The nitric oxide donor JS-K sensitizes U87 glioma cells to repetitive irradiation

Abstract

As a potent radiosensitizer nitric oxide (NO) may be a putative adjuvant in the treatment of malignant gliomas which are known for their radio- and chemoresistance. The NO donor prodrug JS-K (O2-(2.4-dinitrophenyl) 1-[(4-ethoxycarbonyl) piperazin-1-yl] diazen-1-ium-1,2-diolate) allows cell-type specific intracellular NO release via enzymatic activation by glutathione-S-transferases overexpressed in glioblastoma multiforme. The cytotoxic and radiosensitizing efficacy of JS-K was assessed in U87 glioma cells in vitro focusing on cell proliferation, induction of DNA damage, and cell death. In vivo efficacy of JS-K and repetitive irradiation were investigated in an orthotopic U87 xenograft model in mice. For the first time, we could show that JS-K acts as a potent cytotoxic and radiosensitizing agent in U87 cells in vitro. This dose- and time-dependent effect is due to an enhanced induction of DNA double-strand breaks leading to mitotic catastrophe as the dominant form of cell death. However, this potent cytotoxic and radiosensitizing effect could not be confirmed in an intracranial U87 xenograft model, possibly due to insufficient delivery into the brain. Although NO donor treatment was well tolerated, neither a retardation of tumor growth nor an extended survival could be observed after JS-K and/or radiotherapy.

Introduction

Glioblastoma multiforme (GBM) is the most frequent malignant primary brain tumor in adults with a median survival of 16 months despite multimodal treatment. ¹ One reason for the dismal prognosis of GBM is a pronounced radioresistance. ² Nitric oxide (NO) is one of the strongest known radiosensitizing agents. However, its direct application as an antitumor agent is limited as it is a free radical with a half-life of a few seconds in the presence of metal complexes and free radicals.^3,4 Thus, various NO-releasing prodrugs have been developed for the application of NO in different experimental and clinical situations including cancer treatment. First insights for a radiosensitizing effect of NO were shown in V79 Chinese hamster lung fibroblasts using Angeli’s salt in combination with ferricyanide or tempol, S-nitrosoglutathione (GSNO), and 2(N,N-diethylamino)-diazenolate-2-oxide (DEA/NO) under hypoxic culture conditions. ⁵ NO agents, that is, NO-sulindac, isosorbide dinitrate S-nitroso-N-acetyl-penicillamine (SNAP), or sodium nitroprusside (SNP), evoke radiosensitizing effects in different kinds of tumors, including prostate cancer, ⁶ pancreatic cancer, ⁷ and malignant gliomas. ⁸ Kurimoto et al. demonstrated radiosensitizing effects of the NO donors SNAP and SNP in GBM cells by assessing growth inhibition in an in vitro glioma model. However, these two NO donors release NO in a cell-type independent manner. To achieve cell-specific therapeutic NO concentrations, the diazeniumdiolates JS-K and its derivate PABA/NO (O(2) -{2,4-dinitro-5-[4-(N-methylamino)benzoyloxy]phenyl} 1-(N,N-dimethylamino)diazen-1-ium-1,2-diolate) were developed which release NO upon enzymatic metabolism by glutathione-S-transferases (GST). ⁹ These are detoxification type II enzymes overexpressed in many cancers including GBM contributing to their treatment resistance. It has been shown that JS-K has antineoplastic effects in a variety of different tumor types such as prostate cancer, multiple myeloma, and lung cancer.^10–12 Our group demonstrated antiproliferative and cytotoxic effects of PABA/NO in U87 cells and shows a strong chemosensitizing effect in GBM cells for temozolomide. ¹³ Furthermore, we demonstrated antiproliferative and growth inhibitory effects of JS-K in U87 cells in vitro and in a flank tumor model in vivo. ¹⁴ In previous studies, our group could demonstrate that NO released from the proline NONOate (PROLI/NO) permeabilizes the blood–tumor barrier in an intracerebral C6 glioma model in a cyclic guanosine monophosphate (cGMP)-dependent manner, implying a possible beneficial effect of NO donor treatment. ¹⁵ While antitumor effects of NO donors have been described in gliomas, the role of NO donors as radiosensitizers in gliomas is not well understood. The promising antineoplastic effects of NO donors in general and of NONOates in particular encouraged us to investigate possible radiosensitizing effects of JS-K in U87 glioma cells in vitro and in vivo.

