Cold plasma and inhibition of STAT3 selectively target tumorigenicity in osteosarcoma

Abstract

Osteosarcoma (OS) is a malignant type of bone cancer that arises in periods of increased bone formation. Curative strategies for these types of tumors have remained essentially unchanged for decades and the overall survival for most advanced cases is still dismally low. This is in part due to the existence of drug resistant Cancer Stem Cells (CSC) with progenitor properties that are responsible for tumor relapse and metastasis. In the quest for therapeutic alternatives for OS, Cold Atmospheric Plasmas and Plasma-Treated Liquids (PTL) have come to the limelight as a source of Reactive Oxygen and Nitrogen Species displaying selectivity towards a variety of cancer cell lines. However, their effects on CSC subpopulations and in vivo tumor growth have been barely studied to date. By employing bioengineered 3D tumor models and in vivo assays, here we show that low doses of PTL increase the levels of pro-stemness factors and the self-renewal ability of OS cells, coupled to an enhanced in vivo tumor growth potential. This could have critical implications to the field. By proposing a combined treatment, our results demonstrate that the deleterious pro-stemness signals mediated by PTL can be abrogated when this is combined with the STAT3 inhibitor S3I-201, resulting in a strong suppression of in vivo tumor growth. Overall, our study unveils an undesirable stem cell-promoting function of PTL in cancer and supports the use of combinatorial strategies with STAT3 inhibitors as an efficient treatment for OS avoiding critical side effects. We anticipate our work to be a starting point for wider studies using relevant 3D tumor models to evaluate the effects of plasma-based therapies on tumor subpopulations of different cancer types. Furthermore, combination with STAT3 inhibition or other suitable cancer type-specific targets can be relevant to consolidate the development of the field.

Graphical abstract

Highlights

• •Cold Atmospheric Plasma-Treated Liquids at anti-cancer selective doses promote Cancer Stem Cell properties & tumor growth in Osteosarcoma.
• •Expression of antioxidant proteins & STAT3 is cell line dependent & may predic OS cell response to plasma conditioned liquids.
• •3D engineered models predicts cold plasma response in mouse subcutaneous and orthotopic models.
• •Treatment combining STAT3 inhibition & Plasma Conditioned Liquids strongly suppresses OS tumor growth in vivo.

Abbreviations

APPJ

Atmospheric Pressure Plasma Jet

CAP

Cold Atmospheric Plasma

CAT1

Catalase 1

Col1

Collagen Type I

CFU

Colony Formation Units

CT

Computed Tomography

CSC

Cancer Stem Cells

GPX1

Glutathion Peroxidase 1

hBM-MSCs

Human Bone Marrow Mesenchymal Stem Cell

i.b

Intra bone tumor

LUC

Luciferase

NANOG

Nanog Homeobox

nHA

Hydroxyapatite Nanoparticles

OCT4

Octamer-binding transcription factor 4

OS

Osteosarcoma

PTL

Plasma Treated Liquids

RONS

Reactive Oxygen and Nitrogen Species

s.c

Subcutaneous tumor

SOD2

Superoxide Dismutase 2

SOX2

SRY-Box Transcription Factor 2

STAT3

Signal Transducer and Activator of Transcription 3

TFE

Tumorsphere Forming Efficency

1. Introduction

Osteosarcoma (OS) is the most common type of primary bone tumor affecting children and young adults and develops mainly in areas of active growth of long bones [1]. The cornerstone of bone sarcoma management is the wide margin surgical resection of the primary tumor, which is typically accompanied by neo-adjuvant and/or adjuvant chemotherapy (doxorubicin, ifosfamide, methotrexate, cisplatin, etc.) and/or irradiation [2]. Overall, despite the implementation and continuous optimization of multimodal therapies, survival data of OS patients have remained unchanged for the last 40 years. Nowadays, a relevant portion of OS patients do not benefit from current treatments and patients with metastatic disease present a 5 years post-treatment survival of only 20% [[2], [3], [4]]. Thus, OS tumors are still a therapeutic challenge and there is an urgent need to develop effective therapies for these patients.

Cancer Stem Cells (CSC) are a subpopulation of the tumor bulk with an indefinite capacity for division and self-renewal. In addition, these subsets of tumor cells also retain a certain potential to differentiate and originate different tumor subpopulations [5]. Several studies demonstrate that CSC play a pivotal role in OS relapse since they are resistant to cytotoxic drugs through a number of mechanisms such as the adoption of a quiescent state, the expression of drug efflux transporters of the ABC family, the over-expression of anti-apoptotic proteins or the over-activation of DNA damage repair mechanisms [6,7].

In OS, the bone environment seems to be involved in the regulation of self-renewal, differentiation, growth, drug resistance, and/or metastatic potential of CSC subpopulations [1,6,8,9]. Unfortunately, most cancer research has been done in 2D cultures, which lack the 3D tissue microenvironment; as a result, less than 5% of the effective drugs tested in 2D models and preclinical mouse models succeed in clinical trials [[10], [11], [12], [13]]. As an alternative to 2D models, bioengineered tumor models are being developed to more closely mimic the human 3D tumor microenvironment and today are considered as invaluable tools to test new therapies against OS [13]. With regard to 3D bioengineered OS models, the sponge-like scaffolds fabricated using natural materials are especially striking. These bone-like scaffolds lead the growth of OS cells in the surrounding extracellular matrix to bring cell adhesion, proliferation, migration, cell-to-cell interaction and overall activate osteomimicry in vitro [14]. Accordingly, OS models are able to mimic features relevant to the clinics, like the acquisition of drug resistant phenotype to current drugs, or to promote niche-dependent phenotypes associated with poor prognosis, including CSC propagation or epithelial-to-mesenchymal transition [13,14].

In the last decade, Cold Atmospheric Plasmas (CAP) have been approved for medical use both in the USA and in Europe as novel anti-cancer therapy for different indications [15,16]. CAP is an ionized gas composed of a wide variety of Reactive Oxygen and Nitrogen Species (RONS), ions, electrons, metastable particles, electromagnetic fields and weak UV and VIS radiation and at near ambient temperature [17]. CAP can be applied directly to cell cultures [18,19] and living tissues to produce the anti-proliferative effects [15,20]. As these effects depend largely on RONS such as H₂O₂, NO₂⁻, ONOO⁻, these RONS can be transferred to Plasma Treated Liquids (PTL) [[21], [22], [23], [24]], to obtain similar effects to those observed in CAP-treated cell cultures [25], or to allow local injection in tumors located in internal organs like OS [26]. It is also known that treatment with PTL induces cell death mediated by oxidative stress [27,28]. The concentration and type of RONS generated are highly dependent on the treatment conditions (gas flow, treatment time, etc.) and the chemical/biochemical composition of the treated liquid [29]. Therefore, the final concentration of RONS can be modulated to increase the therapeutic window in CAP and PTL treatments [25]. Interestingly, both CAP or PTL have been shown to preferentially eliminate cancer cells rather than healthy cells, opening the door for a selective, side-effect-free cancer therapy [24,[30], [31], [32]]. Despite these promising results, we recently observed that the stemness phenotype was promoted in one type of OS cell line, which could threaten all the advances made in the field [33]. Up to now, the practical totality of research related to anticancer effects of CAP and PTL has been done in 2D cultures and in a few in vivo models, so overall, the impact of these prospective therapies against CSC remains unexplored.

Regarding the promotion of stemness in sarcomas, several works have identified Signal Transducer Activator of Transcription 3 (STAT3) as the main pro-stemness and pro-tumorigenic factor induced by micro-environmental cell types in OS cells [[34], [35], [36]]. The over-expression and aberrant activation of STAT3 is considered crucial in development and progression of OS [37]. STAT3 is one of the transcription factors more frequently activated in OS cell lines [38,39], sarcoma initiating cells [40], as well as in OS patient samples [41]. Likewise, STAT3 inhibitors are suggested to reverse chemo-resistance to fully eliminate drug-resistant subpopulations [[42], [43], [44]]. In this sense, the reduction of STAT3 activity in OS cell lines and in vivo models resulted in decreased metastases and enhancement of the effectiveness of chemotherapeutic drugs [37].

Considering all of the above, our hypothesis is that while PTL can foster tumor resistance through enhancement of CSC properties, the combined targeting of OS by PTL and inhibition of STAT3 could result in a selective and efficient approach for this malignancy. Our strategy is to employ bioengineered 3D OS models generated from a variety of OS cell lines as well as in vivo xenografts as relevant preclinical models of OS to investigate the effects of PTL on CSC subpopulations in OS, and to evaluate whether the undesired pro-stemness signals possibly mediated by PTL can be prevented by the inhibition of STAT3 signaling.

