P-215 Biomarkers of Neoplasm in Ulcerative-Colitis

Abstract

BACKGROUND: Colorectal cancer (CRC) remains a serious complication in ulcerative colitis (UC). Biomarkers that identify patients at greatest risk for this outcome could increase the effectiveness of current surveillance methods. Our previous work compared single nucleotide polymorphisms (SNP) and methylation between UC patients who developed CRC (UC-CRC) and UC patients who did not develop CRC (UC-controls). In that work, we studied methylation in DNA extracted from tissue obtained not only in the cancer section, but also from the non-adjacent, non-neoplastic tissue. These initial studies yielded both a potential "risk" SNP in the tumor necrosis factor alpha gene (TNFa) plus 3 markers with significant alterations in methylation status, RUNX3, MINT1 and COX2. The aims of the current study were to: 1.) validate methylation markers in earlier neoplasms (low-grade dysplasia, LGD), 2.) determine if TNFalpha is a marker of "risk" to develop any neoplasm and 3.) determine if these markers were associated with a LGD-non-progressor phenotype. METHODS: We identified/reviewed 200 UC-LGD patients for potential inclusion. After chart review and histological confirmation, 162 were selected for testing. For each sample, formalin-fixed, paraffin-embedded sections were cut both from the dysplastic area (LGD-DYS) as well as a non-adjacent, non-neoplastic area (LGD-NAN). DNA was extracted using a modified Gentra protocol and 500 ng bisulfite treated using the EZ DNA Methylation Gold Kit (Zymo Research). TNFa genotyping was completed using TaqMan assay (Applied Biosystems). Methylation testing for RUNX3, MINT1 and COX2 was done using MS-PCR as previously described by our group. Comparisons were made between LGD-NAN, UC-controls (prior data) and UC-CRC cases (prior data). LGD-NAN results were also combined with UC-CRC data and compared to UC-controls. The subset of LGD patients who did not progress to CRC in long-term follow up (LGD-NP) was compared to UC-CRC and UC-control data. Statistical analysis was completed using JMP. RESULTS: Table 1 summarizes the Validation (A) and the Non-progressive LGD (B) analyses. P-values and sample numbers for each comparison are provided. *Validation of methylation as a marker for concurrent neoplasm in non-adjacent mucosa: RUNX3, MINT1 and COX2 have altered methylation compared to UC-controls, even in patients with early neoplasm in the form of LGD. * TNFa-308 is associated with neoplasm in UC: Comparisons between UCcontrols and any neoplasm were significant. Comparisons within neoplastic groups (UC-CRC, LGD-DYS and LGD-NP) were not significant. * Methylation in non-progressor LGD: The only variation between UC-controls and LGD-NP-NAN was methylation status of COX2, with LGD-NP-NAN more likely to be unmethylated than UC-controls. In comparisons between LGD-NP and UC-CRC (either NAN or DYS), COX2 was not significant but both RUNX3 and MINT1 were. CONCLUSIONS: Results in Table1A confirm our previous findings that TNFa-308 (G>A) is a risk SNP and that this methylation panel can be used to increase standard surveillance sensitivity. Although the sample set is small, the data for LGD-NP suggest that altered methylation patterns may be informative in guiding treatment of LGD. In addition, the alteration of COX2 between LGD-NP-NAN and UC-controls may suggest a possible target for chemoprevention.