Deficiency of 15-LOX-1 Induces Radioresistance through Downregulation of MacroH2A2 in Colorectal Cancer

Despite the importance of radiation therapy, there are few radiation-related markers available for use in clinical practice. A larger catalog of such biomarkers is required to help clinicians decide when radiotherapy should be replaced with a patient-specific treatment. Arachidonate 15-lipoxygenase (15-LOX-1) enzyme is involved in polyunsaturated fatty acid metabolism. When colorectal cancer (CRC) cells were exposed to radiation, 15-LOX-1 was upregulated. To verify whether 15-LOX-1 protects against or induces DNA damage, we irradiated sh15-LOX-1 stable cells. We found that low 15-LOX-1 is correlated with radioresistance in CRC cells. These data suggest that the presence of 15-LOX-1 can be used as a marker for radiation-induced DNA damage. Consistent with this observation, gene-set-enrichment analysis based on microarray experiments showed that UV_RESPONSE was decreased in sh15-LOX-1 cells compared to shCon cells. Moreover, we discovered that the expression of the histone H2A variant macroH2A2 was sevenfold lower in sh15-LOX-1 cells. Overall, our findings present mechanistic evidence that macroH2A2 is transcriptionally regulated by 15-LOX-1 and suppresses the DNA damage response in irradiated cells by delaying H2AX activation.


Introduction
Colorectal cancer (CRC) is the third most common cancer and the third leading cause of cancer-related deaths worldwide. Surgery is generally recommended for its treatment, but in practice, radiation therapy and chemotherapy are concurrently implemented. Radiation therapy is administered to more than half of all cancer patients worldwide to reduce the tumor size. It is especially important for CRC because it can be performed on the anus in situ and, thus, prevents or delays

Radiation Induces Cell Death and Upregulates 15-Lox-1 Expression
To determine whether 15-LOX-1 expression is regulated by radiation, we irradiated CRC cell lines at 2.5, 5, or 10 Gy. First, we observed cell death upon irradiation. Twenty-four hours after irradiation, cleaved PARP levels ( Figure 2A) and Annexin V-positive cell numbers ( Figure 2B) were increased, as demonstrated by Western blotting and flow cytometry, respectively. Next, we determined the mRNA and protein levels of 15-LOX-1. Real-time PCR and immunocytochemistry (ICC) results showed that 24 h of irradiation significantly upregulated 15-LOX-1 in DLD-1 and HCT8 cells ( Figure 2C,D). However, the 15-LOX-1 levels in HCT116 and HT29 cells were only slightly increased, especially at the protein level, as evidenced by the ICC results. Taken together, these results indicate that radiation induces 15-LOX-1 expression and causes cell death regardless of the p53 status. However, the degree of 15-LOX-1 induction was different in each cell line. A higher

Radiation Induces Cell Death and Upregulates 15-Lox-1 Expression
To determine whether 15-LOX-1 expression is regulated by radiation, we irradiated CRC cell lines at 2.5, 5, or 10 Gy. First, we observed cell death upon irradiation. Twenty-four hours after irradiation, cleaved PARP levels ( Figure 2A) and Annexin V-positive cell numbers ( Figure 2B) were increased, as demonstrated by Western blotting and flow cytometry, respectively. Next, we determined the mRNA and protein levels of 15-LOX-1. Real-time PCR and immunocytochemistry (ICC) results showed that 24 h of irradiation significantly upregulated 15-LOX-1 in DLD-1 and HCT8 cells ( Figure 2C,D). However, the 15-LOX-1 levels in HCT116 and HT29 cells were only slightly increased, especially at the protein level, as evidenced by the ICC results. Taken together, these results indicate that radiation induces 15-LOX-1 expression and causes cell death regardless of the p53 status. However, the degree of 15-LOX-1 induction was different in each cell line. A higher induction of 15-LOX-1 was observed in DLD-1 and HCT8, whose p53 states and radiation sensitivities did not match. induction of 15-LOX-1 was observed in DLD-1 and HCT8, whose p53 states and radiation sensitivities did not match.

