Dianne Newbury: Studying speech and language development on Robinson Crusoe Island
Dianne Newbury and Pia Villaneuva, Image courtesy of Dianne Newbury
Click here to listen to the full podcast episode
Head out 600 miles into the Pacific Ocean from the coast of Chile, either two days on a boat or an hour in a tiny propeller plane, and you’ll get to the Robinson Crusoe Island. Formerly known as Mas a Tierra and renamed in 1966, this small mountainous island is a remote tropical paradise known mostly for the fact that it’s said to be the inspiration behind Daniel Defoe’s novel Robinson Crusoe, hence the name.
Three centuries ago, an impetuous sailor Alexander Selcraig, known as Selkirk, ended up marooned on the island in a failed mutiny bid. After surviving for around five years, he was rescued and made it back to England, where the stories of his experiences caught Defoe’s eye. And the rest, as they say, is history.
Today, Robinson Crusoe Island is far-flung holiday destination for intrepid travellers, populated by a small island community of around 600. One day, about fifteen years ago, a woman named Pia Villanueva came to the island on holiday. And the rest, as they say, is history.
To tell the story, I’m joined by Dr Dianne Newbury, senior lecturer and principal investigator in the molecular genetics of speech and language at Oxford Brookes University.
Dianne: She happens to be a keen diver and the diving on the island is absolutely amazing because it's so remote. I mean, if you look in the sea, you can see the fish. I'm not a diver, but I didn't have to go diving, I could just see them snorkelling. So she went out there to dive on some shipwrecks, which were around that area because of where it's located. And while she was on the island she was talking to some of the people who live there and they were asking her, what do you do? And she said, oh, I'm a researcher, but I'm also a speech and language therapist. And somebody said, oh, that's amazing. I'm really worried about my child's speech development. Would you be able to take a look and just let me know whether I should be worried or not?
Dianne: And she said, yeah, well I'm actually on holiday, I'm not working here. And they were like, oh, please, please. So she said, yeah, well, why don't you drop by my room in the morning? And I'll have a quick chat and I'll see what I think. And she said the next morning when she opened up her room in the hostel that she was staying, there was literally a queue down the road of parents saying, can you look at my child also, my child also. So she actually organised to go back for a proper professional visit to the island. And it was then that she realised just how prevalent speech and language impairments were.
Kat: That's incredible, a real sort of busman's holiday.
Dianne: Yeah.
Kat: So how did you start to get connected with such a remote island in the middle of nowhere, or where does the scientific story of this island start?
Dianne: Some researchers in Chile had been working on the island and they published a paper describing the children who live on the island. The children on the island have a very high incidence of speech and language impairment. So they have difficulties learning to use speech and to understand things that are said to them.
Dianne: And so our collaborators at that time, they were just researchers in Chile, they were interested in this because they're speech and language therapists and they published a paper saying, well, this is interesting. These islanders have a particularly high incidence. And they were trying to find out why, whether that was something to do with schooling on the island, or the situation of living on the island or with something more genetic.
Dianne: And I'm a geneticist, my research involves finding genetic variations that contribute to speech and language impairments. So when I saw this, I emailed the lead researcher and said, you know, we could do a study and we could find out whether it is genetic or not. And she emailed me back and said, that would be amazing. So that was how we got involved.
Kat: So what was the next step for you? How did you start investigating what was going on here at a genetic level?
Dianne: So lots of things can run in families, but that doesn't necessarily mean they're genetic and lots of things can be found within certain populations. But again, that doesn't mean that they are genetic and that's what she was really interested in. Was it something to do with living on the island being so isolated or was it something to do with the way that the school was taught? Although the school on the island follows the same curriculum as mainland Chile.
Dianne: So she had already collected a lot of speech and language information about the children and the kinds of difficulties that they have. But we were able to couple that with collecting DNA from the children. And then what we do is initially we look within families. So she went back to the island and all the children who she knew had speech and language impairment, she went and interviewed all of their family members and asked them about their speech and language development and whether they had any difficulties.
Dianne: And if something runs in a family, then what you see is the family members of people who have speech and language impairments will also be affected themselves. More so than family members of children who don't have speech and language impairment. And that's exactly what she saw. So that indicates that it runs in families. But again, that doesn't mean that it's genetic necessarily because some things run in families. Like which football team you support tends to run in families, but we've no evidence that that's genetic, that's more to do with a shared environment.
