Genome-wide profiling of 5-hydroxymethylcytosines in circulating cell-free DNA reveals population-specific pathways in the development of multiple myeloma

Abstract

Multiple myeloma (MM) and its precursors monoclonal gammopathy of undetermined significance (MGUS) and smoldering myeloma (SMM) are 2-3 times more common in African Americans (AA) than European Americans (EA). Although epigenetic changes are well recognized in the context of myeloma cell biology, the contribution of 5-hydroxymethylcytosines (5hmC) to racial disparities in MM is unknown. Using the 5hmC-Seal and next-generation sequencing, we profiled genome-wide 5hmC in circulating cell-free DNA (cfDNA) from 342 newly diagnosed patients with MM (n = 294), SMM (n = 18), and MGUS (n = 30). We compared differential 5hmC modifications between MM and its precursors among 227 EA and 115 AA patients. The captured 5hmC modifications in cfDNA were found to be enriched in B-cell and T-cell-derived histone modifications marking enhancers. Of the top 500 gene bodies with differential 5hmC levels between MM and SMM/MGUS, the majority (94.8%) were distinct between EA and AA and enriched with population-specific pathways, including amino acid metabolism in AA and mainly cancer-related signaling pathways in EA. These findings improved our understanding of the epigenetic contribution to racial disparities in MM and suggest epigenetic pathways that could be exploited as novel preventive strategies in high-risk populations.

Keywords: 5-hydroxymethylcytosine; Epigenetic modification; Multiple myeloma; Racial disparity.

Fig. 1

Genome-wide profiling of 5hmC from cfDNA derived from EA and AA patients with MM and its precursors. Genome-wide 5hmC was profiled in patient-derived plasma cfDNA samples using the 5hmC-Seal and the next-generation sequencing. The 5hmC-Seal data summarized for gene bodies were the primary targets for differential analysis between MM and its precursors (MGUS + SMM, i.e., MGUS and SMM combined) in all samples, using multivariable logistic regression models, controlling for age, sex, and self-reported race/population. In addition, we performed differential analysis between EA and AA patients with MM only. A The captured 5hmC-Seal reads in cfDNA are more abundant in gene bodies relative to the flanking regions and depleted at the promoter regions, based on the GENCODE annotations (hg19). TSS: transcription start site; TES: transcription end site. B The captured 5hmC-Seal reads are enriched in histone modifications marking enhancers (H3K4me1 and H3K27ac) derived from B-cells and T-cells compared with other tissue types. The annotations for H3K4me1 and H3K27ac were obtained from the Roadmap Epigenomics Project. The standard error is shown as the error bar. C The heat map shows the top 63 differential gene bodies between MM and its precursors in the combined EA and AA patients. D Shown are the enriched KEGG pathways among the top 500 differential gene bodies between MM and its precursors in the combined EA and AA samples. The X-axis represents the ratio between the number of differential genes and the total genes in a given pathway. E The Co-expression Network Enrichment Analysis was performed for differential gene bodies between EA and AA to provide further biological insights. Specifically, three modules (Module 1: 254 genes; Module 2: 156 genes; and Module 3: 75 genes) are shown from the modular gene co-expression analysis using the top 500 differential gene bodies between AA and EA patients with MM as the input. NES: normalized enrichment score. F–H Shown are the protein–protein interaction networks constructed for the co-expression and/or interaction modules identified from differential gene bodies between EA and AA patients with MM

Fig. 2

Differential analysis reflects population-specific 5hmC signatures and pathways between MM and its precursors. The 5hmC-Seal profiles summarized for gene bodies were compared between MM and its precursors (MGUS + SMM, i.e., MGUS and SMM combined) in AA and EA samples, separately, using multivariable logistic regression models, controlling for age and sex. A The heat map shows the top 36 differential genes between MM and its precursors in EA patients. B The heat map shows the top 36 differential genes between MM and its precursors in AA patients. C Shown is the number of shared differential gene bodies between MM and its precursors in individual populations (AA vs. EA), compared with the null distribution. The blue line represents the mean of a null distribution generated by permutation (N = 10,000). The red line represents the observed number of shared differential gene bodies. D The Venn Diagram shows the number of shared gene bodies (top 500) between MM and its precursors in EA, AA, and the combined samples. E Shown are the enriched population-specific KEGG pathways for the top 500 differential gene bodies between MM and its precursors in EA and AA