5-Aza-2′-deoxycytidine (5-Aza-dC) and Trichostatin A (TSA) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Antibodies recognizing histone H3 and acetylated histone H3 were from Millipore (Lake Placid, NY, USA). Antibody recognizing HDAC1 were from Santa Cruz. Antibodies recognizing HDAC2, HDAC3, DNMT1, DNMT3a and DNMT3b were from Abcam. Human colorectal carcinoma cell line HCT116 and SW620 cells were purchased from ATCC. All cells were cultured in media designated by ATCC for propagation and maintenance. Cells were incubated at 37°C incubator with 5% CO₂ and tested to ensure absence of Mycoplasma.

Tumor cells were irradiated by using an RS-2000 Biological X-ray Irradiator (Rad Source Technology, Suwanee, GA, USA). Cells were irradiated at a dose rate of 2 Gy/min for 2.5 min by setting irradiator voltage and current at 160 kV and 25 mA. During irradiation, the cells were maintained in recommended media and kept on ice. Following irradiation, the culture media was replaced with the fresh media.

HCT116 cells (p53 WT) (45) and SW620 cells (p53 mutant) (46) were irradiated (5 Gy), TSA (125 nM) or 5-AZA-dC (20 μM)- treated. Cell surface staining of tumor and normal cells were performed using the following primary labeled antibodies; 4-1BBL-PE, OX-40L-PE, Fas-PE, DR4-PE, and DR5-APC and the appropriate isotype matched controls (BioLegend. San Diego, CA, USA). Surface staining was performed in cell staining buffer for 1 h on ice. Stained cells were acquired on a BD Fortessa flow cytometer (BD Pharmingen, San Diego, CA, USA). Dead cells were excluded from the analysis based on scatter profile. Isotype control staining was <5% for all samples analyzed.

Cell lysates were electrophoresed through polyacrylamide gels and transferred onto nitrocellulose membranes. Membranes were incubated with antibodies specific for histone H3 (Cell Signaling Technology, Danvers, TX, USA), acetylated histone H3 (Millipore, Lake Placid, NY, USA), HDAC1 (Santa Cruz Biotechnology), HDAC2, HDAC3, DNMT1, DNMT3a and DNMT3b (Abcam, Cambridge, MA, USA), then revealed with secondary horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology) and HRP-conjugated anti-mouse IgG (Promega, Madison, WI, USA).

ChIP assays were performed as previously described in Cacan et al (30). Briefly, HCT116 cells were irradiated (5 Gy) or TSA (125 nM)- treated. Cells were then cross-linked with 1% formaldehyde and cell nuclei were isolated and concentrated. Cell nuclei were sonicated using a Bioruptor to generate an average of 500–700 bp of sheared DNA. Sonicated lysates were then precleared with salmon-sperm/agarose beads (Upstate) and 5% of the total lysate was stored as input for normalization. Half of the remaining lysate was immunoprecipitated with control antibody, and the other half was immunoprecipitated with 5 μg of indicated antibody overnight at 4°C. Following an additional 2-h immunoprecipitation with salmon-sperm/ agarose beads, all samples were washed and DNA was isolated. Isolated DNA was quantified by real-time PCR on an ABI PRISM 7900 (Applied Biosystems, Foster City, CA, USA) using the following primers and probe for CD95/Fas: forward, 5′-TCG AGG TCC TCA CCT GAA G-3′, reverse, 5′-TGC ACA AAT GGG CAT TCC T-3′ and probe, 5′-CCA GCC ACT GCA GGA ACG CC-3′; for DR5: forward, 5′-CCC AAG TGC CTC CCT CAA-3′, reverse, 5′-CCG GGC TGT GGT TTG TTT C-3′ and probe, 5′-CCC CAA GTT TCG GTG CCT GTC CT-3′; for 4-1BBL: forward, 5′-CAG CAA GAG ACC CAC AGA GAT-3′, reverse, 5′-CTC CCT GCT GTC TCT CCA T-3′ and probe, 5′-AGA GCG GCA GAG AGG GAG AAC C-3′; for OX-40L: forward, 5′-GGG TTC AAT GTT TAG TTA CA-3′, reverse, 5′-GGG TCA CTT GGT AAA GAT A-3′ and probe, 5′-GCT GCT ACA TTA ACG GAT GAT GAT T-3′; and for GAPDH: forward, 5′-AAT GAA TGG GCA GCC GTT A-3′, reverse, 5′-TAG CCT CGC TCC ACC TGA CT-3′ and probe, 5′-CCT GCC GGT GAC TAA CCC TGC GCT CCT-3′. Values generated from real-time PCR reactions were calculated based on standard curves generated, were run in triplicate reactions and were analyzed using the SDS 2.0 program.

