Improved Killing of Human High-Grade Glioma Cells by Combining Ionizing Radiation with Oncolytic Parvovirus H-1 Infection

Purpose. To elucidate the influence of ionizing radiation (IR) on the oncolytic activity of Parvovirus H-1 (H-1PV) in human high-grade glioma cells. Methods. Short term cultures of human high-grade gliomas were irradiated at different doses and infected with H-1PV. Cell viability was assessed by determining relative numbers of surviving cells. Replication of H-1PV was measured by RT-PCR of viral RNA, fluorescence-activated cell sorter (FACS) analysis and the synthesis of infectious virus particles. To identify a possible mechanism for radiation induced change in the oncolytic activity of H-1PV we performed cell cycle analyses. Results. Previous irradiation rendered glioma cells fully permissive to H-1PV infection. Irradiation 24 hours prior to H-1PV infection led to increased cell killing most notably in radioresistant glioma cells. Intracellular levels of NS-1, the main effector of H-1PV induced cytotoxicity, were elevated after irradiation. S-phase levels were increased one day after irradiation improving S-phase dependent viral replication and cytotoxicity. Conclusion. This study demonstrates intact susceptibility of previously irradiated glioma-cells for H-1PV induced oncolysis. The combination of ionizing radiation followed by H-1PV infection increased viral cytotoxicity, especially in radioresistant gliomas. These findings support the ongoing development of a clinical trial of H-1PV in patients with recurrent glioblastomas.


Introduction
Malignant gliomas have remained a malignancy with an extremely unfortunate prognosis [1,2]. Recent improvements of standard therapies including, when feasible, surgical resection followed by radio-chemotherapy have only extended the 50% survival from 12 months to 16 months [3][4][5]. Long-term survival is rare, only 5% of patients are alive after 5 years. As a consequence, alternative therapies have to be investigated.
One new strategy is the use of replication competent oncolytic viruses that specifically target and destroy tumor cells while leaving normal cells intact. A number of candidate oncolytic viruses for glioma therapy are currently under investigation including genetically modified Herpesviruses [6], Adenoviruses [7], or Poliovirus [8], and wildtype viruses such as Reovirus [9], Vesicular-stomatitis virus [10], and Measles virus [11]. We previously reported the efficient killing of glioma cells of human and rat origin by Parvovirus H-1 (H-1PV), a single stranded nonenveloped DNA virus of 5.1 kb. H-1PV induced lytic infection of glioma cells even when the cells were resistant to agents inducing apoptosis [12,13]. In animal experiments, H-1PV infection of rats bearing large intracranial gliomas led to tumor regression and prolonged survival (Geletneky et al., accepted, 2010). The natural host of H-1PV is the rat; however, the virus can efficiently infect cells of other species including humans. In contrast to some other wildtype viruses under investigation for glioma-treatment, H-1PV does not cause any pathology in rodents or humans [14].
Radiation therapy prolongs survival in patients with malignant gliomas and is used as standard treatment of primary high-grade glial tumors [15]. However, as high grade gliomas are resistant to radiation therapy and a clear dose-limitation exists due to cytotoxic effects on the surrounding brain tissue, this treatment modality is not curative and strategies to improve radiation efficiency are under investigation. The combination of radiation therapy with the oral alkylating agent temozolomide has already proved to be superior to either therapy alone and has become the standard of care for the majority of patients with newly diagnosed glioblastoma multiforme [3,5]. Recent studies of the role of radiation therapy for recurrent gliomas that were already irradiated as part of the primary treatment demonstrated some effect when radiation was applied as a boost to smaller tumor regions [4]. Therefore, depending on the individual situation of the patient, radiation therapy can also be an option for recurrent tumors, but methods to augment the limited therapeutic effect would clearly be beneficial.
Clinical trials with different oncolytic viruses were able to demonstrate the safety of this novel therapeutic approach; however, the positive therapeutic effects were restricted to individual patients [16]. Therefore, the combination of oncolytic viruses with standard therapeutics has become one important focus to improve viral cytotoxicity. Radiation therapy is a part of therapeutic protocols for the majority of malignancies, and an enhanced effect of the oncolytic activity of viruses by radiation therapy could be observed for tumor cells of various histology [17]. This is of particular interest for gliomas, as both treatments, radiation therapy and virotherapy, are primarily designed as regional therapies. For glioma cells, the oncolytic effect of Herpesvirus [18], Adenoviruses [19], Reovirus [20], and Measles virus [21] was shown to be enhanced by IR.
The aim of this study was to assess the influence of IR on the oncolytic activity of H-1PV in glioma cells. The possible interaction of IR with H-1PV oncolytic virotherapy could be twofold: (i) as the use of an oncolytic virus in glioma patients would preferably include the treatment of recurrent tumors originating from previously irradiated remaining tumor cells, it has to be shown whether previous radiation therapy would interfere with viral oncolysis or replication in pretreated gliomas and (ii) administration of radiation therapy together with H-1PV oncolytic virotherapy could lead to improved efficacy of either treatment alone. We therefore investigated the treatment of early-passage glioma cells with IR before or after infection with H-1PV and assessed cytotoxicity, viral replication, and treatment-induced changes of the cell cycle. These findings are important to define patient populations for a clinical trial of oncolytic virotherapy of malignant gliomas and to possibly use H-1PV to increase radiation efficacy in this dismal tumor entity.