2. Results

2.1. OS cells display different levels of sensitivity to PTL

Two different CAP jets - a home-made Atmospheric Pressure Plasma Jet (APPJ) and kINPen - were employed to treat DMEM for different periods of time and obtain PTL with increasing levels of RONS (Fig. 1A and B). Among the variety of RONS formed by CAP in the cell culture medium, here we measured two relevant long-lived ones: H₂O₂ and NO₂⁻. The concentration of both NO₂⁻ (Fig. 1A) and H₂O₂ (Fig. 1B) increased in a treatment time dependent manner. kINPen produced significantly more NO₂⁻ than APPJ especially at long treatment times (Fig. 1A), while both plasma devices produced similar levels of H₂O₂ (Fig. 1B).

Fig. 1.

Anti-proliferative effects of PTL in monolayer cultures. (A-B) kINPen and APPJ were used to treat cell culture medium for 5–240 s to obtain PTL. The concentration of nitrites (A) and hydrogen peroxide (B) in these PTL was measured immediately after CAP treatment using untreated DMEM as blank. Data are presented as mean and error bars represent the SD (n = 4). Asterisks indicate statistically significant differences between both types of RONS (***p < 0.0005; **p < 0.005; *p < 0.05; two-way ANOVA). C) Cell viability (WST-1 assay) of the indicated OS cell lines measured 72 h after exposure to PTL obtained using kINPen (left) or APPJ (right). IC50 values for each cell type are shown. Error bars represent the standard deviation of four independent experiments. D) Cell viability (WST-1 assay) of healthy hBM-MSCs (bars) measured 72 h after exposure to PTL produced with APPJ. Cell viability is expressed relative to the corresponding untreated control and is the mean and SD of four independent experiments (*p < 0.01; ****p < 0.001; one-way ANOVA). E) Illustration of the main intracellular antioxidant enzymes created using BioRender.com (F) Western Blotting showing the levels of the indicated proteins in five OS cell lines and healthy hBM-MSCs. (G-J) Correlation between the IC₅₀ values of OS cells treated by PTL using kINPen for 72 h and their levels of STAT3 (G), CAT1 (H), GPX1 (I) and SOD2 (J). Signal intensity ratios of each protein relative to β-actin are shown in Y axis. The r values indicate Pearson's correlation coefficient with corresponding p-value (P) and R square (R²) for correlation.

To investigate the cytotoxic potential of PTL, we used it to treat monolayer cultures of five human OS cell lines (143.B, SaOS-2, U-2 OS, MG-63 and G-292) (Fig. 1C). Dose–response cell viability experiments measured after 72 h of incubation showed that all OS cells were more sensitive to the PTL generated by kINPen (IC₅₀ values between 15.33 and 40.59 s) than that produced by APPJ (IC₅₀ values between 24.94 and 73.61 s). In any case, the relative sensitivity of the different OS cell lines to PTL is similar in both cases, being 143.B and SaOS-2 cells the most sensitive ones (Fig. 1C). In parallel, non-malignant human Bone Marrow Mesenchymal Stem Cells (hBM-MSCs) showed greater resistance to PTL (Fig. 1D) than OS cells (Fig. 1C) under the same culture conditions. Indeed, hBM-MSCs exposed to PTL-30 s displayed stimulated cell viability, whereas their cell viability decreased only after treatments with PTL obtained after CAP exposures higher than 120 s (and thus, with high concentrations of RONS) (Fig. 1D).

The different sensitivity of cancer cells to PTL could be related to differential expression of factors with antioxidant activity (Fig. 1E). We therefore analyzed the levels of Glutathione Peroxidase-1 (GPX1), which scavenges Reactive Oxygen Species (ROS) like OHOO⁻ and H₂O₂ to form H₂O; mitochondrial Superoxide dismutase (SOD2), an enzyme that eliminates O₂⁻; and Catalase 1 (CAT1), which acts as H₂O₂ scavenger. Western blotting analysis of these enzymes confirmed that OS cell lines display different profiles of antioxidant defenses and STAT3 accumulation (Fig. 1F). Compared to healthy hBM-MSCs, OS cell lines showed higher levels of GPX1 and CAT-1, while SOD2 was expressed at comparable levels in hBM-MSCs and OS cells. Regarding STAT3 there is a highly variable pattern of expression between all types of cells, with 143.B expressing lower levels than hBM-MSCs and MG-63 cells showing the highest levels of this transcription factor (Fig. 1E and Fig. S1A). Interestingly, a lower expression of STAT3 (Fig. 1G) or CAT1 (Fig. 1H) significantly correlated with a higher sensitivity (lower IC50) of OS cell lines to PTL. Besides, the opposite trend was observed for GPX1 (Fig. 1I) and no correlation was observed between SOD2 expression and the cytotoxic effect of PTL (Fig. 1J). These findings suggest that OS cells differ significantly in their profile of expression of STAT3 and antioxidant factors, which is likely to impact on the cytotoxic potential of PTL for each cell type.

2.2. PTL promote CSC properties in OS

Unlike more differentiated cancer cell subpopulations, CSC have been reported to display a similar redox regulation to that of non-malignant stem cells [45]. This prompted us to investigate whether PTL treatments may affect CSC viability. First, we used APPJ to generate PTL containing increasing concentrations of RONS and used it to treat monolayer cultures of 143.B, MG-63 and G-292 cells for 72 h. Immediately after this treatment, cells were grown under tumorsphere culture conditions to assess the self-renewal ability of the surviving cells (Fig. 2A). Importantly, we found that all OS cells pre-treated with PTL displayed an enhanced ability to self-renew and form tumorspheres (Fig. 2B and C). This phenomenon was especially remarkable for G-292 and MG-63 which displayed increased Tumorsphere Forming Efficiency (TFE) up to PTL-60 s treatment. Only treatments with the higher concentrations assayed (PTL-120 s for 143.B and MG-63 and PTL-240 s for G-292) were able to significantly decrease the TFE of OS cells. According to the increase observed in tumorsphere formation, we found that PTL doses below PTL-60 s also enhanced the expression of well-known stemness factors, such as SOX2, NANOG and OCT4 in a dose response manner in OS cells (Fig. 2D).

Fig. 2.

Effects of PTL on the tumorsphere-forming potential of OS cells. A) Monolayer cultures of 143.B, G-292 and MG-63 cell lines were treated with PTL generated with APPJ for different treatment times (from 15 to 120 s) for 72 h in monolayer. Following this treatment surviving cells were recovered and let to grow under tumorsphere-forming conditions for 10 days. Illustration created using BioRender.com. B) Representative optical microscope images of the tumorspheres formed after this period. C) TFE represented as number of tumorspheres formed (size ≥ 70 μM) in each condition relative to the respective untreated controls. Data are shown as the mean and SD of 6 replicates. Asterisks indicate statistically significant differences between the indicated series and the untreated controls (*p < 0.01; **p < 0.001; ****p < 0.0001; two-way ANOVA). D) RT-qPCR analyzing the expression of the stemness genes SOX2, OCT4 and NANOG in MG-63 cells 72 h post-treatment. GAPDH levels were used as a housekeeping gene. mRNA fold change is expressed relative to the corresponding untreated control and is the mean and SD of three independent experiments (**p < 0.01; ***p < 0.001; ****p < 0.0001; two-way ANOVA). E) Western Blotting showing the levels of the indicated proteins after being treated for 72 h with PTL generated by exposure to increasing APPJ treatment times in 143.B and MG-63 cells. β-Actin levels were used as loading control in Western blotting experiments. F) Western Blotting showing the levels of GPX1, STAT3 and phospho-STAT3 at basal levels in shRNA expressing cells. β-Actin were used as loading control (G-H) Analysis of CSC-related properties in 143.B and MG-63 cells transduced with lentiviral vectors expressing two specific shRNAs to knockdown GPX1 or a control shRNA G) Representative images of the tumorspheres formed by shRNA expressing cells H) Number of tumorspheres formed in each condition. Data are shown as the mean and SD of 3 replicates. Asterisks indicate statistically significant differences between the indicated series and the untreated controls (*p < 0.01; two-sided t-student).