The Absence of 15-Lox-1 Decreases Radiation Sensitivity
To investigate the function of radiation-induced upregulation of 15-LOX-1, we generated stable cell lines, using a sh15-LOX-1 expression vector in DLD-1 cells ( Figure 3A). The expression level of 15-LOX-1 in each clone was first measured by Western blotting; qRT-PCR was then conducted to confirm the 15-LOX-1 level in the selected clone (Clone no. 1: shCon stable cell line; Clone no. 3: sh15-LOX-1 stable cell line). The reason for selecting DLD-1 is that 15-LOX-1 is well induced by radiation, as shown in Figure 2C, and this cell line is a mutant for p53. We, thus, used DLD-1 cells to clarify the role of 15-LOX-1 in cells that are not affected by p53 signaling. Next, we investigated whether deficiencies in 15-LOX-1 expression are associated with a radiosensitive phenotype through a colonyforming assay for Clone No. 1 and No. 3 stable cells. Interestingly, sh15-LOX-1 cells were more resistant to IR than shCon cells ( Figure 3B). We obtained the same results by using Western blotting, caspase-3/7 assay, and cell cycle analysis ( Figure 3C-E). The conditions were the same 24 h after irradiation, and each experiment was conducted as described in Materials and Methods. Apoptosis,

The Absence of 15-Lox-1 Decreases Radiation Sensitivity
To investigate the function of radiation-induced upregulation of 15-LOX-1, we generated stable cell lines, using a sh15-LOX-1 expression vector in DLD-1 cells ( Figure 3A). The expression level of 15-LOX-1 in each clone was first measured by Western blotting; qRT-PCR was then conducted to confirm the 15-LOX-1 level in the selected clone (Clone no. 1: shCon stable cell line; Clone no. 3: sh15-LOX-1 stable cell line). The reason for selecting DLD-1 is that 15-LOX-1 is well induced by radiation, as shown in Figure 2C, and this cell line is a mutant for p53. We, thus, used DLD-1 cells to clarify the role of 15-LOX-1 in cells that are not affected by p53 signaling. Next, we investigated whether deficiencies in 15-LOX-1 expression are associated with a radiosensitive phenotype through a colony-forming assay for Clone No. 1 and No. 3 stable cells. Interestingly, sh15-LOX-1 cells were more resistant to IR than shCon cells ( Figure 3B). We obtained the same results by using Western blotting, caspase-3/7 assay, and cell cycle analysis ( Figure 3C-E). The conditions were the same 24 h after irradiation, and each experiment was conducted as described in Materials and Methods. Apoptosis, caspase-3/7 activation, and G2/M arrest occurred less in sh15-LOX-1 than in shCon stable cells. Lastly, cell viability was assessed 24 and 48 h after irradiation in these stable lines, using the WTS assay. The difference of cell viability that did not occur at 24 h after irradiation was significantly confirmed caspase-3/7 activation, and G2/M arrest occurred less in sh15-LOX-1 than in shCon stable cells. Lastly, cell viability was assessed 24 and 48 h after irradiation in these stable lines, using the WTS assay. The difference of cell viability that did not occur at 24 h after irradiation was significantly confirmed at 48 h after irradiation. Inhibition of cell viability by radiation occurred less in sh15-LOX-15 cells than in shCon cells. Together, these results indicate that lack of 15-LOX-1 decreases radiosensitivity in CRC cells.

MacroH2A2 is Transcriptionally Regulated by 15-Lox-1 in Crc Cells
To determine how 15-LOX-1 regulates radiosensitivity, we performed an Affymetrix expression microarray analysis of the stably transfected cell lines in triplicate. Each cell line was exposed to 10

