A prospectus on innovations in the plasma treatment of cancer

Cold atmospheric plasma (CAP) emerges as a possible new modality for cancer treatment. More broadly, CAP has been tested in various applications such as disinfection, wound healing, dentistry, and cancer therapy.^1–6 High level understanding of the CAP interaction with cells and tissue depends on the notion that chemical elements of the CAP are potentially toxic, such as reactive oxygen species (ROS), which might promote a “plasma killing effect,” while others such as reactive nitrogen species (RNS) could produce a “plasma healing” effect. Forming various combinations of these species might provide a great potential for activation of specific signaling pathways in cells. CAP treatment possesses powerful lethal capabilities against tumor cells both in vitro and in vivo, and just as importantly, the normal counterpart cells have been shown to be less sensitive to the same CAP treatment.⁷ All these aforementioned effects are believed to be related to the plasma chemistry. While different species are produced as a result of plasma treatment, the role of other plasma effects such as charged particles and electric field still remains elusive.

What makes plasma unique is its ability to self-organize and to form coherent structures. These coherent structures could instantaneously modulate electric field, ROS/RNS, and charged particles. Thus, as a result of self-organization, plasma adaptivity to specific cells through tailoring its composition in situ might be possible. This unique feature makes plasma interaction with cells intrinsically selective.

The 2017 Ronald C. Davidson Award for Plasma Physics recognized the emerging field of plasma medicine and, in particular, plasma application in cancer therapy. However, the mechanism of plasma action and the precise role of the plasma effect in curing cancer are still subject to intense investigation. An analysis of potentially important plasma effects is the key objective of this Prospectus.

CAP action on living tissue is attributed to ROS and RNS formation. It has been suggested that RON/RNS (or RONS) produced by the plasma are related to the chemical, photodynamic, and radiation effects. Many RONS generated in plasma are also active components in cell biology.⁸ It has been argued that analogy between cellular and plasma-generated RONS represents the major logic of plasma application in medicine including cancer therapy and that many species produced directly or indirectly by plasma will function in cell biology as endogenous species.⁸ RONS play a central role in “redox” or oxidation–reduction biology.⁹ 

The overall schematics of plasma interaction with cells are shown in Fig. 1. This is essentially a multi-scale process spanning from the initial burst at the timescale of nano and micro seconds (depending on specific plasma device) to seconds, followed by the time scale of minutes which is related to RONS formation and transport across the cellular membrane and finally triggering various cellular pathways at the time scale of hours and days.

Importance of RONS warrants some remarks regarding their function. Watson¹⁰ suggested that ROS are “a positive force for life” due to their role in apoptosis—an internal program leading to cell death. At the same time, ROS are also well recognized “for their ability to irreversibly damage key proteins and nucleic acid molecules (e.g., DNA and RNA).” Typically, the normal level of ROS is maintained by the anti-oxidant system. It was noted that “The vast majority of all agents used to directly kill cancer cells (ionizing radiation, most chemotherapeutic agents, and some targeted therapies) work through either directly or indirectly generating reactive oxygen species that block key steps in the cell cycle”. The hypothesis is that the effect of ROS on cell development depends on the level of ROS.¹¹ The low level of ROS supports cell proliferation and helps to maintain cell functionality, while the high level of ROS causes oxidative stress leading to cell death. Healthy cell function is preserved by antioxidant system that maintains ROS level at the tolerable level.

As an example, let us consider a cancer cell whose abnormal metabolism causes aberrant high level of ROS.¹¹ To survive, a cancer cell mutates to regulate ROS. However, raise of the intracellular ROS level might cause irreparable DNA damage.^12,13 Recall that the level of ROS in cancer cells is near the limit at which cell death occurs. At the same time, the ROS level in the corresponding normal cells is generally lower.¹³ Thus, selectivity toward tumor cells is achieved when the anti-cancer therapy produces ROS near the “threshold” between levels of ROS in normal cells and cancer cells. Such an approach seems to be relevant to the CAP based anti-cancer therapy. It was argued that ROS produced by CAP might lead to cancer cells death by damaging the function of intracellular regulatory factors.^14,15 To that end, multiple studies demonstrated that biologically active neutral short- and long-living ROS molecules are produced by CAP including OH, O, O (1D), O₂ (1Δg), O₃, HO₂, and H₂O_2.¹⁶ RNS species such as NO₂⁻, NO⁻, and NO⁺ are generated directly during the discharge in a gas phase and in the plasma-activated media.^7,17

Among other RNS, nitrite oxide (NO) plays a very significant role in cellular processes. In particular, NO is an important factor in the electron transport chain. It is also known that NO might affect the electron transport system by attacking cytochrome oxidase.¹⁸ In turn, breaking of the electron transport will increase the generation of superoxide reacting with NO to form peroxynitrite.¹⁹ These processes suggest that the possible role of plasma technology in biomedical application could be, i.e., plasma potentially can affect the cellular chemical balance by generating RONS. In that sense, plasma might be creating a reservoir for some critical species such as NO. In addition, it should be pointed out that CAP is very different from chemotherapy and radiotherapy, though many chemotherapy and radiotherapy also increase intracellular RONS stress and further kill cancer cells. The primary difference is that CAP itself is the source of RONS including hydrogen peroxide, nitrite, nitrate, and nitric oxide.