To further confirm the pro-stemness potential of plasma, we used kINPen to treat MG-63 cells both with PTL or an equivalent direct CAP flow under the same operation parameters (Fig. S2A). We found that for treatment times between 15 s and 60 s, both treatments significantly increased the TFE of OS cells in all assayed conditions (Fig. S2B–C). Notably, this was also coupled to a strong induction of the expression of SOX2, NANOG and OCT4 in most of the assayed conditions (Fig. S2D). These data indicate that only high doses of PTL generated after plasma treatments above 120 s were able to inhibit the pro-stemness features. These long CAP treatment times producing PTL with high concentrations of RONS are non-selective and in addition to being cytotoxic to tumor cells, they may also affect non-malignant hBM-MSCs (Fig. 1D).

To gain insight about the molecular mechanisms involved in the cytotoxic effect of PTL, we analyzed the impact of a 72 h treatment on the expression of antioxidant factors and STAT3 activation in the highly sensitive 143.B cell line and in the more resistant MG-63 cells (Fig. 2E & Fig. S1B). We found that 143.B and MG-63 cells showed a biphasic modulation of the expression and phosphorylation of STAT3 upon the treatment with increasing doses of PTL. In 143.B cells, STAT3 was moderately up-regulated by the lower dose of PTL-15 s and underwent a dose-dependent down-regulation when treated with higher doses. In the case of MG-63, STAT3 up-regulation was much more intense and dose-dependent until PTL-30 s, followed by a sharp down-regulation at higher doses (Fig. 2E & Fig. S1B). Importantly, the dynamics of STAT3 expression and activation induced by PTL correlated with the stemness potential of OS cells. Thus, the doses of PTL that show a more intense tumorsphere formation were those that produce a more intense up-regulation/activation of STAT3 both in 143.B (PTL-15 s) and MG-63 (PTL-30 s) (Fig. 2C and E). Regarding the antioxidant factors, we also observed a biphasic modulation of GPX1 although the peaks of maximum up-regulation occur at higher doses than those observed for STAT3 (30 s in 143.B cells and 60 s in MG-63 cells). On the other hand, PTL treatments did not induce a clear modulation profile of SOD2 and CAT1 levels in OS cells (Fig. 2E & Fig. S1B).

In a recent study employing a prostate 3D engineered model, PTL was found to induce a decrease of GPX1 expression [46]. Similarly, here we found that PTL decreased the levels of GPX1 in a dose response manner 24 h post-treatment in OS cells (Fig. S1D). Thus, we aimed to study whether the role of antioxidant factors, such as GPX1, is of interest in mediating CSC properties in OS. The analysis of shRNA-mediated GPX1-depleted models of 143.B and MG-63 cells (Fig. 2F and Fig. S1C) reflected the complexity of the role that the antioxidant factors play in the modulation of stemness. On the one hand, depletion of GPX1 in MG-63 cells induced a significant increase in the tumorsphere-forming activity (Fig. 2G and H) associated to an increase of the phospho-STAT3 levels (Fig. 2F & Fig. S1C). On the other hand, not significant modulation of either tumorsphere-forming potential nor phospho-STAT3 levels were observed in GPX1-depleted 143.B cells (Fig. 2F–H). Altogether, these results suggest that, although the modulation of antioxidant factors may play a role in certain cases, the activation of STAT3 seems to be a more pivotal regulator of the pro-stemness effect induced by low-dose PTL treatments.

2.3. PTL may promote OS tumor growth in vivo

To examine the effect of PTL in vivo, we selected Ringer's Saline as liquid to generate PTL due to its compatibility with future clinical uses. Previous works reported that CAP treated Ringer's Saline displayed efficacy towards OS organotypical ex vivo cultures [46] and that Ringer's Lactate is able to efficiently deliver RONS from CAP into solid tumor xenografts [47,48]. We employed APPJ treatment times of 300 and 900 s to obtain PTL (PTL-300 s and PTL-900 s) containing suitable levels of RONS for the treatment of tumor xenografts. These PTL were injected intratumor to treat mice bearing subcutaneous (s.c.) or intra-bone (i.b.) 143.B + luciferase (LUC) xenografts (Fig. 3A). Compared to the control group, a daily treatment of the s.c. model with PTL-300 s did not have any impact on tumor growth (Fig. 3B). However, the treatment with PTL-900 s produced a significant increase of tumor growth both in the s.c. (Fig. 3B) and i.b. (Fig. 3C) models. Consistent with the tumor growth curves, µ-Computed Tomography (μCT) scanning revealed greater areas of bone destruction in i.b. xenografts treated PTL-900 s compared to control series (Fig. 3D). The results of tumor growth were also confirmed by the analysis of the luciferase activity of tumor cells during the treatment of both, the s.c. (Fig. 3E and G) and the i.b. (Fig. 3F and H) xenografts. In any case, we did not observe a significant loss of weight or other adverse effects in the animals during the treatments (Fig. 3I and J).

Fig. 3.

In vivo effect of PTL. A) Scheme of the production of PTL and the treatment of s.c. (^A) and orthotopic i.b. (^B) OS xenografts. APPJ was used to treat Ringer's Solution for 300s or 900s to obtain PTL. 143.B + LUC xenografts were allowed to grow for 16 days before being randomly assigned (n = 6 per group for the s.c. model and n = 3 for the i.b. model) to receive daily intra-tumor treatments with either, vehicle (Ringer's Saline as control), PTL-300 s or PTL-900 s for 18 days. Graphics created using BioRender.com. B-C) Curves representing the relative mean tumor volume of s.c. (B) and i.b. (C) xenografts during the treatments. (D) μCT scanning of tumor tibiae. 3D reconstruction of a representative healthy tibia (left), an untreated xenograft (control; middle) and a PTL-900s -treated xenograft (right). (E-F) In vivo bioluminescence of s.c. (D) and i.b. (E) tumors generated by luciferase-expressing 143.B + LUC cells in a 3 mice-cohort of the indicated series at day 14 after the beginning of the indicated treatments. (g-H) Individual radiance values represented as luciferase activity relative to day 0 values for each mouse. (I-J) Mean body weight of mice during the treatment of s.c (I) and i.b. (J) xenografts. Data are shown as the mean and error of the mean. Asterisks indicate statistically significant differences between the indicated series and the untreated controls (*p < 0.05; **p < 0.005; ***p < 0.0005; ****p < 0.00005; two-sided t-test).

2.4. The inhibition of STAT3 signaling fosters the cytotoxic potential of PTL and arrests its pro-stemness effects

Given the prominent role that the activation of STAT3 may play in the pro-stemness effect of PTL, we explored whether the inhibition of this factor may reverse the pro-tumorigenic effects induced by low/medium doses of PTL in OS cells. To achieve the inhibition of this factor we used S3I-201, a small molecule capable of specifically blocking STAT3-mediated transcriptional activity by inhibiting its dimerization. We found that S3I-201 induced a partial reduction of total STAT3 and phospho-STAT3 (S727) levels both following 0.5 and 6 h of treatment. This level of inhibition was further enhanced by a combined treatment of PTL and S3I-201, which produced an almost complete inhibition of STAT3 phosphorylation after 6 h of treatment (Fig. 4A-B). In addition, the levels of the antioxidant factors CAT1 and GPX1 were also modulated by the STAT-3 inhibitor and by the combined treatment, with a consistent down-regulation of both factors after 6 h of combined treatment with PTL and S3I-201 (Fig. 4B). The whole Western Blotting quantifications can be found in Fig. S3. In agreement with the inhibition of antioxidant factors, the levels of intracellular ROS detected in these OS lines after the combined treatment were significantly higher than those detected following PTL or S3I-201 treatments separately (Fig. 4C).

Fig. 4.