MacroH2A2 is Transcriptionally Regulated by 15-Lox-1 in Crc Cells
To determine how 15-LOX-1 regulates radiosensitivity, we performed an Affymetrix expression microarray analysis of the stably transfected cell lines in triplicate. Each cell line was exposed to 10 Gy. We focused on the H2AFY2 gene from the set of genes downregulated >sixfold ( Figure 4A). We validated our microarray results by evaluating the macroH2A2 mRNA levels in the stable cell lines by qRT-PCR and protein levels by Western blotting (macroH2A2 is the protein encoded by the H2AFY2 gene) ( Figure 4B,C). The macroH2A2 level was found to be reduced in sh15-LOX-1 stable cells with or without radiation. The 15-LOX-1 inhibitor PD146176 downregulated macroH2A2 alongside 15-LOX-1, independent of irradiation ( Figure 4D). We also assessed whether transient downregulation of 15-LOX-1 using RNAi could downregulate macroH2A2. We found that macroH2A2 RNA and protein levels decreased upon transient inhibition of 15-LOX-1 alone ( Figure 4E). However, there was no change in the level of macroH2A2 mRNA when 15-LOX-1 was overexpressed through transfection of DLD-1 and HCT8 cells with the pFlag 15-LOX-1 vector ( Figure 4F). These results indicate that macroH2A2 is transcriptionally regulated by 15-LOX-1 in CRC cells that significantly upregulate the 15-LOX-1 level upon irradiation.

Reduction in macroH2A2 Function Caused by 15-Lox-1 Transcriptional Downregulation Is Involved in Radioresistance through Suppression of the Radiation Response
Although the function of macroH2A2 in response to DNA damage has not been clearly identified yet, we searched for a connection between macroH2A2 function and radiation response, using gene set enrichment analysis (GSEA) of the microarray data. GSEA was performed on expression data, including those regarding 706 genes, using the conditions FDR < 0.01 and |FC| ≥ 0.5. The analysis showed that the UV response was decreased in sh15-LOX-1 stable cells ( Figure 5A). Therefore, we assessed the levels of the DNA DSB marker phosphorylated H2AX (Ser139, γH2AX) during the indicated periods in the stable cell lines. H2AX was activated by 10 Gy after 24 h in shCon, but not sh15-LOX-1 cells ( Figure 5B). It appears that the reduction in 15-LOX-1 activity was associated with a decreased DNA damage response. To investigate the relationship between DNA damage response and 15-LOX-1-regulated macroH2A2 function, DLD-1 cells were transfected with siRNA against macroH2A2 (simacroH2A2) and exposed to 10 Gy 24 h after the transfection. The generation of γH2AX foci by radiation was decreased 24 h after irradiation in simacroH2A2-transfected cells ( Figure 5C). To further determine the extent of DNA damage relative to the expression level of macroH2A2, cells were transfected with simacroH2A2 and exposed to radiation 24 h after transfection. The DNA damage of the cells was evaluated by DNA fragmentation ( Figure 5D) and comet assays ( Figure 5E) 24 h after irradiation. Radiation-induced DNA damage was found to be attenuated upon silencing of macroH2A2 ( Figure 5D,E). In addition, we assessed apoptotic cell death markers by Western blotting. Compared with the DLD-1 cells transfected with siCon, the decrease in macroH2A2 activity in siRNA-treated cells attenuated IR-induced apoptotic cell death ( Figure 5F). The results indicated that reduced macroH2A2 level upon 15-LOX-1 downregulation increases radiation resistance by suppressing the radiation response.

Reduction in macroH2A2 Function Caused by 15-Lox-1 Transcriptional Downregulation Is Involved in Radioresistance through Suppression of the Radiation Response
Although the function of macroH2A2 in response to DNA damage has not been clearly identified yet, we searched for a connection between macroH2A2 function and radiation response, using gene set enrichment analysis (GSEA) of the microarray data. GSEA was performed on expression data, including those regarding 706 genes, using the conditions FDR < 0.01 and |FC| ≥ 0.5. The analysis showed that the UV response was decreased in sh15-LOX-1 stable cells ( Figure 5A). Therefore, we assessed the levels of the DNA DSB marker phosphorylated H2AX (Ser139, γH2AX) during the indicated periods in the stable cell lines. H2AX was activated by 10 Gy after 24 h in shCon, but not sh15-LOX-1 cells ( Figure 5B). It appears that the reduction in 15-LOX-1 activity was associated with a decreased DNA damage response. To investigate the relationship between DNA damage response and 15-LOX-1-regulated macroH2A2 function, DLD-1 cells were transfected with siRNA against macroH2A2 (simacroH2A2) and exposed to 10 Gy 24 h after the transfection. The generation of γH2AX foci by radiation was decreased 24 h after irradiation in simacroH2A2-transfected cells ( Figure 5C). To further determine the extent of DNA damage relative to the expression level of macroH2A2, cells were transfected with simacroH2A2 and exposed to radiation 24 h after transfection. The DNA damage of the cells was evaluated by DNA fragmentation ( Figure 5D) and comet assays ( Figure 5E) 24 h after irradiation. Radiation-induced DNA damage was found to be attenuated upon silencing of macroH2A2 ( Figure 5D,E). In addition, we assessed apoptotic cell death markers by Western blotting. Compared with the DLD-1 cells transfected with siCon, the decrease in macroH2A2 activity in siRNA-treated cells attenuated IR-induced apoptotic cell death ( Figure 5F). The results indicated that reduced macroH2A2 level upon 15-LOX-1 downregulation increases radiation resistance by suppressing the radiation response.