Dianne: So the idea that you might go with your family, to the football on a Saturday, and so you all support the same team. So the next step in order to show that it's genetic is to actually go and collect DNA samples. So she went out there and she took a nurse with her, collected blood from all of the children. And that was shipped back to mainland Chile where the DNA was extracted and then sent to us in Oxford. So it was quite a long journey.
Kat: So you didn't get to go out there and do that bit then?
Dianne: I didn't get out there to do that bit, but I have been out there since.
Kat: Oh, so at least you did get to go.
Dianne: Yeah.
Kat: So when you got the samples, presumably you're gonna look at the DNA sequence there. How did you start trying to piece together what was going on with the samples that you had and the people that were living there?
Dianne: So the idea is, if something is genetic, it can be genetically caused, if you like. So when we think about things like cystic fibrosis, we know that that can be caused by a change in someone's DNA sequence, and if you inherit that change, then you develop the disease.
Kat: That's because it's the instruction, it makes a molecule in your cells and if that gene is not working, then the molecule is not working. So your cells aren't working very well and you get the disease, right?
Dianne: Yeah, so our DNA is changing all the time. When you have children, you pass on your DNA, but that DNA changes and over someone's lifetime, you will accumulate changes in your DNA sequence. The majority of those changes don't do very much. They might not even affect the way that genes function or they affect them very subtly. But occasionally there will be a change that has a really big effect upon a particular gene.
Dianne: And if that gene is really important to the cell function, then we call that a mutation, and that is likely to cause a disease. In actual fact, we don't think what's happening in speech and language impairments in the general population, we don't find that there are single changes in the DNA that directly cause speech and language impairments. There've been one or two described, but in general, we think what's happening is these small changes that have very subtle effects upon the gene function. They accumulate over time and you might inherit a particular combination of changes which then, for whatever reason, put you at a higher risk of developing speech and language. So it's a bit different from a kind of clinical disease where you can say there was one change, it directly causes a disease here. We're saying there's an accumulation of changes and they put you at a high risk of the disorder.
Kat: One of the metaphors that I like to use for something like this is, you can't directly equate like the complexities of the brain and our neural processes directly to a computer, but it's a little bit like that where you've just got whole load of different things and they can be slightly wired in different ways and the dials can be tweaked up and down and you can have two connections in here or three connections or no connections or loads of connections. And everyone's kind of built and wired slightly differently. Directed by, to some extent but not completely, by their genes and the variations they've got.
Dianne: Yeah. So you could change the entire motherboard and that would have a very big effect upon the way that the computer is working, or the entire processor, and that would be akin to a medical disease, a clinical disease. Whereas what we're looking at is more like you say, changing some of those connections. But together they can actually have a relatively big effect upon the way the computer is working overall.
Kat: I generally dislike computer metaphors, but you also think, well then it's about the software that you're running through it and what you're doing with it. And then that's what brings in all the things like the education and the opportunities for development that you have.
Dianne: Yeah, and I think that, as you say, what you're doing with it is a really good analogy for things like speech and language, because there are some, we call them speech and language disorders, but actually what's happening is the way that these children use speech and language and approach speech and language is just different from the way that other people may approach it. So in some circles, things like dyslexia, they don't like to be called a disorder. That is actually, it's just a different way of doing things.
Kat: Let's go back to what you actually found when you started looking at the population. So how did you start looking at the relationships between genes and genetic variations and the families and their language abilities? How did you start to try and tease out what might be going on?
Dianne: So, because we don't really have much idea about the kinds of genes that might be important for speech and language development, we took what we call an unbiased approach. So that's essentially, we looked at every single chromosome in every single individual and we compared them between people who had speech and language disorders and people who didn't and said, are there any differences that we can see between the chromosomes?
Dianne: And we do that by almost sampling the chromosomes at set points that we know vary between individuals and then using that information we can build up, it's almost like a picture of the way that the chromosomes are inherited, because we had families, the way that the chromosomes are inherited within families and across the population. And it's almost trying to rebuild; well, if there was a variation on this chromosome, we can see from our data that it's been inherited in this way, through the different families. And would that fit with the patterns of language impairment that we're seeing?
Kat: So, what did you find when you started to look at everyone and all their chromosomes, what kind of patterns emerge?
