The human genome contains around 20,000 genes with two copies per gene, one inherited from each parent. Changes in the genome including gene breaks, rearrangements, and extra gene copies are frequent in many diseases, most commonly in cancer. DNA breaks occur often throughout the genome, but are usually repaired with little-to-no long-lasting impact. However, sometimes, very specific regions in our genome can break apart and rearrange incorrectly. Some cells will also contain extra copies of specific genes (three or more copies vs two). Recent studies identified a handful of enzymes that can control the ability of cells to generate extra copies of very specific genes commonly altered in cancer. These enzymes influence genes by adding or removing chemical tags - known as methyl groups (one carbon atom bonded to three hydrogen atoms) - on histone tails (the proteins which our DNA wraps around to help organize DNA). These ‘writer’ (adder) and ‘eraser’ (remover) enzymes control the presence or absence of these methyl chemical marks on histones and act as a counterweight to each other, keeping the stability of the DNA in check, also known as epigenetic regulation. 

 

In infant, adult, and therapy-associated leukemia, the MLL gene is commonly rearranged and often has more than two copies, leading to highly aggressive diseases that are difficult to treat. However, it was unknown exactly how cells rearrange this gene or increase its copies. Another common variation in leukemia is deletion of a specific region of our genome that contains a gene encoding a methyl “eraser” enzyme, KDM3B. With these two points in mind, we hypothesized that deletion or suppression of the KDM3B gene directly stimulates MLL gene copy gain and rearrangement. To look at MLL copy number and rearrangement status within cells, we used an assay that allows us to visualize the gene using fluorescence. Fluorescent labels mark each end of the MLL gene in different colors. If the gene is not rearranged, then they will be directly next to or on top of one another when visualized. If it is rearranged, then the fluorescent labels will be separated. To assess copy number, we counted the number of fluorescent markers per cell, expecting a normal number of two per cell. 

 

To test whether loss of KDM3B directly promotes MLL copy gain or rearrangement, we blocked or inhibited the activity of the KDM3B enzyme. We found that loss or inhibition of KDM3B specifically caused MLL to undergo copy gain and rearrangement. As KDM3Bis an eraser of histone methylation, we wondered if blocking or inhibiting an opposing “writer” enzyme of this methylation, called G9a, would counteract the MLL alterations. When we blocked or inhibited G9a before blocking KDM3B, the MLL copy gains and rearrangements did not occur. This was important because it suggested that the histone methylation status itself is controlling whether MLL gains extra copies or is rearranged. We then wondered what the change in histone methylation could be doing to the gene in order to encourage these changes to occur. A particular protein that binds to the MLL gene called CTCF had previously been suggested to promote MLL alterations, although it was not directly proven. When we blocked or inhibited KDM3B, we found that CTCF was no longer binding to the DNA encoding the MLL gene. Furthermore, we found that if we reduced the amount of CTCF, MLL copy gains and rearrangements occurred. Therefore, reduced CTCF binding on the MLL gene stimulated the MLL alterations. 

 

Doxorubicin (Dox) is a chemotherapy commonly used to treat many cancers, including leukemias. Unfortunately, Dox is also associated with aggressive therapy-associated leukemia that have these MLL alterations. We hypothesized that Dox may be driving therapy-associated leukemia by promoting MLL alterations through reducing KDM3B and CTCF levels. We tested this by treating cells with Dox, and found that CTCF and KDM3B protein levels were reduced, leading to increased MLL copies and rearrangement. The effect of Dox on the MLL gene could be completely rescued by inhibiting the methyl “eraser” enzyme, G9a, which provides the first way to control this process. 

 

In conclusion, direct epigenetic regulation controls MLL gene alterations. Our study suggests that inhibiting G9a before treating patients with Dox could prevent the MLL alterations that associate with therapy-associated leukemia. These findings suggest that additional epigenetic “writer” and/or “eraser” enzymes proteins may be working in concert with one another to control different genes throughout the genome, which are also driving other cancers.