S3I-201 fosters the cytotoxic potential of PTL and arrests its pro-stemness effects. (A) Western blotting showing protein levels of the indicated proteins in 143.B (left) and MG-63 (right) treated or not with 100 μ S3I-201, PTL-300 s (DMEM w/o pyr) or the combination of both for 0.5 or 6 h. (B) Protein quantification as shown as the ratio of each protein intensity/β-Actin for each sample (143.B left and MG-63 right) relative to control 6 h post-treatment. β-Actin levels were used as loading control. (C) Intracellular ROS levels in 143.B and MG-63 cells treated or not with 100 μM S3I-201, PTL-30 s or the combination of both for 2 h. Values were normalized to the respective controls. D) Real-time proliferation (normalize cell index) measured in 143.B (left) and MG-63 (right) cells treated or not with 100 μ S3I-201, PTL-5 s or the combination of both. Black arrows indicate the beginning of the treatments (t = 24 h). Cell index was normalized to the initial time (t = 0h). Data are presented as the mean and the SD of two biological replicates. Readings were done in duplicates using xCELLigence. E) Scheme of the protocol employed to analyze the effects of PTL (generated from CSC medium) and/or S3I-201 on tumorsphere growth. Treatments were initiated after 3 days of culture in CSC medium and the formation of tumorspheres and the level of cell death was measured 7 days later. Created using BioRender.com. F) Representative images of tumorsphere cultures and Live/Dead assays (live and dead cells display green and red fluorescence respectively). G). Quantification of the number of tumorspheres formed by the indicated cell lines treated with or not with 100 μM S3I-201, PTL-300 s or the combination of both. Unless otherwise specified, data are presented as the mean of n = 6 biological replicates. Error bars represent the SD, and asterisks indicate statistically significant differences between the controls and the indicated series (*p < 0.01; **p < 0.001; ****p < 0.0001, one-way ANOVA). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Next, we analyzed the effect of this combination in the proliferation and cell survival of OS cells. Real-time cell proliferation assays showed that the treatment with 100 μM S3I-201 was able to slow down the proliferation rate of 143.B cells while it did not affect the growth of MG-63 cells. Relevantly, when combined with a low dose of APPJ-generated PTL-5 s that did not modify the growth rate of OS cells, S3I-201 induced a strong inhibition of cell proliferation in both cell lines (Fig. 4D). Similar effects were observed in four OS cell lines treated with S3I-201 in combination with low dose PTL (PTL-15 s) generated with the kINPen plasma jet (Fig. S4A). Of note, the combination of PTL and S3I-201 did not have any impact on the cell viability of healthy hBM-MSCs (Fig. S4B).

To study whether STAT3 inhibition may abrogate the pro-stemness effects induced by PTL, we evaluated the ability of S3I-201 alone or in combination with PTL to prevent the growth of pre-formed OS tumorspheres (Fig. 4E). To better mimic CSC-growing conditions, we used full CSC medium to obtain PTL. Since the CSC medium contains pyruvate which is a strong ROS scavenger [25], the amount of RONS generated by the CAP treatment in this medium is sensibly lower that obtained in a standard medium without pyruvate (Table S1). Therefore, longer CAP treatments (of 300 s) of CSC medium were needed to obtain PTL with a similar level of RONS to that observed in PTL-60 s by treatment of DMEM (Table S1). In line with our previous findings (Fig. 2), we found that the self-renewal ability of OS cells increased significantly upon exposure to PTL in both cell lines. However, S3I-201 was able to induce cell death and significantly reduce tumorsphere formation, and the combined treatment further fostered this cytotoxic effect and completely abolished tumorsphere growth in both cell lines (Fig. 4F-G).

2.5. STAT3 inhibition sensitizes bioengineered 3D OS models to PTL

The bone extracellular matrix plays a relevant role in the resistance of bone sarcomas to anti-tumor therapies due to its role in limiting the diffusion of therapeutic agents [9]. PTL was found to promote CSC properties in an 3D OS model engineered using MG-63 cells [33]. Here, we confirmed this higher resistance in a 3D environment by generating tumor models with SaOS-2, G-292 and U-2 OS seeded in Collagen Type 1 (Col1)/ Hydroxyapatite Nanoparticles (nHA)-FITC bone-like scaffolds (Fig. 5). Immunofluorescence staining of the cytoplasm (F-actin detection) and nuclei (DAPI staining) revealed that OS cells were uniformly distributed within the scaffold and in the close proximity to the nHA (FITC green staining) (Fig. 5A). All OS cell lines seeded in these bone-like scaffolds were able to proliferate for at least 9 days (Fig. 5B) and activate osteomimicry phenotype in the OS cells. This was determined by the up-regulation of the expression and/or enzymatic activity of genes related to osteogenic differentiation and the acquisition of the OS phenotype, such as RUNX2, SPP1 and ALPL (Fig. S5A-B). These data indicate that these 3D models reproduce the phenotype of OS in vitro better than 2D cell cultures.

Fig. 5.

Effects of PTL on 3D OS models. A) Immunofluorescence detection of F-Actin (orange) and DAPI (blue) in 6-day cultures of the indicated OS cells seeded in Col1/nHA-FITC scaffolds (green). Representative images of axial z-stack sections of the scaffolds are shown. B) Cell proliferation was determined by Picogreen assay measured after 3, 6 and 9 days of culture and relativized to day 0 values (4 h post-seeding). C) Western blotting analysis showing the expression of the indicated proteins in semi-confluent monolayer cultures (2D) and 3D model cultures (day 9) of the indicated cells lines. D) Protein quantification of GPX1, CAT1 and SOD2 in monolayer and corresponding 3D models on the indicated cell lines. Data are shown as ratio of protein intensity/GAPDH intensity. Bars indicates the amount of protein in 3D models relative to monolayer cultures for each cell line. GAPDH levels were used as loading control. E) Cell viability (Picogreen assay) of 2D and 3D models after 72 h of culture following treatment with the indicated doses of PTL generated by treating Ringer's saline with kINPen. Error bars represent the standard deviation of at least three independent experiments. asterisks indicate statistically significant differences with day 0 values in (b) (*p < 0.01; **p < 0.001; ****p < 0.0001 two-way ANOVA) and between 2D and 3D cultures in (d) (****p < 0.0001; two-sided Student's t-test). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The expression levels of antioxidant proteins revealed that different OS cell lines grown in bone-like scaffolds displayed increased levels of GPX1, SOD2 and CAT1 compared to the corresponding 2D cultures (Fig. 5C and D). In line with this finding, we found that OS 3D models were much more resistant than monolayer cultures to the effect of PTL generated by kINPen (Fig. 5E). Moreover, PTL induced a dose-dependent up-regulation of the pluripotency genes SOX2, OCT4 and NANOG in 3D models generated with SaOS-2, G-292 and U-2 OS cells (Fig. S6A).

With this in mind, we assayed whether S3I-201 (100 μM) was able to enhance the anti-tumor activity of PTL-240 s and to counteract its pro-stemness signals also in 3D models. First, we found that the expression of the anti-oxidant factors CAT1, GPX1 and SOD2 was consistently inhibited by S3I-201 and especially by the combined treatment of this inhibitor and PTL (Fig. S7A). Live/Dead assays revealed that S3I-201 largely enhanced the ability of PTL to induce cell death in the three different OS tumor engineered models (Fig. 6A). The quantification of cell viability in these assays showed that, while the treatment with S3I-201 alone was not cytotoxic, combination with this inhibitor was able to reduce the viability of OS cells treated with PTL from 50-70% to 10–20% depending on the cell line assayed (Fig. 6B). This enhanced anti-proliferative activity of the combination was also observed in experiments where OS cells were recovered from the tissue-engineered 3D models after the treatments and allowed to form colonies in Colony Formation Units (CFU) assays. In this case, the treatment with S3I-201 alone was able to reduce the clonogenic ability of OS cells, although, the combined treatment showed a greater effect than S3I-201 or PTL alone in all models assayed (Fig. 6C and D).

Fig. 6.

S3I-201 fosters the anti-tumor potential of PTL on 3D OS engineered tumor models. (A-B) Effects of PTL and/or S3I-201 in the survival of 3D OS models. G-292, SaOS-2 and U-2 OS cells were cultured for 6 days in Col1/nHA scaffolds and treated with the indicated doses of PTL (generated by treating Ringer's saline with kINPen plasma jet) alone or in combination with S3I-201 (100 μM) for 72 h. (A) Representative images of Live/Dead assays after the indicated treatments, showing live cells and dead with green and red fluorescence, respectively. Scale bars = 200 μm. (B) Quantification of the anti-proliferative effects (Picogreen assays) of S3I-201, PTL and combination of both in the indicated 3D models. (C-D) CFU assays performed with OS cells recovered from the indicated 3D models treated with 100 μM S3I-201, PTL-240 s or the combination of both for 72 h. Representative pictures of CFU assays (C) and the quantification of the colonies formed in each condition (D) are shown. E-F) Tumorsphere forming potential of G-292 cells recovered from 3D models treated with 100 μM S3I-201, PTL-240 s or the combination of both for 72 h. Representative images of tumorspheres (E) and quantification of the TFE (F) are shown. In all experiments, data are represented as mean and standard deviation (n = 3). Asterisks indicate statistically significant differences with the control group or between the indicated groups (**p < 0.001; ***p < 0.0001 ****p < 0.00001 two-way ANOVA). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Furthermore, to assay the ability of the combined treatment to target CSC subpopulations in a 3D environment, we analyzed the tumorsphere forming ability of G-292 cells recovered from bone-like scaffolds treated with 100 μM S3I-201, PTL-240 s or a combination of both for 72 h. Similar to that observed after the treatment of monolayer cultures (Fig. 2), PTL treatment significantly promoted the formation of tumorspheres. On the other hand, S3I-201 was able to reduce the potential to form tumorspheres and this anti-stemness effect was significantly fostered in combination with PTL (Fig. 6E and F). In line with these findings, we found that S3I-201 was able to abrogate the over-expression of SOX2, OCT4 and NANOG induced by PTL in several OS cell lines (Fig. S7B).

