This article was taken from the September issue of Wired UK magazine. Be the first to read Wired's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online
Belinda Kembery was pregnant with her second child when she realised something was wrong with Robbie, her first. "He didn't quite get to walking, and then, when he was just over a year old, he stopped sleeping well. He would wake in the night and it would be as though he'd just had a shot of caffeine... he would be rocking backwards and forwards, very agitated, and it would be very hard to calm him down."
Kembery, 41, a solicitor before her marriage, is sitting in a spacious, shining kitchen conservatory in Clapham, south London, a room so calm and tidy you would think it had never had children in it at all. "The next sign we noticed was that he would be sitting up or crawling around, and he would suddenly fall over. That's why we put him in a bike helmet. One day, watching him very closely, we realised he was having a blackout. I went to the GP. We had weeks of appointments - brain scans, blood tests, lumbar punctures. The day we had the diagnosis, I was 22 weeks pregnant." The child she was then carrying is now a healthy, seven-year-old boy.
But Robbie, she was told, had Batten disease, caused by the malfunction of a single gene - CLN1, on chromosome 1, in Robbie's case. Normal copies code for a protein that helps to break down fatty molecules - lipofuscins - within brain and nerve cells.
Without it, the cells are choked in fat and die. The first symptoms are seizures; then comes blindness, dumbness, paralysis and worse.
Robbie would never walk, never speak, and by three he could not see nor swallow. The weakened muscles of his diaphragm gave him acid reflux so terrible that it stripped the enamel from his teeth.
Eventually, an operation closed off the top of his stomach. Today, he is fed through a tube. "Nothing at all had prepared us for this diagnosis," Belinda says. "In the hospital, they asked me to count the number of seizures he had had, and I got to 100 and stopped counting - and that was before lunchtime. Poor little thing. He was really suffering. They told us he would be in a wheelchair by two. "The things that newborn babies are able to do, like smile and have a hand grip - that is all he has retained. We do get feedback: he loves music. He loves sitting in the garden in the sunshine and being read to. "But Batten disease is quite black and white to me. I would not want to bring a child into the world with Batten, for one good year and then eight years of..."
The disease is recessive, so carriers who have a normal copy of the gene need know nothing of it. Only in Finland, where one in 20,000 of the population is a carrier, is it sufficiently common that people have heard of it. In Britain, perhaps one child a year is born with it. But when Robbie was conceived, the random luck of meiosis and recombination gave him two defective copies of the CLN1 gene. The odds were against it. As with any recessive disease on that chromosome, there was a 75 per cent chance of a child without symptoms who would have ended up with at least one properly functioning copy of CLN1.
Four years ago, when Robbie was five, the Kemberys' geneticist told them of a new test that could screen embryos for defective genes such as Batten. The test relies on the extraordinary nature of very early embryos: researchers can divide a four-cell embryo, produced by IVF, into two two-cell embryos, and both will develop into normal individuals. They can even remove the outer coatings to push cells from different embryos together, and the resulting mosaic will develop into an individual. So taking one or two cells from an eight-cell embryo does not much hinder the development of those that remain into a baby; and the cells removed can have their DNA tested. If it turns out to be free of the disease, the embryo is implanted in the mother's womb, as with normal IVF; if faulty, then - in the clinical term that researchers use - it is flushed away.
The technique is known as pre-implantation genetic diagnosis, or PGD. Until recently, its use has been rare, complicated and expensive. In Britain, it has been approved for only a very narrow range of single-gene conditions such as Batten. Belinda and Jonathan Kembery used it to have a healthy child to complete their family: their third son, David, was born free of the disease in September 2006.
So far, the use of PGD has been relatively uncontroversial: its use to eliminate embryos carrying genes for such conditions as Batten disease has been seen - religious objections aside - as a triumph of science and medical understanding over some of life's random cruelties. But as our understanding of the genome grows, such techniques offer wider possibilities for selecting genetic traits deemed worthy of elimination - not all necessarily fatal. In January, University College Hospital in London announced that a baby girl had been born there who had been screened, as an embryo, for the form of BRCA1 gene which greatly heightens a woman's lifetime risk of developing breast cancer, from six to 80 per cent.
This was the first known case of screening for a gene which merely affects the probability of contracting an illness. BRCA1 may not be the death sentence that CLN1, the infantile Batten gene, is, yet some women have resorted to double mastectomies rather than run the risk. So, if a couple at risk decided to eliminate the chance that the gene might be passed on, who could blame them?
