γ-Radiation Promotes Immunological Recognition of Cancer Cells through Increased Expression of Cancer-Testis Antigens In Vitro and In Vivo

Background γ-radiation is an effective treatment for cancer. There is evidence that radiotherapy supports tumor-specific immunity. It was described that irradiation induces de novo protein synthesis and enhances antigen presentation, we therefore investigated whether γ-radiation results in increased expression of cancer-testis (CT) antigens and MHC-I, thus allowing efficient immunological control. This is relevant because the expression of CT-antigens and MHC-I on tumor cells is often heterogeneous. We found that the changes induced by γ-radiation promote the immunological recognition of the tumor, which is illustrated by the increased infiltration by lymphocytes after radiotherapy. Methods/Findings We compared the expression of CT-antigens and MHC-I in various cancer cell lines and fresh biopsies before and after in vitro irradiation (20 Gy). Furthermore, we compared paired biopsies that were taken before and after radiotherapy from sarcoma patients. To investigate whether the changed expression of CT-antigens and MHC-I is specific for γ-radiation or is part of a generalized stress response, we analyzed the effect of hypoxia, hyperthermia and genotoxic stress on the expression of CT-antigens and MHC-I. In vitro irradiation of cancer cell lines and of fresh tumor biopsies induced a higher or de novo expression of different CT-antigens and a higher expression of MHC-I in a time- and dose-dependent fashion. Importantly, we show that irradiation of cancer cells enhances their recognition by tumor-specific CD8+ T cells. The analysis of paired biopsies taken from a cohort of sarcoma patients before and after radiotherapy confirmed our findings and, in addition showed that irradiation resulted in higher infiltration by lymphocytes. Other forms of stress did not have an impact on the expression of CT-antigens or MHC-I. Conclusions Our findings suggest that γ-radiation promotes the immunological recognition of the tumor. We therefore propose that combining radiotherapy with treatments that support tumor specific immunity may result in increased therapeutic efficacy.

Introduction c-radiation or radiotherapy is one of the most widely used treatments for cancer [1]. Irradiation induces death of tumor cells [2,3], but there is accumulating evidence that adaptive immunity significantly contributes to the efficacy of radiotherapy [4]. For example, irradiated tumors in patients and in mice are more often infiltrated by leukocytes than the unirradiated tumors [5,6,7] and very recent studies in preclinical models showed that the efficacy of radiotherapy depends on the presence of CD8 + T cells [8]. The fact that tumors are targeted and controlled by CD8 + T cells is suggested by the increased tumor incidence in immunosuppressed patients [9,10,11] and by the fact that tumor-specific immunity can be detected in cancer patients [12,13,14,15]. As the recognition of tumor cells by CD8 + T cells depends on the presentation of tumor-associated antigens (TAAs) in the context of MHC-I molecules, the often-heterogeneous expression of TAA and/or MHC-I within a tumor negatively impacts on the efficacy of tumor-specific immunity. In the present study we asked the specific question whether irradiation induces or up-regulates the expression of a prominent group of TAAs, the so-called CTantigens. The CT-antigens form an extended family of antigens that are expressed in a large variety of malignancies but are absent from healthy tissue except for the testis and placenta [16,17]. Cancer patients often develop spontaneous immune responses towards CT-antigens, which illustrates their immunogenicity [18]. Due to their immunogenicity and restricted pattern of expression, CT-antigens are considered promising targets for immunotherapy in cancer patients [19,20]. We observed that irradiation induced a higher or a de novo expression of different CT-antigens as well as an up-regulation of MHC-I expression in multiple cancer cell lines and in fresh, ex vivo irradiated tumor biopsies. Importantly, comparison of paired tumor sections obtained from sarcoma patients before and after irradiation showed up-regulated or de novo expression of MHC-I and CT-antigens and the concomitant increase of infiltrating CD8 + T cells, suggesting that irradiation mobilizes local, tumor-specific immune responses. Furthermore, our findings indicate that a combination of radiotherapy and active immunization with relevant CT-antigens may be a treatment modality with higher efficacy compared to either therapy alone.

