Electrotransfer of plasmid DNA radiosensitizes B16F10 tumors through activation of immune response

Murine melanoma cell line B16F10 (American Type Culture Collection, Manassas, VA, USA) was cultured in advanced minimum essential medium (AMEM, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 5% fetal bovine serum (FBS, Gibco), 10 mM/L L-glutamine (GlutaMAX, Gibco), 100 U/mL penicillin (Grünenthal, Aachen, Germany) and 50 mg/mL gentamicin (Krka, Novo mesto, Slovenia) in a 5% CO₂ humidified incubator at 37°C.

The plasmid with tissue specific promoter for endothelial cells, encoding shRNA for silencing endoglin (pET-antiCD105; TS plasmid) was used in experiments as the therapeutic plasmid.¹⁸ The control plasmid, encoding shRNA with no homology to any gene in the mouse genome and with constitutive CMV promoter, was used as a negative control (pControl).²⁵ Amplification of both plasmids was performed in a competent E.coli (JM107; Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA).

All plasmids were isolated using JETSTAR 2.0 ENDOTOXIN-FREE Plasmid MEGA Kit (Genomed, Löhne, Germany) and diluted in endotoxin free water to a concentration of 1 μg/μL (in vitro experiments) and 4 μg/μL (in vivo experiments). Concentrations of plasmids were measured with a spectrophotometer at 260 nm (Epoch Microplate Spectrophotometer, Take3 Micro-Volume Plate, BioTek, Bad Friedrichshall, Germany) and purity of plasmid was determined by agarose gel electrophoresis and measurements of the absorbance ratio at 260 and 280 nm.

All animal experiments were conducted in accordance with the guidelines for animal experiments of the EU Directive and the permission obtained from the Ministry of Agriculture and the Environment of the Republic of Slovenia (Permission No. 344011/2015/16), which was given, based on the approval of the National Ethics Committee for Experiments on Laboratory Animals).

Female C57Bl/6 mice, 6-8-week old, purchased from Envigo Laboratories (Udine, Italy), were used in the study. Before the experiment, mice were subjected to an adaptation period of 2 weeks. Animals were maintained under specific pathogen-free conditions at a constant room temperature, humidity and a 12 h light/dark cycle. Food and water were provided ad libitum. For induction of subcutaneous tumors, a suspension of 1 × 10⁶ B16F10 cells in 0.1 ml of physiological saline was injected subcutaneously into the back of the mice. The animals bearing tumors of 40 mm³ were randomly divided into experimental groups and subjected to a specific experimental protocol. The tumor measurements were completed when the tumors reached 350 mm³, and mice were humanely sacrificed.

Experimental groups and the number of animals in each of them were as follows and as described in table Table 1: triple injection of endotoxin-free water alone (control group; CTRL) or of pControl or TS plasmids alone, or in combination with triple application of electric pulses alone (3 × EP) or combined with plasmids (3 × GET (pControl); 3 × GET (TS)). Furthermore, the remaining groups were also the ones in combination with irradiation (IR) and other therapies described above, which are three injections of plasmids (pControl + IR; TS + IR) and in combination with electric pulses (3 × EP + IR; 3 × GET (pControl) + IR; 3 × GET (TS) + IR).

In vivo GET of plasmid into subcutaneous tumors was performed 3 times every second day (on days 0, 2 and 4). 12.5 μL (4 μg/μL) of plasmid (150 μg in total) in endotoxin-free H₂O was injected intratumorally 10 min before 8 square electric pulses with a voltage-to-distance ratio of 600 V/cm, a pulse duration of 5 ms, and a frequency 1 Hz were applied. Electric pulses were generated by electric pulse generator ELECTRO CELL B10 (Betatech, L’Union, France) and delivered through 2 parallel stainless steel electrodes with 2 or 4 mm distance between them, depending on the tumor volume. After 4 pulses, electrodes were turned for 90° for 4 additional pulses to assure GET to entire tumor.

Tumors were locally irradiated with a single dose of 15 Gy on day 1 from the beginning of the experiment, at a dose rate of 2.16 Gy/min, using a Darpac 2000 X-ray unit (Gulmay Medical Ltd., Shepperton, UK) operating at 220 kV, 10 mA, with 1.8-mm aluminum filtration. During irradiation, mice were restrained in special lead holders with apertures for irradiation of the tumors. Due to the fixed size of the apertures, some healthy tissue (3 – 5 mm of skin surrounding the tumor) was exposed to the irradiation as well.