HCT116 and SW620 cells were plated or irradiated (5Gy) and then incubated for 48 h. Genomic DNA was extracted using the E.Z.N.A. kit according to the manufacturer's protocol. Using the EZ-DNA Methylation-Direct ™ Kit (Zymo Research, Irvine, CA, USA), 750 ng of DNA was bisulfite converted per the manufacturer's instructions. The primers for detection of unmethylated Fas gene were 5′-ATA GGA ATG TTT ATT TGT GTA ATG A-3′ (forward) and 5′-CAA AAT CAA AAA CAA ACT CAC AAA-3′ (reverse) to generate a 186-bp product. The cycling conditions were 2 min at 94°C, followed by 34 cycles of 30sec at 94°C, 30 sec at 48°C and 30 sec at 72°C. To detect the methylated Fas gene, 5′-GGG ATA GGA ATG TTT ATT TGT GTA AC-3′ (forward) and 5′-GAA ATC AAA AAC GAA CTC ACG A-3′ (reverse) were used to generate a 188-bp product. The cycling conditions were 2 min at 94°C, followed by 34 cycles of 30 sec at 94°C, 30 sec at 52°C and 30 sec at 72°C. The primers for detection of methylated β-actin gene were 5′-TTT TGC AGG GTT CAC CCT CCT G-3′ (forward) and 5′-CGA GCA TCC CCC AAA GTT CAC AA-3′ (reverse). The cycling conditions were 2 min at 94°C, followed by 34 cycles of 30 sec at 94°C, 30 sec at 58°C and 30 sec at 72°C.

Statistical differences between groups were calculated using Student's t-test and calculated at 95% confidence. Values represent mean ± SEM of three independent experiments. The p-values <0.05 are indicated by one asterisk (*). The p-values <0.005 are indicated by two asterisks (**). The p-values <0.0005 are indicated by three asterisks (***).

Epigenetic regulation and suppression of death receptors have been reported in lung, medulloblastoma, and melanoma tumor cell lines (17,47,48), and we have recently reported that expression of OX-40L and 4-1BBL co-stimulatory molecules appear to be sensitive to epigenetic modifications in CRC cells (9). Collectively, these proteins are important in influencing the sensitivity of tumor cells to antitumor immune cells as well as providing positive signals to these responding cells. Interestingly, the expression of these molecules has also been reported to be modulated following irradiation of tumor cells (9,10), however, whether radiation is altering expression of these molecules through specific chromatin remodeling enzymes remains unknown. We first sought to determine if pharmacologic inhibitors of histone deacetylation and DNA methylation could alter the expression of multiple death receptors in colorectal cancer cells in a manner similar to previous observations of co-stimulatory molecules (9). The HDAC inhibitor, TSA, and DNMT inhibitor, 5-Aza-dC, were used to inhibit HDACs and DNMTs activity, respectively. HCT116 cells (p53 WT) treated with 125 nM TSA for 2 days, or with 20 μM 5-Aza-dC for 3 days, were compared to cells irradiated with 5 Gy and cell surface protein expression of DR4, DR5, Fas, 4-1BBL and OX-40L was quantified by flow cytometry. Both IR and 5-Aza-dC treatment significantly increased cell surface protein expression of DR4 in HCT116 colorectal carcinoma cells while TSA resulted in a less robust increase that was not significant (Fig. 1A). HCT116 cells uniformly express DR5 (100% positive) and Fas (>98% positive) on their surface (Fig. 1B and C) and there was no significant decrease in the frequency of cells expressing DR5 and Fas. However, changes in the median fluorescence intensity (MFI) values show a significant change in the expression (density) levels of both DR5 and Fas following radiation, TSA or 5-Aza-dC treatment (Fig. 1, insets B and C). Interestingly, radiation increased both Fas and DR5 MFI more robustly than TSA or 5Aza-dC in these cells. Inhibition of DNA methylation robustly increased expression of all three death receptors with a maximum increase of 3-fold seen in Fas (Fig. 1C). We previously reported that inhibition of HDACs by TSA and treatment with 10 Gy IR could influence expression of co-stimulatory molecules on HCT116 tumor cells. We further evaluated HCT116 cells to determine how significant these differences were and if they occurred at a dose of radiation <10 Gy. In addition, having previously observed increased expression of co-stimulatory molecule mRNA following inhibition of DNMT using 5-Aza-dC in these cells, we wanted to determine if this treatment could also increase surface expression of the protein. Inhibition of DNMT, HDACs and 5 Gy irradiation significantly increased 41BBL surface expression (Fig. 1D), while only IR significantly increased OX-40L surface expression. While treatment with TSA or 5-Aza-dC appeared to mildly increase OX-40L expression in HCT116 cells, the difference was not significant (Fig. 1E). These data suggest that the contribution of DNA methylation and histone acetylation on the regulation of these two co-stimulatory genes is different.