Ethics Statement
The ethics committee "Ethical committee of the canton of Zurich" specifically approved this study (Study No: EK-1017). Primary human cells. Human foreskin keratinocytes were obtained as a gift from Dr. Onur Boyman (Department of Dermatology, University Hospital Zurich) and were cultured in Keratinocyte serum free medium (K-SFM; Invitrogen), supplements (EGF Human Recombinant and Bovine Pituitary Extract; Invitrogen), L-glutamine and antibiotics as described [21]. Normal human dermal fibroblasts (NHDF) were obtained from PromoCell GmbH (Heidelberg, Germany). Human microvascular endothelial cells (HMEC-1) were obtained as a gift from Dr. Therese Resink (Department of Biomedicine, University Hospital Basel) and were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 5% FBS (Sigma-Aldrich) and antibiotics. Human lung fibroblasts (MRC-5) were obtained as a gift from Dr. Giancarlo Marra (IMCR, Zurich) and were originally obtained from Coriell Cell Repositories (Camden, NJ) and cultured in MEM medium (Invitrogen), containing 15% FBS, L-glutamine and antibiotics. ZT-821 cells (primary renal cells) were established in our lab from a single cell suspension of healthy kidney tissue. The NHDF and ZT-821 cells were cultured in RPMI 1640 medium (Invitrogen) containing 10% FBS (Sigma-Aldrich), L-glutamine and antibiotics.

Treatment of cells
Ionizing radiation. Cells and fresh tumour biopsies were exposed to c-radiation from a 60 Co source. The doses of irradiation used are described in each experiment.
Hyperthermia. Cells were incubated at 42uC for 1 h or 5 h and cultured subsequently for 72 h at 37uC. The treated and nontreated cells were then processed for analysis by RT-qPCR.
Hypoxia. Cells were incubated under hypoxic conditions (1% O 2 vol/vol) in a hypoxic workstation (InVivoO 2 -400, Ruskinn Technology, Leeds, UK) for different time points (8 h, 48 h or 72 h). The treated and non-treated cells were then processed for analysis by RT-qPCR.
Genotoxic stress. Cells were treated at 37uC with 1 mg/mL cisplatin (Sigma-Aldrich Corp. St. Louis, MO) for 24 h or with 15 mM etoposide (Sigma-Aldrich) for 1 h or with 150 mM bleomycin (Sigma-Aldrich) for 1 h. This time point was considered as T 0 . At the end of the treatment, the medium was replaced with fresh culture medium without the drug. The cells were then cultured for additional 72 h at 37uC. Control cultures were treated under similar experimental conditions in the absence of the drug.
Small-molecule protein kinase inhibitors. A stock concentration of 10 mM in 100 % dimethyl sulfoxide (DMSO) of specific inhibitors for ataxia telangiectasia mutated (ATM), KU 55933 (Tocris bioscience, Ellisville, MO) and DNA-dependent protein kinase catalytic subunit (DNA-PKcs), NU7441 (Tocris bioscience) were diluted to a working concentration of 10 mM in 1% DMSO. The inhibitors were added 1 h before irradiation, and were left in the culture throughout the experiment. Control cultures were treated under similar experimental conditions in the absence of the drug with 1% DMSO.

Tumor biopsies
Biopsies were obtained from the University Hospital Zurich. All patients underwent surgery as part of their standard treatment and signed the informed consent. The ethics committee "Ethical committee of the canton of Zurich" specifically approved this study (Study No: EK-1017). Immediately upon resection, biopsies were cut into multiple pieces of 2-4 mm 3 . The pieces were randomized in two batches and put into culture medium. One batch was irradiated with a single dose of 20 Gy, whereas the other batch served as control. The tumor pieces were cultured for 72 h following ex vivo radiation and the gene expression was determined by RT-qPCR analysis. Paired biopsies from sarcoma before and after irradiation were collected for another purpose and therefore the time points of collection after radiotherapy varied between 2-6 weeks. The tumors were irradiated in vivo with a linear accelerator and received a dose of 50-64 Gy.

