Low Dose Decitabine Treatment Induces CD80 Expression in Cancer Cells and Stimulates Tumor Specific Cytotoxic T Lymphocyte Responses

Lack of immunogenicity of cancer cells has been considered a major reason for their failure in induction of a tumor specific T cell response. In this paper, we present evidence that decitabine (DAC), a DNA methylation inhibitor that is currently used for the treatment of myelodysplastic syndrome (MDS), acute myeloid leukemia (AML) and other malignant neoplasms, is capable of eliciting an anti-tumor cytotoxic T lymphocyte (CTL) response in mouse EL4 tumor model. C57BL/6 mice with established EL4 tumors were treated with DAC (1.0 mg/kg body weight) once daily for 5 days. We found that DAC treatment resulted in infiltration of IFN-γ producing T lymphocytes into tumors and caused tumor rejection. Depletion of CD8+, but not CD4+ T cells resumed tumor growth. DAC-induced CTL response appeared to be elicited by the induction of CD80 expression on tumor cells. Epigenetic evidence suggests that DAC induces CD80 expression in EL4 cells via demethylation of CpG dinucleotide sites in the promoter of CD80 gene. In addition, we also showed that a transient, low-dose DAC treatment can induce CD80 gene expression in a variety of human cancer cells. This study provides the first evidence that epigenetic modulation can induce the expression of a major T cell co-stimulatory molecule on cancer cells, which can overcome immune tolerance, and induce an efficient anti-tumor CTL response. The results have important implications in designing DAC-based cancer immunotherapy.


Introduction
A major challenge in cancer immunotherapy is immune evasion by cancer cells [1]. During tumor development and progression, tumors build up an immune suppressive network, including tumor associated myeloid cells and various regulatory T cells [2,3]. Cancer cells themselves are genetically unstable; they can downregulate major histocompatibility complex (MHC) class I molecules [4,5] and lose the expression of tumor antigens [6,7,8]. In addition, cancer cells do not normally express key co-stimulatory molecules such as CD80, but rather express some co-inhibitory molecules that render tumor antigen specific T cell tolerance [9]. All these factors prevent the induction of an efficient T cell response to tumors. Thus, overcoming immune evasion is of great importance in cancer immunotherapy.
Epigenetic evidence suggests that in cancer cells, some key immune stimulatory molecules are regulated by DNA methylation in their promoter region. Some well known tumor antigens such as cancer testis antigens (CTAs) are almost exclusively regulated by DNA methylation [10,11,12,13,14,15]. MHC class I and its antigen presentation machinery have also been shown to be regulated by DNA methylation [16,17,18,19]. In addition to CTAs and MHC molecules, there is also evidence that adhesion molecules [16,20] such as ICAM-1 and LFA-3, and the costimulatory molecules [19,20] such as CD40 and CD86 can be regulated by DNA methylation in cancer cells. Thus, demethylating agents that can upregulate expression of tumor antigens, MHC class I, and adhesion/co-stimulatory molecules in cancer cells should be useful in enhancing tumor immunogenicity and their susceptibility to immune destruction. Indeed, there is a body of evidence that suggests demethylation treatment can dramatically increase cancer cell susceptibility to destruction by T cells [11,15,17,21]. However, there is no direct in vivo evidence that demethylation treatment of cancer leads to a specific anti-tumor T cell response.
Decitabine (DAC), a DNA demethylating agent [22], has recently emerged as a potent therapeutic for the treatment of preleukemic hematological disease-MDS [23,24], established leuke-mia [25,26,27] and advanced lung cancer [28]. Low dose DAC can cause sustained anti-tumor effects even after discontinuation of treatment [24,29,30], suggesting that an active immune response may be induced in the treated patients. To determine whether DAC treatment can induce anti-tumor immune responses in vivo, we studied the effect of low dose DAC treatment in mice with established T cell lymphoma EL4 tumors. We found that transient and low dose DAC treatment resulted in the induction of an antitumor cytotoxic T lymphocyte (CTL) response that mediated tumor regression. DAC-induced CTL response appears to be elicited by the induction of CD80 expression on EL4 cells. In addition, we also found that a transient, low-dose DAC treatment can induce CD80 gene expression in a variety of human cancer cells. This study provides the first evidence that epigenetic modulation can induce the expression of a major T cell costimulatory molecule on cancer cells, which can overcome immune tolerance, and induce an efficient anti-tumor CTL response.

