BackgroundThe prognostic value of aberrant DNA methylation of cell-free circulating DNA in plasma has not previously been evaluated in diffuse large B cell lymphoma (DLBCL). The aim of this study was to investigate if aberrant promoter DNA methylation can be detected in plasma from DLBCL patients and to evaluate this as a prognostic marker. Furthermore, we wanted to follow possible changes in methylation levels during treatment. Seventy-four patients were enrolled in the study, of which 59 received rituximab and CHOP-like chemotherapy. Plasma samples were collected from all patients at the time of diagnosis and from 14 healthy individuals used as controls. In addition, plasma samples were collected during and after treatment for surviving patients. In total, 158 plasma samples were analyzed for DNA methylation in the promoter regions of DAPK (DAPK1), DBC1, MIR34A, and MIR34B/C using pyrosequencing.ResultsAberrant methylation levels at the time of diagnosis were detected in 19, 16, 8, and 10 % of the DLBCL plasma samples for DAPK1, DBC1, MIR34A, and MIR34B/C, respectively. DAPK1 methylation levels were significantly correlated with DBC1 and MIR34B/C methylation levels (P < 0.001). For the entire cohort, 5-year overall survival (OS) rates were significantly lower in the groups carrying aberrant DAPK1 (P = 0.004) and DBC1 (P = 0.044) methylation, respectively. DAPK1 methylation status were significantly correlated with stage (P = 0.015), as all patients with aberrant DAPK1 methylation were stages III and IV. Multivariate analysis identified DAPK1 as an independent prognostic factor for OS with a hazard ratio of 8.9 (95 % CI 2.7–29.3, P < 0.0007). Patients with DAPK1 methylated cell-free circulating DNA at time of diagnosis, who became long-term survivors, lost the aberrant methylation after treatment initiation. Conversely, patients that maintained or regained aberrant DAPK1 methylation died soon thereafter.ConclusionsAberrant promoter methylation of cell-free circulating DNA can be detected in plasma from DLBCL patients and hold promise as an easily accessible marker for evaluating response to treatment and for prognostication. In particular, aberrant DAPK1 methylation in plasma was an independent prognostic marker that may also be used to assess treatment response.Electronic supplementary materialThe online version of this article (doi:10.1186/s13148-016-0261-y) contains supplementary material, which is available to authorized users.

Global hypomethylation has been linked to disease progression in several cancers, but has not been reported for Diffuse Large B Cell Lymphoma (DLBCL). This study aimed to assess global methylation in DLBCL and describe its prognostic value. Mean LINE1 methylation, a validated surrogate measure for global methylation, was measured in DNA from 67 tumor biopsies. Additionally, cell-free circulating DNA (cfDNA) in plasma samples from 74 patients was tested to assess the feasibility of global hypomethylation as a biomarker in liquid biopsies. LINE1 methylation was assessed using a commercially available kit, based on pyrosequencing of PCR amplified bisulfite-treated DNA. Global hypomethylation was detected in a subset of cases and was associated with poor overall survival in both tumor biopsies (P = .001) and cfDNA (P = .009). It was the strongest risk factor in multivariate analysis in both biopsies (HR: 10.65, CI: 2.03-55.81, P = .005) and cfDNA (HR: 11.87, CI: 2.80-50.20, P = .001), outperforming conventional clinical risk factors. Finally, hierarchical cluster analyses were performed for the cfDNA samples using previously published gene-specific methylation data. This analysis shows that global hypomethylation co-occurs with other epigenetic abnormalities, including DAPK1 promoter hypermethylation. In conclusion, we have shown that global hypomethylation is strongly associated with poor survival in DLBCL both when present in tumor biopsy DNA and when detected in plasma cfDNA, and has potential for clinical application as a prognostic biomarker.