Cancer cells are an unruly bunch. Instead of listening to signals telling them to differentiate into specific cell types to form tissue, they ignore these signals and grow uncontrollably to form tumours.

Most cancer treatments take a brute force approach: kill all cells that are dividing. Although such treatments effectively kill cancer cells, they also kill non-cancerous cells in the process of normal growth — for example, hair follicle cells. This makes the identification of drugs that target only cancer cells critical to the effective treatment of cancer and to improving patient quality of life.

Selectively targeting cancer cells is precisely what scientists at the Dana-Farber Cancer Research Institute in Boston have reported in the advanced online format of Nature this September.

A consortium of scientists led by James Bradner at Dana-Farber has identified the chemical inhibitor of a protein that is responsible for NUT midline carcinoma, a rare cancer of squamous epithelial cells, which are cells on the outer layer of skin. The tumours produced in NMC are the product of uncontrolled proliferation due to a defect in a protein called BRD4.

Normally BRD4 interacts with chromosomes to help turn growth genes on and differentiation genes off. This interaction is tightly regulated and only occurs when cells should be growing. But in the case of NMC, the interactions between BRD4 and the chromosome are deregulated due to the rearrangement of two chromosomes and resultant gene fusion that produces BRD4-NUT, a chimeric protein.

BRD4-NUT constantly sticks to accessible parts of the chromosome, resulting in prolonged expression of growth genes, which ultimately results in tumour formation near the head, neck, pelvis, or other areas of the “midline.” At present, there is no effective treatment for NMC, and death almost always occurs within months of diagnosis. Although NMC currently accounts for only seven to 18 per cent of poorly differentiated carcinomas, many physicians suspect it is often not recognized and diagnosed properly.

Scientists at Dana-Barber used the 3D structure of the BRD4 protein to identify and create chemicals that might be able to bind the protein and potentially inactivate it, eliminating the devastating effects of the BRD4-NUT chimera.
alt text

They tested these chemicals on BRD4 and found that its function was inhibited when combined in a test tube with the chemical (+)-JQ1. The function of the BRD4-NUT chimera was also inhibited. Patient tumors were also grafted onto mice, creating NMC models in which to test for overt toxicity of (+)-JQ1. Experiments using these mice showed that treatment with (+)-JQ1 resulted in smaller tumours, and that treated cancer cells began programmed differentiation into squamous tissue. In addition to halting growth and promoting differentiation, (+)-JQ1 had no obvious toxic side effects, indicating its potential therapeutic value for treatment of NMC in humans.

But how broadly applicable would this treatment be, considering the rarity of NMC? Perhaps one of the most exciting results of this work from Bradner’s group is the potential to identify chemical inhibitors of proteins that are similar to BRD4. According to Dr. Corey Nislow, assistant professor at the Banting and Best Department for Medical Research and the Department of Molecular Genetics at U of T, Bradner and colleagues have “opened the door to a new target set and class of proteins.”

Because there are over 40 genes in humans encoding proteins that are similar to BRD4, the potential for these proteins to be targets for drugs similar to (+)-JQ1 is far-reaching. This is particularly important because of the approximately 30,000 proteins in the human body, only a small fraction are predicted or known to be “drugable” — or in other words, can have their functions altered by chemical drugs. Nislow praised the efforts of Bradner and colleagues for identifying BRD4 as drugable, which could potentially lead to clinical treatments for NMC as early as one to two years from now.

What about side effects? If BRD4 is a protein normally needed to instruct cells to grow at the correct time, how will a chemical inhibitor affect the growth of normal, non-cancerous cells? In other words, would this treatment really target only cancer cells?

Nislow hypothesized that, because there are approximately 40 proteins similar to BRD4, perhaps one of those could ‘cover shift’ for an inactivated BRD4, taking advantage of the robustness of human protein-interaction networks. Of course, multiple stages of testing will be required before potential side effects are identified. But for now, there will be a few less cancer cells winning the battle.