Materials and methods

In vitro experiments

Human U87 cells (HTB-14), fibroblasts (CRL-1634), and astrocytes (CRL-8621; ATCC) were cultured in Dulbecco’s Minimal Essential Eagle Medium (DMEM; Gibco, USA) supplemented with 10% fetal calf serum (FCS) and penicillin (100 U/mL)/streptomycin (100 mg/mL) under normoxic conditions (95% air, 5% CO₂, and 37°C).

The nitric oxide donor JS-K (O2-(2.4-dinitrophenyl) 1-[(4-ethoxycarbonyl) piperazin-1-yl] diazen-1-ium-1,2-diolate, CAS 7054 32-12-8) was dissolved in 100% dimethyl sulfoxide (DMSO) in a concentration of 5.2 mM. U87 was incubated with JS-K doses of 0–15 µM for 4 h. The JS-K solvent control (DMSO ⩽ 1%) was adjusted to the highest JS-K concentration used in these experiments. A cesium source emitting γ-radiation was used for radiotherapy (RT; 3 × 2 Gy) with a 24-h interval between each RT dose (Isotopen Diagnostik CIS, Germany).

Immunocytochemistry for GST isoforms and γH₂AX

After permeabilization and blocking with 10% normal goat serum paraformaldehyde (PFA)-fixed U87 cells, astrocytes and fibroblasts were incubated with primary antibodies against GST isoform α (anti-GSTA1-1, 1:100; Calbiochem, Darmstadt, Germany) or anti-γH₂AX (Ser139, 1:500; Cell Signaling Technology, Danvers, USA) overnight at 4°C. After three wash cycles, they were incubated with secondary antibodies (goat α-rabbit Alexa 568, 1:400; Invitrogen, Karlsruhe, Germany) for 1 h at room temperature. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich, Munich, Germany). Expression of GST isoform α was assessed using fluorescence microscopy. The number of γH₂AX foci was quantified using confocal microscopy (Olympus FV1000MPE, Tokyo, Japan) 24 h after the end of treatment. Three visual fields on each coverslip were evaluated and the mean number of γH₂AX foci/cell was compared between treatment groups. All experiments were performed in triplicates and repeated three times (n = 3).

Semiquantitative polymerase chain reaction of GSTα

Total RNA was prepared from U87 cells, astrocytes, and fibroblasts using the RNeasy mini kit according to the manufacturer’s instructions (Qiagen, Hilden, Germany). Complementary DNA (cDNA) was generated from 1 µg of total RNA in a volume of 30 µL using M-MuLV reverse transcriptase (Thermo Fisher Scientific, Waltham, USA) and 100 pmol of hexameric primers. Semiquantitative polymerase chain reaction (PCR) was performed with Taq polymerase and buffers provided from Thermo Fisher Scientific. Specific primers were used for GSTα4 (5′-AGTTGGTACAGACCCGAAGCA-3′ forward; 5′-AGTTCCAGCATCCAGTGTCC-3′ reverse) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; 5′-GGCCTCCAAGGAGTAAGACC-3′ forward; 5′-AGGGGTCTACATGGCAACTG-3′ reverse) as endogenous control. A first cycle of 3 min at 95°C was followed by 30 s at 95°C, 30 s at 56°C, and 40 s at 72°C for 40 cycles and finished with 72°C for 10 min.

5′-Bromodeoxyuridine incorporation assay

The 5-bromo-2′-deoxyuridine (BrdU) incorporation assay (Roche Diagnostics, Mannheim, Germany) was performed according to the manufacturer’s instructions 4–72 h after treatment. The percentage of cells exhibiting genomic BrdU incorporation was assessed using fluorescence microscopy (Zeiss Axiovert 135, Oberkochen, Germany). Three coverslips were evaluated in each group with three repetitions.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay

Cell viability was determined 4–72 h after treatment with JS-K (0–15 µM) and/or RT (3 × 2 Gy) using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assay as previously described. ¹⁴ JS-K concentrations ranged from 0–15 µM. The amount of viable cells was determined photometrically at 570 nm (reference wavelength 630 nm).