RNA extraction and preparation of cDNA
Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA) and was subsequently digested with DNAse I (Invitrogen). The concentration and purity was evaluated using the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). 150 ng of RNA was reverse transcribed using the high-capacity cDNA Reverse Transcription Kit (Applied Biosciences, Foster City, CA). The cDNA was either used immediately for RT-qPCR reactions or stored at -20uC until use. All kits were used according to the manufacturer's instructions.

RT-qPCR
The expression of CT-antigens and MHC-I was analyzed using TaqMan gene analysis primers and TaqMan 1x universal master mix (Applied Biosystems) on a RotoGene cycler (Corbett Research, Sydney, NSW). The reaction mixture (10 mL) consisted of 1 mL cDNA, 3.5 mL water, 0.5 mL primer and 5 mL TaqMan 1x universal master mix. The following cycle conditions were used: 2 min 50uC, 10 min 95uC, 45 cycles of 15 s at 95uC, 1 min 60uC. The change in expression levels of CT-antigens and MHC-I is given as the fold increase in expression by comparing the delta-Ct values of the treated to that of the non-treated samples after normalization to 18S rRNA. Ct values . 38 cycles were interpreted such that the gene is not expressed and a fold expression , 2 was considered as no change. Each of the breast cancer and osteosarcoma cell lines were tested in two independent experiments and the lung carcinoma and the prostate carcinoma cell lines in one. Each sample was loaded in duplicates/triplicates for each of the RT-qPCR experiments.  Table S4). Three individuals performed the scoring blindly and independently for the MHC-I stainings and two individuals performed the scoring blindly for the other stainings.

Results
In vitro c-radiation up-regulates transcripts of CTantigens and MHC-I molecules Breast, osteosarcoma, lung and prostate cancer and normal primary cells were exposed to a single-dose radiation of 20 Gy, after 72 h the cells were harvested and analyzed for the expression of CT-antigens and MHC-I by RT-qPCR. Most of these cancer cell lines expressed undetectable or very low levels of CT-antigens under standard culture conditions. In contrast, c-radiated cells showed de novo or up-regulated expression of various CT-antigens in a randomized fashion ( Figure 1A-D). In a panel of the four cancer cell line types, c-radiation was found to have the most profound effect on the breast cancer and osteosarcoma cell lines and the least on the prostate cancer cell lines. The up-regulation of MHC-I and CT-antigens upon c-radiation seems to be a specific feature of malignant cells, as this was not observed in any of the normal primary cells that we tested here (ZT-812, human foreskin keratinocytes, HMEC-1, MRC-5, NHDF and MCF 10A, Figure  S1A, S1B). The melanoma cell line SK-MEL-37 is known to express all tested CT-antigens under normal culture conditions and was therefore included as the positive control. Irradiation of this cell line resulted in a clear up-regulation of all the CT-antigens and MHC-I (data not shown). We detected an increased expression of CT-antigens and MHC-I as early as 24 h after irradiation in some cases and a steady increase in gene expression was observed up to 96 h after irradiation. As we observed substantial cell death at 96 h after irradiation, we performed all further analyses at 72 h after irradiation. As radiotherapy can be given as a single high dose (20 Gy) or as fractionated doses, we   CT-antigens upon irradiation that increased with time (Figure 2A, 2B), thus confirming the data obtained by RT-qPCR. Irradiation of the normal breast cell line MCF10A did not result in increased NY-ESO-1 or CT7 protein levels ( Figure 2B).
To study the effect of c-radiation on the surface expression of MHC-I at the protein level, we irradiated multiple cancer cell lines (breast carcinoma and osteosarcoma) and normal primary cells of different origins, followed by flowcytometric detection of surface MHC-I, and observed that irradiation resulted in a timedependent increase of MHC-I surface expression on different cancer cell lines ( Figure 2C, 2D) but not in the normal primary cells ( Figure S1B).