Mice
C57BL/6 and CBF1 mice were purchased from Beijing Vital-River Lab Animal Technology Co. Ltd. C57BL/6 congenic strain mice B6.SJL-Ptprc a Pepc b /Boy-M (CD45.1 + ) were provided by Dr. Chunfeng Qu (State Key Laboratory of Molecular Oncology, Cancer Institute/Hospital, Chinese Academy of Medical Sciences, Beijing, China). All mice were maintained at Chinese PLA General Hospital (CPGH) in specific pathogen-free conditions. All animal experiments were performed in accordance with national and institutional guidelines for animal care and were approved by the animal use and care committee, CPGH.

Cell Culture and EL4 Subclone Establishment
All cell lines were originally obtained from American Type Culture Collection (ATCC). EL4 cells (mouse T cell lymphoma/ leukemia cell line) were cultured and maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% heatinactivated equine serum (Hyclone), 100 mg/ml penicillin/streptomycin and L-glutamine. The cells were incubated at 37uC in 5% CO 2 . EL4 subclones were established using limiting dilution in 96well round-bottomed tissue culture plates based on CD80 expression. Two EL4 subclones, i.e. EL4 C45 with high expression of CD80, and EL4 C3 with no expression of CD80, were used for this study. The human leukemia cell lines K562 (cell line derived from a patient with chronic myeloid leukemia at blast crisis), U937 (acute myelogenous leukemia, M5), THP-1 (acute myelogenous leukemia, M5), NB4 (acute myelogenous leukemia, M3), Kasumi-1 (acute myelogenous leukemia, M2), Molt-4 (T-cell acute lymphoblastic leukemia), Hut-78 (T cell lymphoma) and Raji (Burkitt's lymphoma) were grown and maintained in RPMI-1640 (Hyclone) medium supplemented with 10% fetal bovine serum.   In vitro and in vivo Treatment with Decitabine Decitabine (DAC, Xian-Janssen pharmaceuticals Ltd, China) was dissolved in phosphate-buffered saline (PBS) (pH 7.4) to obtain 100 mM stocks and stored at 220uC. For in vitro studies, DAC was added to cell culture medium to a final concentration of 0.25 mM for 72 hours. The same concentration of Cytidine (Sigma) in PBS or PBS only was added to cells as control treatment. 24 hours after treatment cells were harvested for further study. For in vivo studies using DAC, mice with established EL4 tumors were injected with DAC (1.0 mg/kg body weight in 200 ml PBS) or PBS i.p. once daily for 5 consecutive days. Mice were sacrificed 7-10 days after completion of drug treatment and the tumors excised were processed for tumor infiltrating lymphocytes (TIL) analysis.

Reverse Transcription-PCR (RT-PCR)
Total RNA was extracted from DAC-treated or vehicle-treated EL4 cells and other human leukemia and lymphoma cells using TRIzol reagent (Invitrogen) according to manufacturer's instruction. RT was performed using Reverse Transcription System (Promega) on 1 mg of total RNA, and PCR amplifications were then performed using primers shown in Table 1.Simultaneous amplification of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene using primers for mouse (forward 59-GATGCCCCCATGTTTGT-39; reverse 59-CCGTTCAGCTCTGGGATGA -39) and humans (forward 59-GAGTCAACGGATTTGGTCGT-39; reverse 59-TTGATTTTGGAGGGATCTCG-39) was used as an internal control for the amount and integrity of the RNA analyzed. All PCR products were resolved on agarose gels and visualized using ethidium bromide staining.