Nuclear translocation and accumulation of p53

Nuclear translocation and accumulation of p53 in U87 cells indicating the induction of apoptotic and/or necrotic cell death were analyzed using the p53 Cell-Based Activation/Translocation Assay Kit (Cayman, Ann Arbor, USA) according to the manufacturers’ instructions after 2 h of JS-K treatment (0–5 µM) and or RT (2 × 2 Gy). Nutlin 3 was used as positive control (1:200). p53 translocation was assessed by fluorescence microscopy.

Immunoblotting

The amount of uncleaved PARP1 was determined by western blot 4–72 h after treatment with 5 µM JS-K and/or 3 × 2 Gy. Cells were homogenized with lysis buffer (Biomol, Hamburg, Germany). The protein concentration was determined according to Bradford. ¹⁶ Untreated cells were used as control. An amount of 30 µg of total protein was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes. Blots were blocked in 5% fat-free milk powder and incubated with primary antibodies against uncleaved PARP1 (1:2000; Cell Signaling Technology) overnight at 4°C. After incubation with secondary antibodies, signals were visualized by enhanced chemoluminescence reagents (PerkinElmer, Waltham, USA). GAPDH (1:10,000; Abcam, Cambridge, UK) was used as reference protein. The signal ratio uncleaved PARP1/GAPDH was quantified densitometrically using ImageJ software.

Microscopic analysis of signs of mitotic catastrophe

Morphological changes of U87 cells were investigated using DAPI staining to quantify the amount of cells undergoing mitotic catastrophe after treatment wit JS-K (0–7.5 µM) and/or repetitive irradiation 4–72 h after the end of treatment. Cells with nuclei that showed multiple fragmentations and a nuclear diameter >20 µm were considered positive for MC according to the criteria described by Firat et al. ¹⁷ Evaluation of cell nuclei was performed using fluorescence microscopy (Zeiss Axiovert 135). For each point in time and each treatment group, three culture coverslips were evaluated. All experiments were repeated three times. The percentage of cell nuclei showing MC criteria per coverslip was calculated for all treatment groups.

In vivo xenograft experiments

Institutional guidelines for animal welfare and experimental conduct were followed for all animal experiments which were approved by the Institutional Animal Care and Use Committee and the Regional Administrative Authority under protocol G13/041. The animals were housed in individually ventilated cages (IVC) and received food and water ad libitum; 6-week-old immunodeficient nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice (Charles River Laboratories, Germany) underwent stereotactic implantation of 1 × 10⁵ luciferase-transfected U87 glioma cells into the right striatum under isoflurane anesthesia. Mice were randomly assigned to four groups: (1) DMSO-Pluronic Micelles PEG123 control (n = 9), (2) JS-K 6 nM/g body weight (BW; n = 7), (3) irradiation (n = 9), and (4) JS-K in combination with irradiation (n = 12). For the in vivo experiments, JS-K was formulated in Pluronic Micelles P123 (BASF, USA). Treatment was started 29 days after tumor inoculation when measurable tumors were visible on magnetic resonance imaging (MRI). Cranial irradiation was carried out with a Rad Source RS2000 γ-irradiator with lead shielding of the body using a dose of 2 Gy/day, 5 days per week to reach a total dose of 40 Gy (4 weeks). JS-K was applied via tail vein injection at a concentration of 6 nM in DMSO-PEG123/g BW daily. In case of combined treatment, JS-K was applied directly before the irradiation. For treatment, animals were anesthetized by Medetomidine/Atipamezole 0.25 µL/g s.c. injection. Tumor size was determined once a week by MRI. Images were acquired on a 7 T small bore scanner (BioSpec 70/20 USR; Bruker, Karlsruhe, Germany) using a mouse quadrature volume coil. Each animal was placed on a mouse bed (Bruker) and anesthetized with 1.5% isoflurane at a rate of 0.8 L/min of air during MRI, and breathing and temperature were monitored and kept as stable as possible. Images were acquired using Paravision 5.1 and a cryogenically cooled mouse brain quadrature coil (Bruker). T2-weighted multislice relaxation enhancement (RARE) scans (repetition time (TR)/echo time (TE): 4270/10.6 ms, in-plane resolution 59 × 59 mm) were used to localize the tumor and to perform volumetric measurements. Tumor volumes were calculated using the freely available MIPAV software tool (NIH, Bethesda, USA) by drawing volumes of interest (VOI) on contiguous slices. Mice were euthanized according to predefined criteria.

