The monkey in the mirror
The unfortunately photographed blobfish
Top left: Sculpture of a dead blobfish after it has been depressurised from bringing it up to the surface (Image by Grufnik on Flickr)
Main image: Closely related blob sculpin as it looks alive and at its usual pressure (Image by NOAA)
When it comes to good looks in the natural world, the blobfish isn’t exactly at the top of the list. More formally known as Psychrolutes marcidus, its cartoonish big nose, flaccid skin and almost human-looking sad face have seen the blobfish named the World’s Ugliest Animal, more than a little unfairly.
I first encountered the blobfish a few years ago when I went to visit Joanna Wysocka, Professor of Chemical and Systems Biology at Stanford University in California, while I was researching my first book, Herding Hemingway’s Cats, my quest to understand how our genes work. Alas, this story never made it into the book, but I’m sharing it with you now, because I think it’s fascinating.
On her computer screen, Wysocka pulls up a picture of the fish, sitting alongside photos of a human, a chimp, a gecko and more. These are all animals that scientists would class as vertebrates, meaning they all have a backbone. As far as Wysocka’s concerned, they share a much more interesting and instantly recognisable feature: a face. Each one of them has two eyes, a nose and a mouth, roughly in the places we’d expect to see them. And for someone like her who’s studying how our genes build our bodies, the face is a good place to start.
“You don't have to convince people about variation when you talk about faces,” she tells me, pointing to her own finely defined features under neat blonde curls, in comparison to my snub nose and chubby cheeks. “There’s probably more variation in our faces than anywhere else, and it’s relatively easy to measure.”
It may be easy on a scientific level, calculating the length of your nose or the arch of your brow. But although we’re all made according to the same basic plan, very subtle changes can make all the difference to what we look like- and how other members of our species perceive us. Yet while it’s easy to spot this diversity among our fellow humans, it’s a lot harder to tell how this recognition might be at work in other animals.
“Do mice look different from each other?” I ask her out of curiosity.
“Of course! And if we took detailed measurements we would see the same level of variation as in people.”
“But if you asked mice, would they say, ‘yeah, that mouse over there looks hot?’”
“Well, there’s an evolutionary aspect to this. We have a whole part of the brain that is geared to facial recognition so it is possible that we as a species are better attuned to recognise it, and mice aren't as focused on the face as humans are.”
I take that as a no and ask her instead how our genes and the switches that control them, known as enhancers, have evolved to make something as complicated and unique as our face.
“Right now we can start to unpick that because we finally have the technology to do so,” she says. “Enhancers are the most dynamically utilised part of the genome, and there's no question about it.”
She explains how two very closely related cell types might use a lot of the same switches to turn on the genes that they both need, but there could still be a couple of thousand enhancers that are only used in one cell or the other. And it’s this unique combination of switches that gives cells, tissues and ultimately body parts their individual identity.
What’s exciting is that we’re now at a time when we can actually look inside cells and see which enhancers they’re using at any given timepoint, by using new techniques to see exactly where certain transcription factor proteins are stuck to the DNA of the switches.
Rather than just studying one at a time, it’s now possible to look at thousands of switches at once to see if they’re set to ON or OFF in any particular cell type as it responds to signals coming from the cells around it. During development, when a single fertilised egg cell grows into a new organism, these are the signals that tell a cell it’s going to be nose, eyeball, liver, blood, or anything else.
“Because we have not one but two thousand enhancers, we suddenly gain the statistical power to start asking is there any language - any words, any grammar to the enhancers that we see are active,” Wysocka tells me.
“And because we know exactly which genes are getting turned on, as well as the repertoire of transcription factors that are there, we can start making quite good educated guesses as to which exact factors orchestrate the transition. And then of course we need to start thinking, OK how do those factors connect to the signals coming in from outside the cell, because of course it's a changing environment that then dictates the change in which switches get used.”
In this way - changing signals from neighbouring cells altering the settings on the switches - the right genes come on at the right time to build a baby. As we now know from DNA sequencing data, our actual genes – the bits in our genome that encode functional elements like proteins - are pretty much identical to a chimp’s. But straight away you can see that a chimp’s face is very different from our own, even though it’s made using the same biological ingredients. Wysocka is trying to find out how that happens.
“I guess I’m starting from the grandiose question, what makes our faces human?” she says, laughing slightly at how pompous it sounds.
The answer to that question begins with a small group of cells called the neural crest, which are found early on in development in every single animal that has a face. In humans they’re only present from just three to five weeks into a pregnancy, and these little things are incredible.
They start off boringly enough as part of a tube running through an embryo’s part-formed body, which eventually goes on to become the brain and nervous system. But then they peel off and start migrating all over the place, from head to toes. Some of them become melanocytes - the pigment cells in our skin that turn browner when we’re exposed to the sun. Others become nerves in the gut. But, most importantly for our story, they also give rise to the bones and cartilage that give our face its features.