But once we move away from trying to eliminate genes which kill to trying to eliminate genes which merely disadvantage their bearers, where do we stop? Science has identified genes that affect practically every part of our lives. These range far beyond physical conditions known to have genetic components - such as susceptibility to heart disease, Alzheimer's, diabetes and many cancers - to character traits such as depression, political fervour, and even popularity, which twin studies suggests also has a large genetic component. Certainly skin, eye and hair colour are all under the influence of fairly well-understood gene complexes.
So what colour would you like your baby girl's eyes to be?
A few months ago, an American fertility clinic briefly asked exactly that question. The Fertility Institutes, which already offer sex selection via PGD via their clinics in New York and Los Angeles, announced that they intended to use PGD to make "predictive genomics" available to parents.
The clinics' announcement, posted on their website on December 12 last year, was to the point: "We are pleased to announce the pending availability of a greatly expanded panel of available genetic tests that may be combined with our world renown [sic] aneu-ploidy and gender selection testing. For the first time ever, patients having genetic screening for abnormal chromosome conditions in their embryos will be able to elect expanded testing that can greatly increase the odds of achieving a healthy pregnancy with a preselected choice of gender, eye color, hair color and complexion, along with screening for potentially lethal diseases, screening for cancer tendencies (breast, colon, pancreas, prostate) and more... These services will be offered in both Los Angeles and New York."
The announcement provoked a swift media backlash, with ethical and medical experts outraged at the notion of "designer-baby shopping" or "a Build-A-Bear-style baby". On March 2, the institutes' website announced that the programme had been voluntarily suspended. "An internal, self-regulatory decision has been made to proceed no further with this project," it stated. "Though well intended, we remain sensitive to public perception and feel that any benefit the diagnostic studies may offer are [sic] far outweighed by the apparent negative societal impacts involved."
But ethical arguments are beside the point. Anyone who understands the science knows that what the clinic was proposing was just impossible.
The arithmetic simply doesn't work. Suppose that in 20 years' time we understand the genetic basis of intelligence. And suppose that it turns out that there are ten or 12 genes which act in concert to produce in a child an ability to develop great intelligence. Even assuming that the parents had all the desirable variants between them, they would need to sort through at least 1,024 embryos to hope to find the perfect one. But that is a wholly impractical figure. Even after her ovaries have been artificially stimulated, a woman will not normally produce more than ten eggs in any course of IVF - not all of which will be fertilised. "Provided IVF is successful, it's now feasible to select an embryo for a single gene trait," says Sir John Sulston, who won a Nobel Prize for his work on sequencing genomes. "For instance, one without the dangerous variant of BRCA1 carried by a parent. On the other hand, in order to select height or intelligence, we need to find out much more about the interactions of the numerous relevant genetic variants. And even if we do, we shan't be able to get enough embryos to select a desired combination."
So every time you test for a particular gene or chromosomal arrangement that may be inherited equally from either parent, you halve the number of embryos available - on average, half of them won't have it. So for every additional feature on the chromosome that you select, you need to double the number of embryos to test.
If parents just want a girl, half of their embryos will, on average, be suitable. If they want a girl who does not have the dangerous BRCA1 variant, then one in four embryos will be suitable.
Add one more test - say, for a gene known to predispose the carrier to heart disease - and you need eight embryos. The next test requires 16, and so on. Most of the traits we are interested in are influenced by ten, or 20, or 50 locations on the chromosome - no one knows yet. But we do know that there isn't "a gene" for almost anything except very rare, very terrible diseases such as Batten - for which PGD is already licensed.
So whatever sensational claims are made by private clinics, it's a mathematical fact that a woman will never produce enough eggs in her own lifetime to allow her anything like the number of choices she would need to create the mythical "perfect" baby.
But if modern genetics won't let us pre-select babies, perhaps it makes it possible to understand the enormous consequences of the exchanges of little bits of chromosomes when eggs are fertilised.
The more that scientists discover about genes and their effects, the more they want to know, even if all we can do is work around them. For scientists, knowledge is fascinating in itself - and there is always the hope that if they understand what is happening at a genetic level, they can eventually mitigate its consequences in our bodies and minds. So if they discover a genetic predisposition to, say, depression, although they are unable to create an embryo that eliminates it, they could well recommend particular drugs which may be specially effective for that individual in later life.