Statistical analysis

All experiments were done in triplicates and repeated three times. All data are presented as mean ± standard deviation (SD) and analyzed by two-tailed Student’s t-test with unequal variance. Differences were considered statistically significant at *p < 0.05, **p < 0.01, and *** p < 0.001. Survival data were analyzed by Kaplan–Meier survival statistics.

Results

U87 cells overexpress the GST isoform α

As JS-K is metabolized by the GST isoforms α and π with decreasing affinity, ⁹ the expression was assessed in U87 cells and non-neoplastic cells using immunocytochemistry. U87 cells showed a stronger expression of the GST isoform α compared to non-neoplastic astrocytes and fibroblasts, thus rendering glioma cells good candidates for a cell-type-specific NO donor therapy (Figure 1(a) and (b)). Overexpression of GSTπ has been described previously by our group. ¹³

Figure 1.

(a) GSTα is overexpressed in U87 cells (lower panels) compared to astrocytes (upper panels) and fibroblasts (middle panels); immunocytochemistry for GSTα in cell nuclei counterstained with DAPI (20× magnification). (b) Semiquantitative PCR of GSTα in astrocytes, fibroblasts, and U87 cells. (c) JS-K reduces the proliferation of U87 cells and enhances the antiproliferative effect of RT (upper panel: JS-K monotherapy, lower panels: JS-K + RT). Immunocytochemistry (48 h): BrdU-positive U87 cells (green) and cell nuclei counterstained with DAPI (20x magnification). (d) Quantification of BrdU incorporation assays (4–72 h, controls set to 100%, n = 3).

JS-K and irradiation reduce proliferation of U87 cells

The time-dependent response of U87 cells to single radiation doses of 2, 5, and 10 Gy was initially tested showing no biological effect using 2 and 5 Gy and only a minor reduction of cell proliferation after 10 Gy at late time points after treatment, thus confirming their strong radioresistance (data not shown). Therefore, repetitive irradiation with 2 Gy at 24 h intervals (3 × 2 Gy) was chosen for these experiments as a surrogate for the treatment schedule used in humans.

Repetitive irradiation with 3 × 2 Gy reduced the number of proliferating cells significantly compared to untreated cells 4 h after the end of treatment (Figure 1(b) and (c)). JS-K monotherapy resulted in a dose-dependent decrease of proliferation activity 4 h after the end of treatment with a maximum decrease when using 7.5 µM JS-K. Combination of JS-K and repetitive irradiation resulted in a significantly stronger antiproliferative effect compared to the corresponding monotherapies consistent with a JS-K-mediated radiosensitization. This effect could be observed at all JS-K concentrations (2.5–7.5 µM) and was evident throughout the experimental period (4–72 h).

JS-K enhances cytotoxic effect of irradiation in U87 cells

As a monotherapy, irradiation did not lower cell viability 4–72 h after the last treatment compared to the untreated controls (Figure 2(a)). In combination with single-dose or repetitive irradiation, JS-K acted as a strong radiosensitizer. The combination of single-dose irradiation of 2 Gy, which had no effect on cell viability at all, with 10 µM JS-K resulted in a significant reduction of cell viability (data not shown). A significant reduction of cell viability after combined treatment could be detected 4–48 h after the end of treatment using 5 and 10 µM JS-K (Figure 2(b)). Monotherapy with 15 µM JS-K leads to an almost complete loss of cell viability; hence, no additional radiosensitizing effect could be detected.

Figure 2.