The therapeutic potential in vivo was assessed by measuring the tumor size every second day and calculating tumor volume according to the formula for ellipsoid: V=axbxc π/6, where a, b and c represent perpendicular tumor diameters.^21,26 The tumor growth curves were drawn as arithmetic means (AM) with bars representing standard errors (SEM).

The tumor growth delay for each experimental group was calculated as the difference in tumor doubling times of experimental and control groups. Tumor doubling time is the number of days in which the initial tumor volume (40-50 mm³) doubles. Mice that remained tumor free for 70 days were termed tumor free and local tumor control was deemed to have been achieved (Table 1). The weight of the mice was followed as a general index of systemic toxicity, and acute skin reaction in the whole irradiated field around the tumor was evaluated as described elsewhere.¹⁷

After 70 days, when mice were designated as tumor free, they were challenged with a subcutaneous injection of 1 × 10⁶ B16F10 cells in 0.1 ml of physiological saline in the right flank. Mice that at least 20 days after the challenge remained tumor free were marked as resistant to secondary challenge (Table 1, Figure 2). The growing tumors were measured twice a week and when volume of 150 mm³ was reached mice were sacrificed and tumors were collected for further histological analysis as described below.

After therapies, at day 6, three mice from each experimental group were sacrificed. The tumors were excised, fixed in IHC zinc fixative (BD Biosciences, San Diego, CA, USA) and embedded in paraffin. Six consecutive 2-μm thick sections were cut from each paraffin block and stained as followed. To estimate the percent of the area of tumor necrosis, the first section was stained with hematoxylin and eosin. The other five sections were used for immunohistochemical (IHC) staining to evaluate percentage of hypoxic cells, cells in apoptosis, immune cells, proliferating cells and the number of blood vessels. To determine hypoxic cells, rabbit polyclonal antibodies against HIF-1-alpha (ab2185, Abcam, Cambridge, MA, USA) at dilution 1:3500, were used. In addition, apoptosis was evaluated with help of cleaved Caspase-3 (Asp175., Cell signaling Technology, Danvers, MA, USA) at dilution 1:1500, whereas immune cells (NK and CTL) were stained with help of Granzyme B (ab4059, Abcam) at dilution 1:1250. For staining proliferating cells, rabbit monoclonal antibodies against Ki-67 (clone SP6, Thermo Fisher Scientific) at dilution 1:1200 were used. The last section was stained for determination of the number of blood vessels, by using primary rabbit polyclonal antibodies against CD31 (ab28364, Abcam) at dilution 1:1000. For these sections, a peroxidase-conjugated streptavidin–biotin system (Rabbit specific HRP/DAB detection IHC kit, ab64261, Abcam) was used as the colorogenic reagent followed by hematoxylin counterstaining. From each slide and each feature (apoptosis, hypoxia, proliferation, vascularisation and immune cells), five randomly selected viable parts of each tumor were observed and captured under the light microscopy, by DP72 CCD camera (Olympus, Hamburg, Germany) connected to a BX-51 microscope (Olympus) under 40× magnification (numerical aperture 0.85). The viable parts were analyzed in blind fashion and the results were presented as the percent (hypoxia, apoptosis, proliferation) of the cells or the number of cells (immune cells) or the structures (vascularization) positive to IHC staining. The percentage of necrosis was contributed to the tumors as whole and was also evaluated in blind fashion, as previously described.¹⁸

In addition, tumors that grew up to 150 mm³ after secondary challenge of tumors, were excised, fixed and embedded in paraffin, as described above. Furthermore, from each tumor sections were cut and stained with hematoxylin and eosin to determinate morphological changes of tumor cells.

All data were tested for normality of distribution with the Shapiro-Wilk test. The differences between the experimental groups were statistically evaluated by one-way analysis of variance (one-way ANOVA) followed by a Holm-Sidak test for multiple comparison. A P-value of less than 0.05 was considered to be statistically significant. SigmaPlot Software (Systat Software, Chicago, IL, USA) was used for statistical analysis and graphical representation.