Real Time PCR
Real-time PCR was performed using an ABI 7900-HT sequence system (PEApplied Biosystems, Carlsbad, CA, USA) using previously determined conditions [31]. SYBR Premix Ex Taq PCR kit (Takara Bio, Inc.) and the same primers used for RT-PCR were employed for amplifying mouse P1A, Mela, Magea4, CD80, CD79b, CD74, CD48, CD300a, CD3eap, CD274, CD247, CD180 and GAPDH genes. Each sample was assayed in triplicate, and the experiments were repeated twice. The relative amount of mRNA was calculated by plotting the Ct (cycle number), and the average relative expression for each group was determined using the comparative method (2 2ggCt ).

Tumorigenesis
C57BL6 mice were injected with various doses (most often using 1610 4 cells/mouse, s.c.) of DAC-treated or control EL4 cells. Tumor volumes were measured as ab 2 /2 (a = length; b = width) every two to three days. To induce mouse leukemia, mice were injected with 2610 4 EL4 cells/mouse i.v.
In vivo CD8+ and CD4+ T Cell Depletion C57BL/6 mice were injected with 400 mg of anti-CD4 (clone GK 1.5; BioXcell, NH, USA) or anti-CD8 (clone 53-6.72; BioXcell, NH, USA) monoclonal antibodies i.p. every 4 days immediately after treatment with DAC. The control mice were treated with purified rat IgG2b or IgG2a, respectively (BioXcell, NH, USA). The depletion effects were assessed at the endpoint of the experiment by staining of spleen and tumor cells for CD4 and CD8 using anti-CD4 (RM4.4) and anti-CD8 (5H10-1) mAbs (eBioscience), followed by flow cytometric analysis.  FCS) on ice for 30 min. Cells were fixed in 1% paraformaldehyde in PBS. For detection of intracellular cytokines, cells were stimulated in vitro with PMA (100 ng/ml) and ionomycin (1000 ng/ml) for 5 h. Golgi Stop (BD Pharmingen, USA ) was added (1/1500) during the last 2 h of incubation. The cells were first stained for the cell surface markers such as CD4 or CD8, followed by a standard intracellular cytokine staining procedure for IFN-c. Cells were collected using a FACSCaliburH flow cytometer and data were analyzed using the FlowJo software (Tree Star, Inc., OR).

Gene Expression Microarray Analysis
Total RNA was extracted from DAC-treated and vehicletreated EL4 cells using TRIzol reagent (Invitrogen) according to manufacturer's instruction. cDNA labeled with a fluorescent dye (Cy5 and Cy3-dCTP) was generated using CapitalBio cRNA Amplification and Labeling Kit (CapitalBio, Beijing, China). SmartArray (CapitalBio, Beijing, China) chips containing about 25000 mouse genes were hybridized with labeled cDNA probes on a GeneChip system. Arrays were scanned with a confocal LuxScan TM scanner and the images obtained were then analyzed using LuxScan TM 3.0 software (Both from CapitalBio, Beijing, China). Significance Analysis of Microarrays (SAM) was performed to determine differentially expressed genes.

Bisulfite Sequencing
The methylation status of the CpG dinucleotides within two regions (oligo 1: nucleotide 2792 to 2335, oligo2: nucleotide 2151 to +258), relative to the 59 end of the CD80 gene, was analyzed. Genomic DNA was prepared from EL4 subclones EL4 C3 (CD80 2 ) and EL4 C45 (CD80 + ), vehicle or DAC-treated EL4 cells using the Wizard Genomic DNA Purification Kit (Promega Inc, USA). For subcloning and sequencing of individual alleles, bisulfite-treated genomic DNA was amplified using the Epitect The PCR products were gel purified and cloned into the pGEM-T vector (Promega Inc, USA). The inserted PCR fragments of individual clones were sequenced using an ABI PRISMDNA sequencer (Applied Biosystems, USA).

