Manel Esteller works at the Spanish National Cancer Institute in Madrid. Brona talks to him about new epigenetic drugs for cancer, epigenetic landscapes and star-shaped chromosomes

B: Cancer isn’t only caused by mutations in DNA. It’s also an epigenetic problem, right? I understand that DNA methylation has something to do with this…

M: DNA methylation by itself is not healthy or harmful. What is harmful is to have an excess or a defect. We as human beings need a certain level of DNA methylation. Without this we might suffer diseases like cancer, cardiovascular disease and maybe Alzheimer’s.

DNA methylation is like the brakes on a car, because it stops genes from running, stops gene activity. When we look at cancer cells under the microscope, we see lots of broken chromosomes

B: The cell can still exist with broken chromosomes?

M: Not only can it exist, it can excel. Where breaks occur in tumour-suppressor genes, this can confer a selective advantage to the cell [in terms of becoming immortal]. This is Darwinian evolution at the level of cells

B: Except that cancer cells might eventually kill their owner…

M: Tumour suppressor genes can also get switched off because of DNA methylation. But there are epigenetic drugs which can try to repair epigenetic problems. There are two kinds. One type puts a foot down on the accelerator of tumour-suppressor genes, or rather releases the brakes by wiping off DNA methylation

B: So the tumour suppressor genes can switch back on again and do their protective job?

M: Yes and these drugs have been approved for the treatment of leukaemia and other malignant diseases. There are other kinds of drugs that work at the level of histones. While there are tumour-suppressor genes that are switched off by DNA methylation, there are others that get affected by modifications to histones

B: When you use these drugs to switch on tumour-suppressor genes, are there other genes that get switched on too?

M: Well, there are two main genes important for cancer: oncogenes and tumour-suppressor genes. Oncogenes are already switched on, so it’s not possible to enhance them any further. We haven’t noticed side-effects on other genes

B: So these drugs have fewer side-effects than traditional chemotherapy?

M: Yes, they are less toxic than classical chemotherapeutics and considering they are stripping methyl off the DNA, we’ve seen fewer than expected effects

B: So how widely are these drugs being used?

M: So they’re being used initially to treat leukaemia and lymphomas. From there treatments will be extended to solid tumours, breast cancer, colon, lung, etc

B: So when the tumour-suppressor genes get switched back on in a tumour, what happens?

M: You can see lots of effects. For example, the cells stop growing and they don’t break off, reducing the risk of metastases…

B: So you don’t get secondary cancers?

M: That’s right. Also, they stop producing their own blood vessels and if you look at them under a microscope, they begin to look like normal cells again

B: So this is post-treatment in humans. You can see tumour cells returning to a normal personality again?

M: Yes. One thing we’ve discovered for sure is that these DNA demethylating agents switch on these tumour-suppressor genes, but we’ve also found that they switch on small RNAs with tumour-suppressing functions

B: So what do these small RNAs do?

M: They regulate oncogenes. So the idea is that by switching on these small RNAs, we would switch off oncogenes

B: Cool. So are you using a microarray to see which genes switch on in response to drug treatment?

M: Yeah, we’re using microarrays not for gene expression but for the expression of these small RNAs

B: OK. As I understand it a microarray is a like a tiny chip with miniscule probes for every gene in the genome (so like 30 000 for humans)

M: Yes, that’s right. The chips we have don’t have probes for the genes that produce protein, only for the genes that make these microRNAs

B: Ah, so are these small RNAs made from tiny DNA sequences?

M: Yes, 20 nucleotides

B: 20 nucleotides! So, very short

M: Yes, and it matches an oncogene

B: Matches an oncogene?

M: Yes, for example CDK6, this is an oncogene which is overactive in tumours. So the reason why is that the microRNA that regulates this gene has been methylated and is silent. The brakes are on and the oncogenes unleash all their nastiness on the cell. So these DNA demethylating agents are able to restore the expression of these microRNAs

B: Fantastic!

M: Yes, this is a very hot area of research right now

B: So have we looked at patients yet?

M: So far, in none of the patients treated with DNA demethylating agents has anyone tested the expression of microRNAs. It has been done in vitro in cell lines. But we do know that patients have the brakes on their microRNAs. Each oncogene has 2 or 3 microRNAs that regulate its activity. In our genome there are around 500 genes that produce microRNAs

B: So there’s quite a lot of choice. So how many oncogenes do we have?

M: [laughs] that’s the question! No one really knows how many oncogenes we have, but perhaps there are 50 genes with clear oncogenic function. They are expressed at very low levels in normal cells, but in cancers they get overactive. With tumour-suppressor genes it’s the other way around. They work at low levels in normal cells, but get completely switched off in tumours

B: So what do the oncogenes do when they’re not causing cancer?

M: They’re involved in cell division, cell proliferation and cell growth

B: OK, so can I ask you a bit about the way chromosomes are organised in normal cells versus cancer cells? I’d like to know more about this 3D epigenetic landscape that is emerging as a concept

M: Sure, so I mentioned histones, which package up our DNA. They are very important in bringing together regions of the genome that are very far apart on chromosomes, like a magnet…

B: Promoting social interactions between genes?

M: Yes, yes. And in a cancer cell, this interaction is disrupted. Genes that would normally be close in normal cells can be far apart in cancer cells and vice-versa

B: So the difference between normal and cancer cells involves different social interactions between genes

M: Yes, the formation of different chromosomal domains and this is related to aberrant epigenetics, messing up DNA methylation and histone modifications.

There is a disease where people have a mutation in an enzyme that methylates our DNA. When we look at their chromosomes under the microscope, they are shaped like stars

B: Star-shaped?

M: Yes, several of them, because the chromosomes are joined together

B: And this is because they can’t methylate DNA?

M: Right, and they die very young. The syndrome is called ICF and involves a lot of defects in their immune system. Epigenetic alterations frequently cause problems with immune functioning

B: Interesting stuff, so star-shaped chromosomes are not good for you

So there are really lots of ways arising from epigenetics to increase the barrage of anti-cancer drugs

M: Yeah, although the idea behind most of these treatments is to switch on tumour-suppressor genes

B: So are we moving towards a stage where we’ve got enough knowledge about genetics and epigenetics to personalise treatments?

M: We have the potential to do that. So we can assess tumours for both mutation and messed up DNA methylation, and decide which patients will be most sensitive to the different drugs. The best example we have is a paper we published in the New England Journal of Medicine about three years ago on a DNA repair gene, Mgmt that is methylated. This is the best marker in response to therapy with alkylating agents. So if this gene is methylated, the patient will respond well to chemotherapy.

This is something that’s going to make the translation to the clinic soon

B: So what about the expense? I mean we have the NHS in this country, but if you get cancer you’ll get much better treatment if you go private

M: It depends. If you want to do a global epigenomic study, it’s expensive. But with the Mgmt gene for example, it’s very easy. You only have to look at DNA methylation for that one gene. This can be done with a simple kit

B: How much would it cost?

M: One gene can be done in 20 euros.

B: But how do you know which gene to look at?

M: It requires a lot of experimentation. For example, we discovered the Mgmt gene in 1999. Now, eight years later, it is going to be applied to patients