Pharmacoepigenetics

About two decades ago, the first draft of the human genome project - documenting the 3 billion letters in the human DNA code - was published. With a budget in the billions, scientists said this would open up a new vista in medicine, one where disease risks could be predicted, and personalised treatments prescribed based on an individual’s genetic makeup. We’ve certainly made strides in that direction, but, in the intervening two decades, the importance of another mechanism, called epigenetics, that exerts a further layer of control over how active or not our genes are, is increasingly being recognised. And not only does it dictate the activity of genes, that means it also dictates the way we respond to drugs, giivng rise to a whole new field in medicine. Speaking with Chris Smith, Stanford University's Delaney Smith…

Delaney - The field is called pharmacoepigenetics. And basically what scientists in that area are interested in is how our genetic makeup - so our DNA - changes how we respond to drugs that we take. It might not surprise many people that when you take a drug, you might respond differently. Some people are gonna respond well and some others not so well. And it turns out that our genetic makeup is one of the factors that determine how we respond. However, in the past 20 years, we've also figured out that this is not the only thing that matters. And one of the other areas that scientists are interested in is epigenetics: how your DNA gets turned into proteins. But it is not actually part of your DNA sequence itself.

Chris - Historically, when we've made drugs and then given them to people, we've sort of done the pharmacological equivalent of you or I walking into a shoe shop and rather than expecting our feet to be measured, we're just given the first pair of shoes off a shelf and told to wear them, whether they fit our feet or not. And that's why there's this whole idea about trying to personalise medicine a bit more, isn't it? By informing what the drug should look like - what the shoe size should be - based on your genetics. And everyone thought that was the end of it. But now you're saying there is an additional wrinkle here, which is an additional layer of control on the genetics?

Delaney - Yeah, so personalised medicine, as you mentioned, is something that's been popularised in the past 20 years or so. It's still not frequently deployed in every setting, but a lot of people are now, especially for things like cancer treatment, receiving their personalised genome, and your physician is then targeting what treatments are going to be most appropriate. So you're indeed measuring your foot before you put on a shoe. That being said, it seems like that approach is accounting anywhere between 10 and 50% of the variance we're seeing in drug response. So that's not everything. And this additional wrinkle of epigenetics is promising in regards to how much of that additional variance it might be able to explain. So in the shoe analogy, you're now not just measuring the length of your foot, but you could also be looking at the shape of the arch or the width.

Chris - We've got reasonably good at tying genes to diseases and therefore indirectly, in some cases directly, tying genes to treatments. But we certainly don't know all of them; how on Earth can we therefore go the next step and start understanding the epigenetics as well? Do we start with the known - the genes we have got - and then explore the epigenetic space around them? Or are we going to carry on doing shot in the dark type science looking for possible associations, this link to this drug?

Delaney - Right. So we definitely do start on what is known. As personalised medicine has become more popular, we've generated a lot of new data. So you're right in the sense that there's still quite a few genes that are left to be discovered and studied. But, for epigenetics, what we do is we take a lot of this existing data and we look at genes where it's known that variation in them is linked to drug response. And then what we're doing is we're adding a second layer and we're saying, okay, in this gene there are several kind of body parts, if you will, the anatomy of the gene. There's a part that makes the protein and the part next to it that is controlling how much of the gene gets made in the cells in your body. And commonly you'll see epigenetic markers regulate this portion of the gene that's telling your body how much of it needs to get made. So for example, DNA methylation is one type of epigenetic marker, and that is commonly present in these areas of genes that are regulating production. So what we do is we go in using computational approaches and we can then identify genes that have these epigenetic markers and that are associated with drug response. And then try and work our way through and figure out if these epigenetic markers are in themselves affecting the response to the drug.

Chris - Do we know many associations like this yet? Do we have clear cut examples of with this epigenetic mark on this gene, this drug won't work, use this one instead?

Delaney - We do for a very specific subset of all drugs. So these are epigenetic drugs or epi-drugs, and they're drugs that specifically target epigenetic mechanisms. So they're commonly used in cancer treatment and a few other areas, but what they do is they go into your body and they're getting after these epigenetic markers that are already present on your DNA, normally removing them. And for these cases, yeah, we have quite a few studies that show if you have this set of epigenetic markers, this drug may or may not be effective for you. But the area that is still really unknown and important to figure out is the relationship between regular drugs that you and I might take every day, such as aspirinblood thinners, things like that, and how epigenetic marks might affect the efficacy or how well those drugs work in your body.

Chris - Can we filter out the effects though of the microbiome? Because some people have argued that the metabolic knife and fork that the microbiome brings to the table and therefore has impacts on drugs, drug metabolism, and our biochemistry more broadly, is on the one hand a wonderful thing. But on the other hand, it can make it more difficult because if you just analyze the human, you're ignoring this whole extra organ that we haven't considered?

Delaney - Absolutely. The microbiome along with other types of genomes as we now call themincluding like the metabolome and the proteome. So those are things inherent to the person, but also the microbiome, which is a whole other set of DNA set of interactions between the bacteria that live within us and our own systems. All of these things likely have some impact on how we respond to our environment, how we develop disease, and how we respond to treatment. So epigenetics is just one layer, and as these machine learning methods get more advanced, our screening tools develop and we generate more data, especially for the microbiome. The ideal situation would be to be able to take a patient when you are visited in the hospital, sequence their genome, their epigenome, proteome, et cetera, their microbiome, and then create a fully customised treatment plan that optimises all of these different parameters to create a medicine or treatment approach that is most likely to be effective.

Chris - And are we there yet?

Delaney - We are not, but we are moving quickly, especially with the expanding amount of data and expanding amount of computational power. It's something that we could be seeing in the next 10 to 20 years.