Dianne: So we found, well, actually our Chilean collaborators found, that the children had quite variable kinds of speech and language impairment. So some of them would have problems understanding what was said to them. Some would have problems producing language and the problems that they had were very different from one child to the next. So when we looked at the chromosomes, we found that there are a couple of chromosome regions that seemed to be more similar in the children who are affected by language impairment than we would expect by chance alone. And one in particular was chromosome seven, but it's quite a broad region of chromosome seven. So the region itself included a couple of hundred genes. So it didn't allow us to get down to a single gene kind of level.
Kat: That's not that helpful, is it?
Dianne: No, well, if you think it's narrowed down from 23 chromosomes to part of one chromosome. It did allow us to narrow it down.
Kat: Okay, you've gone from twenty thousand genes to a hundred-ish genes. Okay, I'll give you that.
Dianne: So then we thought maybe we need something a bit higher resolution. And so to do something higher resolution, now we are able to sequence until, rather than sampling across the chromosomes we can literally characterise every single base in the DNA sequence.
Dianne: So when we're sequencing people's DNA, because we're sequencing every single base, it gives us much higher resolution. And then it allows us to work out lots of things about the population. So we were particularly interested in speech and language impairments, but it allowed us to see, for example, what kind of genetic background the Islanders had, how related they were to each other and what the genetic background of that particular population looked like. So there are lots of interesting things that you can get from these in-depth studies aside from looking for genetic variations that cause speech and language impairment.
When it comes to small island communities like the Robinson Crusoe folk, people don’t arrive very often and they don’t leave either, so the inhabitants tend to be more closely related to each other than in larger populations. Pia Villaneuva and her colleagues initially suspected that the high levels of speech and language impairment they were seeing in the island’s children was likely due to the relatively small gene pool. So how did these people - and their genes - get there in the first place?
Unlike many islands in the South Pacific, there’s no good evidence to suggest that there was an indigenous population from South America or Polynesia living on the island before the arrival of european seafarers in the 16th century, led by the Spanish explorer Juan Fernandez. But while passing sailors came and went, there wasn’t a really substantial attempt to colonise the island until the end of the 19th century, when Swiss general Alfred von Rodt pitched up with a small group of families to set up home.
Born into a strict religious family, Von Rodt threw off the shackles of his upbringing by joining the army in search of adventure, eventually ending up in the port of Valparaiso on the Chilean coast. One day, an advert in the local press catches his eye: the government is auctioning off the right to rent the island group including what’s then known as Mas a Tierra to the highest bidder.
Von Rodt’s offer of $ 1,500 a year wins, and on April 17, 1877, he officially becomes the Sub-Prefect, Judge and Minister of Customs and Post of the Juan Fernandez Islands. He splashes out a couple of grand on a small boat, fills it with people and supplies, and heads out into the ocean to take possession of his new domain.
In an enthusiastic letter to a cousin back home in Switzerland, Von Rodt writes: “This island will be my homeland, my Switzerland, the ocean will replace the Alps” signing it “Robinson Crusoe II”
Dianne: From the population history and the historical records, we think this was about 60 or 70 people who first colonised the island, and that was in 1877. So that's a fair while ago, but probably only about five or six generations in terms of families. When we have the DNA, because we had this amazing population and historical records that our collaborators had collected, we were able to relate people to each other and see exactly how the DNA sequence might be inherited.
Dianne: And one really interesting thing that we saw was that the majority of individuals affected by speech and language impairment all related back to this one founder family who had first come to the island in the 19th century. So it gives us an idea about what might be going on at the genetic level. One thing might be that that original family had one of these kinds of big mutation events in their genetic sequence, which was directly causing the speech and language impairments and that has subsequently filtered through the generations of the families on the island.
Dianne: The other possibility is that this particular family had lots of these small changes, but because the population on the island is quite constricted, they were kind of maintained within the current population and they put all of the children, a particularly high risk of speech and language difficulties.
Kat: Looking at the genetic data you've got so far, are you starting to hone in on any genes in particular, and any kind of clues as to what might be going on?
Dianne: So that's the nice thing about having the genetic sequence is that we're able to almost align it with the population history and with these great big pedigrees that we had. And we found, as I said, the majority of the individuals related back into this one large family, but it is a really, really big family. We were able to relate 227 individuals within this family over seven generations. So it's not like all 200 live on the island now, we're talking about great-grandparents and great-great-grandparents. So that gives you an idea about how difficult the problem is.
Dianne: If you think we've got 3 billion base pairs of genetic sequence for every single individual who we sequence. And we're trying to then look at how is that inherited between these 227 individuals and almost rebuild the genetic sequence and look at how that might have, if you like, flowed down from those founder families down to the current day population in whom we are able to get language data, because obviously we don't have language data for the previous generations.
