Skip to Main Content
Text size: SmallMediumLargeExtra-Large

Heading in the Right Direction

Research into the genetics of craniosynostosis continues to progress. A recent case in point is the gene TCF12 

May 2013

Mention the word suture, and most people think of surgery and stitches. But the term has an anatomical meaning as well. Sutures are thin bands of fibrous connective tissue that are built into the developing skull just as stitches are built into the design of a soccer ball. There are four major types – coronal, lambdoid, metopic, and sagittal. Each winds along its own pre-patterned path to form supportive borders among the individual skull bones, much like caulk divides the panes of glass in a window. In this configuration, the sutures also prevent the individual bones from interlocking until after the third decade of life into a single sheet of protective bone, which brings a limited capacity to grow.

And grow the head must early on, adding on average 2.4 millimeters in circumference per week in the first year of life alone. Most of these millimeters stem from the growth of the brain and the skull’s coordinated response to expand. Enter the sutures. They contain cells with the innate ability to form new bone. These cells proliferate along the suture, adding new mineralized tissue at the margins of the skull bones. The bones then can grow and thrust further outward, allowing the brain to continue developing normally.

Given the critical growth dynamic between brain and skull, children must have intact sutures to allow craniofacial development to proceed without incident. But in roughly one out of every 2,200 live births, one or more of the sutures prematurely becomes bone and is lost, triggering abnormal, asymmetric skull growth and a condition known by the umbrella term of craniosynostosis.

Craniosynostosis can occur as part of a multifaceted syndrome. More than 100 syndromes have a reported craniosynostosis component. More frequently, the condition is termed non-syndromic, meaning infants are born with only the fused suture(s) and the resulting cranial malformation. Whether syndromic or non-syndromic, all cases are subcategorized according to a combination of clinical features, most prominently the name of the affected suture (say, the coronal suture) and the extent of the closure - unilateral (one side) or bilateral (both sides).

Andrew Wilkie, M.D.

Andrew Wilkie, M.D.

Like all developmental conditions, craniosynostosis varies in its severity. At its worst, craniosynostosis can be devastating. “The abnormal skull growth may be associated with raised intracranial pressure, impaired cerebral blood flow, airway obstruction, impaired vision and hearing, learning difficulties, and adverse psychological effects,” summarized Andrew Wilkie, M.D., a medical geneticist and basic scientist at the Weatherall Institute of Molecular Medicine in Oxford, England, who has studied craniosynostosis for 20 years.

In the March issue of the journal Nature Genetics, Wilkie and colleagues report identifying a gene called TCF12 that, when its encoded protein is produced at abnormally low levels, appears to contribute to a significant proportion of non-syndromic forms of unilateral and bilateral-coronal craniosynostosis. Preliminary analyses suggest that mutations in TCF12 may account for as much as 30 percent of bilateral-coronal craniosynostosis.

The coronal suture runs across the top of the head, from ear to ear. In bilateral-coronal craniosynostosis, both sides of the coronal suture close prematurely. Their shutdown places extra pressure on the other sutures to accommodate the growth of the brain, causing a characteristic developmental malformation in which the back portion of the skull is shortened, with vertical and lateral compensatory growth.

The finding is noteworthy for two additional reasons. One, Wilkie and colleagues employed next-generation exome sequencing to find the TCF12 mutations. An exome is the complete set of protein-coding sequences in a DNA sample. It comprises a little over 1 percent of human DNA. By cutting out the other 99 percent of a DNA sample, exome sequencing allows scientists to cut to the chase and take a detailed look at just the coding sequences for possible alterations. It’s rapid, comprehensive, detailed, and relatively inexpensive.
Rob Maxson, Ph.D.

Rob Maxson, Ph.D.

Second, the finding offers an excellent example of bidirectional research and its benefits. After identifying TCF12, Wilkie turned to mouse developmental biologist and NIDCR grantee Rob Maxson, Ph.D., at the University of Southern California. Maxson and colleagues performed a key experiment that nailed down the role of TCF12 in bilateral-coronal craniosynostosis.

Science Spotlight recently spoke with Drs. Wilkie and Maxson to provide some perspective on the finding, exome sequencing, and the ongoing progress in craniosynostosis research.

