It’s been a long road, but the role of epigenetics in cancer research is more important than ever.

For the past few decades, researchers have viewed cancer as a genetic disease, with many focused on determining the relationship between the molecular biology of mutations and cancer. This research, which is focused on sequence-based heritable changes, has only been ongoing since the 1980s. At the same time, the field of epigenetics, which is focused on non-sequence-based heritable changes, was brewing.

“The field of epigenetics was really a dark horse,” says Peter Laird, PhD, associate professor of biochemistry and molecular biology at the University of Southern California, Keck School of Medicine Norris Comprehensive Cancer Center, Los Angeles. “[Epigenetics] was not really actively pursued by many investigators.”

According to Laird, epigenetics helps explain why, for example, our liver cells are not the same, functionally, as our brain cells, despite the fact that both cell types are derived from a single fertilized egg. So, epigenetics plays a role in controlling cell behavior, including that involved in cancer. In fact, many of the early papers in epigenetics tried to describe some of the epigenetic changes found in tumors.

click to enlarge  The principle of Illumina’s Methylation Assay, the most recent tool for determining the methylation pattern of CpG islands, is illustrated here. See text for detailed explanation. (Source: Illumina Corp.)

But up until the mid-1990s—when Laird entered the field—investigators were only looking for the presence of methylation changes in cancer cells. Surprisingly, they had not attempted to look for a causal relationship between cancer and these changes. The relationship did not exist until Laird and others decided to look more directly for the cause and effect. For example, in Laird’s studies, he took noncancer cells, induced DNA methylation and then asked: what is the effect of these changes on cancer?

“So I was flipping things around to help me to see if these methylation changes were just mere byproducts or really causal contributors. I published a paper in 1995 showing that when you inhibit the ability of mice to lay down methylation changes, you can almost completely block intestinal polyps from forming in a mouse model for a human colorectal cancer,” says Laird.

Going through changes
DNA methylation occurs on cytosine residues in DNA promoter regions, especially in CpG islands, which, in most genes in most normal cells, are free of methylated cytosine nucleotides. The methylation event is catalyzed by cellular enzymes called DNA methyltransferases, which are commonly found in cells.

“You can find hypermethylated genes in every type of human cancer that I am aware of, including liquid and solid tumors,” says Stephen Baylin, MD, professor of oncology and medicine and associate director of laboratory research at The Johns Hopkins University, Baltimore, Md. However, he says, in terms of the relationship between methylation and cancer, it is not a matter of whether the cancer cells have methylated genes or not, but rather that the number of genes involved and patterns of methylation vary between cancers.

“Because you know exactly where to look in these genes, it makes detecting this in a sensitive way an extremely productive biomarker strategy for tumor detection,” says Baylin. And he goes on to explain that this strategy would also be good for risk assessment because the epigenetic changes seen in tumors can occur early in the pre-stages of cancer development.

“There’s a lot of promise for DNA methylation as an early detection mechanism,” says Laird. “One could look at abnormal DNA methylation patterns by detecting DNA released from tumors in body fluids such as blood.” But detecting a methylation event is not so easy. Earlier detection methods relied on digestion by a methylation-sensitive restriction enzyme such as HpaII, followed by Southern Blotting. That was in the 1970s.

During the 1980s, the methodology shifted to PCR and cloning, both of which erased the methylation events, so these were quickly scrapped. In the 1990s, however, Laird developed two

click to enlarge The Swi/Snf complex performs its chromatin-remodeling duties, freeing up the DNA for its interaction with activated glucocorticoid receptors (GR). (Source: Gordon Hager, PhD)

technologies for determining methylation patterns in cancer cells, Methyl Light and COBRA, both of which were based on the deamination of methylated cytosines—which converts these cytosines to uracil residues—and identification of the converted sequence by analytical methods.

COBRA—which stands for COmbined Bisulfite Restriction Analysis—was developed by Laird in the late 1990s. The method involves bisulfite treatment of the DNA sample, PCR amplification of the treated DNA, and finally, a restriction analysis that is not methylation-sensitive. The restriction enzyme used in COBRA has a recognition sequence that is affected by the change in sequence induced by bisulfite treatment. In 1999, Laird published a second method, Methyl Light, which involves treating DNA with bisulfite and then amplifying it using specific TaqMan-based PCR reactions to recognize fully methylated versions of the sequence.

Laird says that further advancements based on this technology need to have even greater ability to look at methylation patterns in cancer cells and with greater sensitivity. Luckily, companies like Illumina, San Diego, Calif., have created assays specifically for this purpose. The Methylation Assay Illumina created is an adaptation of one of their earlier assays, The Golden Gate Assay. The main difference: The Golden Gate Assay is a three-probe design, whereas The Methylation Assay is a four-probe design.

“As far as workflow is concerned, they are basically the same…You simply do a bisulfite conversion [of the sample DNA] upfront and then feed that into The Golden Gate workflow platform, which all of our customers are familiar with,” says Vivian Zhang, methylation product manager at Illumina.