JS-K alone or in combination with RT (3 × 2 Gy) reduces cell viability of U87 cells. (a) Cell viability of irradiated cells normalized to untreated controls (4–72 h, n = 5). 4 h control was set to 100% (±SD). (b) JS-K monotherapy and combination with RT (4–24 h, n = 3). Treatment groups as well as irradiation control (gray) are normalized to untreated control (black) set to 100% (±SD of triplicate). (*p ⩽ 0.05, **p ⩽ 0.01 indicate significance of combined treatment to JS-K alone.)

JS-K and irradiation significantly increase the number of DNA double-strand breaks in U87 cells

Irradiation or NO induces cell death by induction of DNA double-strand breaks (DSBs) which can be visualized by γH₂AX immunocytochemistry (Figure 3(a)). In radioresistant U87 cells, no significant increase in the number of γH₂AX foci could be detected in U87 cells that received 3 × 2 Gy compared to untreated controls 24 h after end of treatment (Figure 3(b)). JS-K monotherapy led to a dose-dependent increase of γH₂AX foci, with up to 11.73 foci per cell in the group that received 15 µM JS-K compared to 1.19 foci per cell in the control group (985% of DSB/cell in the untreated control, Ctr set to 100%; p < 0.001). The radiosensitizing effect of JS-K could be confirmed by a significantly higher number of γH₂AX foci compared to the corresponding monotherapies: a combination of 10 µM JS-K and irradiation resulted in 12.22 ± 2.44 foci per cell compared to 9.57 ± 1.53 foci in the JS-K monotreatment group (p = 0.014). Treatment with 15 µM JS-K and 3 × 2 Gy resulted in 18.25 ± 2.14 γH₂AX foci per cell compared to 11.73 ± 2.08 foci per cell in cells that received 15 µM JS-K and no irradiation (p < 0.01).

Figure 3.

(a) Immunocytochemistry for γH₂AX foci (red) in U87 cell nuclei after the treatment with 10 µM JS-K compared to untreated cells (48 h), and cell nuclei was counterstained with DAPI (20× magnification). (b) Quantification of γH₂AX foci compared to untreated controls (24 h, n = 3). JS-K alone and in combination with RT (2 × 2 Gy) induces DNA damage repair and cell death mechanisms. (c) Nuclear accumulation of p53 after JS-K (1–5 µM) and RT (2 × 2 Gy, 48 h); arrows indicate nuclear p53 (20× magnification). (d) PARP1 after RT, JS-K, and combined treatment (4–72 h). (e) Ratio of PARP1/GAPDH over time (n = 3). (f) Morphological changes indicating mitotic catastrophe (MC). Arrow indicates giant, multifragmented cell nuclei counterstained with DAPI (20× magnification). (g) Rate of cells expressing morphological features of MC compared to untreated controls (n = 3).

Increased nuclear translocation and accumulation of p53

In order to avoid covering of a synergistic effect of JS-K and repetitive irradiation by the strong antitumor effect of JS-K itself, low concentrations of JS-K (1, 2.5, and 5 µM) were chosen.

JS-K monotherapy (2.5 and 5 µM) induced a strong nuclear accumulation of p53 compared to the untreated controls as early as 2 h after treatment (Figure 3(c)); 48 h after JS-K monotherapy (2.5 and 5 µM) and after combined treatment, cell nuclei showed strong accumulation of p53 and frequently presented a multifragmented morphology. This result indicates that JS-K-mediated p53 accumulation is an early event in NO-induced cell death in U87 cells.

Levels of uncleaved PARP1 decrease in a time-dependent manner after treatment with JS-K and/or irradiation

Caspase 3-mediated cleavage of poly(adenosine diphosphate (ADP)-ribose) polymerase 1(PARP1), an enzyme involved in DNA-repair, is a crucial event in the course of cell death and occurs in late stages of apoptosis. It may also be cleaved into differential fragments during necrosis without involvement of caspases. ¹⁸ A time-dependent decrease in the level of uncleaved PARP1 could be observed in all treatment groups: repetitive irradiation alone resulted in a decrease of uncleaved PARP1 48 h (p < 0.05) and 72 h (p < 0.01) after the end of treatment (Figure 3(d)). JS-K treatment (5 µM) alone led to a reduction of uncleaved PARP1 24 h (p < 0.05), 48 h (p < 0.05), and 72 h (p < 0.01) after the end of treatment (Figure 3(e)). The observed effect was strongest and started earliest in the group that received combined treatment: 4 h after end of the treatment, expression of PARP1 was decreased to 77.33% of the level in untreated cells (p < 0.05). An even stronger decrease could be detected 24 h (48.53%; p < 0.05), 48 h (43.67%; p < 0.05), and 72 h (15.97%; p < 0.001) after the end of the treatment. Taken together, JS-K monotherapy, repetitive irradiation, and combined treatment result in a time-dependent decrease of uncleaved PARP1 suggesting cleavage during cell death by apoptosis and necrosis.