It has been established that one of the important species formed in course of plasma interaction with cell culture media is hydrogen peroxide (H₂O₂).^20,21 To that end, a model based on aquaporin (AQP) which is well known H₂O₂ channels on the cell membrane has been proposed.^22,23 Multiple studies inform that in general cancer tissues express more AQP channels than corresponding normal tissues. According to the proposed hypothesis,^22,23 after plasma treatment, H₂O₂ transport across the cellular membrane of cancer cell is higher than that of normal cell. Such a differential H₂O₂ consumption rate might be the possible mechanism of the selective anti-cancer action of CAP. As such, the rise of intracellular ROS as a result of ROS diffusion across cell membrane correlates with the rise of extracellularly plasma-originated ROS.

The high-level model of plasma interaction with cells and selective action on the cancer cells is shown schematically in Fig. 2. Current understanding of CAP action could be attributed to several cellular factors i.e., higher expression of AQPs by cancer cell membrane and lower expression of anti-oxidant enzyme (catalase) in cancer cells.^22,23

Reaction of cells to ROS excess is well documented. It is established that catalase is the major enzyme that controls concentrations of H₂O₂ in both cancer and normal cells.²⁴ For instance, a recent paper correlated the rate of H₂O₂ removal and activity of catalase for 15 cancer cell lines and 10 normal cell lines. This study showed that H₂O₂ produced from the oxidation of P-AscH- is a principle mediating factor in selective targeting of the cancer cells. More importantly, it was demonstrated that normal cells have a higher constant of H₂O₂ removal than that of cancer cells. This trend is illustrated schematically in Fig. 3. Granted, H₂O₂ is known to be very strong oxidant, but it has a slow reaction rate with the majority of biomolecules. As such, hydrogen peroxide accumulates in cells and the aforementioned process of H₂O₂ removal by catalase becomes critical process for cell survival.²⁴ 

Overall, one can summarize the above discussion by quoting Halliwell who suggested that “we understand more now about the importance of RONS than we did 10 years ago, but not exactly how they are acting.”²⁵ He concluded that RONS can be good and bad and thus it is very difficult to modulate RONS by employing some targeted “antioxidant” dosage.

One more aspect of CAP treatment should be pointed out. Recently, it was discovered that cancer cells instantly generate high (μM) levels of H₂O₂ during the direct CAP treatment of cancer cells.²⁶ These results are shown in Fig. 4. One can see that plasma treatment can modulate the generation of H₂O₂ by breast cancer cell lines (MDA-MB-231), a pancreatic cancer cell line (PA-TU-8988T), and a brain cancer cell line (U87MG). On the other hand, the normal fibroblast cell (WTDF) displays very weak H₂O₂ generation (not shown). In addition, the direct interaction between cells and the plasma jet is a prerequisite for the cell-based H₂O₂ generation. If the plasma jet did not directly touch the cells, the cell-based H₂O₂ generation will not occur.²⁶ These unique features provide a strategy to discriminate specific tumorous cells or tissues from normal cells or tissues.

In this section, we briefly describe the utility of CAP application in cancer by using a brain tumor therapy as an example. To study the effect of CAP on brain tumor, a micro-CAP device has been developed. This device was directly applied to the glioblastoma tumor by utilizing an implanted endoscopic delivery system as shown in Fig. 5.²⁷ Tumor volume was assessed by the bioluminescence imaging in real time [as shown in Fig. 5(b)]. These experiments suggested that the tumor volume in the case of a control (i.e., helium only treatment) increased by about 600% after two days. At the same time, CAP treatment led to decrease in the tumor volume by about 50% as shown in Fig. 5(c).

It was mentioned in the Introduction that the plasma adaptation presents a unique platform for optimizing the plasma therapeutic potential. To this end, the formation of coherent plasma structures in plasmas could allow adaptation to externally imposed conditions, such as different types of cells and tissues. Plasma adaptivity to specific cells through tailoring its composition in situ might be possible due to plasma self-organization. In this section, we will discuss briefly the self-organization of plasmas.