GET of either plasmids (pControl, TS) to melanoma tumors had statistically significant antitumor effectiveness compared to untreated tumors, which resulted in 8.2 ± 2.5 and 8.6 ± 3.0 days of tumor growth delay, respectively (Table 1). However, GET (TS) resulted in 44% of tumor free mice and 75% of them were resistant to secondary challenge, whereas in the GET (pControl) group no tumor free mice were obtained. Histological analysis (Table 2) of GET (TS) group demonstrated reduction of vascularization (14.9 ± 1.1%) and proliferating cells (49.5 ± 3.8%), whereas hypoxia (46.3 ± 3.0%) levels were increased and statistically significant to at least pertinent control groups (CTRL, pControl, TS and 3 × GET (pControl)). The levels of necrosis, apoptosis and number of infiltrating immune cells in the tumors were comparable in both GET treatment modalities (Figure 1) and statistically significant to at least CTRL, pControl and TS. These results indicated the vascular targeted effects of the GET (TS), used for silencing endoglin.

Irradiation monotherapy with 15 Gy had minor effects on radioresistant melanoma tumor. The tumor growth delay was moderate in the groups of irradiation alone or in combination with injection of plasmids (from 0.7 ± 0.3 to 2.8 ± 1.2 days) and up to 11% of tumor free mice were obtained in the group of irradiation in combination with injection of plasmid pControl (Table 1). The results of histological analysis (Table 2, Figure 1) indicates an immunological effect of the irradiation alone or in combination with plasmids injection, since the number of infiltrating immune cells in the tumors was statistically significantly higher in comparison to control groups (CTRL, pControl, TS) and groups applying electric pulses (alone or combined with injection of plasmids (GET (pControl) and GET (TS)), and the tumor free mice were resistant to secondary challenge. Furthermore, irradiation alone caused the reduction in proliferating cell levels (53.3 ± 3.5%) and in tumor vascularization (30.1 ± 2.3%), whereas apoptosis (18.4 ± 1.8%) and hypoxia (30.2 ± 2.4%) levels were elevated. All of these results were statistically significant compared to CTRL, pControl and TS groups.

The application of electric pulses to tumors in combination with tumor irradiation resulted in better antitumor effectiveness than observed in each of these two treatment modalities alone. The tumor growth delay in combined treatment was 3.2 ± 1.3 days and 22% of tumor free mice were obtained (Table 1). The histological analysis of combined treatment demonstrated statistically significant higher levels of apoptosis (31.0 ± 2.0%) and hypoxia (51.7 ± 2.6%) in comparison to the groups of irradiation and electric pulses alone. Among these three groups no statistically significant differences were observed in the levels of necrosis. Furthermore, the analysis of levels of proliferating cells, the number of tumor blood vessels and immune cells infiltrating in the tumors, resulted in similar and moderate effectiveness of the combined treatment (3 × EP + IR) and irradiation (alone or combined with injection of plasmids), which statistically significantly differed from group of electric pulses alone. Nevertheless, in this combined treatment modality (3 × EP + IR) no tumor free mice were resistant to secondary challenge of tumors.

The groups combining GET of plasmids (pControl, TS) and irradiation resulted in pronounced therapeutic effectiveness, with up to 89% and 88% of tumor free mice, respectively, and from those up to 63% and 57% of mice was resistant to secondary tumor challenge, respectively (Table 1). The histological analysis, of tumors excised 6 days after treatment (Table 2, Figure 1), indicated on similar mode of action of these combined treatment modalities (3 × GET (pControl or TS) + IR), since the elevation of necrosis (65.0 ± 7.6%, 79.2 ± 4.2%), number of immune cells (45.8 ± 3.0%, 46.5 ± 2.1%), reduction of tumor vascularization (18.9 ± 2.0%, 12.8 ± 1.0%) and proliferation (46.0 ± 4.8%, 46.7 ± 4.8%), did not differ between these groups. Only two statistically significant differences between the therapeutic groups combining GET (pControl, TS) and irradiation were observed which were the level of hypoxia (58.0 ± 2.8%, 66.7 ± 1.6%), and apoptosis (38.7 ± 3.1%, 50.3 ± 2.4%).