Signs of MC increase by JS-K treatment in combination with repetitive irradiation

Treatment-dependent changes in nuclear morphology indicating MC, that is, exhibition of multiple fragmentation and nuclear diameter >20 µm were analyzed by DAPI staining of the cell nuclei (Figure 3(f)). ¹⁷ Repetitive irradiation (3 × 2 Gy) showed no significant increase in MC features compared to the untreated controls at any point in time (Figure 3(g)). JS-K monotherapy resulted in a time- and dose-dependent increase in MC features reaching its maximum when using 7.5 µM JS-K 24 h after the end of the treatment: 15% of the treated cells showed signs of MC compared to cells with spontaneous MC in the untreated control group (<0.5%). This effect could be observed at all assessed points in time and was most prominent after 24 h. Combination of repetitive irradiation with any of the used JS-K concentrations resulted in a significant increase in MC compared to JS-K monotherapy.

In vivo application of JS-K alone and in combination with RT does not improve overall survival in an orthotopic U87 xenograft model

Daily tail vein injections of JS-K using a concentration of 6 nM/g BW dissolved in DMSO-Pluronic Micelles PEG123 were well tolerated by the animals without inducing any systemic side effects. Radiation with a daily dose of 2 Gy was equally well tolerated by the mice. Higher daily RT doses (3 and 5 Gy) were tested previously but were poorly tolerated by the animals (unpublished data).

MRI data analysis revealed no significant difference in average tumor volume over time between the three treatment groups and the controls (Figure 4(a)). Figure 4(b) shows that neither JS-K monotherapy (42.6 ± 4.2 days, p = 0.80) nor the combination of JS-K and irradiation (41.3 ± 5 days, p = 0.65) led to a prolonged survival compared to the PEG123 controls (42.1 ± 3.2 days). All tumors grew considerably within the study period. U87 tumors appeared compact, homogeneous, and well-delineated (Figure 4(c)).

Figure 4.

The radiosensitizing effect of JS-K was studied using an orthotopic U87 xenograft mouse model. (a) Fold change of tumor volume measured by MRI analysis for all experimental groups. Groups were normalized to the initial tumor volumes at the beginning of the treatment on day 29 (*p < 0.05). (b) Kaplan–Meier survival analysis of mice treated with RT (2 Gy/day, 5 days/week) and/or JS-K (6 nM/g BW) and of untreated controls. Arrows on the x-axis indicate dates of MRI study. (c) MRI images of U87 tumors in one representative animal of each group over the treatment period (days 0–15). Arrows indicate the tumor border.

Taken together, although well tolerated, the strong in vitro radiosensitizing effect of JS-K could not be reproduced in vivo.

Discussion

Increasing the efficacy of RT using targeted NO delivery is the experimental strategy in this study to overcome the strong radioresistance in GBM. Release of NO by the diazeniumdiolate JS-K has potent antitumor effects and sensitizes U87 cells to irradiation in vitro by inhibition of proliferation and induction of DNA DSBs leading to MC as the dominant form of cell death. U87 cells were selected as an established model system in glioma research with known molecular features and reproducible tumor growth in vivo and because they express the tp53 wt gene.^19,20 GBM patients with inactive p53 benefit more from RT than patients with wild-type p53. ²¹ This is probably due to induced cell-cycle arrest and efficient DNA-repair response elicited by DNA damage in the latter. The molecular influence of JS-K on primary cell lines as well as established U87 and LN229 cells was shown by our group in a previous study.^22,23 We demonstrated the dose-dependent cytotoxic effect of JS-K leading to necrotic cell death by MC with profound metabolic changes. Previous experiments of our group also showed that single irradiation doses up to 10 Gy or repetitive doses of 2 Gy did not reduce the viability of U87 cells in vitro as shown by MTT viability assays. These results confirm the high radioresistance of U87 cells and emphasize these cells as good candidates to analyze putative radiosensitizing features of NO and the molecular basis of its antitumor effect. The antitumor effect of JS-K could be confirmed in other glioma cell lines in vitro.^14,22,23