Some scientists think that this little cluster of migratory neural crest cells is one of the keys to the evolution of backboned vertebrate animals like humans, chimps, blobfish and all the rest. Known as the ‘new head’ hypothesis, or cephalisation, the theory goes that by evolving a face with a mouth and a solid skull to house your brain, you can start preying on simpler creatures. It’s thought that this played a big part in helping our fishy ancestors evolve away from the water and onto dry land
Stuff of nightmares
The jaw of a sea lamprey (Image by Drow_male via Wikimedia Commons)
Very simple fish called lancelets, thought to be similar to the kinds of creatures that were around hundreds of millions of years ago - don’t have a face as we’d recognise it. Instead, their long body ends in a nubby bump and a few waving tentacles. They don’t have proper neural crest cells as far as anyone can tell. Neither do lampreys, which are the fishy stuff of nightmares - just a long body with layers of horrible rasping teeth round a gaping circular mouth. No neural crest, no face.
Turning to her computer to pull up a presentation, Wysocka tells me about an interesting experiment from 2003 by Jill Helms and her postdoc Richard Schneider, proving that the neural crest cells are the facemakers. To do it, they manipulated embryonic birds growing in their shells - something that’s impossible to do in the secret darkness of the mammalian womb.
“It’s really amazing. They took neural crest cells from a quail early in development and then transplanted them from a quail to a duck, and they got a duck with a beak that looks like a quail’s - they call it a quck. And a duail is the opposite of that, a quail with a duck bill.”
These Frankenstein-esque experiments revealed that the neural crest cells have a certain amount of autonomy over what they make. Regardless of the fact that the quail cells were in a duck’s face, so all the molecular signals they were receiving would have been duck proteins, they still went right ahead and made a quail beak. The programme for what they were going to become had already been put into their genes - the switches were already set.
Looking at the images of the quck and duail on Wysocka’s screen, there’s a question I have to ask, and it’s not a pleasant thought.
Quck
Artist’s impression of what an adult duck with a quail’s beak might look like.
“So as a hypothetical experiment, if you took human neural crest and put it into a chimpanzee they would have a human-shaped face, and vice versa?”
“Provided that we could get the embryo to survive, in principle that's exactly what we would expect based on those experiments. Of course, it could only be a thought experiment, but how do you start addressing this question?”
How indeed. It turns out, Wysocka had an idea…
Obviously, transplanting neural crest cells between embryonic chimps and humans would be deeply unethical, as well as technically implausible as it’s almost impossible to get hold of human neural crest cells, even from aborted foetuses (the usual source for human embryonic tissue), due to the early time that they appear and their transitory nature. So instead, Wysocka is using an alternative approach: stem cells. But these aren’t just any ordinary stem cells - these are induced pluripotent stem cells.
It works like this. A couple of years ago, Japanese scientist Shinya Yamanaka discovered that adding a cocktail of four specific transcription factor genes into adult cells could make them forget all the biological decisions they’d made in their lifetime and turn them back into the kind of stem cells that are found in a very early embryo. Then, by adding other transcription factors, the newly-naive cells can be persuaded to go down different routes, such as becoming brain cells, muscle, or anything else.
Not only has this revolutionary discovery won Yamanaka a Nobel Prize and opened the door to all kinds of potential medical applications (new liver, anyone?), but it’s proved to be a hugely valuable tool for researchers studying the early stages of development.
Wysocka and her team have taken skin cells, called fibroblasts, from humans and chimps, then used the Yamanaka reprogramming technique to ‘brainwash’ them and convert them into neural crest-like cells. To all intents and purposes, at least from their tests, these seem to be a pretty good lab model for the impossible-to-get real neural crest. And, in theory, any differences in the genetic switches between these human and chimp neural crest cells should explain why we don’t look like our monkey cousins.
She pulls up some data on the screen - a bunch of slightly confusing graphs that reflect whether particular gene switches are set to ON or OFF in humans or chimps, by looking at whether certain DNA sequences are occupied by transcription factor proteins or carry other key identifying features.
“So, in both species we have sequences that stand out,” she says, tapping on the screen at points on the graph that stand out from the general trends. “These are enhancers outliers which are used differently in either chimp or humans.”
“And those are the interesting ones!” I gasp.
“Yes, that's what we think - that's where the magic happens. And there are about 2,000 that are more active in humans, and 2,000 that are more active in chimps, where they have at least three times stronger activity in their respective species. And there are about 200 in each species where it's almost like black and white – boom! – active in one, not active in the other.”
Her conclusion is that some of these enhancers are acting like simple light switches - they’re either ON or OFF, depending on whether they’re in a human or a chimp’s cells. But many more are like a dimmer dial, making their associated genes somewhat more or less active in the two species.
Over time, evolution has fiddled with the settings on these switches as humans and chimps diverged from our last common ancestor - some kind of proto-monkey that existed about 8 million years ago - creating the human faces we see all around us and the distinctive facial features of modern chimps.