There are a large variety of patterns and coherent structures in Nature. They range from microscopic organisms to galaxies. The overarching principle driving various pattern formations can be traced to the so-called “principle of self-organization.”²⁸ According to this principle, any dynamic system evolves towards a state of equilibrium.²⁹ It was pointed out that systems in the non-equilibrium state having exchange of energy and matter could display spontaneous self-organization.³⁰ I. Prigogine in the Nobel Prize Lecture suggested that self-organization of chemical reaction can occur so that the distribution of reactive particles near instabilities is not random.³¹ In other words, it was suggested that “chaos gives rise to order!.”³¹ 

It is known that plasmas can be also subjected to self-organization, i.e., spontaneous transition from a state with homogeneous distribution to a pattern.³² In fact, multiple examples of discharge self-organization have been documented including dielectric barrier discharge,³³ high frequency discharge,³⁴ gas flow stabilized discharges,³⁵ resistively stabilized discharge,³⁶ and various discharges having liquid electrodes.³⁷ In a recent review, pattern formation in discharges and their relation to near electrode phenomena were summarized.³⁸ 

One of the best-known and widely used theoretical models describing self-organization is the reaction-diffusion model. In fact, this model has been used to describe various patterns in electric discharges (see Ref. 38 and references therein). More advanced models are based on the drift-diffusion approach. Such approaches allow physics-based description of discharge self-organization. For instance, computationally pattern formation in glow and arc discharges using the 2D drift-diffusion approximation was studied.^39,40 To that end, an average current density as a control parameter in the case of glow discharge was used.

Self-organization at various discharge modes has been a focus of recent study.⁴¹ It was demonstrated that complex structures can be formed at the plasma-liquid medium interface dependent on discharge current as shown in Fig. 6 (Ref. 41). One can notice that in this particular setup, the discharge current changes drive the pattern formation above the dioinized water (DI) as shown in Fig. 6(d). Recall that four specific types of discharge modes can be noticed. In particular, stage I is the low current discharge (glow discharge) having a single filament. Discharge current increase (stage II) leads to temperature raise of the tungsten cathode and high heat radiation [see Fig. 6(c)]. Stage III is an unstable intermediate state in which discharge oscillates between two (II and IV) stages displaying either high heat radiation (as typical in the stage II) or the multi-filament pattern (as typical in the stage IV). It was observed that the multi-filament discharge (stage IV) is generally stable.

It should be pointed out that the effect of different discharge modes on the production of RONS and resulting treatment of cancer cells was analysed (Refs. 36 and 42). It was observed that the selective cancer treatment is possible at some discharge conditions (Ref. 41). Utilizing the transition between discharge modes, one can invision a possibility to control treatment using plasma and adaptation of plasma treatment to a particular area or specific cell type.

Here we outline new possible ideas related to the proposed adaptive CAP (ACAP) as shown schematically in Fig. 7. ACAP device relays on real time monitoring plasma-living tissue interaction and adaptation of plasma parameters based on a feedback mechanism.^43,44 Plasma device feedback control and managed delivery of plasma dose were described in a recent paper.⁴⁵ As a result of differential action of ACAP on cancer and normal cells, it might be possible to manage selective plasma treatment of cancer cells. Schematic concept of ACAP is shown in Fig. 7 in which both direct and feedback-based adaptations are illustrated. In the case of a direct feedback or self-adaptation, plasma interaction with cells could lead to self-organization via transition between various discharge modes.⁴¹ As an example of feedback-based ACAP, recent work⁴⁶ utilized a RealTime^TM assay and provided evidence that the cellular response to plasma treatment can be monitored in real time that is a pre-cursor for adaptation.

It should be pointed out that plasmas with fixed and well characterized properties are suited well for adaptive approach via the feedback mechanism outlined above. On the other hand, the self-adaptive approach relays solely on plasma discharge self-organization as a result of interaction with various cells and tissues. It is very important to note that non-equilibrium systems are strongly non-linear. As such, one expects that significant changes of plasma parameters will occur in response to a small variation of external parameters. Indeed, this can be viewed as an important asset and plasma-self-organization opens an opportunity for application in medicine. In that respect, significant advance in our understanding of the plasma self-organization and related plasma chemistry is necessary and research in this directed is warranted.

Plasma uniqueness can be utilized in the adaptive therapeutic platform. Several aspects of such a system including plasma chemistry modulation, discharge mode control, and efficient feedback system warrant detailed analysis. To this end, in situ recording of the plasma chemical composition using non-invasive techniques is of paramount importance. Such approaches might be based on sufficiently sensitive laser based diagnostics such as laser induced florescence.

Biomedical approach based on the adaptive cold atmospheric plasma could potentially revolutionize therapy by introducing personalized treatment. The idea behind this is that medical treatment could be tailored to meet peculiarities associated with patient's genetic makeup. By applying any drug including plasma, one can expect a unique response dependent on the person's genotype. As such, in the case of plasma treatment, the same plasma composition might lead to a different response. This response can be tailored and targeted by utilizing the adaptive plasma approach outlined above.

Granted that further progress in plasma application in medicine requires a significant effort in understanding the plasma induced microbiology, chemistry, and engineering. However, it is clear that there is an important role for plasma physics to support such progress. In particular, understanding plasma discharge self-organization, the mechanisms driving transition between different discharge patterns and the development of the plasma devices having multiple discharge modes need to be considered.

The authors acknowledge Barry Trink, Jonathan Sherman, Dayun Yan, Xiaoqian Cheng, Zhitong Chen, Eda Gjika, and Li Lin for valuable contributions to the concept of adaptive plasma. This work was supported by the National Science Foundation, Grant Nos. 1465061 and 1747760.