As a science writer, I could use Charles Dickens’ classic line, “It was the best of times, it was the worst of times” to describe the state of just about any area of research. But the best-and- worst-of-times subhead was very much the case writ large in craniosynostosis research about five years ago. Why?

Wilkie: I think it’s just the ebb and flow of science. In November 1993, I can remember that I was in my first week here at Oxford when I opened the new issue of the journal Cell. I scanned the table of contents and saw that Rob [Maxson] and his group had discovered the first gene, MSX2, involved in craniosynostosis. The discovery helped to spark a true golden age in craniosynostosis research, energizing other groups to keep hammering away to find other genes involved in syndromic craniosynostosis. During the next decade, the results were quite spectacular. We could point to mutations in four genes – FGFR2, FGFR3, TWIST1, and EFNB1 – and estimate that they accounted for nearly a quarter of craniosynostosis.

But the progress then slowed? 

Wilkie: By the mid 2000s, the investigative well had run dry. It wasn’t unique to research on craniosynostosis. This was a theme in other developmental conditions. At issue was the family linkage studies that brought the four genes that I just mentioned and other key finds had been so successful that they exhausted their investigative resources. That meant few family cohorts and syndromes remained to be studied, and the field stood at a crossroad. Important genes certainly remained to be found on the non-syndromic side that would help to fill in the missing blanks in our understanding of suture development. But linkage studies wouldn’t work to find them. We lacked the means to continue pushing forward.

And the linkage studies wouldn’t work because of the complexity of the genetics in non-syndromic craniosynostosis?

Wilkie: That’s right. Non-syndromic craniosynostosis is exceedingly challenging for a number of reasons. First, there are no genes in the human genome that are specifically dedicated to the development of a cranial suture. The genes that are expressed in the suture are used in the development of all sorts of other bodily organs. For a child to develop craniosynostosis via a genetic mechanism, a huge amount of arbitrariness is involved.

What do you mean?

Wilkie: What I mean is to reach a developmental stage where craniosynostosis can occur, the embryo already has passed through a huge developmental sieve in which the underlying gene mutation could have undermined development in a number of other ways. So, even though the mutation ultimately would have given rise to craniosynostosis, the embryo never reaches that stage of development. Something else proves fatal first.

Another issue is in about one out of every three infants who end up passing through the developmental sieve, their non-syndromic craniosynostosis arises from a new mutation. In other words, the affected child is the proband, or source of the condition. That means collecting DNA samples from other family members becomes meaningless. Linkage analysis won’t work to narrow down the location of the altered gene.

And that’s where the arrival of exome sequencing in 2009 proved so timely.

Wilkie: Exactly. Exome sequencing allows us to drill down and look directly at the coding sequences. We can compare them to known coding sequences and pull out the possible variants for further study. Ten years ago, I doubt people thought it would be possible to sequence an exome or full genome in the manner that is quite routine today. The next-generation technology came upon us very quickly, and it just keeps improving and enabling needed discoveries.

That’s why I say craniosynostosis research now is entering a second golden age of discovery. In the March issue of Nature Genetics, we published a second paper on another craniosynostosis gene, ERF. We also found it with exome sequencing. In November 2012, we published another exome discovery in the American Journal of Human Genetics. That comes to three genes that we’ve identified by this means. Without this technological leap forward, these discoveries wouldn’t have been made. We’d still be standing at a crossroad.

Getting the Genetics Right

In preparing for our conversation, I heard your laboratory identified TCF12 as a candidate gene for craniosynostosis about a year ago but didn’t realize its significance at first. Is that true?

Wilkie: It’s absolutely true. Let me give a little of the backstory. I’m a clinical geneticist, and a few years ago I was working with three families that, genetically speaking, really bugged me. Each had multiple family members with craniosynostosis, and that told me there had to be a genetic cause. But I couldn’t find anything, and I had thrown the investigative book at them. I finally decided to tackle the problem by exome sequencing.

But there was one family in particular that caught your eye.

Wilkie: That’s right. In one of the families, two siblings had bilateral-coronal craniosynostosis. All of their genetic tests had come up negative for alterations in known genes. So we examined the parents very carefully for any subtle craniofacial changes that might be potentially informative. Nothing. I assumed this was a textbook case of a recessive gene from both parents being passed on to their two children. And that got me thinking about something else.

