Demethylation of Cancer/Testis Antigens and CpG ODN Stimulation Enhance Dendritic Cell and Cytotoxic T Lymphocyte Function in a Mouse Mammary Model

Background. Cancer/testis antigens (CTAs) are ideal targets for cancer immunotherapy in virtue of their restricted expression profile in normal tissues. However, CTA-targeted immunotherapy has been rather disappointing clinical setting for CTAs are downregulated by cytosine-phosphate-guanosine (CpG) methylation in their promoter regions, so that tumor cells have low immunogenicity. Methods. We reinduced mouse CTA P1A through demethylation process and generated P1A-specific cytotoxic lymphocytes (CTLs) by immunizing BALB/c (H-2d) mice with dendritic cells pulsed with a P1A-specific peptide and CpG oligodeoxynucleotide (ODN) immune adjuvant. Results. We found that demethylation and CpG ODN immune adjuvant stimulation facilitated DC maturation and enhanced the allogenic capacity of P1A-specific CTLs against target cells both in vitro and in vivo. Conclusions. Our results suggested that CTA induction and immune adjuvant stimulation is a feasible strategy in cancer immunotherapy.


Background
Immunotherapy has long been considered as a potent anticancer therapy due to its powerful tumor inhibition [1]. Although immunotherapy was indeed able to induce cytotoxic lymphocytes (CTLs) that recognize tumor antigens in vitro, it did not show equally potent antitumor activity in vivo, due to the lack of either tumor cell immunogenicity [2] or tumor-associated antigens for specific recognition by immune cells [3]. Therefore, major challenges, including activating antitumor immune cells and enhancing tumor cell immunogenicity to facilitate effective recognition, occurred at present.
CpG oligodeoxynucleotide (ODN), extensively investigated as an immune adjuvant [4], is recognized by toll-like receptor 9 (TLR9) and induces immune responses, during which effective presentation of tumor-specific antigen (TSA) to T lymphocytes is essential for subsequent activation of CTLs that ultimately eliminate tumor cells. However, tumorspecific CTLs do not work well due to the lack of tumor cell immunogenicity.
Cancer/testis antigens (CTAs) are considered a promising class of tumor antigens for therapeutic cancer vaccines on account of their restricted expression profiles in normal tissues and also due to the testis being an immunoprivileged site [5]. Recently, adoptive T-cell therapy has achieved prospective clinical results, including cancer regression in metastatic melanoma patients [6]. However, most of the initial cancer immunotherapies against CTAs resulted in poor clinical outcome owing to immune tolerance [7], with one main reason that CTA hypermethylation in the process of cancer transformation [8], that might downregulate or silence gene expression in cancer cells, makes the recognition of immunotherapeutic vaccines or cancer-specific CTLs difficult.
In consequence, we hypothesized that reinduction of CTAs by demethylation and concomitant stimulation by CpG ODN could induce an effective immune response against cancers expressing the related CTAs. In this study we would reinduce mouse CTA P1A with 5-aza-2-deoxycytidine (5-aza) and employed P1A peptide and immune adjuvant CpG ODN to stimulate DC maturation and the proliferative, as well as cytotoxic activities of T lymphocytes for the purpose of developing a new useful procedure for immunotherapy.

Cell Lines and Cell
Culture. A20 (mouse B-lymphoid tumor), CT26 (mouse colon adenocarcinoma), and 4T1 (mouse mammary carcinoma) cell lines were purchased from Cell Culture Center, Chinese Academy of Medical Science. Cells were cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) plus 1% penicillin/streptomycin and 1% glutamine. Cultures were maintained in a 5% CO 2humidified incubator at 37 ∘ C, and experiments were performed with cells in the exponential growth phase.

Treatment of Tumor Cells with 5-Aza
In Vitro. A20, CT26, and 4T1 cells were treated with 5-aza with a final concentration of 3 mol/L and cultured for 48 hours as described above. Control cultures were grown under similar experimental conditions but without 5-aza.

DC Maturation by CpG ODN and P1A-Specific Peptide
Pulsing. Bone marrow (BM) monocytes isolated by Ficoll density gradient centrifugation were then cultured in RPMI 1640. On day 3, GM-CSF (20 ng/mL) and murine IL-4 (10 ng/mL) were added. After 7 days' culture, CpG ODN, P1A peptide, or CpG ODN (10 g/mL) + P1A peptide (10 g/mL) were added and further cultured for 3 days. The cells and culture supernatant were separately collected for further cell phenotype and cytokine level assays. Supernatant IL-6 and TNF-levels were assayed using a mouse IL-6 enzymelinked immunosorbent assay (ELISA) kit and a mouse TNF-ELISA kit (Diaclone, Besançon, France) according to the manufacturer's instructions.