Discussion
Although the relationship between p53 and the radiation response is well-known [25][26][27][28], p53 is not the only factor that controls the radiosensitivity of cells. Previous studies have suggested that 15-LOX-2 and 12-LO are also involved in regulating radiosensitivity both in head and neck and prostate cancers [22][23][24]. Even though radiation therapy is essential for the timely treatment of CRC to prevent anus amputation, few biomarkers are available to predict CRC patient response to radiation therapy. We hypothesized that 15-LOX-1 could be a regulator of radiosensitivity. In our clonogenic cell survival assay, radiosensitivity was found to be correlated with the expression level of 15-LOX-1 protein and p53 mutational status (Figure 1).
Histones are commonly known as nuclear proteins that are involved in DNA repair, DNA replication, and chromosomal stability. MacroH2A2 is one of the several variants of histone H2A. Until this study, macroH2A2 was known to function only in X inactivation [29,30]. This study offers the first demonstration of a relationship between macroH2A2 and the DNA damage response. Although we do not yet understand in detail how macroH2A2 modulates the DNA damage response, γH2AX regulation of macroH2A2 was identified as a novel mechanism ( Figure 5). To better understand the specificity of this regulation, we plan to evaluate the expression levels of DNA DSB sensors, members of the MRN complex (MRE11, Rad50, and NBS1), and ATM in cells deficient for macroH2A2. We will also confirm the binding of these genes because the γH2AX signal does not appear until ATM binds to the MRN complex.
If 15-LOX-1 primarily functioned as a transcription factor, the mechanism by which it regulates macroH2A2 would be more obvious. However, most 15-LOX-1 is found in the cytoplasm and not in the nucleus. Several transcription factors, such as c-Myb, Oct-1, and Oct-2, are known to regulate macroH2A2. We hypothesized that 15-LOX-1 regulates macroH2A2 transcription via these promoterbinding proteins. The protein levels of Oct-1 and Oct-2 were found to decrease in sh15-LOX-1 stable cells, si15-LOX-1 transient cells, and PD146176-treated cells (data not shown). In addition, Oct-2 transcription is regulated by binding to the Oct-2 promoter of PPAR-gamma activated by 15-LOX-1. How these genes correlate with each other will be assessed in the future.
Additionally, the gene set enrichment analysis (GSEA) also suggested that the markers of the epithelial-to-mesenchymal transition (EMT) were significantly upregulated in sh15-LOX-1 stable cells. Especially N-cadherin, which is a hallmark of EMT, was upregulated in sh15-LOX-1 stable cells irrespective of irradiation (Supplementary Figure S1). When these cells were exposed to radiation, their N-cadherin level seemed to decrease slightly but remained higher than that in shCon stable