What’s that?

Wilkie: I had noticed that the parents were from the same part of the UK. I figured they had to be distantly related. With the exome sequencing, I thought we had a fighting chance of finding a small, mutated segment of DNA that the children had inherited from both parents.

Did the exome sequencing do the trick?

Wilkie: It did. But, getting back to your original question, here’s where the delay surfaced. Remember our working model was a recessive mode of inheritance. When Steve Twigg, a post doc in my lab who worked on the project, analyzed the family’s exome sequences, a TCF12 nonsense mutation popped up but didn’t make the cut. Why? Although the mutation was in both affected siblings, Steve determined that it was heterozygous. In other words, the mother was the source, and, of course, she was fine. So, based on our model, he correctly discounted the mutation and moved on to the next possibility.

You eventually got the gene, though.

Wilkie: Through good fortune. About a year later, Vikram Sharma, a surgical student, presented data during a lab meeting from a completely different exome-sequencing project that involved seven patients with bilateral-coronal craniosynostosis. His data showed that two of the seven had changes in the TCF12 gene. Steve, of course, was sitting there and taking in the data. He thought back to his project and the family, and the gene rang a bell. Once the meeting finished, he hurried back to his office and found that indeed he had identified this TCF12 mutation in the two children.

Interestingly, after this same lab meeting, lab member Aimée Fenwick looked at her data and found further TCF12 mutations. So, in twenty-four hours, we went from literally not knowing about TCF12 to having three separate projects that had pulled up mutations in the gene. So, we knew immediately that TCF12 was an important gene that clearly would be involved in a major proportion of craniosynostosis.

So the cautionary tale is: Get the genetic model right at the front end.

Wilkie: Exactly. We thought the condition was recessive. It was dominant with non-penetrance. The answer had been sitting right in front of us, and we missed it. That happens with deep-sequencing projects. They can generate dozens, if not hundreds, of gene variants. All potentially could be causative. So, you’ve got to look at the data in the right way and recognize the 99 out of 100 that are just red herrings that will lead you blithely down the wrong path. And that’s often difficult to do.

The Mouse Literature

When a gene involved in a human disease is identified, scientists typically turn next to the mouse literature to glean leads into its possible function in the body. Was that your next step?

Wilkie: Yes, we looked at the mouse. But the mouse literature was unhelpful in this case.

Why’s that?

Wilkie: The mouse geneticists had shown that if both copies of the Tcf12 gene are inactivated, the mouse pups have terrible immunity because of a failure to make fully normal B and T cells. So their focus was completely on immunology and offered no insight into the effects of losing just one copy of the gene. But, for all intents and purposes, losing one copy in people seems to be the norm and results in craniosynostosis.

But the research literature did provide a crucial piece of information.

Wilkie: Oh definitely. We discovered that TCF12 encodes a protein that functions as a transcription factor, meaning it binds to DNA to help regulate gene expression. The TCF12 protein belongs to a specific class of transcription factors called helix-loop-helix proteins.

Helix, loop . . .?

Wilkie: I know it’s a mouthful. But the name actually is quite descriptive, like a track-and-field coach saying a hop, a skip, and a jump. The proteins have a core structure but also an arm that consists of a helix, a loop, and a helix. Knowing that TCF12 is a helix-loop-helix protein immediately got us thinking about a possible mechanism.

Involving the protein TWIST1?

Wilkie: That’s right. About six years ago, a protein called TWIST1 was found to be associated with sagittal and coronal craniosynostosis. TWIST1 also is a helix-loop-helix transcription factor. We know that its core protein binds to the DNA, while the nearby helix-loop-helix part forms a binding partnership, or dimer, with a second protein. Here’s the point to remember. The second protein could be either another copy of TWIST1 or another helix-loop-helix protein.

And your prediction was the second protein is TCF12?

Wilkie: Exactly. We predicted that TCF12 interacts with TWIST1. It’s one of those things that once the answer presents itself, it seems very obvious. The TWIST1-TCF12 heterodimer essentially sticks like a pair of scissors between the groove of DNA to regulate gene expression. When the scissors don’t work properly, the genetic program malfunctions.

How did you test your prediction?

