Association between long interspersed nuclear element-1 methylation levels and relapse in Wilms tumors

Author(s)

de Sa Pereira, Bruna M. ; Montalvao-de-Azevedo, Rafaela ; Faria, Paulo Antonio ; Silva, Neimar de Paula ; Nicolau-Neto, Pedro ; Maschietto, Mariana ; de Camargo, Beatriz ; Soares Lima, Sheila Coelho

• Document type: Journal article
• Source: CLINICAL EPIGENETICS; v. 9, p. 7-pg., 2017-12-12.

Abstract

Background: Wilms tumor (WT) is a curable pediatric renal malignancy, but there is a need for new molecular biomarkers to improve relapse risk-directed therapy. Somatic alterations occur at relatively low frequencies whereas epigenetic changes at 11p15 are the most common aberration. We analyzed long interspersed element-1 (LINE-1) methylation levels in the blastemal component of WT and normal kidney samples to explore their prognostic significance. Results: WT samples presented a hypomethylated pattern at all five CpG sites compared to matched normal kidney samples; therefore, the averaged methylation levels of the five CpG sites were used for further analyses. WT presented a hypomethylation profile (median 65.0%, 47.4-73.2%) compared to normal kidney samples (median 71.8%, 51.5-77. 5%; p < 0.0001). No significant associations were found between LINE-1 methylation levels and clinical-pathological characteristics. We observed that LINE-1 methylation levels were lower in tumor samples from patients with relapse (median methylation 60.5%) compared to patients without relapse (median methylation 66.5%; p = 0.0005), and a receiving operating characteristic curve analysis was applied to verify the ability of LINE-1 methylation levels to discriminate WT samples from these patients. Using a cut-off value of 62.71% for LINE-1 methylation levels, the area under the curve was 0.808, with a sensitivity of 76.5% and a specificity of 83.3%. Having identified differences in LINE-1 methylation between WT samples from patients with and without relapse in this cohort, we evaluated other prognostic factors using a logistic regression model. This analysis showed that in risk stratification, LINE-1 methylation level was an independent variable for relapse risk: the lower the methylation levels, the higher the risk of relapse. The logistic regression model indicated a relapse risk increase of 30% per decreased unit of methylation (odds ratio 1.30; 95% confidence interval 1.07-1.57). Conclusion: Our results reinforce previous data showing a global hypomethylation profile in WT. LINE-1 methylation levels can be suggested as a marker of relapse after chemotherapy treatment in addition to risk classification, helping to guide new treatment approaches. (AU)

• FAPESP's process: 15/06281-7 - CHARACTERIZATION OF THE EPIGENETIC REGULATION IN HUMAN SOLID PAEDIATRIC TUMOURS
• Grantee: Mariana Camargo Maschietto
• Support Opportunities: Scholarships in Brazil - Young Researchers