Evolution and resulting tumor heterogeneity is currently under investigation for many malignancies since it may explain resistance of tumors to therapies. Pronounced intra-tumor genetic variation has been recently appreciated for solid tumors and leukemias, including chronic lymphocytic leukemia (CLL). Heterogeneous epigenetic alterations, such as DNA methylation, have the potential to add complexity to the leukemic cell population. Studies of the CLL methylome have revealed an abundance of genomic loci that display altered DNA methylation states, including methylation marks showing high prognostic significance. Despite the ubiquity of these epigenetic alterations, the mechanisms and impact of changes to the tumor epigenome in CLL are currently undefined.

Here, we have used Illumina 450k arrays and next-generation sequencing to evaluate intra-tumor heterogeneity and evolution of DNA methylation and genetic aberrations in 80 cases of CLL, with 30 cases evaluated at two or more time points. CLL cases exhibit vast inter-patient differences in intra-tumor methylation heterogeneity. Genetically clonal cases maintain low methylation heterogeneity, resulting in up to 10% of total CpGs existing in a monoallelically-methylated state throughout the tumor cell population. Cases with high levels of methylation heterogeneity display a significantly shorter treatment-free time window preceding first therapy (median difference 11 vs. 49 months, P<0.01), coincident with unfavorable prognostic markers (IGHV unmutated, P<0.01; ZAP70 demethylated, P<0.05). Increasing methylation heterogeneity correlates with advanced genetic subclonal complexity (P<0.001). Intriguingly, a longitudinal evaluation reveals that selection of novel global DNA methylation patterns is observed only in cases that undergo genetic evolution. The level of methylation heterogeneity and presence of a genetic subclonal driver mutation in early time points are significantly associated with methylation evolution, signifying that heterogeneity indicates the presence of active evolution occurring within the tumor population. Independent genetic evolution without broad alterations to DNA methylation is uncommon and is associated with low-risk genetic alterations (e.g. deletion of 13q14). Cases showing high levels of methylation evolution display a significantly shorter event-free time window following first therapy (median survival 9 vs. 110 months, P<0.0001).

This study articulates the novel finding of epigenetic and genetic coevolution in leukemia and highlights the dominant role of genetic aberrations in the selection of developing methylation patterns. As epigenetics plays a key role in determining cellular phenotypes, we propose that parallel alterations to the genome and epigenome endow expanding subclonal leukemic populations with novel attributes which contribute to acquired therapy resistance. This work also advocates a benefit of monitoring DNA methylation heterogeneity and evolution during CLL disease course.


Kipps:Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding. Stilgenbauer:Roche: Consultancy, Research Funding, Travel grants Other; Mundipharma: Consultancy, Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.