To confirm that these changes are not limited to HCT116 cells and could occur similarly in another CRC cell line, we evaluated surface expression in treated SW620 (p53 mutated) colorectal cancer cells. SW620 cells were irradiated, treated with TSA or 5-Aza-dC and surface protein expression was measured by flow cytometry. Consistent with observations in HCT116 cells, the treatments upregulated the surface expression of DR5, Fas and 4-1BBL most robustly in SW620 cells (Fig. 1G–I). Interestingly, while TSA treatment did appear to increase expression of DR4 and OX-40L the difference was not significant (Fig. 1F and J). Furthermore, we observed a much more robust increase in the expression of OX-40L 48 h post-IR than previously observed at 24 h post-IR using a higher dose. In contrast to the significant increase in surface expression of 4-1BBL seen with all three treatments, only IR and 5-Aza-dC induced a significant increase in the expression of OX40L. While the magnitude of 4-1BBL increase was similar between IR-treated cells and cells treated with the epigenetic drugs, the increase in OX-40L was much lower in drug-treated cells than observed in irradiated cells. Overall, these data indicate that expression of these genes collectively is most consistently increased following exposure to radiation in CRC cell lines. Moreover, we observed that TSA and 5-Aza-dC treatments were also able to significantly enhance expression of DR5, Fas and 4-1BBL across the cell lines.

Our data indicate that expression of death receptors and co-stimulatory molecules can be influenced by drugs that inhibit epigenetic activities in the cell, and that radiation also increases expression of these genes in CRC cell lines. We observed a significant increase in the expression of DR5, Fas, 4-1BBL and OX-40L following IR in both colorectal cancer cells. We previously reported that 10 Gy IR increases histone acetylation at the 4-1BBL promoter (9). To explore whether histone modifications also regulate death receptor expression post-IR, we first assessed the levels of histone acetylation at the promoters of DR5 and Fas by ChIP assays in both untreated and irradiated HCT116 cells. TSA inhibits HDACs activity and TSA-treated HCT116 cells were used as a positive control for promoter acetylation. As expected, we observed a robust increase in histone acetylation status following TSA treatment. Histone H3 acetylation levels significantly increased at both the Fas and DR5 promoters following 5 Gy radiation as compared to untreated control cells (Fig. 2A and B), while we observed similar amounts of total H3 histone binding at Fas, DR5 promoters (Fig. 2D and E). GAPDH was evaluated in parallel since this gene is stably and constitutively expressed at high levels in most tissues and cells. In contrast, to the increased acetylation we saw at the death receptors promoters, we observed no change in the levels of acetylated histone H3 at the GAPDH promoter between untreated and irradiated HCT116 cells (Fig. 2C), nor any change in the total H3 histone binding (Fig. 2F). As GAPDH is gene that is constitutively on and thus in an open, or acetylated, conformation; we also evaluated a gene that would be expected to be closed and remain closed in tumor cells. For this purpose, we looked at the global histone H3 and acetylated histone H3 levels at the promoter of CD4 and did not observe any alteration in acetylated histone H3 levels at this promoter following radiation treatment (data not shown). These data suggest that radiation is specifically causing promoter acetylation of Fas and DR5 and not uniformly across all genes.

We previously reported that 10 Gy IR significantly increased H3 acetylation at the 4-1BBL promoter in CRC cells and wanted to determine if the same occurred at the promoter of another effector T cell co-stimulatory molecule. For this we assessed the levels of histone acetylation at the promoter of OX-40L by ChIP assay. Again, TSA-treated HCT116 cells were used as a positive control for promoter acetylation. Since we have already reported enhanced histone acetylation at 4-1BBL promoter following 10 Gy IR, here we used 4-1BBL promoter acetylation as a positive control but also to determine if 5 Gy IR was sufficient to alter histone acetylation status of this promoter. We found that histone H3 acetylation levels significantly increased at the 4-1BBL promoter following 5 Gy radiation as compared to untreated control cells (Fig. 3A). Interestingly, we observed a 2-fold increase in histone acetylation level at the OX-40L promoter following 5 Gy IR, however the difference was not significant (Fig. 3B). Thus, we saw no significant change in OX-40L surface expression following TSA treatment in either HCT116 nor SW620 cells (Fig. 1E and J), and no significant increase in acetylation at the promoter for this gene. These data suggest that, in contrast to 4-1BBL expression, OX-40L expression is not strongly regulated by histone acetylation. As before, we observed no change in the overall levels of H3 and total H3 histone binding was similar at 4-1BBL and OX-40L promoters (Fig. 3C and D), suggesting that the increased acetylation at these promoters is not simply due to increased overall histone H3 levels in the cells. Overall, our data suggest that radiation increases the expression of some co-stimulatory molecules and death receptors by increasing promoter histone acetylation. This would facilitate transcription initiation by loosening interactions between the histones and DNA; however, the regulation of these genes by histone acetylation shows variability across the genes. We further determined the total cellular histone H3 and acetylated histone H3 protein levels in HCT116 cells. TSA-treated cells were used as positive control for histone acetylation and we found that 5 Gy IR did not alter total cellular H3 or AcH3 levels in HCT116 cells (Fig. 3G). These data suggest that IR induces histone acetylation in a gene specific manner.