2.6. STAT3 inhibitor enhances the anti-tumorigenic effect of PTL in vivo

To validate the previous findings regarding the combined treatment with PTL and S3I-201 on OS development in vivo, we generated orthotopic i.b. 143-B + LUC xenografts and treated them by intratumor injection daily with PTL (Ringer's saline treated with kINPen plasma jet for either 600 s or 1200 s) and combined or not with 5 mg/kg S3I-201 three times a week for 18 days (Fig. 7A). The level of RONS detected in PTL-600 s generated by kINPen was higher than that of PTL-900 s generated by APPJ (Table S1). Nevertheless, this concentration of RONS was still in a range that promoted tumor growth (similarly to the observations in Fig. 3), as reflected by the increase in tumor volumes (Fig. 7B) and by measuring the luciferase activity at day 14 after initiating the treatment (Fig. 7C). However, PTL containing higher concentrations of RONS, PTL-1200 s, was able to slightly reduce tumor growth (Fig. 7B) and significantly decrease luciferase activity (Fig. 7C). Importantly, although S3I-201 alone did not have any impact on tumor growth, this inhibitor was able to abrogate the pro-tumor effect of PTL-600 s (Fig. 7B and C). Moreover, the combined treatment of S3I-201 and PTL-1200 s was able to significantly inhibit tumor growth at the experimental end-point (Fig. 7B) and the luciferase activity of tumors at day 14 (Fig. 7C). Interestingly, we did not observe any significant loss of weight or other adverse effects during the treatments (Fig. 7D). Altogether, these results indicate that the combination of a PTL containing a cytotoxic level of RONS with the inhibitor of STAT3 S3I-201 is an efficient treatment for OS that is able to eliminate CSC subpopulations and to inhibit tumor growth in vivo.

Fig. 7.

In vivo effect of PTL combined with S3I-201. A) Scheme of the procedure. The kINPen plasma jet was used to produce PTL from Ringer's saline. Orthotopic (i.b.) 143.B + LUC xenografts were allowed to grow until tumors reached a mean volume of 100 mm³ (approximately 16 days) and were randomly assigned (n = 6 per group) to receive daily intra-tumor treatments with either, vehicle (Ringer's Saline as control), PTL-600s or PTL-1200 s; oral doses of 5 mg/kg S3I-201 three times a week; or the combination of PTL treatments with S3I-201 for 18 days. B) Curves representing the relative mean tumor volume of xenografts during the treatments. C) Radiance values represented as the luciferase activity relative to day 0 values. D) Mean body weight of mice during the treatments. Data are shown as the mean and error of the mean. Asterisks indicate statistically significant differences with the control series or between the indicated series (**p < 0.01; ***p < 0.001; ****p < 0.0001; two-sided t-test). E) Scheme illustrating a possible mechanism of action of S3I-201 to enhance the cytotoxic effect of PTL by lowering the threshold of resistance to ROS in CSC subpopulations. Illustrations created using BioRender.com.

3. Discussion

New cancer treatment modalities are needed to improve the outcome of patients with difficult-to-treat cancer types such as OS. In the last decade, CAP, which is the fourth state of matter, has been explored with promising anti-tumor results in preclinical and in some clinical studies [49]. To apply this technology to living tissues, it is necessary to employ medical devices that generate cold plasma upon applying an electrical discharge to a gas that becomes partially ionized. This excited CAP operates at ambient pressure and biocompatible temperatures. CAP can be applied directly to superficial tumors [15], but to treat inner tumors like OS, an alternative has been developed to transfer the cytotoxic agents from CAP to liquids and allow local injection to the tumor [26,50]. To do this, CAP is applied to treat a liquid, which results in the formation of RONS within this liquid, that is designated as PTL [29,51]. This PTL, which retains the cytotoxic potential of CAP, could be used to treat cell cultures or tumors. Although both CAP and PTL have been proposed as selective anti-tumor therapies, it should be noted that the amount of RONS varies based on the type of plasma device, time, volume of liquid, distance to the target, and overall, the biochemical composition of the liquid [29]. Moreover, the concentration and composition regarding the kind of RONS in PTL is critical to ensure anti-tumor selectivity [25]. So far, the complexity of the components of the plasma, as well as the variety of experimental settings reported, limits comparison of data from different laboratories [52]. Different types of PTL have been suggested as potential therapies against OS. However, most in vitro studies conducted up to now have been performed on cancer cell monolayers [[53], [54], [55], [56], [57], [58], [59], [60]], which lack the 3D tumor microenvironment and therefore, may not accurately predict in vivo responses. To the best of our knowledge, to date the literature in the field has not undertaken the effects of plasma on CSC properties. These cells with self-renewing and pluripotency properties may be especially relevant in the management of OS, where therapy failure has been ascribed to its high CSC content and the presence of bone matrix that acts as barrier of drug diffusion [8,61].

Increasing evidence supports that CSC are highly resistant to RONS. Thus, CSC can: i) over-express antioxidant enzymes, ii) acquire a quiescent state, iii) display a greater ability to repair DNA and iv) adopt a high plasticity state to overcome mitochondrial damage induced by oxidative stress [45,62,63]. Here we show that PTL generated with long CAP treatment times, which contain high levels of RONS, are able to eliminate the ability of OS cells to grow as tumorspheres (Fig. 2B and C). However, these long CAP-treatment times are cytotoxic also to non-malignant hBM-MSCs and thus lose the anti-tumoral selectivity of PTL (Fig. 1D). Some reports have shown that CAP can increase the proliferation of healthy stem cells [25,64,65]. Here we confirm that low-dose PTL increase the proliferation of hBM-MSCs (Fig. 1 D & Fig. S4B) and the stemness potential of OS cells (Fig. 2 & Fig. S2). Accordingly, our data revealed that selective doses of PTL (targeting OS cells and not affecting non-malignant hBM-MSCs) resulted in a potent up-regulation of pluripotency genes and the increase of the percentage OS cells able to grow forming self-renewing tumorspheres (Figs. 2 and 4F). Importantly, we found that the induction of CSC properties is independent of plasma jet (Fig. 2 & Fig. S2) and the type of PTL employed (Fig. 2 Fig. 4F & Fig. S6), pointing out to the concentration of reactive species as key parameter (Table S1).

So far, the specific vulnerabilities of cancer cells and CSC to PTL are not well defined, being the main hypothesis, the defective redox defenses of cancer cells compared to normal cells [66]. RONS may favor the promotion and expansion of subpopulation of CSC which, in a paradoxical phenomenon, are able to maintain low levels of intracellular ROS in comparison with non-CSC subpopulations [67]. Among different pathways that may be activated by ROS, STAT3 signaling seems to play a pivotal role in the promotion of stemness in cancer cells [[68], [69], [70]]. In addition, previous studies reported that cold plasmas activate STAT3 in healthy stem cells [71]. Here, we found that the expression of CAT1, GPX1, SOD2 and STAT3 in OS is cell line dependent and the profile of expression of these factors may predict the response of tumor cells to PTL. Indeed, our results suggest that those lines expressing low levels of STAT3 are more sensitive to PTL treatment (Fig. 1F and G). In addition, low-dose PTL treatments induced much greater STAT3 activation in resistant MG-63 cells than in more sensitive 143.B cells and these dynamics of STAT3 activation correlated with the tumorsphere-forming potential of OS cells (Fig. 2C and E).