Mixed Lymphocyte Reaction and CD80 Blockade
T cells harvested from spleen and lymph nodes of EL4immunized BALB/c mice were used as responder cells. DAC or vehicle-treated EL4 cells irradiated with X-ray (14 Gy) were used as stimulator cells. Responder cells and stimulator cells were incubated in 96-well plates in a ratio of 8:1, 4:1, 2:1 and 1:1, and cultured at 37uC at 5% CO 2 for 6 days. T-cell proliferation was assessed with Cell Counting Kit-8 (Dojindo, Kumomoto, Japan) according to the manufacturer's instructions. Briefly, after incubation for 6 days at 37uC, 10 mL water soluble tetrazolium salt (WST)-8 was added to each well, and the cells were incubated for another 4 to 6 hours at 37uC. The absorbance of the sample at 450 nm was measured. For CD80 blockade, anti-CD80 (16-10A1, eBioscience, San Diego, CA) or a control isotype matched IgG, (eBioscience, San Diego, CA) were added into co-culture wells at a final concentration of 5 mg/ml.

IL-2 and IFN-c ELISA
Supernatants of mixed cultures were harvested 72 h after coculture and were analyzed for IL-2 and IFN-c using ELISA kits (eBiosciences, San Diego, CA).

Statistics
Statistical analyses were conducted using GraphPad Prism Software (GraphPad Software, Inc., USA). Student's t test was used to compare tumor volume differences between two groups. For comparison of mouse survival, the Kaplan-Meier survival analysis and log-rank test were used. A P value less than 0.05 was considered significant.

DAC Treatment Inhibits EL4 Cell Tumorigenesis and Leads to Regression of Established Tumors in C57BL/6 Mice
To test whether DAC treatment affects tumorigenicity of cancer cells in mice, EL4 cells were treated with either DAC (0.25 mM) or Cytidine (0.25 mM) or vehicle (PBS) for 72 hours. We included Cytidine as a control agent since DAC is the analogue of Cytidine. In addition, a previous study has revealed that nucleotide (Thymidine) treatment is toxic to EL4 cells [32]. Viable EL4 cells from the above mentioned treatments were then injected into  Figure 1A). EL4 tumors were established in C57BL/6 mice only when very high numbers of DAC-treated EL4 cells were used (8-16 fold higher doses) ( Figure 1B). Intravenous injection of vehicle-treated EL4 cells (1610 4 cells/mouse) led to the development of EL4-induced leukemia in CBF1 mice and caused death of most mice; in contrast, no mice receiving this number of DACtreated EL4 cells died of leukemia ( Figure 1C). To test whether in vivo injection of DAC could inhibit the growth of established tumors, EL4 cells were injected into C57BL/6 mice s.c. When tumors grew to a size of about 566 mm, these mice were treated with DAC or PBS i.p. daily for 5 consecutive days. In contrast to the progressive tumor growth in vehicle treated mice, DAC treatment caused continuous tumor regression even after DAC treatment was stopped ( Figure 1D). Thus, DAC-treated EL4 cells show reduced capacity in tumorigenesis, and transient DAC treatment of mice with established tumors leads to continuous tumor regression.