Irradiation of cancer cells enhances T cell recognition in vitro
In order to determine whether irradiation-induced up-regulation of CT-antigens and MHC-I results in increased recognition by antigen-specific CD8 + T cells, we measured degranulation [23][24]

Ex vivo irradiation of fresh tumor biopsies induces upregulation of CT-antigens and MHC-I
To expand our findings to clinically relevant material, we collected fresh tumor biopsies from 23 different cancer patients, cut those into at least 50 small pieces, and randomized them into two experimental groups. We irradiated one group of biopsies with 20 Gy, whereas the others served as control, and analyzed the expression of CT-antigens and MHC-I 72 h later by RT-qPCR and immunohistochemistry. These results confirmed the findings we obtained with cancer cell lines, that c-radiation often induced increased expression of CT-antigens and/or MHC-I ( Figure 4A, 4B, Table S1). Importantly, our results suggest that heterogeneous expression of CT-antigens and/or MHC-I can become more homogeneous upon radiotherapy, which obviously supports immunological control of the tumor.

In vivo irradiation of human sarcoma results in increased expression of CT-antigens and MHC-I and in lymphocyte infiltration
Finally, we compared the expression of CT-antigens, MHC-I and the infiltration by lymphocytes in 15 paired paraffin-sections obtained from sarcoma patients before and after radiotherapy by immunohistochemistry. We found that radiotherapy resulted in

Other forms of stress have no impact on the expression of CT-antigens or MHC-I in vitro
Stress and environmental alterations induce changes in the transcriptional profile in order to cope with those assaults [25,26,27]. As irradiation induces DNA-damage, we investigated whether the up-regulated expression of MHC-I and CT-antigens would also occur after exposure of cell lines to other treatments that induce DNA damage such as cisplatin, etoposide and the radiomimetic drug bleomycin. MDA-MB-469 and SK-MEL-37 cells were treated separately with each of the DNA-damaging agents and RNA levels were monitored at different time points following treatment. Our results show that none of these agents up-regulated the expression of CT-antigens and MHC-I ( Figure  S3B-D). However, the hallmark genes pS15-p53 and FANC-D2 were up-regulated following treatment, indicating that these agents induced stress at the concentration used ( Figure S3A). Furthermore, we tested whether other types of cancer-related stress, including elevated temperatures or low oxygen levels, induced increased expression of CT-antigens and/or MHC-I. We found that none of these treatments impacted on the expression of CTantigens and MHC-I whereas the signature gene CA9 was upregulated ( Figure S4A, S4B). Together, these results suggest that the increased expression of CT-antigens and MHC-I by cancer cell lines is a specific response to c-radiation and does not occur after exposition to other inducers of various stress response pathways.
The ATM and DNA-PK signaling pathways are dispensable for c-radiation-induced expression of CTantigens and MHC-I The activation of DNA-damage repair checkpoint pathways as a response to genotoxic insult helps to maintain the genomic integrity in mammalian cells [28]. DNA damage triggers the activation of various serine/threonine protein kinases, which constitute the primary transducers in the signaling cascade and of which, ataxia telangiectasia mutated (ATM) and DNA-dependent protein kinases (DNA-PKcs) are of utmost importance [29]. The ATM protein kinase is a critical intermediate in a number of cellular responses to c-radiation and other forms of stress [30]. We thus investigated whether the up-regulation of CT-antigen and MHC-I expression in response to c-radiation depends on the activation of ATM and/or DNA-PKcs. We irradiated MDA-MB-469 and SK-MEL-37 cells with a single dose of c-radiation (20 Gy) in the presence or absence of specific inhibitors of ATM (KU55933) and DNA-PK (NU7441). 10 mM of the inhibitor was added 1 h before irradiation, and was left in the culture throughout the experiment. We observed that ATM and DNA-PK as well as p53 were phosphorylated in response to c-radiation, which was prevented by the specific inhibitors ( Figure 6A, 6B). However, neither ATM-nor DNA-PK inhibition prevented the up-regulation of CT-antigen and MHC-I expression ( Figure 6A, 6B, S5A, S5B). These results indicate that the ATM and DNA-PK DNA-damage signaling pathways are not crucially involved in the up-regulation of CT-antigen and MHC-I expression upon cradiation.