Methods and Materials
2.1. Cell Culture. Human short term cultures derived from histologically confirmed glioblastomas (NCH-82, NCH-89, NCH-307) and a gliosarcoma (NCH-37) were established and characterized at the Department of Neurosurgery, Heidelberg, Germany as described previously [12]. NCH-307 is a recurrent glioblastoma cell line that had been irradiated in vivo prior to resection of the recurrent tumor. NCH-37, NCH-89, and NCH-307 were tested at low-passage numbers <30; NCH-82 was tested at a passage number of 100. The ethylnitrosourea-induced rat glioma cell line RG2 was previously shown to be highly susceptibly for H-1PV [12] and was used for virus-titration experiments. All cells were grown in DEME (Sigma-Aldrich, Steinheim, Germany) supplemented with 10% FCS (BiochromKG, Berlin, Germany) and 1% antibiotics (penicillin/streptomycin; Gibca, Invitrogen Corporation, Karlsruhe, Germany) in a 5% CO 2 humidified atmosphere at 37 • C. (IR). Radiation of cell cultures was performed at room temperature in a linear accelerator (Siemens Mevatron KD2, 6-MV photons) at doses of 5 Gy,10 Gy, or 20 Gy as indicated for the respective experiment (dose rate: 3 Gy/min; distance from source to flask: 95 cm). Control cells were transported to the accelerator but not exposed to IR (0 Gy). NCH-307 recurrent glioblastoma cells were not reexposed to IR in vitro.

H-1 Virus Production and Infection.
H-1PV was propagated in human NB324K cells, and purified as described previously [22]. Monolayers of all glioma cell cultures were infected under standard conditions: cells were inoculated with H-1PV diluted in PBS at a multiplicity of infection (MOI) of 5 plaque forming units (PFU) per cell or 100 PFU/cell as indicated for the respective experiment. After 60 minutes, virus suspensions were removed, cells were washed with medium, and cultures were allowed to grow. The corresponding mock-infected cultures were subjected to the same procedure using virus-free PBS instead of virus suspensions.
In vivo irradiated NCH-307 cells were seeded at 3 × 10 4 cells/well into 12-well dishes and infected (MOI = 5 PFU/cell, 100 PFU/cell or mock-infection) 24 hours postseeding. The MOI of H-1PV was calculated based on counts of living cells immediately prior to infection. Cells were harvested 3 days after H-1PV infection/mock-infection and counted with an electronic counter (Casy, Schaerfe System, Reutlingen, Germany) in triplicate and results were confirmed with the typan blue exclusion test as described previously [12].