Dianne: Looking at that, we were able to find that there are a few changes in the DNA sequence that we don't typically see in other populations outside of the island or in other Chilean populations. So there were nine changes that had never, ever been described before in any other DNA sequence that had been reported at that time. So then we went and looked further at our other DNA samples that we had available and looked particularly just at these nine different sequence variants. And we found that one in particular was much, much more common in the children who had language impairment than those who didn't.
Kat: So what is this mystery secret? Where is it? And do we know anything about if it's in a gene, near a gene, any kind of insights.
Dianne: So in order to make that problem a bit easier to solve, initially what we did was we said, okay, we're only going to look at changes in those 3 billion base pairs if they directly occur within a gene. So there's lots of the DNA sequence that doesn't actually form a gene or code for a protein,
Kat: Okay. By lots, it's like 95%. So you're gonna make your job a lot easier aren't you if you're only looking at a little bit,
Dianne: Yeah, so we can go from 3 billion down to a much smaller number just by looking at those coding variants. And the other thing is that because we know the way that genes are coded, if we've got the DNA sequence, we can say this change would have this effect upon the protein sequence in this gene. And we know the function of this gene is this. So it makes our life a lot easier. But even just looking at those, we saw on average about 40,000 DNA variants per individual.
Dianne: So we still have to sift through quite a lot of data but, as I said, the way that we approached it was, we said, okay, we're just going to look at those ones that directly change the gene sequence and have never ever been reported before. So we know these are super, super rare variations. And this one variation that we found occurred in a gene, which has quite a catchy name of Nuclear Transcription Factor, X-Box Binding Like 1 or NFXL1.
Kat: NFXL1, okay. I mean, It's a nickname. It will do. What's it doing in language?
Dianne: That is the million dollar question, We know that this protein turns genes usually off actually, it's a transcriptional repressor. We know that it turns genes off, particularly during, when the cells within the embryo are differentiating or turning into other cells. So as the embryo develops, certain types of cells will change into other types of cells to grow the embryo. And we know that this is when the NFXL1 gene is particularly important in turning genes on and off. But what we don't know is the genes that it turns on and off and how that's important in presumably the brain development and how that is important in language. So there's still lots and lots of questions to answer. So aside
Kat: So aside from the genetics and the biology and the neuroscience of it, there's the deeper question of what can you do for the children that are on the islands? You know, when you turn up and there's a queue round the door to try and see a speech therapist that's come on holiday, what can actually be done to support the children now that you know that there is this high incidence of speech and language problems.
Dianne: So at the genetic level, we're still a little way from understanding what is causing the language impairment at the cellular and the molecular level. If we know that it would probably help us to understand why these children are struggling with speech and language, but in just the fact that we've demonstrated that the children do have a particularly high incidence of speech and language impairments, that allowed us to lobby the local government and the island does now have a resident speech and language specialist who can help them and give speech and language therapy where it's needed. So that's one immediately feel good story.
Dianne: The other thing that we've been able to do is we're actually able to visit the island and to start, we call it a reading the language intervention study. So this essentially is giving these children specialist support in terms of speech and language and seeing whether that has an impact upon the way that their speech and language develops.
Dianne: So this again was in collaboration with Chilean researchers and also with researchers in Oxford, at St. John's college and the department of education. We were able to send four full time tutors to the island, and they spent time with all the children who wanted help. They spent four 30 minute sessions with those children per week, specifically helping them with their reading and their language skills.
Dianne: So we were able to see 68 children and we were out there for 30 weeks. So we were out there for the best part of a year, helping these children, and we measured their speech and language at the beginning and then measured it at the end and said, well, is this making a difference? And by comparing those children to other children who didn't receive the speech and language intervention, we were able to see that it did make a difference on the way that they were using language by the end of the intervention study.
Dianne: So again, that's a feel good story because it shows that even though the children, genetically, may be at a high risk of speech and language impairments, giving them that extra support with the things that they need at the time that they need it can help them to catch up.
Kat: Broadening this out, obviously the Robinson Crusoe island, isn't the only place in the world where there are children with difficulties in their speech and their language development. What have you managed to discover about the underlying genetic causes and also what you can do to help children in other parts of the world who are struggling with their speech and their language development?
