, using NTP in combination with existing therapies. 12,52,54,[56][57][58][59][60]  Not only did our NTP treatment modulate the tumor, but it also recruited more DCs into the TME, and the anti-cancer T cells were more activated and less exhaustive. This is highly desirable as the immunosuppressive environment induced by melanoma is known to elicit cytotoxic lymphocyte dysfunction. 70 In fact, recent breakthroughs with immune checkpoint inhibitors, including anti-PD-1 and anti-CTLA-4, aim to restore CD8 + T-cell function, though systemic administration of these inhibitors is often associated with heavy adverse effects. In our study, not only were more activated CD8 + and non-regulatory CD4 + T cells found in the TME following local NTP treatment, but these activated cells were also found in the TDLN. Furthermore, the CD8 + T cells and the NK cells in the TME were modulated to exhibit a higher cytotoxic capacity, as indicated by elevated anti-tumor functions of T cells and NK cells in the TME is advantageous, though elucidation of the mechanism is still needed.
Another crucial non-malignant cell type of the TME is the cancerassociated fibroblast (CAF), which has been implicated to be particularly detrimental for the outcome of melanoma therapy. 70 CAFs are highly involved in tumor angiogenesis, remodeling of the extracellular matrix, and secretion of pro-tumorigenic and immunosuppressive factors. 70 Indeed, STAT3-associated resistance against BRAF-inhibitors has been closely tied to the secretion of fibroblast growth factor 2 (FGF2) signaling. 66 In addition, CAFs also present a physical barrier around the tumor, which can impede therapy. The effect of NTP on CAFs has only recently been proposed and studied. 19 Finally, the practicality of NTP treatment of patients must also be considered as this technology moves toward clinical translation.
Under research laboratory settings with stationary in vitro cancer models (e.g., monolayer cell cultures, 3D spheroids), treatment parameters are easily-and well-controlled, but much of this control is lost when treatment is moved to the hospital setting due to practicality and treatment subject. Currently, NTP therapy relies heavily on clinician judgment and experience, as these devices are handheld and operated by the clinician. 26,27 Consequently, this leads to large variability which is further amplified when the treatment area is significantly larger than the NTP applicator and translation across the lesion is necessary. Furthermore, spatial displacement of the treatment area due to the breathing motion of the patient, a phenomenon referred to as respiratory-induced tumor motion, 73 can also affect off-target treatment and NTP therapy response. This is also observed in our in vivo study, in which the respiration of the mouse can change the distance between the tumor and the DBD applicator. While the NTP energy per pulse is relatively the same when the application distance is between 1 and 4 mm, it is dramatically reduced at larger deviations ( Figure 1h). Indeed, this is crucial as we have previously identified that NTP treatment energy may be an essential link to treatment response, 32 and therefore, precise control over NTP therapy is necessary.
Some of these challenges are not exclusive to NTP and are also faced by radiation technology. Therefore, we can begin to address these challenges by incorporating techniques and experiences from the field of radiotherapy. In fact, our laboratory is looking to incorporate real-time tumor tracking and robotic control with NTP therapy.
This is a methodology used in radiotherapy that applies active compensatory motion of the radiation source with the tumor to counteract any tumor wandering. 74,75 By applying advances in tumor tracking technology and precision robotics, we may be able to address these challenges and reduce the variability of NTP therapy associated with both the operator and the patient.

| CONCLUSION
It is clear that the field of "plasma oncology" has just started to scratch the surface into understanding how NTP affects not only the tumor but also the TME and the cancer-immunity cycle at large. Here, we provided a detailed examination into how NTP treatment affects downstream, anti-cancer immune responses over time in vivo. We also characterized the electrical and thermal properties of NTP therapy in situ over a range of application distances. These results showed that NTP was thermally and electrically well-tolerated and provided insight into the stability of local NTP application. Studies into NTP characterization in more clinically-relevant settings will further facilitate translation of this technology, and more in-depth investigations into the holistic effects of NTP treatment is critically required. This will shed light on the potential niche for NTP treatment in oncotherapy via strategic combination with other treatment options.

| NTP treatment
A microsecond-pulsed dielectric barrier discharge (DBD) system previously described, was used for all NTP treatments. 33 Briefly, a microsecond pulser (Megaimpulse Ltd., Russia) generated a 30 kV output pulse with rise time fixed within 1-1.5 μs and a pulse width of 2 μs.
The frequency of the pulses was fixed at 700 Hz and treatment was performed for 5 consecutive days. The applicator of the system was a copper electrode, covered with a quartz dielectric, and was connected to the output of the microsecond pulser. The applicator was held by hand above the tumor ($1-3 mm) for treatment. Here, an electrically safe plasma was created in direct contact with the tumor, and the surrounding gas and tissue were not significantly heated.

| Power measurements
To determine the plasma power during in vivo treatment of tumors, voltage and current was measured on the skin surface of a mouse during NTP treatment (Figure 1d). A mouse was sacrificed on the day of measurement and the hair was removed before treatment. The NTP applicator was fixed at a specified distance above the skin surface using a z-positioner, and NTP was generated at 700 Hz. Voltage was measured using a 1000X high-voltage probe (P6015A, Tektronix), and the current was measured with a current monitor (4100, Pearson Electronics, Inc.).
The voltage and current waveforms were recorded on an oscilloscope (DSOX1102G, Keysight) with a 50 ns sampling rate. Voltage and current was used to determine instantaneous power (P(t)): P(t) = V(t) Â I(t). The energy per pulse (ε pulse ) was calculated: where Δt was 50 nanoseconds and the number of recorded samples (n) was 2000. In order to account for displacement current, the energy per pulse from displacement current (ε pulse(displacement) ) was also measured when the high-voltage pulse was applied to the NTP applicator 30 mm above the mouse. This value was subtracted from all other energy measurements. Taken together, the energy per pulse reported here (ε pulse ) is: ε pulse = ε pulse(discharge) -ε pulse(displacement) . The energy per pulse was measured over a range of application distances (1-10 mm).

| Thermal imaging
The thermal images shown in Figure 1 34 An FDR ≤ 0.05 was set as the threshold for affected pathways.

| Flow cytometry analysis
Characterization of tumor-infiltration lymphocytes (TILs) was per-  and Goat anti-rat IgG (H + L) AlexaFluor594 secondary antibodies were also used at 1/500 dilutions. All tumor slides were imaged on the same day, as detailed above, and batch processed with our software in order to reduce variability. TxRed (FOXP3) and 0.3 for GFP(CD3). Masked pixels that did not overlap with either nuclei or cytoplasm were removed. Similar to the analysis for DAPI, binary objects were indexed using connected component analysis and small objects (less than 20 pixels) were filtered. In this case, large objects greater than 500 pixels were also filtered. Cells that contained fluorescent signal overlapping with either the nuclei or the cytoplasm were counted and those that contained both fluorescent markers were considered to be double positive.

| Statistical analysis
Statistical differences for this study were analyzed using the linear

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.