Statistical
Analysis. The numbers of living cells were estimated as means and standard deviations of three independent assays. Cell viability (in % + /− standard error) was defined as the number of living treated cells over the number of living untreated cells multiplied by 100; standard error was calculated using the Gaussian law of error propagation. Statistical analysis was performed using a two-way ANOVA Journal of Biomedicine and Biotechnology 3 with virotherapy and IR as independent factors. Comparative analyses between groups were performed using post-hoc analysis. The SPSS software package (SPSS Inc., Chicago, IL) was used to perform statistical analysis.  In short, cells were fixed with 4% paraformaldehyde and 100% methanol and permeabilized with 0.1% Triton-X-100 (Sigma-Aldrich, Taufkirchen, Germany). To identify NS-1 protein, probes were blocked with fetal calf serum and incubated on ice for 30 minutes with a polyclonal rabbit-anti-NS-1 antibody (SP8, courtesy of N. Salome, DKFZ, Heidelberg, Germany) in a concentration of 1: 25. The FITC-conjugated secondary goat-anti-rabbitantibody (Jackson ImmunoResearch, Suffolk, UK) was incubated in a 1:250 dilution for 20 minutes on ice. Probes were analyzed for intracellular NS-1 content by measuring of fluorescence intensity, using a cytometer (FACScan flow cytometer, Becton Dickinson, Heidelberg, Germany) at an excitation wavelength of 525 nm. The data were analyzed with the aid of a software program (FlowJo, Tree Star, Olten, Switzerland) with dead cells gated out using pulse processing. A cell was determined as NS-1-positive when its fluorescence intensity (FL1-H) was greater than a certain threshold value of 5% of false positive mock infected cells.

Susceptibility of Irradiated Glioma Cells to Infection with H-1PV.
As previous radiation therapy of tumor cells induces genetic alterations that could interfere with the susceptibility and efficiency of H-1PV infection, we infected glioma cells with a delay of 9 days after IR (late infection) when the cells had reentered the cell cycle. The goal of this experiment was to mimic and evaluate the possibility of H-1PV virotherapy of recurrent tumors after completion of radiation therapy as part of the initial standard treatment. In addition to testing tumor cells from primary gliomas (NCH-37, NCH-82, NCH-89), we also included NCH-307 cells that were established from a recurrent glioblastoma that had been irradiated several months prior to secondary surgical resection.

Combination of IR and H-1PV Infection.
In initial experiments, the effect of radiation therapy or H-1PV infection alone was examined prior to testing combination  Figure 2. Comparative analyses between groups revealed the following: in all glioma cell cultures (NCH-37, NCH-82, NCH-89), combination of H-1PV infection 1 day after IR was significantly (P < .05) more effective than IR alone (Figure 2). Compared with H-1PV infection alone, combination treatment was significantly (P < .05) more effective after previous IR with 5 Gy, 10 Gy, or 20 Gy in NCH-37 cells and after previous IR with 20 Gy in NCH-82 cells. When the order of treatments was reversed and H-1PV infection was performed 24 hours prior to IR, combination treatment only led to significantly (P < .05) improved cell killing in NCH-37 when compared to IR alone, but not when compared to H-1PV infection alone or in the other cell lines tested.

Long-Term Effects of IR Followed by H-1PV Infection.
Even though high-dose radiation of NCH-37, NCH-82, and NCH-89 cells with 20 Gy or infection with H-1PV was highly cytotoxic, approximately 2 weeks after single treatment with IR or H-1PV alone, all cell lines resumed to proliferate from surviving clones, albeit at a much reduced rate (Table 1). Thus, neither IR nor H-1PV infection alone was able to eradicate all tumor cells. In contrast, when glioma cell cultures were treated with the combination of IR (20 Gy) and H-1PV infection (MOI = 5 PFU/cell) 24 hours after IR, no surviving tumor cells could be observed on day 21 p.i. or at later time points after treatment in any of the tested cell cultures (NCH-37, NCH-82, NCH-89) indicating long-term efficiency of combination treatment (Table 1 and Figure 3). The experiment was confirmed in triplicate in all cell cultures.

Replication of H-1PV in Human Glioma Cells after IR.
Replication of H-1PV in infected glioma cells was tested (i) by the presence of viral NS-1-specific RNA by RT-PCR, (ii) by the detection of the viral protein NS-1 as the main effector of parvoviral cytotoxicity [14] by FACS analysis, and (iii) by the release of infectious viral particles into the supernatant of cell cultures.