Multiple reasons underlie the strong therapy resistance of GBM cells: these include the accumulation of mutations that result in the increased expression of antiapoptotic and tumorigenic factors, deregulation or block of apoptosis inducing pathways, upregulation of enzymes essential for DNA repair, that is, O(6)-methylguanine-DNA methyltransferase (MGMT), or cell detoxification, as GSTs, and hypoxic conditions within the tumor area.

Different strategies to overcome resistance to treatment or an insufficient drug delivery into the tumor area have been evaluated by in vitro and in vivo experiments. One of them, the application of NO, is a promising approach to increase the efficiency of the multimodal therapy. Antitumor effects of NO include induction of DNA DSBs, direct transcriptional and translational changes, and reversible enzyme inactivation by post-translational protein modifications due to S-nitrosylation or nitration or simply the improvement of the tumor blood flow and vascular permeability.^24,25 However, direct application of NO is limited by its high chemical reactivity and eventual detrimental vasodilatory side effects. ²⁶ The use of enzyme-activated NO donor drugs such as JS-K might offer a possibility to ensure a cell-type-specific and efficient release of NO inside the tumor cells reducing detrimental effects in the surrounding tissue. JS-K releases NO upon enzymatic activation by GSTs. The overexpression of GST isoforms is a common event in different malignancies including GBM and is known to promote chemoresistance.^27,28 We could demonstrate elevated intracellular levels of GST isoform α in U87 cells compared to non-neoplastic astrocytes and fibroblasts and a stronger cytotoxic effect in tumor cells. These findings were in line with previously published data of our group emphasizing the overexpression of GSTs in different primary GBM cell lines. ¹⁴

GST-dependent NO release from JS-K exerts a cytotoxic and antiproliferative effects in U87 cells in a time- and dose-dependent manner.^14,22

In contrast, irradiation has, if any, only mild effects on proliferation and cell viability in vitro. Combination of repetitive irradiation with JS-K concentrations up to 7.5 µM intensified and prolonged this inhibitory effect. Combining the NO prodrug JS-K with repetitive irradiation reduced the cell viability significantly stronger in a dose- and time-dependent manner than JS-K or irradiation alone. Radiosensitizing effects mediated by NO have been described for numerous malignant tumors including GBM so far. ⁸ Nevertheless, this study is the first to report sensitization to repetitive irradiation by the enzyme-activated NO donor drug JS-K in GBM. The mechanisms of NO-mediated radiosensitization are diverse and include activation of proapoptotic proteins, ²⁹ mimicking an oxygen-like structure and the creation of highly reactive oxygen species. ²⁴

We assume that the radiosensitizing effect of JS-K is probably based on the induction of DNA DSB as a direct effect of the NO release. The number of DNA DSB induced by repetitive irradiation is low and only leads to a temporary arrest of the cell cycle in order to repair this damage.^30–32 If cells are treated with repetitive irradiation in combination with JS-K, an effective DNA damage repair seems no longer possible, thus cells undergo cell death. The number of induced DNA DSB cells and the inactivation of enzymes related to repair mechanisms such as PARP1 ultimately lead to increased cell death. NO derivates, that is, peroxynitrite, probably contribute to the damage of cell compounds resulting in the inhibition of proliferation and/or cell death.