Relevant to the involvement of STAT3 and anti-redox enzymes in the promotion of stemness, a protective antioxidant feedback mechanism has been reported in which ROS-activated STAT3 signaling induce the expression of antioxidant factors such as GPX1, SOD2 or CAT1 [72]. In addition, the activation of pathways leading to the up-regulation of GPX1 and CAT1 has been shown to mediate ROS-induced stemness properties in CSC of different types of tumors [73,74]. Our results also support the existence of this feedback mechanism, since we show that the repression of the CSC phenotype in OS cells following STAT3 inhibition was associated with the inhibition of GPX1 and CAT1 expression (Fig. 4). Closing this feedback loop, we also found that the depletion of GPX1 led to increased levels of phospho-STAT3 and tumorsphere formation in the cell line, MG-63, showing a higher level of resistance against PTL treatment (Fig. 2F–H). In this sense, the anti-stemness activity of some STAT3 inhibitors was associated to the inhibition of CAT1 expression and the concomitant accumulation of intracellular ROS [75]. Altogether, our results suggest that the activation of STAT3 signaling plays a relevant role as a pro-stemness and resistance mechanism, able to restrain ROS levels in response to PTL treatment.

To confirm this prominent role of STAT3 in mediating PTL-induced stemness in OS, we analyzed the effect of inhibiting STAT3 with S3I-201, a small molecule capable of specifically blocking STAT3-mediated transcriptional activity by inhibiting its dimerization. A concentration of S3I-201 that showed only a marginal/moderate effect on the proliferation of 2D cultures and no effect on 3D models, was able to significantly reduce clonogenic potential and the self-renewal ability of both 2D and 3D models (Fig. 4, Fig. 6). These findings strongly suggest that S3I-201 may specifically target OS CSC subpopulations. Importantly, we found that S3I-201 was able to sensitize 3D models to PTL (Fig. 6). In this sense, the combined treatment was much more effective than each monotherapy alone in decreasing the cell viability of different OS monolayer cultures, tumorspheres (Fig. 4) and 3D models (Fig. 6 & Fig. S7) without affecting the proliferation of healthy stem cells (Fig. S4B). Moreover, S3I-201 counteracted the induction of the pro-stemness signals mediated by PTL. Mechanistically, the combined treatment produced a consistent down-regulation of STAT3, anti-oxidant factors and pluripotency genes in 2D cultures (Fig. 4A and B) and 3D models (Fig. S7). The reduced expression of this antioxidant defenses coupled to the increased accumulation of intracellular ROS in OS cells observed after the treatment with PTL and S3I-201 may contribute to explain the enhanced anti-tumor potential of this combined treatment (Fig. 4). This is in line with previous studies reporting OS CSC were effectively eliminated using STAT3 inhibitors in vitro and in vivo [37].

In this work, we examined for the first time the effect of treating tissue-engineered OS 3D models and OS xenografts with PTL. Our data indicate that the presence of a bone-like matrix in 3D models protects OS cells from the lethality of PTL observed in 2D cultures in G-292, U-2 OS, SaOS-2 cells (Fig. 5D). These differences in the response to PTL could be related to the higher expression of CAT1, SOD2 and GPX1 proteins observed when the cells are grown in a 3D environment (Fig. 5C and D). This increased level of resistance was also observed in the in vivo setting (Fig. 4). While the treatment of 2D cultures with PTL-240 s was completely cytotoxic for OS cells, a similar treatment of OS-bearing mice with PTL-300 s resulted totally insufficient to inhibit tumor growth (Fig. 3). Moreover, we found that the treatment of s.c. and i.b. OS xenografts with PTL containing moderate levels of RONS (PTL-900 s generated from APPJ plasma jet) resulted in a significant boost of tumor growth (Fig. 4). Only using a PTL containing large concentrations or RONS (PTL-1200 s from kINPen) we observed a decrease in tumor growth (Fig. 7B) as was reported for similar PTL treatments in other types of cancer [47,48]. Therefore, these data suggest that there is a threshold of RONS concentration below which a given PTL may promote tumor growth, and only PTL containing levels of RONS above this threshold may be effective in inhibiting tumor growth. However, this high-dose PTL may not be selective and may be toxic also for healthy tissues (Fig. 7E). Our data also show that low-dose PTL ensures anti-tumor selectivity in vitro, but this treatment is not enough to eliminate CSC properties or in vivo growth. Therefore, it is plausible to speculate that CSC are more efficient in eliminating RONS and we need to apply higher non-selective doses of PTL to be able to eliminate these subpopulations (Fig. 2). Thus, the pro-tumor effect of PTL containing levels below this threshold could be related to the pro-stemness effect of PTL and may be favored by the presence of a suitable microenvironment.

These promising results prompted us to assay the anti-tumoral activity of combined treatment in vivo. Our results confirmed that combined treatments were more efficient than PTL alone also in the in vivo setting (Fig. 7). Firstly, the undesirable effects observed after the treatment with PTL-600 s were efficiently abrogated by S3I-201. Secondly, the mice group that received high dose PTL-1200 s and S3I-201 showed an effective inhibition of tumor growth. Relevantly, this is the first study that applied PTL in mice bearing ectopic or orthotopic OS xenografts. Our data indicate that only high doses of plasma-generated RONS may have a partial anti-tumor activity. Since these doses are likely to affect healthy tissues, it is necessary to find combinatory therapeutic strategies able to avoid the unwanted pro-stemness and pro-tumor effects of PTL. Altogether, our results suggest that the inhibition of STAT3 activity may foster the anti-tumor activity of PTL in OS cells through the alteration of the antioxidant machinery, resulting in a lowering of the threshold of resistance to PTL of the CSC subpopulations (Fig. 7F).

4. Conclusions

This is the first study unveiling undesirable pro-tumoral functions by PTL in monolayer, 3D tumor-engineered models, CSC enriched cultures and xenografts. 3D environments greatly limit the effectiveness of PTL, promoting the survival of CSC and tumor growth in vivo. We found that the concomitant inhibition of STAT3 can counteract the adverse pro-stemness and pro-tumor effects of PTL by interfering with the antioxidant defenses of OS cells. The data presented here clearly showed that the combination of STAT3 inhibitors with CAP-based therapies could be considered a promising therapeutic option to eliminate the resistance of CSC to oxidative stress in OS.

5. Methods

5.1. Cell culture

Cell lines and drugs. MG-63, 143.B, G-292, SaOS-2 and U-2 OS human OS cells and hBM-MSCswere obtained from the ATCC. MG-63, 143.B and U-2 OS were maintained in DMEM, while SaOS-2 and G-292 were grown in McCoy's A5 Medium with 1,5 mM l-glutamine (Gibco™, Carlsbad, CA, USA) and hBM-MSCs were cultured in Advanced DMEM. All these cell culture mediums were supplemented with 10% of FBS, penicillin/streptomycin (50 U/mL and 50 μg/mL, respectively) and 1 mM sodium pyruvate, all from Gibco™. According to previous works [25,52], each cell type was cultured in DMEM without sodium pyruvate 2 h prior plasma treatments. 143.B cells expressing luciferase and GFP (143.B + LUC) were generated by lentiviral transduction as previously described [40]. All cell lines were maintained and cultured at 37 °C in a 95% humidified atmosphere containing 5% of CO₂. S3I-201 (cat. No S1155, Selleckhem) was prepared in 10 mM stock solutions in sterile DMSO, maintained at −20 °C and brought to the final concentration before use.

Lentiviral shRNA down-regulation of GPX1 expression. The knockdown of GPX1 was performed using two different sequences of shRNA under the control of a CMV promoter using previously described protocols [69]. 143.B and MG-63 cells were transduced with lentiviral particles from the SMARTvector Human Lentiviral shRNA collection (Dharmacon, Lafayette, CO, USA): shGPX1-I: Clone Id: V3SVHS07_6205961; Antisense: AACAGGACCAGCACCCATC and shGPX1-II: Clone Id: V3SVHS07_5483591; Antisense: TCCCGCAGGAAGGCGAAGA. A non-target shRNA (Dharmacon) was used as shRNA control. Selection of cells stably expressing GPX1 shRNAs and control shRNA started 72 h post-transduction using puromycin at 20 μG/mL for 72 h. The depletions of GPX1 were confirmed by western blotting in 143.B and MG-63 cells.