One of the first scientists to study how the whole genome affects an embryo's life chances is Professor Alan Handyside, based at the London Bridge Fertility, Gynaecology and Genetics Centre. He recently began clinical trials of a new embryo-screening process that he calls "karyomapping". His work has been made possible by huge advances both in technology and knowledge since the first human genome sequence (for which Sulston won his Nobel Prize) was published in 2001.
The first human-genome sequence was meant as a reference. It was not the sequence of any particular human being. This is important to keep in mind, because everyone differs from the basic human genome and in these differences - haplotypes - lie much of our medical individuality. Since 2001, a global effort has been underway to examine our genetic differences. By sequencing a hundred people from different ethnic groups and then datamining their genomes to match them to common diseases, it has been possible to find haplotypes that correspond to a higher risk of heart disease, many cancers and diabetes. Beyond that there are indicators for many mental illnesses - not just depression, but also schizophrenia and autism-spectrum disorders.
Researchers have created high-resolution maps of most of the important areas of the human genome and there are affordable micro-arrays - large-scale testing devices - which will test hundreds of thousands of haplotype sequences to locate anyone tested on these maps. Because haplotypes close to each other on a chromosome tend to travel together down the generations, it is possible to make confident predictions even about ones which have not been sampled, providing we have sampled their near neighbours.
This approach vastly expands our knowledge of the genome without requiring detailed understanding of how it works, which is a much slower and more difficult process.
For Handyside, this offers a new prospect of detailed knowledge about the likely characteristics of any embryo. "I do foresee in the near future offering this to parents to examine them for haplotypes, for example, related to mental illness and autism," says Handyside, a quiet, bearded man whose accent retains traces of his Tyneside roots. "Parents are going to be very interested in saying, 'Am I at risk?' "We have always known that if you have a family history of breast cancer, or stroke or heart disease, that makes you likely to have this. That's just a fact of life. But what microarrays will let us do is say, 'You actually inherited those risky areas of the genome from that side of your family that has breast cancer.' "So we might run the array and say, 'You've got a deletion here which can be associated with this.' We would be able to tell you whether that part of that chromosome has been passed on to the embryo."
Handyside has been working with embryos for all his scientific life. He was originally a mouse biologist, but, he points out, the techniques of sequence-based medicine do not change much between species: "The difference between mice and humans is only about 100 genes."
He worked under Martin Evans, who later won a Nobel Prize for the isolation of embryonic stem cells in mice; in 1986, Handyside and his team published simultaneously in Nature with Evans the news that they had grown a line of mice with a gene deliberately knocked out, so that it could not properly metabolise uric acid. In humans, this produces Lesch-Nyhan disease: afflicted children chew compulsively at their lips and fingers.
These mice are produced by mixing stem cells that have had the gene knocked out into very early embryos, which then grow into "mosaic" animals in which some cells are descended from the original embryo and others come from the intruded stem cells. If the gonads grow from the intruded cell, the mouse will produce offspring in which the gene has also been knocked out. Identifying these as embryos was the first use of PGD in any species.
By 1990 , Alan Handyside was working with Robert (later Lord)
Winston at the Hammersmith Hospital in London, where they performed the first successful PGD in humans, testing the sex of eight cell embryos in order to make sure that only female ones were implanted in a woman who carried a mutation which is lethal in males. By 1992, they were ready to perform a more subtle version of the technique, and one of their patients successfully produced a baby that had been specially selected as an embryo because it did not carry a gene for cystic fibrosis.
Gradually, the technique became widely accepted - if not by religious bioethicists - though it remained rare and expensive.
Guy's Hospital in London, the largest British centre, has carried out 200 such procedures in the last decade, compared to roughly 40,000 IVF births in that time.
But recent advances in sequencing technology and a more sophisticated understanding of the ways that genes work seem to Handyside to point towards a future when PGD will be able to reveal a great deal more about embryos and their likely fates if they are born. The technology has been advancing as fast as computers have been growing more powerful - and Moore's Law states that processing power doubles every 18 months. "When I started, you could sequence 10,000 base pairs in three years, and that was enough to earn you a PhD," he says. "Now we can do 100 million base pairs in seven hours; Nature has predicted that by the year 2050 we will have enough computing power to sequence the six billion base pairs of everyone on Earth."
This means that, even if it is impracticable to change embryo genetics much, we can understand a great deal more about them.
Instead of looking at single genes, which mostly affect very rare conditions, it is now possible to look at some of the combinations of genes which affect much more widespread conditions, those which vary over a spectrum.