What’s so surprising is that most of the differences in the settings are extremely small. Around 50 of the switches that Joanna and her team have looked at have major changes between chimps and humans - whole chunks of DNA that have gone missing over time in one species or the other. But most of them sit in DNA regions that are very similar, differing by just a few letters. A few thousand spelling mistakes, scattered in our genetic switches, are the difference between making a chimp’s features and human face.
But then things got really interesting.
When Wysocka and her team looked at the DNA sequences of the switches that turn on genes in the neural crest cells, they noticed that although they were similar in humans and chimps, they didn’t seem to be there at all in other vertebrates, even though they have faces too. And when she looked closer still, she noticed something strange.
Endogenous retroviruses
The difference between a chimpanzee face and a human face.
Many of these DNA sequences looked a lot like what are known as endogenous retroviruses or ERVs, usually classified as junk DNA, with small tweaks and changes that have crept in over time. These endogenous retroviruses are the remnants of ancient viral infection that have lingered in the genome. When they’re fully functional, retroviruses are able to hop in and out of the genome, but gradually lose this ability over time. But in the meantime, they can shake things up in the genome, causing new alterations and activating genes, depending on where they end up.
The particular ERVs that Wysocka and her team found invaded the genome of our proto-monkey ancestor, way back in evolutionary history before humans and chimps split. So they might have played a role in the evolutionary journey that separated primates from other mammals. But since then, they’ve been at work driving an evolutionary wedge between our two species, creating new control switches for genes that are turned on in the face-making neural crest cells.
As Wysocka explains, “These viruses are junk that invaded our DNA a long time ago, and our genome has to protect itself from the viral activity. Yet later we are co-opting these elements, using them to evolve human-specific features. So, I think it's kind of cool.”
“Are you saying it’s the viruses that made us human?” I ask.
“Exactly! But probably to evolve many of these features involve multiple genes and you need multiple changes - small changes at lots of enhancers. It's still about the combination, the cocktail of genes, but nevertheless I think it's really quite cool that we see mostly tiny changes and a lot of them overlapping with these virus elements.”
All of this helps to explain why even though we have virtually the same underlying protein-coding genes as a chimp, we're not chimps and we don’t look like them. It’s also an interesting example of how nature co-opts any kind of things it finds lying around if they look like they might be useful.
We’re carrying a whole load of crap around in our genome – with up to 8% of the human genome estimated to be various kinds of endogenous retroviruses, plus a bunch of other junk - and we just keep taking it with us through life. But, occasionally, a few of these old viruses happen to be in the right place at the right time to become new genetic switches that mix things up and make something different.
Maybe a couple of letters of DNA get changed, creating an attractive site for the gene-reading machinery to sit down and start reading a nearby gene or increasing its activity, which could slightly change how the organism turns out. Over aeons of time these subtle changes spread through the population and become the norm. Rinse and repeat over millions of years, and you’ll eventually get from a proto-monkey to two separate species - humans and chimps - that both look very different but are essentially made of the same stuff.
“It's not such a leap to say that probably if you evolve multiple enhancers that are human-specific near a particular gene you would dial its activity levels up. And then you would see a particular feature dialled up in human faces and dialled down in non-human faces.”
“What about people with genetic syndromes that affect their faces?” I ask. “Do we know if this kind of thing is happening in them?”
“That's exactly right - we’re learning a lot from humans as well as animal models.”
She explains that there are more than 700 known syndromes that affect the face and skull. In most cases they’re caused by a fault in one copy of a protein coding gene, slashing the amount of that particular protein in a cell by half. But Wysocka believe that alterations in genetic switches also have an important part to play. In fact, one recent paper from her team has shown that changes in genetic switches located nearly a million and a half basepairs away from a gene called Sox9 are responsible for a facial shape disorder called Pierre Robin sequence, or PRS.
But while syndromes like PRS that have a significant effect on the face and skull are an extreme manifestation of these underlying genetic changes, much smaller alterations in genetic control switches could be making a big difference, if not in biological terms then certainly socially.
“This is remarkably common. A 50 per cent reduction of gene dosage causes major face changes. So it's really not that crazy to think that 20 per cent increase or decrease in activity of a gene, or even less, will have an impact on the shape of the face.”
There’s a lot more work to be done to figure out exactly how subtly turning the dials up and down next to different neural crest genes affects how an animal’s face turns out. But it’s likely that even very small changes could have a noticeable impact, especially for a species like ours that is highly attuned to recognising - and judging - each other based on our appearance.
As we wrap up our chat, I ask Wysocka if she thinks she’s getting closer to figuring out how we build our faces.
“When I think that I may be understanding something, I think I'm getting an idea about how this works, then we get the next result and I realise we have no clue! There’s so much more complexity. I think that peeling off the layers of complexity and putting them back together to actually understand how the system works, that's really challenging.”