Tumorsphere assay. Cultures of sarcoma cells can be enriched in CSC due to the ability of these subpopulations to grow as floating tumorspheres [40,[74], [75], [76], [77]]. OS cell lines were plated at a density of 1.5 × 10³ per well in UltraLow Costar 6-well plates (Corning) in serum-free CSC medium containing 2 mL of DMEM-F12+Glutamax (Gibco), B-27/VitA Supplement (1:50; Life Technologies), Heparin (1:1000; Sigma), human EGF (20 ng/ml; GoldBio), human bFGF (10 ng/ml; GoldBio) and 1% sodium pyruvate. In addition, fresh aliquots of EGF and bFGF were added every three days. Rounded spheres (diameters ≥ 70 μM) formed after 10 days of culture were counted and images were captured using an optical microscope.

Tissue-engineered OS 3D models. Bone-like sponges composed by Col1/nHA were prepared as described in Ref. [78]. For some experiments, nHA were labelled with FITC to produce Col1/nHA-FITC scaffolds as previously explained [78]. Col1/nHA scaffolds were sterilized by immersion in 96% ethanol and washed twice in sterilized distilled water for 30 min. After washing, scaffolds were dried with sterile paper and 20 μL of FBS was added to hydrate them under cell culture-like conditions for 1h before cell seeding. Subsequently, 20 μL of culture medium containing 3 × 10⁵ cells were seeded on top of each scaffold and they were maintained at 37 °C, 95% humidity and 5% CO₂ during 1h to increase cell attachment. Finally, the scaffolds were transferred to non-adherent 6-well plates containing 4 mL of culture medium, which was replaced every 3 days.

Fluorescence imaging and immunofluorescence. Immunofluorescence staining of tissue-engineered OS 3D models was performed as previously described [78]. Briefly, Col1/nHA scaffolds containing 6-days cultures of OS cells were transferred to 24-well plates with DPBS, washed 3 times with DPBS and fixed with 1 mL of 4% paraformaldehyde (Sigma-Aldrich) for 20 min at room temperature. Then, the scaffolds were washed twice with DPBS, permeabilized with 0,1% Triton X-100 (Sigma-Aldrich) in DPBS for 5 min at 4 °C, washed 30 min with DPBS-0.1% Tween 20 and incubated with the antibody Alexa Fluor™ 546-conjugated anti-Phalloidin (Invitrogen, 1:400 in DPBS-0.1% Tween 20) for 1 h in the dark. Following this, the scaffolds were washed with DPBS-0.1% Tween 20 for 30 min in agitation and the whole scaffolds were mounted with Fluoroshield Mounting Medium with DAPI (Abcam, #ab104139). The resulting fluorescence was analyzed using a Zeiss laser scanning microscope.

3D Proliferation. The evaluate cell proliferation, dsDNA content was employed to estimate the number of cells. Col1/nHA scaffolds with cells adhered for 4 h (day 0) or after 3, 6 and 9 days or culture were collected in 1.5 mL Eppendorf tubes and stored at −20 °C until analysis. For DNA extraction, AllPrep DNA/RNA Mini Kit (Qiagen, 80204) was employed. dsDNA was determined using the Quant-iT™ PicoGreen ® dsDNA Kit (Invitrogen, #P11496) as previously described [78] and values were represented as fold change relativized to day 0.

ALP activity. Col1/nHA scaffolds at 0, 3, 6 and 9 days of culture were collected, mechanically disaggregated and digested with 0.5 mL of M-PER™ buffer (ThermoScientific). Then, samples were frozen and stored at −80 °C until analysis. ALP activity was analyzed using the SensoLyte® pNPP Alkaline Phosphatase Assay Kit (BioNova S.L.) following manufacturers indications. Samples were diluted 10–100 times to avoid signal overflow and values were relativized to the dsDNA content. Finally, values were represented as fold change relativized to day 0.

Western Blotting. Proteins levels were evaluated in control and after treatment with PTL and/or S3I-201 in monolayer and tissue-engineered OS 3D models. Whole cell protein extraction and Western blot analysis were based in previously described protocols [75]. Frozen cell pellets and 3D constructs were digested using the M-PER digestion buffer complemented with Halt protease and Halt phosphatase inhibitor cocktails (ThermoFischer). The resulting protein extracts were quantified using the Bradford assay (BioRad) and 30 μg of proteins of each sample were separated by SDS-PAGE and blotted onto nitrocellulose. Primary antibodies used in these analyses are summarized in Table S. 2.). Infrared fluorescent signals in Western blotting analysis were detected and quantified using Odyssey Fc imaging system and the software Image J. The signal intensity (pixel density) of the background was subtracted from the signal of each spot, next, normalized signal intensity was calculated by dividing each spot value over the corresponding housekeeping protein for each sample.

RT-qPCR Analysis. The gene expression was evaluated in 3D proliferation and in monolayer and tissue-engineered OS 3D models after treatment with PTL and/or S3I-201. RNA was isolated using the AllPrep DNA/RNA Mini Kit (Qiagen, 80204). One microgram of RNA was used for each RT reaction using the Maxima First Strand cDNA Synthesis Kit for qRT-PCR, with dsDNase (Thermo Scientific). The gene expression was assessed by qPCR using Fast SYBR™ Green PCR Master Mix (Qiagen). GAPDH was used as a housekeeping gene. PCR conditions for all genes analyzed were previously described [78]. Primer sequences used are shown in Table S2.

Plasma-Treated Liquids. Two CAP devices were employed to generate PTL: APPJ operating with Helium (5.0 Linde, Spain) was used in a jet design with a single electrode as described elsewhere [79]. The electrode was connected to a commercial high voltage power supply from Conrad Electronics (nominally 6 W power consumption). The discharge was operating with sinusoidal waveform at 25 kHz with (U) ∼ 2 kV and (I) ∼ 3 mA. Helium flow in the capillary was regulated to 1L/min through a MassView flow controller (Bronkhorst, Netherlands). In addition, we also employed another atmospheric plasma jet, kINPen (kINPen® IND, Neoplas tools GmbH, Greifswald), which has been approved for clinical use [15]. kINPen consists of a hand-held unit that produces a plasma discharge at atmospheric pressure, using Argon as gas to generate the discharge, in our work at a flow rate of 3L/min. This radiofrequency device consists of a pin-fed electrode in a quartz dielectric capillary with an outer ring electrode connected to the ground. The distance between the tip of the CAP jets and the surface of the liquid was fixed at 10 mm and treatment times from 5 to 1200 s were employed for both CAP devices.

Concentration of RONS in PTL. Here, two different RONS were investigated: NO₂⁻ and H₂O₂. The determination of Nitrites was performed using the Griess reagent and the concentration of hydrogen peroxide was determined by Amplex™ Red Hydrogen/HRP Peroxide Kit (cat.no A2218, Invitrogen) according to a previous protocol [52]. The concentrations of nitrites and hydrogen peroxides of the PTL employed in this work are summarized in Table S1.

Effects of PTL on monolayer cultures. PTL was generated in 24-well plates using 1 mL DMEM, high glucose (without glutamine, phenol red and pyruvate) as previously described [52]. To evaluate the antitumor effects of PTL or direct CAP treatment, a WST-1 (Roche) cell proliferation assay was performed according to a protocol described elsewhere [52]. Briefly, cells were seeded in a 24-well plate at a density of 3 × 10⁴ in 1 mL of culture medium. Following different treatments times, the culture medium was replaced after different incubation times with the WST-1 working solution (18 μL/mL). Untreated cultures were used as controls.

Then, the effect of PTL on the proliferation ability of OS cells, alone or in combination with S3I-201, was analyzed using the xCELLigence system (ACEA Biosciences, Inc, San Diego, CA, USA). A suspension of 1 × 10⁴ cells in 500 μL of culture medium was seeded in specially designed microtiter plates containing interdigitated gold microelectrodes. 24 h after cell seeding, we replaced 400 μL of culture medium with 400 μL PTL-5s containing or not 100 μM S3I-201. Real-time proliferation, measured as cell impedance changes (cell index), was monitored by the xCELLigence system every hour until the end of experiment. In addition, intracellular levels of ROS were measured using DCFH-DA in MG-63 and 143.B cells with PTL-30 s containing or not 100 μM S3I-201, following a previously described protocol [26].

Effects of PTL on tumorsphere formation. To evaluate the effect of PTL on the abundance of CSC, we pre-treated bulk monolayer cultures prior to assay their ability to form tumorspheres as previously described in [40]. Briefly, 3 × 10⁴ cells were seeded in 24-well plates in monolayer for 24 h with 1 mL of DMEM w/o sodium pyruvate (GibcoTM) and incubated for 72 h with PTL generated from DMEM w/o sodium pyruvate treated with CAP at increasing times. Then, viable remaining cells (determined by trypan blue staining) were recovered, filtered through a 70 μM cell strainer to avoid aggregates, and cultured in tumorsphere culture conditions for 10 days. The TFE for every PTL treated culture was calculated as follows: TFE = number of tumorspheres formed in pre-treated PTL cultures/number of tumorspheres formed in untreated cultures.