Histone acetylation is dynamically regulated in cells by the opposing actions of histone acetyltransferases (that add the acetyl functional group to histones), and HDACs (that remove them). We have observed an increase in histone acetylation levels at promoters of co-stimulatory genes and death receptors following radiation treatment of cancer cells. To determine the potential mechanism that could be responsible for the enhancement of histone acetylation following radiation treatment, we sought to evaluate the binding of HDACs to the promoter region of Fas and 4-1BBL since we observed a robust and significant increase in histone acetylation levels at the promoter region of both of these genes. HCT116 cells were irradiated (5 Gy) and ChIP assays were performed to determine binding of class I HDACs to the promoter regions of Fas, 4-1BBL and control gene, GAPDH. We observed that HDAC2 and HDAC3 proteins bind with decreased frequency to the Fas promoter in radiation-treated cells compared to untreated cells (Fig. 4B and C), and the change in HDAC2 binding was more significant than that observed with HDAC3. We did not observe any alterations in the binding of HDAC1 to the Fas promoter between control and irradiated cells (Fig. 4A). Similarly, we observed a significant decrease in binding of both HDAC2 and HDAC3 to the 4-1BBL promoter, and no change in binding of HDAC1 to the promoter. These data suggest that HDAC2 and HDAC3 play a role in regulating Fas and 4-1BBL transcription following radiation of tumor cells. Moreover, radiation mediated decrease in HDAC2 and HDAC3 binding to the Fas and 4-1BBL promoters is gene specific, as there was no significant change in the level of HDAC association with the GAPDH promoter between irradiated and control cells (Fig. 4G–I). To determine if radiation was simply reducing the overall cellular levels of HDAC2 and HDAC3, we performed western blot analysis. HDAC1, HDAC2 and HDAC3 expression levels were similar in both radiation-treated and untreated cells (Fig. 4J). Together, these data indicate that radiation increases expression of Fas and 4-1BBL molecules, in part, by decreasing HDAC2 and HDAC3 accumulation at the Fas and 4-1BBL promoters in colorectal cancer cells.

DNMTs are responsible for methylating DNA and for silencing genes. Recent studies suggest that radiation treatment alters expression of some target genes by changing DNA methylation status of these genes (49). We observed that treatment of CRC cells with 5-Aza-dC or IR increased surface expression of Fas and 4-1BBL in both cell lines evaluated (Fig. 1), suggesting that expression of these molecules is regulated, in part, by DNMTs. However, the role of methylating enzymes in expression of co-stimulatory molecules and death receptors in response to radiation has not been investigated. To determine whether the Fas promoter is methylated in colorectal cancer cells, we utilized methylation specific PCR. We observed that Fas promoter is methylated in HCT116 and SW620 colorectal cancer cells and 5 Gy radiation decreased DNA methylation at Fas promoter in both cell lines (Fig. 5A and B). To determine if radiation treatment causes changes in DNA methylating enzyme status at co-stimulatory and death receptor genes, the human colorectal cell line HCT116 was irradiated and DNMTs binding to the Fas and 4-1BBL promoters was assessed by ChIP assays. We observed a similar binding of DNMT3a and DNMT3b to the Fas promoter following radiation treatment (Fig. 6B and C). However, we detected a significant decrease in DNMT1 binding to the promoter region of Fas (Fig. 6A). Radiation induced similar changes at the 4-1BBL promoter, and there was no change in the binding of either DNMT3a and DNMT3b (Fig. 6E and F), and a decrease in the binding of DNMT1 (Fig. 6D) following radiation treatment. Conversely, 5 Gy radiation did not cause any alteration in the binding of DNMTs to the promoter of the housekeeping gene, GAPDH (Fig. 6G–I). Further, we determined that overall cellular DNMT1, DNMT3a and DNMT3b levels were unchanged by 5 Gy radiation (Fig. 6J), suggesting that radiation was specifically reducing the binding of DNMT1 to these promoters. These data suggest that DNMT1 also contributes to regulation of Fas and 4-1BBL expression post-IR.