Generation and Activation of P1A CTL.
The stimulation effects of DCs on T cell proliferation and cytotoxic activity were determined using a previously described method [12]. T lymphocytes were separated by T-cell separation columns and used as response cells. The DCs were exposed to CpG ODN and P1A-specific peptide for 3 days and used as stimulating cells. The response cells and stimulating cells were cocultured for 96 hours at a ratio of 20 : 1. The cells and the culture supernatant were then separately collected for further experiments cytokine level assays. Granzyme B and perforin levels in the supernatants were measured using the mouse granzyme and mouse perforin ELISA kits (Diaclone).

Cytotoxic Activity of P1A CTL: Apoptosis Assay.
To measure the cytotoxic activity of P1A-specific CTLs on 4T1 cells, we detected apoptosis using the Annexin V method. Briefly, demethylated 4T1 cells (10 6 ) were cocultured with P1A-specific CTLs at indicated effector-to-target (E/T) ratios for 16 hours and then stained with Annexin V and 7-AAD for detection by flow cytometry. were maintained and handled under aseptic conditions with 12/12-hour light-dark conditions and had free access to food and water during the study. The in vivo antitumor effect was evaluated with xenografttransplanted BALB/C mice according to the previously reports [13]. 4T1 cells were transplanted into the flank of mice. On day 6, the mice bearing 4T1 cells were randomly divided into four groups ( = 6 per group), and 5-aza (1.0 mg/kg) was injected intraperitoneally twice a day for 5 day. On day 13, 5 × 10 5 specially treated CTLs were injected into the tail vein. Tumor volumes were measured every other day during the whole experiment period and calculated using the reported equation [14]. On day 21, the mice were euthanized, and the tumors were weighed for further analysis. Tumor inhibition rate was calculated as inhibition rate = ( − )/ ×100%, where " " and " " represent the average weights of the control and treatment groups, respectively.

Statistical Analysis.
All data from in vitro and in vivo experiments were analyzed by Pearson 2 tests or Student's t-tests, and values of < 0.05 were considered statistically significant.

DNA Methylation and Gene Expression of CTA P1A in
Cancer Cell Lines. We firstly detected the DNA methylation status of P1A gene in A20, CT26, and 4T1 murine cells using a methylation-specific PCR. After 5-aza treatment, both methylated and unmethylated products were shown in Figure 1(a). We further measured P1A expression levels by RT-PCR and analyzed whether 5-aza-induced P1A mRNA expression was dose-and time-dependent. As shown in Figures 1(b), 1(c), and 1(d), P1A mRNA was undetectable without previous 5-aza, while after exposure to 5-aza, P1A mRNA was induced in 4T1 cells with dose-and timedependent pattern. EL9611, a murine erythroblastic leukemia cell line expressing a high level of P1A [15], was used as positive control. In our studies, 5-azacytidine exposure for 48 h at different concentrations is not cytotoxic to the cells (data not shown).
We also investigated the relationship between the duration of 5-aza exposure and P1A gene expression. We found that a 6-hour 5-aza exposure with a final concentration of 3.0 M was sufficient to induce P1A expression; longer exposure time could increase CTA expression until 48 hour (Figure 1(d)). These results demonstrate that 5-aza can induce P1A expression in a dose-and time-dependent manner in 4T1 cells.

Effect of P1A-Specific Peptide and CpG ODN on DC Phenotype and Function.
We assessed the phenotypes of DCs pulsed with IL-4 coupled with GM-CSF, P1A-specific peptide, CpG ODN, or P1A-specific peptide + CpG ODN. As shown in Figure 2(a), the number of DCs positive for CD80, CD86, and CD11c after induction with P1A peptide + CpG ODN were 64.35 ± 8.2%, 85.41 ± 10.5%, and 83.61 ± 6.4%, respectively, which were significantly higher than those in DCs treated with either P1A-specific or CpG ODN alone ( < 0.05) or IL-4 and GM-CSF ( < 0.01). When mouse BM cells were further cultured in the presence of P1A-specific peptide and CpG ODN for 3 days, IL-6 and TNF-(both markers associated with mature DCs) were secreted into the culture media. As shown in Figure 2(b), IL-6 and TNF-levels were significantly higher in P1A-specific peptide + CpG ODN cells than those in either the P1A-specific peptide or the CpG ODN alone groups ( < 0.01), suggesting that P1A-specific peptide and CpG ODN stimulated DC maturation.

Effect of CpG ODN-and P1A-Specific Peptide-Treated DCs on Allogenic T Lymphocyte Proliferation and Cytotoxic
BioMed Research International Tumor weight in the P1A-specific peptide + CpG ODN treated group was significantly lower than that of the control groups. * < 0.05, * * < 0.01 versus Control; # < 0.05, ## < 0.01 versus group treated with P1A-specific peptide or CpG ODN.
Activity. We cocultured lymphocytes isolated from H-2L drestricted BALB/C mouse spleen with DCs at 20:1 for 96 hours and measured granzyme B and perforin secretion by DC-stimulated CTLs (Figure 3). Both were markedly increased upon treatment with P1A-specific peptide + CpG ODN, illustrating that P1A-specific peptide and CpG ODN enhanced CTL cytotoxicity through increase of granzyme B and perforin release.

