Vardhman Rakyan - Nature meets nurture

Kat - You're listening to the Naked Genetics podcast with me, Dr Kat Arney. Still to come, our Gene of the Month is truly titanic. But first, it's time to turn to the eternal struggle - the battle between nature and nurture. Over recent years, scientists have used huge genome-wide association studies, or GWAS, to find hundreds of genetic variations linked to a huge range of traits and diseases. But they're still missing something. And we know even less about how factors in the environment influence how our genes are turned on and off - something known as epigenetics. Now Vardhman Rakyan and his team from Queen Mary University of London have discovered that tiny chemical tags - called DNA methylation - on repetitive genes known as ribosomal or rDNA, could be passing information from one generation to the next, building a bridge between the effects of nature - our genes - and nurture, the environment. 

Vardhman - I suppose the core question is, what determines a person's phenotype, the way they look and the diseases they're susceptible to? In recent years, probably the biggest advance in terms of understanding phenotypes has been at the level of genetics.

Kat - That's the DNA that you've got?

Vardhman - Yes, that's exactly right. This has been made possible by these very large scale studies where people have looked across the entire genome and looked at a lot of genetic variants in lots and lots of people. They have indeed found for many diseases and phenotypes that there are genetic variants associated with diseases such as type 2 diabetes, type 1 diabetes, cancers and so on. I personally think that these genome-wide association studies as they're called GWASs have been very successful. Although one big surprise after these studies have been conducted over the last 7 or 8 years is that almost every case, we can still explain only a small proportion of the heritability in straightforward terms, it means how much of a person's specific phenotype of disease determined by genetic variation. There are precise definitions but operationally, that's fine.

Kat - So, for some diseases, we know that a certain proportion of it should be in the genes but when you start looking for it, you don't find that much.

Vardhman - In a manner of speaking, yes. Now, people are calling this the missing heritability. So, where is this missing heritability? People are now looking for rarer variants. So for example, there could be only one person out of a thousand who has a particular variant. So we need to do as people are doing that more sequencing, looking at more individuals. But there are suggestions that the missing heritability may lie elsewhere. It could be epigenetics or it could be in other parts of the genome that we simply have not looked at. There's a lot of the genome that we have not looked at such as the repetitive portion of the genome. So, a lot of repeat elements, retro elements, and ribosomal DNA - ribosomal DNA codes for the ribosome, the protein production machinery of the cell.

Kat - They're like the sort of the molecular chefs that make the stuff that makes our cells.

Vardhman - Yes. The surprising thing about rDNA is that there are many copies of it spread across different chromosomes. That's true in humans and true in all mammals. As an example, I could have just 40 copies of it, of rDNA and you could have 200 copies of the rDNA.

Kat - Does that mean that I'm loads better at making proteins than you are?

Vardhman - Surprisingly, not. A lot of these copies are silenced, are epigenetically silenced in fact. So, this has been known for a long time that the number of copies can vary amongst individuals. What is not that well-known though is each individual copy within a person or within a single genome can vary genetically. So, in the mice that we looked at, these are inbred mice that is, they've been bred together to try and eliminate genetic variation which can be a confounder in a variety of studies. So, this was a surprise that even within an inbred mouse which will have many copies of rDNA just like humans would - even with an inbred mouse, individual copies within a single mouse will be genetically different.

Kat - What was the idea for this study? What were you trying to look for?

Vardhman - What we were really interested in is how the maternal in utero environment influences the offspring's phenotype.

Kat - So basically like what your mum's tummy does to a baby.

Vardhman - Yeah, kind of like that.

Kat - Yeah, because a foetus is gestating in there, it's not in a completely sealed box. It's inside a female.

Vardhman - That's right. So it's exposed to the mum's environmental condition depending on whether the mum eats or doesn't eat, and other factors like smoking. So, what we were interested was in this process called developmental programming. In utero, when a foetus is exposed to various environmental factors, how does that affect the phenotype postnatally after the baby is born, as an adult, and so on? And there had been lots of suggestions and a lot of excellent work has been done by groups in the UK and overseas that have shown that in some cases, epigenetic marks in the offspring can be changed as a result of the mother's environment.

Kat - So, this is things happening to the mum that are somehow being written in in these marks in the baby's DNA.

Vardhman - Yeah. So, another way of putting it is these epigenetic changes represents a molecular memory of what the foetus was exposed to. So, that was the driver for our project.

Kat - So what did you do? What did you look at?

Vardhman - In our model, we looked at the effect of protein. We compared pregnant mums on a normal protein diet which is 20 per cent protein versus mothers on a slightly lower protein diet, 8 per cent protein. The offspring look healthy, there's no major effects. One big phenotype though is the offspring of low protein mums were approximately 25 per cent smaller.

