Methylation analysis of plasma DNA informs etiologies of Epstein-Barr virus-associated diseases

Abstract

Epstein-Barr virus (EBV) is associated with a number of diseases, including malignancies. Currently, it is not known whether patients with different EBV-associated diseases have different methylation profiles of circulating EBV DNA. Through whole-genome methylation analysis of plasma samples from patients with nasopharyngeal carcinoma (NPC), EBV-associated lymphoma and infectious mononucleosis, we demonstrate that EBV DNA methylation profiles exhibit a disease-associated pattern. This observation implies a significant potential for the development of methylation analysis of plasma EBV DNA for NPC diagnostics. We further analyse the plasma EBV DNA methylome of NPC and non-NPC subjects from a prospective screening cohort. Plasma EBV DNA fragments demonstrate differential methylation patterns between NPC and non-NPC subjects. Combining such differential methylation patterns with the fractional concentration (count) and size of plasma EBV DNA, population screening of NPC is performed with an improved positive predictive value of 35.1%, compared to a count- and size-based only protocol.

Conflict of interest statement

Y.M.D.L. is a scientific cofounder and a member of the scientific advisory board for Grail. Y.M.D.L., R.W.K.C. and K.C.A.C. hold equity in Grail and Take2 Health, and receive research funding from Grail/Cirina. Y.M.D.L., R.W.K.C. and K.C.A.C. are cofounders and board members of DRA Company Limited. P.J. is a consultant to Grail and holds equity in Grail and is nominated as a director of KingMed Future. W.K.J.L. is a consultant to Grail and holds equity in Grail. E.P.H. receives fees for serving on an advisory board for Bristol-Myers Squibb and Merck Sharp & Dohme. Y.M.D.L., R.W.K.C., K.C.A.C., P.J. and W.K.J.L. have filed patent applications based on this work.

Figures

Fig. 1

Distinctive plasma EBV DNA methylation profiles among different EBV-associated diseases. Unsupervised hierarchical clustering analysis of plasma EBV DNA methylome for the 29 samples from patients with different EBV-associated diseases. Each vertical bar represents one plasma sample. Each horizontal bar represents the selected 500-bp regions in the viral genome which demonstrated the most variable methylation densities (coefficient of variation > 30%) across all 29 samples. The corresponding methylation density of each region for all samples were represented in different colours. Samples of different patients with the same EBV-associated diseases were clustered together

Fig. 2

Mining of NPC-associated DMRs in the EBV genome. aBamHI restriction map of the EBV genome is shown. b Methylation densities of CpG loci across the EBV DNA genome from the pooled sequencing data of the 15 cases of NPC and the pooled data of the 5 cases of infectious mononucleosis used for mining of DMRs. Each dot (grey and coloured) shows the methylation density at a corresponding CpG site. A DMR was constructed by two or more differentially methylated CpG sites (>80% in the pooled data of NPC and <60% in the pooled data of infectious mononucleosis) within 200 bp. The coloured dots highlight those CpG sites which fulfilled our criteria. The red dots are those CpG sites with the methylation densities greater than 80% in the pooled data of NPC and the blues dots are those with the methylation densities less than 60% in the pooled data of infectious mononucleosis. Viral DNA fragments mapped to the BamHI-W repeat region presented ambiguities in alignment to the exact member of the repeat family. Hence, this region was not used in the final approach

Fig. 3

Methylation-, count- and size-based analyses of plasma EBV DNA in the exploratory sample set. a The EBV DNA methylation scores of the NPC patients and non-NPC subjects with transiently positive and persistently positive results are shown. A cutoff value in the EBV DNA methylation score was defined at 3 standard deviations below the mean of the methylation scores of these 10 NPC patients in the exploratory dataset. The cutoff value of 73.7 is denoted by the red dotted line. b The proportion of EBV DNA reads among the total number of sequenced plasma DNA reads (both human and viral) of the NPC patients and non-NPC subjects with transiently positive and persistently positive results are shown. A cutoff value in the proportion of plasma EBV DNA reads was defined at 3 standard deviations below the mean of the logarithmic values of portion of EBV DNA reads of the 10 NPC patients in the exploratory dataset. The cutoff value of 2.7 × 10−5 is denoted by the red dotted line. c The EBV DNA size ratios of the NPC patients and non-NPC subjects with transiently positive and persistently positive results are shown. A cutoff value was defined at 3 standard deviations above the mean values of the EBV size ratios of all the 10 NPC patients in the exploratory dataset. The cutoff value of 5.0 is denoted by the red dotted line. Source data are provided as a Source Data file

Fig. 4

Methylation-, count- and size-based analyses of plasma EBV DNA in the validation sample set. The cutoffs in the corresponding analyses defined in the exploratory dataset are shown. a The EBV DNA methylation scores of the NPC patients (from both the screening and external cohorts) and non-NPC subjects with transiently positive and persistently positive results are shown. The same cutoff value of 73.7 defined in the exploratory dataset is denoted by the red dotted line. b The proportion of EBV DNA reads among the total number of sequenced plasma DNA reads (both human and viral) of the NPC patients (from both the screening and external cohorts) and non-NPC subjects with transiently positive and persistently positive results are shown. The same cutoff value of 2.7 × 10−5 defined in the exploratory dataset is denoted by the red dotted line. c The EBV DNA size ratios of the NPC patients (from both the screening and external cohorts) and non-NPC subjects with transiently positive and persistently positive results are shown. The same cutoff value of 5.0 defined in the exploratory dataset is denoted by the red dotted line. Source data are provided as a Source Data file