DAC Treatment Induces Anti-tumor CTL Responses
The fact that the transient DAC treatment resulted in persistent tumor regression suggests that DAC treatment could have induced anti-tumor immune responses. To test this possibility, C57BL/6 mice with established EL4 tumors were treated with DAC as described above ( Figure 1D). Seven to ten days later, mice were sacrificed, and mononuclear cells from tumors were stained for CD4, CD8 and NK1.1/CD3, followed by analysis using flow cytometry. As shown in Figure 2A, tumors from DAC-treated mice had significantly increased numbers of CD8 + and CD4 + T cells compared with tumors from vehicle-treated mice; NK1.1 + CD3 2 NK cells were barely detectable in tumors from DACtreated and PBS-treated mice (Figure 2A, B). The tumors from DAC-treated mice contained higher numbers of IFN-c producing CD8+ T cells compared with tumors from vehicle-treated mice ( Figure 2C, D). To determine whether increased T cell responses were responsible for DAC-induced tumor regression, C57BL/6 mice with established EL4 tumors treated with DAC were also treated with anti-CD8 or anti-CD4 antibodies or their relative control antibodies i.p. Anti-CD8 or anti-CD4 treatment resulted in complete depletion of CD8 + or CD4 + cells from spleens and tumors ( Figure 3A, C). Depletion of CD8 + T cells ( Figure 3B), but not CD4 + T cells ( Figure 3D), resulted in more aggressive tumor growth in otherwise protected mice. Thus, DAC treatment of mice with established tumors induced CD8 + T cell-mediated anti-tumor immunity.

DAC Treatment Induces CD80 Expression in Tumor Cells
Since DAC treatment induced CTL responses in EL4 tumors, we hypothesized that DAC treatment can up-regulate immune stimulating molecules present on EL4 cells, which can stimulate CTL responses in immunocompetent mice. Earlier studies have reported up-regulation of MHC class molecules and tumor antigens by demethylating agents [11,17]. However, we failed to detect upregulation of MHC class I and class II molecules in DAC-treated EL4 cells (data not shown). To further explore the mechanisms of DAC-induced tumor immunity, we compared the differential expression of genes in DAC-treated EL4 cells and controls using cDNA microarray analyses (GEO accession # GSE38473) and found that DAC treatment significantly altered the gene expression profiling in EL4 cells compared with vehicle treatment ( Figure S1A). A total of 847 upregulated genes were identified in DAC-treated EL4 cells using the SAM analysis ( Figure S1B). Among the upregulated genes in DAC-treated EL4 cells, 3 cancer testis antigens and 9 cluster of differentiation (CD) molecules including CD80 (B7-1), an essential T cell costimulatory molecule, were identified ( Figure 4A). RT-PCR and qRT-PCR also verified their up-regulation ( Figure 4B and 4C).
To determine the persistence and stability of DAC-induced CD80 expression, EL4 cells were cultured in medium containing DAC (0.25 mM) or PBS for 72 hours. DAC-induced CD80 expression in EL4 cells was verified at both mRNA ( Figure 5A) and protein level ( Figure 5B). Kinetic analysis revealed that CD80 expression peaked at 3-5 days after DAC treatment, and was maintained at significant levels for about 3 weeks and then declined to the base level ( Figure 5C). To test whether DAC treatment can induce CD80 expression in vivo, EL4 cells (CD45.2 + ) were injected s.c. into C57BL/6 mice or i.v. into CD45.1 congenic mice. Two weeks later, mice were treated with DAC (1mg/kg body weight, i.p.) for 5 consecutive days. Seven to 10 days later tumor cells from DAC-and vehicle-treated mice were examined for CD80 expression. DAC treatment significantly up-regulated CD80 expression in EL4 cells from s.c. tumors ( Figure 5D, left panel) as well as in EL4 cells from bone marrows ( Figure 5D, right panel). To determine whether DAC-induced upregulation of CD80 is a general phenomenon, 6 human leukemia cell lines and 2 human lymphoma cell lines were treated with DAC (1 mM for 72 hours) followed by RT-PCR analysis. Four out of 5 CD80negative cell lines (U937, THP-1, NB4 and Molt-4) showed DACinduced CD80 expression, while in three cell lines with preexisting CD80 gene expression (K562, Hut-78 and Raji), the induction effect was not obvious ( Figure 5E).