Kat - They're little dinky mice!

Vardhman - They are small. So that's been shown also a number of times. There's one reason why we chose this model because it's been done a number of times and it's a very robust phenotype. So then we wanted to ask, can we see DNA methylation differences that associates with the smaller birth weight? When you take a sort of standard analytical procedures, you map it to the unique part of the genome. That's the easiest way to look at the data and we found nothing. So we then noticed that there was one part of the genome where we did see a very big methylation difference. This was in the part of the genome that is not a unique portion. It seem to be like a fragment of something else. And then when we'd look at this in more detail, we realised it was very similar to rDNA.

Kat - So, it's not the kind of protein coding genes. It's not the honest to God real genes if you want to call them that. It's these ribosomal DNA genes. What difference do you see between the mice that were fed high protein and the mice that were fed low protein?

Vardhman - The mice that were fed with low protein showed more methylation. Essentially, what that means is, with more methylation especially in the promoter region, the region that controls the expression of the gene, that turns gene expression down and this makes sense. You have this foetus developing inside a mum who is not getting enough protein. So the cells of the foetus are thinking, "We're not getting enough protein, so let's tune down protein production and we can do that in some ways by not making as many ribosomes. Hence, we methylate our ribosomal DNA" and that copes with the stressful conditions they're experiencing.

Kat - So this is almost a prediction that, "Okay, we're not getting much protein now. We might not have much protein in the future. Let's not try and do too much molecular cooking with all of this. Let's just wait and see."

Vardhman - Trying to carry out things as they would've been carried out might not be the best strategy here. Let's slow down protein production. We can have a smaller organism, but one that actually lives.

Kat - Because that's the goal really. It doesn't matter if you're a small baby, but as long as you're an alive baby, that works.

Vardhman - Indeed, yes. That's correct.

Kat - Not only are there these epigenetic differences depending on whether the mum gets more protein or not. But there also seem to be genetic differences, differences in the DNA sequence. What influence does that seem to have on the baby?

Vardhman - Okay, now that was a really big surprise. So, when we looked at these rDNA copies further, we realised they were genetically different. Now genetic differences within rDNA in the mouse have been shown before. But for us, we were looking at an inbred strain though. This was really surprising that this had not been shown because we would've assumed that with inbreeding, all these genetic differences would have been essentially bred out. So that's when we found these genetic differences. Essentially, we found two types of rDNA. One, which we can call the A variant because it has a specific nucleotide in the promoter A and another called the C variant because it has a C in that same position. Different copies will have C whereas other copies will have an A. So we compared the A variant to the C variant. To our surprise, most of the epigenetic changes that has increased methylation was occurring at the A variant. What I've told you so far is that within a single genome, there will be some A variants, some C variants. But the next thing we found was that the relative number of A variants to C variants varied from mouse to mouse, from brother to sister. If you were to look at in mice and in fact, even in humans, if you were to look at the genetic variation of the mum and the dad, you can only predict the genetic variation rDNA of the offspring with only a certain degree of probability. You cannot predict with a 100 per cent certainty. This was really surprising. And so then when we investigated further, we found that it was the mice which have more of this A variant rDNA copies that seem to attract more methylation and ended up being smaller. That was the sort of key finding and the key message of our paper.

Kat - So, this is a really incredible example of its nature - it's the DNA sequence - and it's nurture. It's something happening to the mum, a change in their diet, making this change in the babies. Has this kind of really strong link been found before?

Vardhman - Almost any phenotype is due to both genes and environment. In our opinion, this provides a very unique example and I think a very strong example of how genes interact with the environment to influence phenotype. So this is not by any stretch of the imagination a purely epigenetic effect, not at all. What we're saying is that the underlying genetics is absolutely key but that underlying genetics becomes important when exposed to a certain environment.

Kat - There is a lot of discussion about the rise in diseases, things like type 2 diabetes, rises in obesity, and trying to work out how do we link changes in diet, unhealthy diet in parents, in fathers and mothers to what we're seeing now in populations. Do you think that this is the link that we've been looking for?

Vardhman - I wouldn't say it's THE link, but I think it could be important for sure. We'll only really know after we've been able to design proper powerful assays for it. A lot of this, all science and technology, scientific discoveries go hand in hand. So in terms of what it means for humans, we've got a clear idea what we want to do. But the next few years at least, we want to develop our ability to be able to look at it and then I might have hopefully an answer. That's what we're aiming towards.

Kat - Vardhman Rakyan from Queen Mary University of London, and that work was published last month in the journal Science. And if you're interested in finding out more about how traits are passed on down the generations, check out my recent documentary for BBC Radio 4, called Down the Generations.
http://www.bbc.co.uk/programmes/b07krkgt