Discussion
Oncolytic virotherapy is a promising new approach for the treatment of a variety of malignancies including malignant brain tumors. In early phase clinical trials, intracerebral infection of patients with oncolytic viruses of different genera was well tolerated. However, only in a few patients tumor regression and a prolongation of progression-free survival could be demonstrated. This led to numerous attempts not only to further improve the oncolytic activity of viruses but also to investigate the combination treatment of viral infection of tumor cells with established conventional therapies. Radiation therapy is a standard treatment of patients with high-grade malignant gliomas. It is administered locally to the tumor region and the surrounding brain tissue. Recurrences occur in 70 to 80% of cases within 2 cm of the primary tumor site and will therefore develop from cells that were already hit by a radiation dose of up to 60 Gy. As radiation therapy is known to induce long-term changes in cellular genomes, this could potentially lead to an altered effect of virus infection compared with unirradiated primary glioma cells. In addition to other studies, we therefore specifically investigated the oncolytic potential of H-1PV in glioma cells that grew from irradiated clones.
A cell culture established from a recurrent glioma that was irradiated with 56 Gy during the initial treatment of the patient, and that had its origin in the radiation field, was fully permissive to H-1PV infection and cell killing was dosedependent. This finding was confirmed in primary glioma cultures that were irradiated in vitro with a sublethal dose, allowed to regrow, and infected with H-1PV in a time interval of several days. The intact susceptibility of glioma cells to H-1PV infection after IR is of clinical significance as patients with recurrent gliomas who face an even worse prognosis with oftentimes less therapeutic options are prime candidates for experimental therapies. As a consequence, the group of patients with recurrent gliomas is usually the main patient population of early clinical trials of oncolytic virotherapy. However, to our knowledge the response of previously irradiated glioma cells to the oncolytic infection has never been specifically addressed for other oncolytic viruses.
When H-1PV infection was performed early after IR, our data show improved killing of glioma cells, most pronounced in the most radioresistant cell line tested. These findings are in line with studies conducted with other oncolytic viruses that is, Herpesvirus [18], Adenovirus [24], Reovirus [20], and Measles virus [21] that also showed an enhanced efficacy of viral oncolysis in combination with radiation therapy. Whether this effect can also be demonstrated in vivo was beyond this proof of concept study and should be addressed in future experiments.
When radiation treatment was performed one day before H-1PV infection, combination treatment was significantly better in all cell lines than single treatment. Virus infection followed by IR was less efficient in all cell cultures and had a reduced cytotoxic effect. One possible reason for the improved cytotoxicity of H-1PV after IR is the increased level of glioma cells positive for NS-1 expression 24 hours after early infection. Previous studies revealed that NS-1 is the key-mediator of parvoviral cytotoxicity and its expression is strongly S-Phase dependent [14,25]. Cell cycle analyses revealed that in all primary glioma cell cultures tested, the rate of cells in S-phase was increased 24 hours post-IR. Even though improved viral cytotoxicity may depend on several factors, this altered cell cycle distribution supports the finding of increased viral transcription and increased cell killing. As a consequence, when cells were infected when the cell cycle had returned to control levels (late infection), NS-1 expression decreased.
The glioma cell cultures in our study, like other glioma cell lines, were relatively radio-resistant and even after treatment with 20 Gy remaining cell clones continued to grow. Recent data suggests that stem-like cells exist within high-grade gliomas which are radioresistant and capable of initiating tumour regrowth. This is considered to be due to upregulated DNA damage checkpoint pathways [26]. In our system, neither radiation therapy nor H-1PV infection was able to kill all tumor cells. However, combined treatment with a high-radiation dose resulted in complete cytotoxicity in all cell cultures, indicating improved efficacy also in relatively resistant clones. These results may warrant to test whether the combination of radiation and H-1PV infection could also overcome the resistance of glioma cells expressing stem cell markers thereby offering a new treatment opportunity in these therapy refractory cells.
In conclusion, irradiated glioma-cells show intact susceptibility for H-1PV infection with even improved cell killing by combining IR with H-1PV, most pronounced in radioresistant glioma cells. These results further support the ongoing development of a phase-I clinical trial for the use of H-1PV in malignant gliomas, allowing for the inclusion of pretreated patients into the study population.