Vol 3. Issue 39 / December 20, 2004

Small Molecule Shuts Off Gene Expression and Keeps Cancer in Check

By Jason Socrates Bardi

A team of scientists from The Scripps Research Institute and the California Institute of Technology (Caltech) is reporting that a small molecule they have created blocks the replication of a wide variety of cancer cells, including cells derived from leukemia, prostate, pancreatic, cervical, colon, and bone cancers.

Evidence suggests that the compound works by shutting down the human gene that makes the protein histone H4, an essential building block of chromosomes. This is significant because histone H4 has never before been considered a target for anti-cancer therapy. Although there are fourteen genes in the human genome that encode the same histone H4 protein, one of these genes is highly expressed in cancer cells. The active molecule selectively down-regulates this particular histone gene.

Furthermore, the scientists report in the November issue of the journal Chemistry & Biology, the compound can block the growth of tumors in mice without obvious toxicity.

"The compound prevented any further growth of the tumors," says Scripps Research Professor Joel Gottesfeld, who led the research with his collaborator Professor Peter Dervan of Caltech. "It looked [like] the compounds actually caused the [existing] tumors to die."

An Unusual Molecule

The compound is actually a hybrid molecule made by combining a DNA-binding module, called a hairpin polyamide, with the anti-cancer drug chlorambucil. Chlorambucil is a widely used chemotherapeutic drug, which targets DNA relatively nonspecifically. By combining a sequence-specific polyamide with chlorambucil, the researchers were able to target far fewer sites in the human genome.

The hairpin polyamide is based on a naturally occurring toxin called distamycin, which is a tripeptide composed of three unnatural amino acids called N-methylpyrroles. This toxin has the ability to bind to DNA—specifically to any stretch of five A-T base pairs in DNA.

Dervan has been studying distamycin and its derivatives for several years, at first figuring out how it binds to DNA and later working out ways to make it bind to new stretches of DNA. The limitation of distamycin is that it only recognizes A-T or T-A base pairs. A big breakthrough came when Dervan and his colleagues figured out how to make distamycin-based molecules that bind to G-C sequences as well.

The key was to swap out some of the N-methylpyrroles for another unnatural amino acid called an imidazole. By incorporating imidazoles, Dervan found that he could design molecules that could recognize G-C or C-G pairings in DNA, and, in fact, these pyrrole-imidizole polyamides could be designed such that they could recognize any stretch of DNA. Molecules have been created in the Dervan lab that recognize sequences as long as 16 base pairs, sequences that are predicted to be unique in the human genome.

"These molecules have astonishing affinity—sub-nanomolar to picomolar," says Gottesfeld, who started collaborating with Dervan several years ago, after Dervan had made this breakthrough. The goal of the collaboration for Dervan, a chemist, and Gottesfeld, a biologist, was to find ways to apply this basic technology to fighting human diseases.

Gottesfeld and Dervan began by designing compounds with the ability to bind to promoters in the DNA. DNA promoters are stretches of DNA adjacent to a gene to which proteins called transcription factors bind. Promoters play a key role in gene expression because the transcription factors that bind to them are involved in the recruitment of other proteins that lead to the expression of the nearby gene.

The idea is that molecules designed to bind to the promoters might be used to effectively silence a gene by occupying the promoter site and preventing the transcription factors from binding to the DNA. "That will then inhibit the transcription of the gene that requires that [promoter] sequence," says Gottesfeld.

A few years ago, Gottesfeld and Dervan demonstated their first success with this technology when they designed a small molecule that binds to the human immunodeficiency virus (HIV-1) promoter. The HIV promoter is required for the virus's replicative lifecycle, and blocking it effectively blocked HIV replication in infected blood cells, Gottesfeld, Dervan, and their colleagues reported in 1998.

Cancer—The Next Challenge

Gottesfeld and Dervan went no further with the HIV promoter-targeting molecules as a potential therapeutic for HIV-related disease. In practice, such a therapeutic faces the challenge of the virus's ready mutation during its lifecycle, and mutations can easily arise that change the DNA promoter sequence enough to make the mutant virus completely resistant to the effect of the compounds.

However, the HIV work was an important proof-of-principle that they could target DNA promoters with pyrrole-imidizole polyamides. In the last few years, they have looked to apply the same technology to human DNA promoters that are connected to genes involved in diseases like cancer.

They wanted to design a molecule that would silence genes that are upregulated, or highly expressed in cancer cells but not in normal cells. To do this, Dervan and his colleagues synthesized a library of his pyrrole-imidazole polyamides with a variety of specificities and attached to them an anticancer compound called chlorambucil.

Chlorambucil is what is known as an alkylating agent. It chemically modifies DNA by attaching itself covalently to the nucleotide bases. By attaching chlorambucil to a polyamide, Dervan and colleagues created a molecule that when bound to DNA would lead to silencing only the genes that contain binding sites for the polyamide portion of the molecule. Biochemical experiments in the Scripps Research laboratory demonstrated that when such a hybrid molecule lies within the coding region of a gene, the enzyme that copies the DNA sequence into messenger RNA (called RNA polymerase) cannot read through the bound molecule. Importantly, these hybrid molecules are able to enter cells and bind their DNA target sites in the cell nucleus.