Wilkie: In two important ways. First, I got on the phone with our clinical collaborator Irene Mathijssen in the Netherlands. Irene kindly sent us about 50 DNA samples to study from people with bilateral-coronal craniosynostosis. That enabled us to find a lot more mutations in TCF12. It also gave us a better sense of where the mutations tend to occur in the gene. That’s invaluable information moving forward.

Secondly, I had a very clear experiment in mind that one could imagine doing in the mouse. It would be a great complement to the human genetics. So I got on the phone to Rob Maxson at the University of Southern California. Rob and I have worked together for years, basically giving each other hints about genes that might be involved in craniosynostosis. We probe the human genetics, and they do the mouse genetics. As far as I’m concerned, Rob is the go-to guy when it comes to the coronal suture.

Mouse Genetics

You received a phone call from Andrew Wilkie. What was the experiment that he proposed?

Maxson: Andrew proposed that we test in mice the idea that TCF12 and TWIST1 might work together in cranial suture development. The plan was to create a double heterozygote mouse. In other words, the mouse pups are born with one functional and one inactivated copy of both Tcf12 and Twist1. The hope was partially inactivating both genes would produce severe bilateral-coronal craniosynostosis and confirm what he thought was happening in people. I agreed to make the mice, and Andrew impressed upon us the need to work very quickly. The human genetics community is extremely active, and every group is worried about getting scooped. Breeding the mice took some time, but we managed to get the double heterozygote as quickly as possible.

And what did you see?

Maxson: Well, it was quite spectacular. The double mutant had severe bilateral-coronal craniosynostosis in which the coronal sutures were just completely fused in all of the mice.

Did you also breed mice with other genetic combinations of the genes?

Maxson: Yes, we did. The Twist1 knockouts had just partial fusion on one side or the other of the suture. Then the individual Tcf12 homozygote, which isn’t mentioned in the paper, has a mild coronal craniosynostosis. So, the experiment worked the way that we thought it would. Either gene individually has some effect. Put the two together, you get a very severe effect.

How would this helix-loop-helix lead to premature fusion of the coronal suture?

Maxson: We don’t know the answer yet. But, broadly speaking, the answer is written into how skull bones form early in embryonic development. The process starts when progenitor cells exit the neural crest, an area of fetal tissue along the neural tube, and travel to the area above the eye. These cells wait there for a while and then migrate up into the dermis to insert into the leading edge of what, under their direction, will mineralize into bone. The migration of these cells into the leading edge is orchestrated by TWIST1 turning genes on and off.

What can go wrong?

Maxson: What can go wrong is the strict physiological boundaries of this genetic program can go awry during the migration. That is, as the progenitor cells migrate, instead of respecting the coronal suture as a natural boundary and stopping right there, they just keep right on going, migrate into the suture, and form bone. That fuses bone and suture, resulting in bilateral-coronal craniosynostosis.

When you saw the severe bilateral-coronal craniosynostosis in the double heterozygote mouse, did you pick up the phone and call England?

Maxson: No, we did better than that. I sent Andrew an email with pictures.

Wilkie: I was ecstatic. There couldn’t have been a better result from that experiment, especially when you think what could have gone wrong. Creating a double heterozygote could have been lethal. Or, the results might not have been so clear-cut. But you’ve got a case where the interpretation was unequivocal. The Tcf12 heterozygous mouse has normal-looking coronal sutures at a gross level. Then you cross it with a Twist1 mouse heterozygote, and they are just obliterated. I mean, it’s extraordinary. That was a very exciting moment, of course. I opened the email and said, “Amazing!”

From Discovery to the Clinic 

Bringing this full circle, what does this tell you about the families and trying to provide them with better diagnostic care to differentiate between the different types?

Wilkie: Let me give you a clinical vignette. Whilst all of this was going on, I received an email out of the blue from a lady who had been treated at our unit as a child and had been involved in a research protocol. She contacted me almost as a last resort. She’d been to see her local clinical genetics team, and they told her because she’d had both coronal sutures fused as a child, her condition was suspected to have a genetic basis. The implication being, it was likely that her children had a 50-50 chance of inheriting her condition. The trouble was nobody could seem to identify her condition, and she felt backed into a corner. People said, “Your risk is high, but I’m sorry that I can’t advise you any further.” She got in touch with me because she’d been treated at Oxford as a child and wanted to ask me, “What do I do next?”