Dianne: Yes. At the beginning of when we were talking, I was saying, well, it's not caused by one gene, but then all of our studies came back to this one genetic variation at the end. So I should say that that is, we don't think that that one genetic variation causes the speech and language disorder. We think that is just part of the puzzle, but it was maybe one of the bigger parts of the puzzle. So it was easier to detect.
Dianne: So we're still working on the data that we've collected from the island. We're trying to figure out, well, what other genetic variations might be in there. And we're also carrying out similar studies with children in the UK, looking at their DNA, looking at the way that speech and language impairments are inherited within families. And through all of those studies, we've been able to identify maybe five or six genes that have these changes in their sequence that seemed to be more common in people affected by speech and language difficulties than people without.
Dianne: The main finding from those studies is that we need much, much, much bigger samples to be able to pull this picture apart. So if you'd asked me when I first started looking for genes that are implicated in speech and language difficulties, I'd say there's maybe, I don't know, a combination of five or ten genetic variations and if you're unlucky enough to inherit those in a particular combination that will put you a high risk of speech and language disorder. Now we're thinking it's probably more like a combination of a thousand different genetic variations.
Dianne: And that means every single one of those genetic variations really only contributes a very, very small risk if you like. And that makes them much harder to find. So in some of our more recent studies we've been working with researchers from across the world to get together these really big samples where we can look at the genetic data of 30,000 or 40,000 samples.
Dianne: And using those, we are starting to identify, not just a handful of genes, but hundreds of genes that may be important. We're not quite at this point yet, but once we get those lists, hundreds of genes, the idea is then we can look at how do these function during brain development, what are the common functions between them? And that will then tell us more about why particular people might be at a higher risk of speech and language disorders.
Kat: And also given, once you can identify this genetic information, what does that then mean for families and for children who have these particular variations? What can you do for them that will help?
Dianne: There's a couple of things. One is that we've found that there are actually a subset of individuals who do have these genetic mutations that directly cause their speech and language difficulties. And we're starting to identify a lot more of those. So if you're unlucky enough to be in one of those families, then we can almost diagnose the disorder that you have. We can put you in touch with other families who have a very similar disorder caused by variations within the same gene. And in those families we're able to predict with much better accuracy, the kinds of problems that they might have and put support in place earlier.
Dianne: But that really is just for a very small proportion of people. The rest of the people, it's unlikely that we'd ever be able to diagnose, if you like, speech and language impairments. But what we can tell you by looking at the genes that are involved are these sets of genes are particularly involved in the development of this part of the brain, and we know this part of the brain has this particular job during development, and that can give us an idea about what's actually causing the speech and language impairment, and it might help us to better classify the problems that they're having, and to target better therapies at much younger ages.
Dianne: So that's the goal that we all aspire to. I think we're a little way from there yet, but that's what I'd like to think we might be at if you talk to me in another 10, 20 years.
Kat: Did you ever imagine when you started working as a genetics researcher, that you'd end up on a desert island in the middle of nowhere really getting stuck into the very, almost practical, consequences of the research that you'd done?
Dianne: No, So I first started looking at speech and language impairments as part of my PhD, my doctorate. And as I said, if you'd asked me, then I would have said there were maybe five, ten genetic variations. And I thought by the end of my PhD, I probably would've identified one or two. Ask me in another five, ten years and we probably would have solved this, we'll understand this completely. But that was very, very naive. We now understand that it's a much bigger problem to solve, but we are starting to get that and we're starting to understand these things.
Dianne: And I think probably some of the things that I'm the proudest of are the things I didn't foresee. So I would have said, well, I will be able to identify genes that contribute and this is what this will tell us. But I wouldn't have said to you, I will be visiting desert islands or that I would be helping to deliver speech and language intervention programs.
Dianne: And that's only really made possible by our great collaborators and these networks that have been set up. But these are the kinds of things that provide immediate impact for families that are affected. And we're actually, on the basis of the program on the Robinson Crusoe island, we're also developing a speech and language intervention that's being rolled out in Sao Paulo at the moment. So these are the things that I would never ever have seen myself doing.
References:
Genome-wide analysis of genetic susceptibility to language impairment in an isolated Chilean population Pia Villanueva Dianne F Newbury et al Eur J Hum Genet. 2011 Jun;19(6):687-95. Doi: 10.1038/ejhg.2010.251. Epub 2011 Jan 19.
High prevalence of specific language impairment in Robinson Crusoe Island. A possible founder effect (in Spanish) Pía Villanueva et al, Rev Med Chil. 2008 Feb;136(2):186-92. Epub 2008 May 7.