DAC Induces Demethylation of the Promoter Region in CD80 Gene
To determine the underlying mechanism by which DAC induces CD80 expression in EL4 cells, we analyzed the sequence of the promoter region of murine CD80. There were no typical CpG islands in the promoter region of CD80. However, 14 CpG dinucleotides were found in Exon1 and its upstream 800 bp region ( Figure 6A). Two pairs of primers were designed for bisulfite sequencing analysis of two fragments covering nucleotide 2792 to 2335 (F1), and nucleotide 2151 to +258 (F2). We first compared the methylation status of the CpG dinucleotides in the two fragments from two EL4 subclones with or without natural expression of CD80 ( Figure 6B). In F1, 96% of CpG dinucleotides were methylated in CD80 2 cells versus 67% in CD80 + cells. In F2, 92% of CpG dinucleotides were methylated in CD80 2 cells compared to 6% in CD80 + cells ( Figure 6C). DAC treatment of EL4 cells caused a significant increase in demethylation sites in both F1 and F2 regions ( Figure 6D). Among DAC-treated EL4 cells, CD80 + cells exhibited more demethylation sites compared   with DAC-treated CD80 2 cells ( Figure 6E). Thus, CD80 gene expression in EL4 cells is closely associated with the demethylation status of its promoter region and DAC treatment significantly increases demethylation of CD80 promotor.
To further understand the impacts of DAC-induced tumor cell expression of CD80 on T cell activation and effector functions, we performed mixed lymphocyte culture and CD80 blockade experiments. DAC-treated EL4 cells induced stronger alloreactive T cell proliferation ( Figure 8A), and more IL-2 ( Figure 8B) and IFN-c ( Figure 8C) production compared with control EL4 cells. CD80 blockade significantly reduced DAC-EL4 cell-induced T cell proliferation ( Figure 8D) and cytokine production ( Figure 8E, F). Thus, DAC-induced CD80 expression on tumor cells can directly enhance T cell activation and effector functions.