In Vitro Cytotoxic Activity of P1A-Specific CTL on H-
To determine the activity of P1Aspecific CTLs against P1A-expression tumor cells, we cultured target cells (4T1) treated with 5-aza coupled with specially treated CTLs at different effector/target ratios for 16 hour. As shown in Figure 4, the proportion of apoptotic 4T1 cells was significantly higher in the P1A-specific peptide + CpG ODN group than that in the P1A-specific peptide or CpG ODN alone or control groups (IL-4 and GM-CSF, < 0.05). Nearly all the tumor cells died when the effect/target ratio reached 25 : 1. These data demonstrate that the cytotoxic effect of P1Aspecific CTLs was enhanced by P1A-specific peptide and CpG ODN stimulation.

Antitumor Effect of H-2L d -Restricted P1A-Specific CTLs In
Vivo. We then studied the effect of specially treated CTLs on tumor growth using a BALB/C tumor-bearing mouse model. Briefly, 2 × 10 3 4T1 cells were transplanted into mice on day 1. Five days later, 5-aza and specially treated CTLs were administered as mentioned above. As shown in Figure 5, P1Aspecific peptide and CpG ODN-stimulated CTLs significantly inhibited the growth of 4T1 cells in 5-aza-treated BALB/C mice. Compared with the control group (not treated with P1A or CpG ODN), after CTL administration, the tumor volume was barely increased compared with pre-treatment ( Figure 5(a)). Tumor weight in P1A-specific peptide and CpG ODN-stimulated CTLs group was significantly lower than that of the control groups ( Figure 5(b)). The inhibitory rate of H-2L d -restricted P1A-specific CTLs on tumor growth reached a level of 79.8%, whereas the inhibitory rate of the control groups was less than 35.5%.

Discussion
Some tumors allow them to escape from immune surveillance and destruction, that have puzzled the oncologist for a long time [16]. Hence, counteracting tumor escape mechanism is a key issue for successful immunotherapy [17]. Although CTAs are considered to be ideal targets for cancer immunotherapy, their expression levels are often decreased or silenced by DNA methylation in the promoter region of CT genes [18], that make it difficult for antigen-specific CTLs to recognize and kill tumor cells. Recent studies have shown that demethylating agents modulating CTA expression could be a useful strategy helpful for cancer immunotherapy. High CTA expression levels facilitate tumor cell or tissue to be recognized by the body's immune system and ultimately lead to tumor clearance by CTLs. In our study, the P1A gene expression was regained with 5-aza in murine tumor cell lines normally without P1A expression, suggesting that the low/no expression of the P1A gene was due to hypermethylation in the CpG region, that was the key mechanism for immunologic escape and tolerance. However, several current studies reported that cancer vaccines have little efficacy if tumor-specific cytotoxins are not incorporated with immune helpers, such as immune adjuvant or helper peptides [19]. When using single epitope wild-type p53, peptide-specific CD8+ T cells were generated in just one-third of healthy donors or subjects with cancer. In this study, we induced DC maturation with a P1A-specific peptide and CpG ODN and found that mature DC phenotypes were upregulated and cytokine release was significantly enhanced. These results suggest that reinduction of the P1A gene incorporated with immune adjuvant CpG ODN may be a feasible and useful cell-based immunotherapy strategy.
TLR agonists have been widely used in cancer therapy due to their inducement to potent antitumor immune responses [20]. Binding of CpG ODN and TLR9 facilitated their ability to induce DC maturation and promoted the differentiation of T helper (Th0) into Th1 [21][22][23]. Effective CTL activation is the key step in cancer immunotherapy. Mature DCs provide an interface between the innate and adaptive immune systems, acting as antigen-presenting cells [23], and may enhance cytotoxic T lymphocyte activation and function. In this study, we cocultured mouse T lymphocytes with P1Aspecific peptide and CpG ODN-stimulated DCs and found that the function of P1A-specific CTLs was enhanced. Additional experiments showed that P1A-specific CTLs possessed antitumor activity against 4T1 tumor cells both in vitro and in vivo, providing strong evidence that reinduction of P1A CTA and CpG ODN stimulation can be used to enhance the efficacy of cancer immune therapy.

Conclusions
We observed that demethylation and CpG ODNs immune adjuvant stimulation facilitated DC maturation and enhanced the capacity of allogenic P1A-specific CTLs against target cells both in vitro and in vivo. And our results suggest that induced CTA and immune adjuvant stimulation is a feasible strategy for adoptive immunotherapy for cancer.