Researchers have discovered a new category of cancer caused by chromatin recognition gone awry. An aberrant protein that binds to activated DNA-winding proteins drives up gene expression leading to unchecked cell growth, according to a study published online yesterday (May 10) in Nature.

Chromatin wound on histone proteins Image: Eric Smith, DFCI

Several forms of the blood cancer acute myeloid leukemia (AML) are marked by a chromosomal crossover, or translocation, between the genes that code for two protein fragments: a short, common protein motif known as the plant homeodomain (PHD) finger, which is involved in chromatin recognition and gene regulation, and a large protein responsible for trafficking RNA into and out of the nucleus, called nucleoporin-98 (NUP98). How these translocations bring about cancer, however, was largely a mystery.

"We now have a much better sense for why those translocations would be messing up the cell and promoting cancer," Or Gozami, a molecular cancer biologist at Stanford University who was not involved in the study, told The Scientist. The newfound mechanism is probably a common conduit leading to many pathologies, he added. "This is something that's beyond AML."

David Allis, a chromatin biochemist at the Rockefeller University in New York City, and his postdoc Gang (Greg) Wang used mouse bone marrow cells to study the NUP98-PHD finger fusion -- the "funky hybrid protein," as Allis called it. Allis and Wang showed that the PHD domain in the protein fusion acted like "chromatin Velcro," binding to and docking on a "landing pad" formed by a particular triple methylated histone protein, H3K4me3. The PHD finger then served as a "protective cap" to block enzymes that normally remove methyl groups from the histone protein in order to dampen gene expression. In this way, the rogue NUP98-PHD fusion "somehow sets up an epigenetic signature which pretty much correlates with on-ness," Allis said.

Mouse cells with leukemia Image: Greg Wang

Allis and Wang performed microarray analyses, real time PCRs, and chromatin immunoprecipitation assays, and showed that capping the histone with the PHD finger led to substantial and prolonged upregulation of critical developmental genes, including several HOX genes. This epigenetic "on state" is typical of leukemia cancer stem cells.

The researchers also injected mice with bone marrow cells -- either containing or lacking the PHD domain bound to the NUP98 protein -- and found that those with the hybrid fusion died of leukemia within a few months whereas all the control animals without the PHD domain survived for at least a year. Thus, Allis and Wang concluded that the misguided PHD sticks its finger in the wrong place and drives cancer progression through its effect on histone modifications.

"Our cell and animal cancer models clearly demonstrated that the PHD finger harbored in leukemic fusion proteins is indeed a 'reader' of histone mark H3K4me3, and is essential in driving the normal blood progenitor cells into the cancerous ones," Wang told The Scientist in an email. Since there are some 200 different PHD fingers in the human genome, with many linked to a wide range of diseases, Wang proposed that similar cellular "misreadings" of histone modifications might underlie other unstudied pathologies.

Since we know what goes wrong in diseases with harried histones, we can try and fix it, noted Gozani. "We have a target to go after now," he said. Still, the explanation that the PHD finger is modifying only a select few key developmental genes is probably overly simplistic, he added. "I'm positive it's more complicated than that. It's part of what's going on, but I'm sure that there are other regions of the human genome that are probably also being misregulated and therefore contributing to this cancer."


Correction (May 11): A previous version of the story incorrectly stated that a chromosomal translocation occurred between two protein fragments. In fact, the translocation happened between the genes that code for these proteins. The Scientist regrets the error.


Related stories:

Demystifying histone demethylases
[February 2008]

Support for histone code?
[22nd May 2006]

Researchers focus on histone code
[17th September 2001]