Discussion
Although the relationship between p53 and the radiation response is well-known [25][26][27][28], p53 is not the only factor that controls the radiosensitivity of cells. Previous studies have suggested that 15-LOX-2 and 12-LO are also involved in regulating radiosensitivity both in head and neck and prostate cancers [22][23][24]. Even though radiation therapy is essential for the timely treatment of CRC to prevent anus amputation, few biomarkers are available to predict CRC patient response to radiation therapy. We hypothesized that 15-LOX-1 could be a regulator of radiosensitivity. In our clonogenic cell survival assay, radiosensitivity was found to be correlated with the expression level of 15-LOX-1 protein and p53 mutational status (Figure 1).
Histones are commonly known as nuclear proteins that are involved in DNA repair, DNA replication, and chromosomal stability. MacroH2A2 is one of the several variants of histone H2A. Until this study, macroH2A2 was known to function only in X inactivation [29,30]. This study offers the first demonstration of a relationship between macroH2A2 and the DNA damage response. Although we do not yet understand in detail how macroH2A2 modulates the DNA damage response, γH2AX regulation of macroH2A2 was identified as a novel mechanism ( Figure 5). To better understand the specificity of this regulation, we plan to evaluate the expression levels of DNA DSB sensors, members of the MRN complex (MRE11, Rad50, and NBS1), and ATM in cells deficient for macroH2A2. We will also confirm the binding of these genes because the γH2AX signal does not appear until ATM binds to the MRN complex.
If 15-LOX-1 primarily functioned as a transcription factor, the mechanism by which it regulates macroH2A2 would be more obvious. However, most 15-LOX-1 is found in the cytoplasm and not in the nucleus. Several transcription factors, such as c-Myb, Oct-1, and Oct-2, are known to regulate macroH2A2. We hypothesized that 15-LOX-1 regulates macroH2A2 transcription via these promoter-binding proteins. The protein levels of Oct-1 and Oct-2 were found to decrease in sh15-LOX-1 stable cells, si15-LOX-1 transient cells, and PD146176-treated cells (data not shown). In addition, Oct-2 transcription is regulated by binding to the Oct-2 promoter of PPAR-gamma activated by 15-LOX-1. How these genes correlate with each other will be assessed in the future.
Additionally, the gene set enrichment analysis (GSEA) also suggested that the markers of the epithelial-to-mesenchymal transition (EMT) were significantly upregulated in sh15-LOX-1 stable cells. Especially N-cadherin, which is a hallmark of EMT, was upregulated in sh15-LOX-1 stable cells irrespective of irradiation (Supplementary Figure S1). When these cells were exposed to radiation, their N-cadherin level seemed to decrease slightly but remained higher than that in shCon stable cells. Radiation-induced EMT is one of the mechanisms underlying resistance to radiation [31,32]. Therefore, it will be another interesting topic in the future.
Patient-specific treatment is becoming more common for the elimination of unnecessary chemotherapy, but it is used less often for radiotherapy. The discovery and characterization of biomarkers that can predict the efficacy of radiation therapy has the potential to improve cancer treatment. This study suggests that 15-LOX-1 has significant potential as a biomarker indicating the radiosensitivity of CRC cells.

Cell Culture
The human CRC cell lines DLD-1, HCT8, HT29, and HCT116 were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). Cells were grown in RPMI 1640 and McCoy's 5A medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) and L-glutamine, in a humidified chamber, at 37 • C, and under 5% CO 2 .

Western Blot Analysis
Proteins were extracted, using RIPA buffer supplemented with protease and phosphatase inhibitors (Sigma-Aldrich, St. Louis, MO, USA). Equal amounts of proteins were analyzed by SDS PAGE after quantitation through the bicinchoninic acid (BCA, Thermo Fisher Scientific, Hudson, NH, USA) protein assay. Membranes with the transferred proteins were blocked with skim milk and sequentially incubated with primary and secondary antibodies. Chemiluminescent reagents (DoGen, Seoul, Korea) were used to visualize protein levels on X-ray films. Detailed information of western blot can be found at Supplementary Figure S2.

Clonogenic Cell Survival Assay
CRC cells were counted and plated (5 × 10 2 cells per 60 mm dish) in triplicate and exposed to various doses of radiation the next day. Depending on their clonogenic ability, cells were grown for 14-18 days. Cells were finally washed and stained with crystal violet. Colonies containing >50 cells were counted. Specifically, 137Cs-γ-rays (Atomic Energy of Canada) were used for in vitro experiments, at a dose rate of 3.81 Gy/min.

