A validation study of potential prognostic DNA methylation biomarkers in patients with acute myeloid leukemia using a custom DNA methylation sequencing panel

Abstract

Background Multiple studies have reported the prognostic impact of DNA methylation changes in acute myeloid leukemia (AML). However, these epigenetic markers have not been thoroughly validated and therefore are still not considered in clinical practice. Hence, we aimed to independently verify results of selected studies describing the relationship between DNA methylation of specific genes and their prognostic potential in predicting overall survival (OS) and event-free survival (EFS). Results Fourteen studies (published 2011–2019) comprising of 27 genes were subjected to validation by a custom NGS-based sequencing panel in 178 newly diagnosed non-M3 AML patients treated by 3 + 7 induction regimen. The results were considered as successfully validated, if both the log-rank test and multivariate Cox regression analysis had a p-value ≤ 0.05. The predictive role of DNA methylation was confirmed for three studies comprising of four genes: CEBPA (OS: p = 0.02; EFS: p = 0.03), PBX3 (EFS: p = 0.01), LZTS2 (OS: p = 0.05; EFS: p = 0.0003), and NR6A1 (OS: p = 0.004; EFS: p = 0.0003). For all of these genes, higher methylation was an indicator of longer survival. Concurrent higher methylation of both LZTS2 and NR6A1 was highly significant for survival in cytogenetically normal (CN) AML group (OS: p < 0.0001; EFS: p < 0.0001) as well as for the whole AML cohort (OS: p = 0.01; EFS < 0.0001). In contrast, for two studies reporting the poor prognostic effect of higher GPX3 and DLX4 methylation, we found the exact opposite, again linking higher GPX3 (OS: p = 0.006; EFS: p < 0.0001) and DLX4 (OS: p = 0.03; EFS = 0.03) methylation to a favorable treatment outcome. Individual gene significance levels refer to the outcomes of multivariate Cox regression analysis. Conclusions Out of twenty-seven genes subjected to DNA methylation validation, a prognostic role was observed for six genes. Therefore, independent validation studies are necessary to reveal truly prognostic DNA methylation changes and to enable the introduction of these promising epigenetic markers into clinical practice.

MethScore as a new comprehensive DNA methylation-based value refining the prognosis in acute myeloid leukemia

“…Therefore, we designed a custom NGS-based DNA methylation panel comprising of loci/genes from selected studies as well as genes generally connected to AML pathogenesis such as HOX genes and WT1 . Apart from our recent study [ 7 ] that validated individual DNA methylation changes and utilized the same AML cohort as well as the DNA methylation sequencing panel, we now wanted to evaluate all potential epigenetic markers comprehensively at once.…”

“…List of regions targeted by the methylation sequencing panel (according to the Human GRCh37/hg19 genome assembly) is shown in Additional file 1 . This DNA methylation panel was utilized also in our previous DNA methylation validation study [ 7 ]. Apart from the current study, we examined selected individual genes and validated their influence on AML prognosis separately.…”

Role of Biomarkers in the Management of Acute Myeloid Leukemia

“…Another group found that the methylation status of ATP11A, ITGAM, and ZNRF2 served as accurate prognostic markers in AML patients [ 100 ]. Sestakova et al analyzed previously-touted prognostic DNA methylation markers in AML across fourteen published studies, and confirmed methylation of four genes ( CEBPA, PBX3, LZTS2, and NR6A1 ) as predictive for longer survival, as well as two other genes ( DLX4 and GPX3 ) [ 101 ]. Despite some promising results, methylation patterns of these or other genes are not yet used in clinical practice for AML patients.…”