More epigenetics Another major epigenetic event is acetylation of chromatin, which serves to make DNA in these areas accessible to transcription factors, making gene expression possible. But in the same way that cloning and PCR erase methylation events, purifying DNA away from its chromatin proteins destroys the possibility of detecting these chromatin-remodeling events. “There are emerging data that mutations in remodeling proteins are involved in cancers,” says Gordon Hager, PhD, chief scientist in the laboratory of receptor biology and gene expression at the National Cancer Institute, Bethesda, Md. “Mutations in remodeling proteins can cause developmental effects.” And these might lead to cancer. Aberrant silencing of tumor suppressor genes occurs as a result of hypermethylation and chromatin remodeling, says Stephen Baylin, MD, professor of oncology and medicine and associate director of laboratory research at The Johns Hopkins University, Baltimore, Md. By using a combination of epigenetic inhibitors, Baylin’s group can re-express these tumor suppressor genes in a selective way. More important, he can look at the differential re-expression of select genes by microarray analysis. Researchers need two points of evidence for this type of epigenetic research to be clinically relevant, says Baylin. First, they need to detect the mutation. And second, they need to prove that this mutation disrupts the function of a protein relevant to the development or progression of cancer. Luckily for Baylin and others studying the role of epigenetics in cancer, there are methods to do both.

Here’s how it works. The bisulfite treatment converts unmethylated cytosine residues to uracil, but leaves methylated cytosines unchanged. Two pairs of probes called allele-specific oligonucleotides (ASO) and locus-specific oligonucleotides (LSO) are designed for each CpG island in the genome under study, giving the assay high specificity when they hybridize to the treated DNA. A major advantage of The Methylation Assay is that it can multiplex up to 1,536 CpG sites in a single run, allowing it to rival the gold standard for methylation profiling, methylation-specific PCR, which can only measure one gene at a time.

Because the ASO and LSO probes hybridize very near each other in the genome, extension occurs followed by ligation, which links the two probes together and makes them targets for PCR amplification. The ligated products are then amplified by PCR using fluorescently-labeled common primers, and the PCR products are hybridized to Illumina’s Bead array technology, which bears the complementary sequence for hybridizing to the PCR product. The fluorescence from each bound product is measured, and the readout can be used to determine whether a CpG site is fully methylated, semi-methylated, or unmethylated. For The Methylation Assay, the readout is called beta, which is calculated from the fluorescent signal ratio of Cy3 (green) to Cy 5 (red), where all green signal means unmethylated and all red, fully methylated.

Epigenetic targets
Although there are methods for accurately detecting a specific methylated gene, this in no way proves causation of cancer. “When you see the methylation in these genes, you have to take some further steps to show that these genes are causative or have a high degree of function,” says Baylin. In other words, researchers have to prove that the disruption of this function, either genetically through mutation, or epigenetically through DNA methylation or chromatin modification, causes the initiation or progression of cancer. And that’s a big challenge in the field of epigenetics.

The next major challenge is learning how to reverse aberrant gene expression in cancer cells. “There is the potential for using some epigenetic therapy strategies to turn those [tumor suppressor] genes on again,” says Baylin. And, he says, major pharma is getting very interested in working with academics to see whether this approach will work as a cancer therapy.

In fact, the US Food and Drug Administration (FDA) has approved anticancer agents that work by inhibiting epigenetic mechanisms. “We are now looking to use knowledge about the relationship between DNA methylation and cancer development to specifically and effectively treat cancer,” says David Bearrs, PhD, vice president and chief scientist at SuperGen, Salt Lake City, Utah. And with the acquisition of MGI Pharma’s decitabine, SuperGen is making this treatment option a reality.

Decitabine targets DNA methyltransferases— the enzymes that catalyze the addition of a methyl group to the C-5 position on cytosine residues. The drug is a cytosine analog that can become incorporated into DNA in replicating cells in place of normal cytosine. Because it is modified, this analog (2-deoxy, 5-azacytosine) cannot be methylated by the DNA methyltransferase as it comes down the DNA strand. “Our goal is to try to understand how we might be able to affect the methyltransferase-maintenance pathway, even the de novo methylation of cytosine residues during cancer, as a treatment for cancer,” says Bearrs.

With drugs like decitabine, he says, there is the possibility of having pleiotropic effects on tumor cells. That is, targeting more than one protein or pathway. In particular, the expression of tumor-suppressor genes—which is turned off through DNA methylation—has been targeted by hypomethylating agents like decitabine. “So if we can turn those genes back on by inhibiting DNA methylation in cancer cells, we think that we could have a positive effect on reversing the tumorigenic phenotypes.”

So far, the FDA has only indicated decitabine for myelodysplastic syndrome—a precursor to myelogenous leukemia. However, Bearrs says the application of decitabine should not be limited to liquid tumors, and that drugs similar to decitabine are already being investigated for solid tumors. He predicts that these investigational hypomethylating agents will likely be used in combination with a class of compounds known as histone deacetylase inhibitors, which block the epigenetic event, histone deacetylation—another regulatory mechanism that can activate the expression of specific genes.

This article was published in G & P magazine: Vol. 7, No. 3, March, 2007, pp. G2-G5.