The mechanism of cell death induced by JS-K is still under discussion. Liu et al. ³³ described a JS-K-mediated upregulation of apoptotic proteins including caspases and tumor necrosis factor α (TNFα) with concomitant downregulation of differentiation markers such as CD14 in HL60 leukemia cells. Activation of p53, its translocation, and nuclear accumulation also suggest the initiation of apoptotic pathways. Synergistic activation of the tumor suppressor p53, a key player in the apoptotic cascade, after inducible nitric oxide synthase (iNOS) transfection and irradiation, has been identified as a major event in cell death of colorectal carcinoma cells. ²⁹

Increased cleavage of PARP1 occurs in late stages of apoptosis as well as in necrosis as described by Gobeil et al. ¹⁸ and is evoked by JS-K treatment in a multiple myeloma model. ¹¹ Our results showed a prominent cleavage of PARP1 enforced by combined treatment of U87 cells with JS-K and repetitive irradiation in vitro. Although examination of morphological features of the cells revealed characteristic features of MC. MC has been described as the major mechanism of cell death induced by irradiation in solid tumors. ³⁴ Furthermore, fluorescence-activated cell sorting (FACS) analysis (propidium iodide (PI)/Annexin V) of U87 cells treated with JS-K indicates that cells undergoing MC rather belong to necrotic cell populations than to subpopulations undergoing apoptosis. ²² MC as the dominant mechanism of cell death has also been described in U87 cells treated with an integrin-linked kinase (ILK) inhibitor and irradiation. ³⁵ As the cell death events arise relatively late after treatment, there might be a dual mechanism of cell death induction: the induction of apoptosis shortly after treatment and the induction of necrosis at later points in time, that is, MC. Furthermore, the genetic background might influence the cellular decision to undergo either necrosis or apoptosis in response to molecular damage to a great extent.

As mentioned before, GBM is among the most aggressive tumors in humans, mainly due to its strong therapy resistance and privileged location in the brain.^36,37 Application of NO as adjuvant via NO prodrugs such as JS-K might improve the outcome by sensitizing tumor cells to chemotherapeutic drugs and irradiation and second by improving their transport across the blood–tumor barrier. Our in vivo studies could not verify the radiosensitizing effect of JS-K for repetitive irradiation in an U87 orthotopic xenograft mouse model. In contrast to our previous studies in orthotopic intracranial and flank-tumor rat models which demonstrated a moderate effect of NO donor compounds (JS-K, PABA/NO) on tumor growth, cell proliferation, survival, and tumor physiology, no significant in vivo antitumor effect could be observed in this orthotopic U87 xenograft mouse model.^13,14,38,39 Several facts might contribute to this result. First, in this study, JS-K was applied intravenously (i.v.) after packaging in PEG123 micelles to increase its stability in the blood circulation. Dynamic contrast-enhanced (DCE) MRI studies of our group showed that pegylated JS-K reached the blood–tumor barrier after i.v. delivery and enhanced perfusion and permeability in U87 tumors, reduced the mitotic activity in the tumor as shown by Ki67 immunohistochemistry, but failed to influence tumor growth. ³⁸ In spite of being packaged in PEG123, JS-K might be partly hydrolyzed in the presence of glutathione upon intravascular delivery reducing the concentration of JS-K which reaches the blood–tumor barrier. Second, JS-K might be scavenged in other organs, especially during the first-pass effect through the liver exhibiting high GST expression, thus reducing the concentration of JS-K reaching the central nervous system (CNS). Although we could confirm biological activity of NO at the blood–tumor barrier and a reduction of the mitotic index, the intracellular NO doses which were generated after intracellular activation by GST might be too low to induce a potent antitumor or radiosensitizing effect. Analysis by MRI clearly showed that tumor growth of PEG123-treated controls and JS-K-treated or irradiated animals did not differ significantly, thus the overall survival was not improved by JS-K treatment and/or irradiation. Irradiated tumors tended to have a smaller tumor volume, but this did not result in a significant extension of survival. Addition of JS-K to RT did not have any impact on tumor volume or survival. Treatment was initiated 4 weeks after tumor inoculation in order to monitor measurable tumor burden by MRI. Only animals with measurable disease were included in the study. The large tumor burden at the initiation of treatment and the exponential tumor growth over the treatment period might contribute to the lack of tumor control. Studying this new treatment approach early after tumor inoculation may produce different results.

In conclusion, based on the encouraging in vitro results, we assume that by overcoming the obstacles of JS-K stability and efficient delivery of the NO prodrug into the tumor, JS-K might be a promising experimental adjuvant in the treatment of malignant gliomas upon further investigation.