At the same time as the reach of genetic explanations appears to be expanding, so is the definition of "gene"; and this is the focus of Alan Handyside's current work. When the first genomes were sequenced, the division between coding genes and junk DNA seemed to be perfectly clear: 97 per cent of the human genome was meaningless junk and the remaining fragments made genes, even if coding genes were broken up by apparent stretches of nonsense that are not transcribed. But the more we learn, the less of this clarity remains. Bits of the genome that looked meaningless turn out to carry rather important signals.
It is only loosely true that the DNA of the genome codes for anything. It is the machinery of the cell which decides which parts of the genome will be transcribed into RNA and used to make proteins. So as soon as the genome of an organism has been worked out, and the raw DNA mapped, the next step is to analyse all the RNA products the cell makes from DNA, known as the transcriptome, and then all the proteins this RNA codes for, known as the proteome. It's an enormously complicated task.
Depending on how it is transcribed, the same sequence of DNA can code for different proteins; and when these proteins are analysed, they too can differ subtly, even when they are the products of the same gene. And all this complication matters. "We have discovered another layer of genetics in the last few years," says Handyside. "We already had the genome, the transcriptome and the proteome. We have known about epigenetics for some while. But we have now discovered in the last three to five years that there is a very widespread copy number variation."
Both inside and outside the coding regions of genes, there are passages where the DNA just repeats itself for a while, without seeming to do anything. These can vary in length from a couple of hundred base pairs to entire genes and more. The important point, which has only become clear as more genomes are sequenced, is that these repeats vary between individuals, and they affect the way that genes are expressed and the proteins that they code for.
One of the earliest examples was the discovery of a variable-length repeat around ST, a gene that codes for serotonin receptors. Serotonin is one of the most important neurotransmitters in the brain. Almost all modern anti-depressants, such as Prozac, work by boosting the levels of serotonin, even though no one knows quite how or why. What is more, they work differently on different people. Variable-length repeats seem to explain why this happens.
A team under Karl-Peter Lesch in Würzburg, working alongside American researchers from the National Institutes of Health, discovered in 1996 that the length of a repeat sequence outside the gene, but important in its transcription, affected the probability of major depression. Then, in 1999, a Japanese team discovered a variable-length repeat sequence inside the gene. The shorter the repeat sequence that anyone carries, the more likely it is that they will be anxious, prone to mood swings - and unresponsive to the drugs, like Prozac, which boost serotonin levels in the brain by preventing it from being broken down.
But there are many, many more repeat sequences that may be just as significant. "First reports said there are 150 of these; three years later, they had found 1,500," says Handyside. "Now, as we look with more resolution on the microarrays, we are finding more and more."
In fact he believes that it may be these copy number changes, rather than the emergence of whole new genes, which are responsible for the changes that distinguish us from apes, and even mice. Even if there are only a hundred entirely new genes in a human, when compared to a mouse, the varying copy numbers within the genes that we share can change their effects to produce the great physical differences between man and mouse.
Yet if the premise of the "designer baby" remains a scientific impossibility, that's not stopping scientists and philosophers from debating how to improve on nature. The thing about scientific impossibilities is that we can never be certain they will stay impossible: we just have no idea how to make them possible. Skipping lightly over the fact that we can't do it, and have no idea as to whether we will ever be able to do it, the Princeton geneticist Lee Silver, in a book published ten years ago, foresaw still further developments, once the genome was completely understood. Parents then would be able to test for particular character traits and talents - anxiety, skill at music and maths - and choose which balance they wanted for any particular child.
Certainly this would add a new venom to subsequent quarrels between the parents. Instead of blaming the other partner's family for their child's failures, they could blame the genes the other partner chose.
More modestly, Alan Handyside looks forward to a time when it will be possible to generate avatars on a computer for every embryo, showing how they will look as a result of the particular haplotypes identified by a quick genetic scan.
Lee Silver's dream will remain science fiction until some way is found to produce unlimited embryos to choose from. If adult cells could safely and reliably be reprogrammed to work as pluripotent stem cells, and if these stem cells could safely and reliably be coaxed into developing into eggs and sperm, it would be possible in principle for any couple to produce unlimited embryos. This has been demonstrated in mice; but we are still a long way off from anything like this in humans.
In March 2009, Nature published the news that researchers in Scotland and Canada seem to have found a way to generate pluripotent stem cells from adult human beings. Yet years of testing will be needed before the method is known to be safe, and we still have no way of directing these cells to differentiate into eggs or sperm.