Microarray Analysis
Stable cells were irradiated with 10 Gy radiation. After 6 h, RNA was extracted, using the TRIzol reagent. Reverse transcription was carried out, and the Affymetrix Whole Transcript Expression array was used, according to the manufacturer's instructions, for microarray analysis (GeneChip Whole Transcript PLUS reagent kit and GeneChip Whole Transcript Amplification kit). The cDNA labeling (GeneChip WT Terminal labeling kit), hybridization (Affymetrix GeneChip 2.0 ST Array), and slide scanning (Affymetrix, GCS3000 Scanner) were performed according to the manufacturers' instructions.

Cell-Cycle Assay
Cells were harvested, washed twice with PBS, and fixed in 85% ethanol containing 5 mM EDTA. The fixed cells were treated with 20 µg/mL RNaseA, followed by staining with 50 µg/mL propidium iodide (PI, Sigma-Aldrich). DNA content was determined through the use of a Beckman Coulter flow cytometer. Results were processed by using the FlowJo program for the determination of cell numbers in the G1, S, and G2/M phases.

Analysis of 15-LOX-1 Using Flow Cytometry
Cultured cells were trypsinized and harvested in a conical tube. Subsequently, they were fixed in 3.7% paraformaldehyde for 15 min and permeabilized with 90% methanol for 5 min, at 25 • C. Cells were incubated with FITC-conjugated anti-15-LOX-1 (sc-133085 FITC) for 30 min, in an ice bucket protected from light. Unstained cells were used as a negative control.

Immunofluorescent Staining
Cells plated on glass coverslips were exposed to radiation; then, they were fixed and permeabilized with 3.7% paraformaldehyde and 0.5% Triton X-100, respectively, for 15 min each, at 25 • C, and blocked with 3% bovine serum albumin for 30 min, in a covered ice bucket, the following day. After blocking, cells were incubated with FITC-conjugated anti-15-LOX-1 and anti-γH2AX, overnight, at 4 • C. Lastly, cells were washed and stained with Alexa Fluor 488-labeled secondary antibody (only for γH2AX) and DAPI, in consecutive order.

Annexin V Assay
Collected cells were resuspended in the binding buffer and stained with 1.25 µL of Annexin V-FITC and 10 µL of propidium iodide (PI) for 30 min, at 4 • C (ApoScan Annexin V-FITC apoptosis detection Kit, BioBud, Seoul, Korea). The results were analyzed with a BD flow cytometer.

Cell Viability Assay
Cells were grown in 96-well plates and exposed to 10 Gy radiation. To confirm the effect of radiation, cells were incubated for 24 or 48 h. Cell viability was measured by using the EZ-Cytox kit reagent (DoGen, Seoul, Korea). Absorbance at 450 nm was used to analyze cell viability.

Apoptotic DNA Ladder Detection
Cells were lysed with 35 µL of TE Lysis Buffer and then incubated with 5 µL each of Enzyme A and Enzyme B, to digest RNA and protein, respectively. Next, the DNA of the cells was precipitated, using 5 µL of ammonium acetate. The DNA pellet was resuspended in the DNA suspension buffer and analyzed by 1.2% agarose gel electrophoresis. The results were visualized through the use of a transilluminator.

Comet Assay
Cells (1 × 10 5 cells/mL) were mixed with Comet Agarose at a 1:10 ratio and plated onto a Comet Slide. The slide was then incubated at 4 • C, for 15 min, to harden the agarose. Then, it was sequentially immersed in prechilled lysis buffer and alkaline solution for 30 min, at 4 • C. The Comet Agarose gel was stained with Vista Green DNA Dye after electrophoresis (OxiSelect Comet Assay Kit, Cell Biolabs, Inc., San Diego, CA, USA).

Statistical Analysis
Statistical analyses were carried out by using the Prism5 software (GraphPad Software, Inc., San Diego, CA, USA). Results are expressed as the mean ± standard error of the mean (SEM). All the results were evaluated by using unpaired Student's t-test, and a p-value < 0.05 was considered significant.

Conclusions
In this paper, 15-LOX-1 is proposed as a biomarker for the radiosensitivity of which cannot be determined based on the p53 status.