Gottesfeld, Dervan, and their colleagues screened the library of polyamide-chlorambucil molecules looking for those that could affect the growth of cancer cells. They identified one molecule that caused the cancer cells to stop growing and to undergo profound morphological changes—the cells grew very large. The cells were also unable to divide, arresting in the stage of cell division before they are able to separate their chromosomes.

"We think the morphological change is a swelling of the cell nucleus because the cells cannot condense their chromosomes properly," says Gottesfeld.

These molecules must have been silencing some genes that were necessary for the cancer cell's growth and division, Gottesfeld and Dervan reasoned, but which ones? The molecule was designed to recognize a seven base-pair sequence of DNA, and all else being equal, one would expect this molecule to bind to a vast number of sites in the human genome and perhaps silence hundreds of different genes.

However, there is no reason to expect the molecule to interact with every single one of these sequences because many would be inaccessible to it in the cell because of the compaction of DNA in the cell nucleus.

"The chromatin structure of a gene in the cell nucleus dictates whether the site is going to be available," says Gottesfeld. "Our working hypothesis is that it is the level of transcription of a gene that dictates whether it is going to be accessible." Only those genes that are actively being copied into messenger RNA have open chromatin structures, and those genes should be most accessible for targeting.

So they used DNA array technology to see which genes were affected by the active small molecule. They found 20 or so genes that were— what Gottesfeld calls an astonishing result because it was far fewer than he had guessed—and of the 20 genes, only one was down-regulated by more than two-fold.

Silencing this gene alone might have a powerful effect on cancer cells.

A Highly Conserved and Essential Protein

The gene in question encodes the protein histone H4, which is a key component of chromatin, the compact protein and DNA material that makes up the chromosomes inside cell nuclei. An octamer (eight molecules) of histones make up something like a fishing reel—cylinders around which the DNA is wound to keep the chromosomes in the nucleus compact. These structures are called nucleosomes, and nucleosomes are present on about 95 percent of the cell's DNA.

Compacting all the DNA in the nucleus is important for cells and not surprisingly, histones are important proteins in biology. Histones are one of the most conserved proteins in evolution, and in fact histone H4 is so conserved across different species that it differs by only a single amino acid between organisms as diverse as cattle and snow peas.

To prove that silencing of the histone gene was indeed responsible for blocking cell growth, the researchers used another approach, called siRNA, to silence the histone H4 gene. Cells treated with either this siRNA or the small molecule show the same change in shape and growth arrest, proving that this gene was involved in the growth arrest by the small molecule.

Histone proteins are also important in cancer. Since cancer cells are rapidly dividing, they need to overexpress histone proteins so that they can assemble their replicated DNA. Not having enough histones is anathema to cancer cells, as Gottesfeld, Dervan, and their colleagues found that when they tested their compound on cancer cells in the laboratory. The compound blocked the replication of cells, and the cells ballooned because they could not wrap their DNA around histones and keep it compact.

Gottesfeld, Dervan, and their colleagues then asked if the compound also works in living systems. They looked at this question by setting up an experiment that involved injecting laboratory rodents with live cancer cells from metastatic colon carcinoma cells, which were previously treated with the small molecule. A control group was injected with the untreated cancer cells. At the end of the experiment, all the rodents in the control group had developed large tumors, while none of the animals in the experimental group showed tumor growth. In a second experiment, the researchers injected cancer cells into the laboratory animals once again, but this time they waited for tumors to develop prior to injecting the active compound. Those animals that were not treated with the compound developed very large tumors, but the tumors in the animals that received the compound did not develop further.

These results suggest that the gene encoding the histone H4 protein might be a new target for drugs designed to treat human cancers, a particularly rewarding result for Gottesfeld because it brought his career full-circle. He started out his career working on histones and chromatin structure while in the laboratory of James Bonner as a graduate student at Caltech. Bonner was the first person to figure out the amino acid sequence of histone H4, quite an astonishing coincidence.

Current work in the Gottesfeld and Dervan labs will extend these findings to other forms of cancer and will define the mechanism of action of the active molecule further. Gottesfeld says that there is a lot left to be done with this molecule before clinical trials can be contemplated. "We need to know about the bioavailability and potential toxicity of the compound before going further."

This work was funded in both the Dervan and Gottesfeld laboratories by grants from the National Institutes of Health.

To read the article, "Arresting Cancer Proliferation by Small Molecule Gene Regulation" by Liliane A. Dickinson, Ryan Burnett, Christian Melander, Benjamin S. Edelson, Paramjit S. Arora, Peter B. Dervan, and Joel M. Gottesfeld see the November, 2004 issue of the journal Chemistry & Biology or go to: http://dx.doi.org/10.1016/j.chembiol.2004.09.004


Send comments to: jasonb@scripps.edu


Professor Joel Gottesfeld and colleages report on a compound that can block the growth of tumors without obvious toxicity.














A graphic representation of arresting cancer cell proliferation by chemical gene regulation. Art by Ryan Burnett. Click to enlarge.