What did you tell her?

Wilkie: The thing is I knew what her situation was because in the previous month or two before she’d emailed me, we’d identified that she’d gotten a TCF12 mutation. It meant that the advice that she’d been given by the other clinicians was entirely correct. But having identified the causative gene, I had access to transformative information for her. It changed things from “yes, we think there is this risk, but that’s as far as we can take it” to “we know exactly what caused the condition.” It’s a mutation in this gene, and here are the various options available to deal with that information.

Such as what?

Wilkie: Well, they’re quite an interesting mix actually that will vary for different people. One way of looking at this condition is in most individuals it can be fixed with a single operation. So having given her condition a name and having identified its natural history in a series of patients, the fear of the unknown may be mitigated for her. This woman might decide that she is prepared to take the risk of having children. Her children will have a 50-percent chance of inheriting the mutation. However, we now know that inheriting the mutation doesn’t necessarily always mean the child will be affected. As I’ve mentioned, there’s quite a high level of non-penetrance with this condition.

Another issue is we have shown that the majority of affected individuals do well from surgery. Given this information, she might decide to go ahead and take the risk without any tests at all. If she wanted to ensure that one of her children didn’t have to repeat her experience as a child, there are two options. One, she could try prenatal diagnosis, which is higher risk and wouldn’t be everyone’s optimal choice. Two, she could have a pre-implantation diagnosis, in which the whole thing is done in the test tube as in vitro fertilization. In this scenario, only an unaffected embryo would be re-introduced into the uterus. We haven’t counseled the woman yet, and different people will opt for different choices. The important point, though, is she now has the information that will help to empower her to make the decision that is right for her.

From Discovery to New Science

What types of new science does this discovery open up?

Maxson: That’s hard to say at this point. The gene always is the entry point. But what’s nice in this case is TCF12 and TWIST1 work together. That gives us more of the mechanistic story to explore right off the bat, and clearly it is an important one to understand. So there should be much more to come from this discovery.

You’ve been a part of two decades of progress in craniosynostosis research. How do you see the basic-research success feeding into the clinical realm?

Maxson: The unfortunate thing that I’ve discovered about craniosynostosis is it gets started pretty early in development. In the mouse, things begin going wrong by 13 or 14 days after conception, or before the skull is completely present and all of the bones are there. That means that any intervention within this developmental window would involve, first, recognizing that there is a problem and, secondly, trying to do something about it in utero. That’s problematic on multiple levels.

It’s also possible that some of our lessons learned in the laboratory will benefit children after birth. One possibility is to refine surgical repair of the cranial bones. I just attended a joint craniofacial meeting between the University of Southern California and the University of California at San Francisco. Among those in attendance were craniofacial surgeons. They mentioned to us that we should start thinking about how to make an artificial suture that could be implanted into patients. The reason being, a common post-surgical problem is the affected bones grow together again, and a second surgery is needed. An artificial suture might help to solve the problem.

Well, how would you approach it?

Maxson: The first step is to learn to isolate, maintain, and culture sutural cells. That’s a real challenge, but it’s certainly doable. That would allow us to learn a great deal more about their biology and weave this new knowledge into the larger issue of stem cells. One possible approach would be to take fibroblasts from elsewhere in the body, say the skin, and reprogram them to be sutural fibroblasts. Then you might have replacement tissue to implant.

And, of course, as the biology advances over the next several years, it will inform and refine the engineering of an artificial suture.

Maxson: Exactly. The session really got us thinking. Obviously, we don’t know how to engineer a replacement suture. But we decided as a group to proceed with the idea. We’ll arrange another meeting in the near future to think about which members of our multidisciplinary team would need to do what.

The point is, the basic science is starting to get you there.

Maxson: That’s absolutely right. We know enough now about what a suture actually is, how it works, and how it develops. It’s not a totally crazy idea to create an artificial suture. Ten years ago, it would have been a real stretch. Not today, and certainly not tomorrow.

Share This Page

GooglePlusExternal link – please review our disclaimer

LinkedInExternal link – please review our disclaimer


This page last updated: February 26, 2014