While we’ve been captivated by the powerful applications of precision epigenetic editing via deactivated Cas9 (dCas9), we’ve also been mesmerized by the fact that epigenetic editing doesn’t always work as expected. Building on our need to take a step back and look at the bigger picture, new findings from the lab of Alexander Meissner at the Max Planck Institute for Molecular Genetics (Germany) demonstrate that epigenetic editing leaves a mark not only on target loci but also the whole-genome.
To understand the off-target effects specific to epigenetic editing systems, the talented team employed engineered mouse embryonic stem cells (mESCs) to track de novo methylation following epigenetic editing. This mESC model contains a double knockout for the de novo methyltransferases (DNMT3A & DNMT3B) as well as a transiently repressed maintenance methyltransferase (DNMT1). Therefore, their system is depleted of methylation but maintains de novo methylation caused by epigenetic editing, all without the interference of endogenous de novo methyltransferases or a highly methylated genome. Finally, the group utilized dCas9 fused to the catalytic domain of DNMT3A (dCas9-DNMT3A) to introduce de novo methylation.
Here’s what they discovered:
- Whole-genome bisulfite sequencing uncovered that dCas9-DNMT3A leads to a rapid and global increase in DNA methylation levels (16% on Day 0, 28% on day 2, and 47% on day 7; 83% in wild-type)
- This increase tends to occur at sites highly methylated in wild type mESCs and not in regions marked with H3K4me3 that are protected from methylation
- Targeted bisulfite sequencing of the 20kb up- and down-stream of sgRNA target sites revealed that the addition of a single or multiple sgRNAs does not prevent the global increases in methylation
- Interestingly, the target sites for some sgRNAs also exhibit increased methylation in the presence of dCas9-DNMT3A alone, demonstrating that the off-target effects can create a background signal that may appear as an on-target effect without this control
- The introduction of the DNMT3A catalytic domain on its (not fused to dCas9) yields a similar global effect to dCas9-DNMT3A
- Transient transfection of human somatic cell lines (human embryonic kidney [293T] and human breast adenocarcinoma [MCF7]) also generates off-target methylation
Overall, these findings provide some much-needed insight into the unique considerations of designing and interpreting epigenetic editing studies. The study also leaves us wondering whether the global off-target effects occur with designer DNA methylation systems, such as dCas9-Dnmt3a-Dnmt3L or dCas9-SunTag-DNMT3A, as well as the suite of other epigenetic effector domains that can be coupled with dCas9.
Set your sights off target over at Nature Communications, February 2018