The high number of infiltrating immune cells (Table 2) in these tumors indicated on important mode of this therapeutic action; the highest and statistically significantly increased number of immune cells was observed in the groups combining GET of plasmids and irradiation, followed by all the remaining groups that included irradiation (alone, TS, pControl and 3 × EP). In these groups, the immune cells infiltration was comparable and statistically significant to all of the other remaining groups. Furthermore, the immune cells infiltration was the lowest in the groups that included electric pulses (alone or with plasmids alone). Additionally, all of the mice that were tumor free after the therapies including irradiation, either alone or in combination with other modalities, had fur discoloration, known as vitiligo, indicating immune response (Figure 2). This was further confirmed with high number of mice that were resistant to secondary challenge in tumor free mice, with the exception of the group combining electric pulses and irradiation (Table 1). In these groups, in which tumors grew after secondary challenge, the growth rate and histology of the tumors were the same as after the initial induction of tumors (data not shown).

In addition, the safety of the treatment (irradiation alone or in combination with other modalities) was proven, since no body weight loss over 10%, or any other side effects were observed, except for the temporary hair loss in the irradiated area, without skin desquamation (data not shown).

The first direct effect of GET (TS) was on tumor vasculature, which was significantly reduced after the treatment. This vascular effect can be ascribed to the specificity of the plasmid for endothelial cells and to the endoglin silencing. This direct effect on tumor vasculature was also observed in other studies using GET of plasmid for silencing endoglin on melanoma^22,23 and other tumor models.^18,19,21 The vascular targeted effects in these studies were first confirmed in non-endoglin expressing tumors, i.e. murine mammary adenocarcinoma, by endothelial-specific promoter^18,21 and non-specific, constitutive promoter^18–21, that resulted in pronounced antitumor effectiveness. In melanoma tumor model, B16F10-luc, the GET of plasmids silencing endoglin resulted in significant antimetastatic effectiveness.²² Furthermore, GET of plasmid silencing endoglin, with constitutive promoter, performed on small melanoma B16F10 tumor model (4 mm³ at the beginning of therapy) confirmed vascular targeted effects, which resulted in prolonged tumor growth delay and tumor free mice (58%), also due to nonspecific nature of constitutive plasmid that was used in this study silenced endoglin also in melanoma cells.²³ Nevertheless, in the current study, which was done on bigger tumors (40 mm³), in addition to significant increased level of hypoxia, also decreased number of proliferating tumor cells was observed. This was attributed to the second direct effect of GET (TS), that is silencing of endoglin also in melanoma cells, since it is known that plasmid with tissue specific promoters can be leaky and can also be transcribed in non-targeted tissue.²⁷ Therefore, GET (TS) has dual direct effects; primarily by targeting vasculature and secondly by inhibition of proliferation of melanoma cells. These two effects together resulted in 44% of tumor free mice, from which 75% were resistant to secondary challenge of tumor cells.

Nevertheless, high level of infiltrating immune cells in the tumors, indicated also an indirect effect of this treatment, through the stimulation of immune response. This is also supported by the data obtained with GET of the plasmid devoid of therapeutic gene (pControl). In comparison to GET of therapeutic plasmid, TS, no tumor cures were obtained, however, in other measured parameters, infiltration of immune cells into the tumors, levels of necrosis and apoptosis as well as in tumor growth delay no statistically significant differences were obtained between these groups. Thus, to a certain degree, a similar mode of antitumor action can be ascribed to the GET of plasmid DNA, mainly because the values of these parameters were statistically different from pertinent control groups. Antitumor effectiveness of GET of non-therapeutic (control) plasmids was also observed in other studies^9,10, and the authors indicated that effectiveness was due to involvement of immune system. Namely, the plasmid introduced during GET can act as a DAMP that is recognized through different sensors (Pattern Recognition Receptors), leading to the activation of the signal transduction cascade that ultimately triggers the production of type 1 interferons and other cytokines.²⁸ These act as a link between the innate and adaptive immune response²⁹ and can induce the adaptive immune response against the introduced DNA and consequently the transfected cells. Furthermore, similar immune response can also be triggered by the stress³⁰ that is produced during the transfection procedure, like mechanical stress, heat, and ROS, that have all been previously reported after electroporation.^31,32