Alternatively, we explored the effect of PTL on pre-formed tumorspheres. Thus, tumorspheres pre-formed in CSC medium for 3 days were centrifuged (100 G for 10 min) and resuspended in 2 mL of PTL generated using the CSC medium treated with APPJ and cultured for another 7 days. Besides, some cultures were also treated with 100 μM S3I-201 and in combination with PTL. Following this procedure, the LIVE/DEAD™ Viability/Cytotoxicity Kit (Invitrogen, #L3224) was used to assay the viability of tumorsphere cultures, and images were taken under a Zeiss Cell Observer Live Imaging microscope (Zeiss, Thornwood, NY) coupled with a CO₂ and temperature-maintenance system.

Effects of PTL on tissue-engineered OS 3D models. To evaluate the effects of PTL on OS engineered 3D models, OS cells seeded in Col1/nHA scaffolds and allowed to grow for 6 days were placed in 24-well plates and exposed to 2 mL of untreated Ringer's saline or of PTL for 2 h. For these experiments, to produce PTL, 2 mL of sterile Ringer's saline (8.6 g/L NaCl, 0.33 g/L CaCl₂ and 0.3 g/L KCl) in a 24-well plate were treated with kINPen for 30–240 s. Then, 10% of FBS was added to the PTL before putting it in contact with the cells, according to previous works [46,72]. In some experiments, the 3D tumor models were also exposed to 100 μM of S3I-201 alone (diluted in Ringer's Saline) or in combination with PTL-240 s. Afterwards, 3D models and PTL for each condition were transferred to a non-adherent 6-well plate and fresh culture medium was added 1:1. After 9 days in culture, the LIVE/DEAD™ Viability/Cytotoxicity Kit (Invitrogen) was used according to the manufacturer's protocol to monitor live and dying cells using a confocal microscopy. Following this, DNA and RNA was isolated using the AllPrep DNA/RNA Mini Kit (Qiagen, 80204) and dsDNA content was estimated by Quant-iT™ PicoGreen ® dsDNA Kit (Invitrogen, #P11496) to determine cell viability and gene expression was analyzed by RT-qPCR.

In other experiments, the viable remaining OS cells after the treatments were extracted from the scaffold according to a previously described procedure [78] and were subjected to the tumorsphere assay (above) or to CFU. In the colony formation assay, 2.5 × 10³ viable cells (determined by trypan blue exclusion) per well were seeded in 6-well plates and allowed to form colonies for 14 days. Culture medium was replaced every 3 days. Then, cells were fixed with methanol, stained with Crystal Violet and counted as previously described [80].

5.2. In vivo effects of PTL

Effects of PTL produced by APPJ treatment in vivo. To produce the PTL, 1 mL of sterile Ringer's saline (8.6 g/L NaCl, 0.33 g/L CaCl2 and 0.3 g/L KCl) in a 24-well plate was treated with APPJ for 300 s or 900 s (PTL-300 s and PTL-900 s, respectively). To investigate the effects of PTL on ectopic tumors, 6-weeks-old female athymic nude mice (Envigo, Barcelona, Spain) were inoculated s.c. with 100.000 143.B + LUC cells mixed 1:1 with BD Matrigel Matrix High Concentration (Becton Dickinson-BD Biosciences, Erembodege) previously diluted 1:1 in culture medium. Once tumors reached approximately 100 mm³, the mice were randomly assigned (n = 3 per group) to receive the 100 μL of the following daily intra-tumor treatments (18 doses): vehicle (Ringer's Solution), or PTL-300s or PTL-900s. Animals were sacrificed when tumors reached approximately 1800 mm³. In parallel, we also generated orthotopic i.b. xenografts. Thus, mice were anaesthetized with isoflurane and the leg was bent 90° to drill the tip of the tibia with a 25G needle before cell inoculation (50.000 cells in 10 μL of complete DMEM per mouse) using a 27G needle as previously described [81].

Effects of PTL produced by kINPen treatment in vivo. To produce PTL, 1 mL of sterile Ringer's saline (8.6 g/L NaCl, 0.33 g/L CaCl2 and 0.3 g/L KCl) in a 24 well plate was treated with kINPen for 600 s and 1200 s (PTL-600 s and PTL-1200 s, respectively). 6-weeks-old female athymic nude mice (Envigo, Barcelona, Spain) were inoculated i.b. with 143.B + LUC cells (50.000 cells in 10 μL of complete DMEM per mouse) injected into the left tibia of mice [81]. Once the tumors reached a mean volume of 100 mm³ (corresponding to a luciferase activity of approximately 10⁵ photons/s), mice were divided into 6 groups (n = 6 per group) to receive vehicle (corn-oil) or S3I-201 treatment, PTL or the combination of S3I-201 and PTL. The mice were randomly assigned (n = 6 per group) as follows vehicle (Ringer's Solution), PTL-600 s or PTL-1200 s, 200 μL of S3I-201 alone (orally), and S3I-201 combined with PTL-600 s or PTL-1200 s. Each PTL group received daily intra-tumor treatments (18 doses) of 100 μL of PTL. S3I-201 was used in vivo at a dose of 5 mg/kg (diluted in corn oil containing 2% DMSO) according to manufacturer's instructions, and administered 3 times per week orally. Control groups received intra-tumor injections of 100 μL of Ringer's Saline and/or 200 μL of corn-oil containing 2% DMSO orally.

Differences in mean tumor volume between groups were determined using a caliper. Relative tumor volume (RTV) for every xenograft was calculated as follows:

$$RTV=V_t-V_0$$

where RTV is relative tumor volume, V_t is tumor volume at day of measurement and V₀ is tumor volume at the beginning of the treatment.

Bioluminescent in vivo was performed using In Vivo Imaging System (IVIS) Lumina II (PerkinElmer). Anaesthetized mice were injected s.c. with 150 mg D-Luciferin/kg body weight (Melford Laboratories) and monitored in the IVIS 25 min post-injection. Bioluminescence intensity was quantified using Living Image Software. Differences in bioluminescence between groups were determined using a standard size ROI to quantify luciferase signal for every xenograft during the treatments. The evolution of biolumiscence for each tumor was calculated as follows:

$$RelativeLuciferaseActivity=L_t\left(\frac{\frac{\mathrm{p}\mathrm{h}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{n}\mathrm{s}}{\mathrm{s}}}{{cm}^2}\right)-L_0\left(\frac{\frac{\mathrm{p}\mathrm{h}\mathrm{o}\mathrm{t}\mathrm{o}\mathrm{n}\mathrm{s}}{\mathrm{s}}}{{cm}^2}\right)$$

where L_t is photons/s/cm² at day of measurement and L₀, the levels at the beginning of the treatment for each tumor. All experimental animal protocols were carried out in accordance with the institutional guidelines of the University of Oviedo and were approved by the Animal Research Ethical Committee of the University of Oviedo (Ref. 14/2020). Bone destruction was evaluated in the preclinical image laboratory of the University of Oviedo using a μCT system (SkyScan 1174, Bruker, Antwerp, Belgium) following mice sacrifice at day 18.

Statistics. Statistical analysis was performed using GraphPad Prism version 8.0 (Graphpad Software Inc, La Jolla, CA, USA). Data are presented as the mean (± standard deviation or error of the mean, as indicated) of at least three independent experiments. Two-sided Student's t-test was performed to determine the statistical significance between groups. Multiple comparisons of the data were performed using the one-way ANOVA, and two-way ANOVA. p < 0.05 values were considered statistically significant. Pearson's correlation was used for correlation analysis with significance set at 5%.

Author contributions

JT conceived the project and designed experiments; JT, MM, VR, DM, CH and BG performed experiments and analyzed data; JT and MM wrote the manuscript.: JT, VR, DM performed in vivo experiments; AR contributed to the review of the manuscript: CC & RR contributed to the review of the manuscript; supervised the study & obtained funding.

Conflicts of interest

The authors declare that there is no conflict of interest.

Declaration of competing interest

The authors declare no conflicts of interest.

Appendix A. Supplementary data

The following are the supplementary data to this article:

Data availability

Data will be made available on request.