Discussion
In this study we show that DAC, a DNA methylation inhibitor that is currently used for the treatment of MDS, AML and other malignant neoplasms, is capable of eliciting an anti-tumor CTL response in mouse EL4 tumor model. A transient, low dose DAC exposure to EL4 cells in vitro and in vivo induces CD80 expression on cancer cells, which is likely the major cause of the induction of CTL response. Mechanistically, we show that DAC induces CD80 expression in EL4 cells via demethylation of CpG dinucleotide sites in the CD80 gene promoter.
Previous studies have revealed that demethylating agents can restore cancer cell expression of MHC class I and its antigen presentation machinery [16,17,18,19], and induce the expression of tumor antigens [10,11], thereby increasing their susceptibility to destruction by immune effector cells such as CTL. Some studies have also shown that demethylation treatment up-regulates costimulatory molecules such as CD40 and CD86 [19,20]. However, it has not been shown that such a treatment can induce CD80 expression in cancer cells. In addition, there is no direct in vivo evidence that demethylation treatment of cancer leads to a specific anti-tumor T cell response. Thus, this study provides the first evidence that epigenetic modulation of cancer cells leads to the expression of a major T cell co-stimulatory pathway in cancer cells that triggers anti-tumor CTL responses.
In this study we found that DAC treatment induced a potent CTL response characterized by tumor infiltration of high numbers of IFN-c producing CD8 + T cells. Although we also observed increased numbers of CD4 + T cells in tumors, depletion of CD8 + T cells, but not CD4 + T cells resulted in accelerated tumor recurrence. Thus, DAC-induced CTL responses, but not CD4 + T cell responses are mainly responsible for the tumor regression. Although tumor infiltrating NK cells were not induced in tumors from DAC-treated animals, increased NK cells were detected in tumors from mice injected with DAC-induced CD80 + EL4 cells, suggesting that higher levels of CD80 expression on tumor cells may be required for the induction of NK response in tumors.
Decitabine was shown to up-regulate the expression of genes involved in growth control and apoptosis, and down-regulate genes related to the malignant phenotype of cancer cells [33]. Microarray analysis indicates that DAC induces profound gene expression alteration in cancer cells. The up-regulated genes include those encoding for cancer testis antigens such as P1A and MAGE-A1 and many cell surface molecules such as CD80. It is thus likely that many mechanisms, such as elevated tumor antigen expression may also contribute to tumor regression. However, a few lines of evidence suggest that up-regulation of CD80 is the single most important factor that contribute to the T cell responses and tumor regression. First, CD80 is the most potent costimulatory molecule [34]. Classic tumor immunology studies [35,36,37] have revealed that ectopic expression of CD80 on tumor cells has potent effects on the induction of anti-tumor immune responses, mainly anti-tumor CTL response [35,36,37] and sometimes NK response [38]. In addition, tumor expression of CD80 also enhances CTL effector functions and facilitates better tumor destruction [39,40]. Second, in this study we found that in DAC-treated tumor bearing animals, tumor cells strongly upregulated CD80. Third, DAC-induced CD80 + tumor cells, but not CD80 2 tumor cells induced potent anti-tumor CTL responses. Finally, in vitro CD80 blockade experiments suggest that blocking DAC-induced CD80 interaction with T cells directly enhance T cell proliferation and production of cytokines.
CD80 expression is normally tightly controlled, and epigenetic mechanism for the regulation of CD80 expression remains unclear. The promoter region of CD80 gene does not contain a typical CpG island. However, we found that DAC treatment induced dramatic demethylation in the CpG dinucleotides in EL4 cells. Notably, demethylation of the F2 region (nucleotide 2151 to +258) appears to have a more close correlation with the induction of CD80 expression. DAC-induced CD80 expression is also observed in a number of human leukemia and lymphoma cell lines. However, those human cell lines show differential susceptibility, thus suggesting that different cancer cells may have differential expression of epigenetic modulating machinery. Further studies are required to determine this issue.
A transient low dose DAC treatment in mice with EL4 tumors resulting in the induction of tumor specific CTL response has important clinical relevance. Recently, compiling evidence suggests that low dose of demethylating agents can cause sustained anti-tumor effects even after termination of the agent used. For instance, patients with hematological cancer receiving DAC treatment often experience prolonged responses [24,29]. Low dose azacitidine was used to treat patients with advanced nonsmall cell lung cancer that failed chemotherapy, very impressive responses were achieved [28]. In an animal study transient low doses of DAC exert durable antitumor ''memory response'' [30]. Thus, it is likely that immune responses are involved in ''sustained effects'' due to low dose DAC treatment in patients with cancer. Further investigation of the underlie mechanisms will reveal new therapeutic strategies of those incurable cancers.
Taken together, we have demonstrated that transient, low dose DAC treatment in mice with established tumors induces antitumor CTL response and induction of a major co-stimulation pathway. Based on these observations, we propose a mechanistic model depicting the mechanisms of DAC induction of anti-tumor T cell responses. In the tumor microenvironment, the promoter region of CD80 gene in cancer cells is highly methylated, therefore CD80 is not expressed. Lack of co-stimulation signal will result in T cell tolerance. DAC treatment leads to demethylation of the promoter region of CD80 gene in cancer cells, leading to the expression of CD80, which stimulates T cell responses (Figure 9). Given the fact that DAC has already made an impact in patients with hematological malignancies and solid tumors such as lung cancer, treatment of many cancers with transient, low dose DAC or other demethylating agents may result in the desired immune response that lead to tumor rejection in these patients. Figure S1 Gene expression profiling of DAC-treated vs PBS-treated EL4 cells. The mouse SmartArray chips were hybridized with RNA derived from DAC or PBS treated EL4 cells.

Supporting Information
(A) Scatter plot comparing global gene expression profiles between the DAC treated EL4 and PBS treated EL4 cells. (B) SAM plot indicating differentially expressed genes between DAC treated and PBS-treated EL4 cells. 847 genes were up-regulated by DAC. (TIFF)