To target primary tumors, as was done in this study, irradiation is one of the most used treatment modalities.³³ This therapeutic approach alone or in combination with injection of plasmids, in radioresistant melanoma tumor model, resulted in moderate tumor growth delay, which was attributed to decreased proliferation of melanoma cells, increased levels of apoptosis and necrosis. Further histological analysis indicated on the radiation damage of the tumor vasculature, as a second effect of the irradiation, as seen in other studies.^34,35 This vascular damage was less pronounced than in group of GET (TS), although in comparison to GET (TS), irradiation monotherapy resulted in higher infiltration of the immune cells in the tumors. Furthermore, in the group combining irradiation and injection of therapeutic plasmid, one mouse was tumor free, which was also resistant to secondary challenge of tumor. This indicates a priming effect of irradiation combined with introduction of foreign DNA, through boosting the immune response. When irradiated cells die, they release their antigens in the context of the danger signal (DAMPS), which result in the priming of a tumor specific immune response against the released antigens. Therefore, irradiated tumors can sometimes act as a powerful individualized in situ vaccine¹³, which is manifested as the abscopal effect of the irradiation. This was confirmed in in vivo studies indicating that irradiation can induce in vivo priming of T cells to exogenous model antigens engineered to be expressed by tumors.¹³ In melanoma tumor model, priming of antitumor T cells to the model antigen ovalbumin, was more effective when single, 15 Gy dose was used, rather than 3 Gy given in 5 consecutive days.³⁶ Further on, another group also showed induction of antitumor T-cell responses with other antigen expression, when single 20 Gy dose was applied, but not by 5 Gy given 4 times.³⁷

The combination of electric pulses and irradiation exerted moderate antitumor effectiveness that resulted in 22% of tumor free mice. Similar results were obtained in our previous studies on murine mammary adenocarcinoma¹⁸, sarcoma¹⁷ and Ehrlich-Lettre ascites³⁸, where also 20%, 27% and 54% of tumor free mice were observed, respectively. Therefore we can assume that electroporation of tumors contributes to radioresponsiveness of melanoma, most probably due to generation of ROS^32,39 after application of the electric pules, which also resulted in significantly elevated levels of hypoxia in our histological analysis.

The combined treatment modalities (combination of GET (TS or pControl) with irradiation) exerted excellent, highly statistically significant antitumor effect in comparison to control groups, which resulted in 88% and 89% of tumor free mice, respectively, of which 57% and 63% of mice were resistant to secondary challenge of tumors, respectively. The results of histological analysis were similar between groups of combined treatment modality (GET + IR), regardless of the applied plasmid. The only two differences were on the levels of apoptosis and hypoxia. The analysis of the presence of the immune cells showed on the highest number of the immune cells in the groups of GET of plasmids and irradiation, followed by irradiation groups (alone or combined with injections of plasmids or combined with electric pulses). Therefore, it can be presumed that the priming effect of irradiation can be boosted by GET of the plasmid DNA to fully exert vaccinating effect. Additionally, fur discoloration, vitiligo, an immune-mediated destruction of normal melanocytes that has also been recognized as a positive prognostic indicator for treatment response^40,41, was also observed in this study in all the groups where tumor free mice were observed (except 3 × GET (TS)). In addition, majority of the tumor free mice were resistant to secondary challenge of tumors, further indicating on the development of the immune memory. The tumors in experimental groups, which were resistant to secondary challenged of tumor, were therefore radiosensibilized through GET of plasmid, since in the group combining electric pulses and irradiation, no mice resulted resistant to secondary challenge of tumor. The GET of plasmids in combination with irradiation presumably generated many danger signals that collectively define immunogenic cell death.¹³

Furthermore, the immunogenicity of tumor might also play an important role in combined treatment modality.⁴² In the current study we obtained significant antitumor effectiveness and up to 89% of tumors free mice, that were up to 63% resistant to secondary challenge of tumors, when combining GET and irradiation in melanoma tumor model. Furthermore, in our previous study combining the same treatment modality on a different tumor model, murine mammary adenocarcinoma TS/A, we achieved up to 44% of tumors free mice with therapeutic plasmid silencing endoglin and 20% tumor free mice with the control plasmid.¹⁸ We can presume that the boosting of immune response depends on the tumor type, melanoma being more immunogenic that mammary adenocarcinoma TS/A. Nevertheless, the mechanisms of these therapies are not fully elucidated. Currently, we presume that there is involvement of DNA sensors, ROS and specific immune response after irradiation, but further studies are needed.