Discussion
Radiotherapy is one of the most widely used and successful cancer treatments to date [31]. Recent studies described that cradiation leads to a plethora of alterations in the tumor cells [32], including the de novo synthesis of particular proteins and the upregulation of MHC-I expression [1,32,33]. In addition, irradiated tumors are often more infiltrated by leukocytes than the nonirradiated tumors and studies using preclinical models showed that the therapeutic success of high-dose irradiation depends on adaptive immunity [34]. We therefore hypothesized that the expression of a specific class of TAAs, the so-called CT-antigens, is induced by irradiation, thus making tumors more susceptible to recognition by effector T cells. In addition the de novo expression of highly immunogenic CT-antigens may result in tumor-specific immune responses that are not yet compromised by mechanisms of central [35] or peripheral tolerance. Because cancer cells can only be recognized by CD8 + T cells when they express MHC-I and because tumors often express very low levels of MHC-I or are even negative, radiation-induced up-regulation of MHC-I may further support immune recognition. We observed that irradiation induced de novo or up-regulated expression of various CT-antigens and MHC-I in a randomized fashion and independent of the tissue origin of the malignancy. We confirmed this effect using established cancer cell lines, fresh ex vivo irradiated tumors and paired biopsies from sarcoma patients before and after radiotherapy. Studies have shown that the radiation-induced up-regulation of MHC-I makes the tumor cells more susceptible to lysis by CTL in vitro [1,36]. We confirm these data here and show that irradiation of tumor cells not only up-regulates the expression of CT-antigens and MHC-I but also increases their recognition by CD8 + T cells. The up-regulated expression of CT-antigens and MHC-I seems specific for c-radiation, as similar changes in gene expression were not observed upon other treatments that induce DNA-damage, upon hypoxia or hyperthermia. We could not identify the molecular mechanism underlying radiation-induced up-regulation of CT-antigens and MHC-I expression, but we excluded the involvement of the ATM or DNA-PK signaling pathways. Previous studies have shown that the expression of CTantigens is regulated through demethylation of their promoter CpGs [37]. In addition, all CT-antigen genes that are expressed in tumors or testis can be induced in vitro by DNA demethylation [38]. However, when we compared the methylation status of CpGs in the promoter region of NY-ESO-1 of irradiated and nonirradiated MDA-MB-469 and MCF 7 cells, we found no difference (data not shown). This result suggests that the up-regulated expression of CT-antigens and MHC-I upon c-radiation is regulated through another pathway than demethylation of CpG islands in the promoter region. However, as we performed this experiment only with a limited number of cell lines and we analyzed only one CpG region in the promoter of NY-ESO-1, we can't fully exclude that other CpG regions may have been hypomethylated following irradiation.
Our data strongly suggests that irradiation supports immunological control of tumors through the expression of novel tumor-associated antigens to which the immune response presumably is not compromised. In addition, the concomitant up-regulation of MHC-I expression makes irradiated tumor cells more susceptible to tumor-specific CTL. It is currently not known whether the increased infiltration by CD8 + T cells upon irradiation is a direct effect of increased local expression of cognate peptide/MHC-I complexes or whether other factors such as local inflammation or changes in vasculature contribute.
It may be interesting to combine radiotherapy with immunization to maximally exploit the changes induced by irradiation. However, the fact that it seems unpredictable as to which CTantigens will show de novo or up-regulated expression makes it difficult to choose the correct antigen for immunization, at least in those cases where tumor biopsies are not available. Nevertheless, the combination of radiotherapy with treatments that generally stimulate the immune system such as the blockade of co-inhibitory interactions (CTLA-4, PD-1) or mediators (IL-10, TGF-b, IDO) or activation of dendritic cells by innate stimuli may further improve the efficacy of radiotherapy.