Even if these obstacles are overcome, it wouldn't mean that anyone could have the perfect baby. If neither parent has the genes to excel in any particular field, they won't be able to pass such excellence on to their offspring. "We can't all be Beckhams," as Handyside says.
But should we even be dreaming of such futures? Doesn't this sort of tinkering threaten huge risks for a limited benefit?
Not according to Julian Savulescu, an Australian philosopher at the Oxford Uehiro Centre for ethics. He unequivocally backs the quest for more "perfect" children. Savulescu argues that we don't merely have a right to try genetic enhancement, we have a positive duty to do so, both for our children and for society. If it becomes possible to identify genes that will reliably increase intelligence, for instance, then parents should exert themselves to find them.
Savulescu argues in his Principle of Procreative Beneficence that "couples (or single reproducers) should select the child, of the possible children they could have, who is expected to have the best life, or at least as good a life as the others, based on the relevant, available information".
He is also in favour of the use of performance-enhancing drugs in sport, and indeed academic life, providing that these don't cause measurable harm - after all, people drink coffee for the cognitive improvement it brings. Suppose a drug were discovered which reliably raised adult IQ by four or five points and had no side effects. Would it not be the right thing to take it ourselves?
Would it not be responsible parenting to ensure that our children were given an adequate supply? After all, a high IQ can be shown to be good for everybody: more intelligence diminishes crime and increases wealth. So parents have a duty of care to promote intelligence in their children, and society has an interest in seeing that they do.
Savulescu points out that genes influence character traits such as aggression and criminal behaviour, monogamous tendencies, novelty-seeking, schizophrenia and substance addiction. Freedom from heart disease, from cancer, freedom from Alzheimer's - all these are to some extent under genetic control, and are all worth having in themselves, irrespective of whether anyone else has them.
Similarly, there is evidence from studies in the last ten years that some of our basic moral intuitions may be under genetic control: a sense of fairness, and the desire to become politically engaged have both been cited, as have styles of political engagement and even religiosity. At present we are a long way from knowing where on the genome to look for these - but the haplotype-mapping techniques that Alan Handyside enthuses over could in principle be turned towards mapping these traits just as easily as they map diseases. As soon as anyone sees a profit in it, it will be done.
Nor is this vision limited to human genes. Human insulin for diabetics is already produced by genetically engineered microbes.
What if the necessary genes could be spliced into our own bodies?
Already, in an act which casts a chilly light towards the future, human embryos have been produced which express the green fluorescent protein originally derived from jellyfish. If pig hearts or pig kidneys contain genes that would be more useful to us than our own, why not place these in selected embryos?
John Harris, professor of bioethics at the University of Manchester, argues that this is exactly what we have a duty to do, if we safely can. "If we could be improved by an admixture of animal genes, why not?" he asked a seminar last autumn. "If it's not wrong to wish for a bonny, brown-eyed girl with high intelligence," he says, "why is it wrong to do it? I don't think anyone who wishes for an intelligent child of a particular gender is a moral monster."
For Savulescu, the prospects of PGD are not worth worrying about when biotechnology will be put to far more destructive purposes - to engineer deadly gene-based biotech weapons, for instance. "Millions of people will have access to weapons of mass destruction that we haven't even begun to imagine," he says. So why should we be concerned that a handful of rich people will probably demonstrate consumer choice when planning their offspring?
For some philosophers, such as Julian Savulescu and John Harris, choosing a baby's sex through PGD - to avoid genetic diseases, or even cultural stigma affecting one gender - is an ethical advance over sex-selection by abortion. The embryo not implanted can't be said to be wronged, he argues. Sex selection via pre-implantation genetic diagnosis is illegal in the UK, but search the internet for "family balancing" and you will find clinics in Cyprus, Turkey and Lagos advertising for British customers.
Besides, if PGD ever becomes cheap and simple enough for widespread use, which government wouldn't choose to have citizens who were intelligent, calm, long-lived, generous and not going to die of expensive Alzheimer's - or be afflicted by the terrible consequences of unmediated natural reproduction that have befallen Belinda Kembery's son Robbie? "At some stage," says Julian Savulescu, "we will all be reproducing artificially. Nature's way is just too inefficient."
Want more Wired UK magazine? Get your copy every month by subscribing online today
This article was originally published by WIRED UK
