by Kirk, Edwin;
Unfortunately, at high levels, it is toxic to the brain. During pregnancy, the placenta filters the baby’s blood and no harm is done, so that babies with PKU are born with perfectly normal brains. As soon as they start drinking milk, they are taking phenylalanine into their bodies and the damage starts.
In the 1950s, a German doctor, Horst Bickel,4 showed that a diet very low in protein could lower levels of phenylalanine in the blood, with some benefits for people with PKU. Bickel and two colleagues published a paper describing the first successful treatment of a child with PKU.5 The paper, published in the journal Lancet in 1953, makes for somewhat confronting reading by today’s standards. Part of that is the bluntness of the language, which to be fair was usual for the time. ‘She was an idiot and unable to stand, walk, or talk: she showed no interest in her food or surroundings, and spent her time groaning crying, and banging her head’. The paper describes a series of experiments to modify the food the girl was given, ending up with a diet containing almost no natural protein, tiny supplements of phenylalanine to supply the body’s essential needs, and the addition of a special formula that contained all of the other amino acids, except phenylalanine. The effects were dramatic: the child learned to crawl, then to stand; ‘her eyes became brighter, her hair grew darker; and she no longer banged her head or cried continuously’.
[4 By a nice coincidence, Bickel and Guthrie shared a birthday, 28 June, which is now international PKU day.]
[5 Unusually, we know the girl’s name: Sheila. She was diagnosed aged 17 months and, reportedly, her mother begged Bickel to try to find a treatment for her. Against enormous odds, Bickel and his colleagues succeeded.]
So far, so good. But how to be sure it was the treatment that was doing this? Even in 1953, it’s clear that Bickel and his colleagues were aware of the risks of introducing a treatment that only seemed to work. To make sure, they decided to add a large amount of phenylalanine back into the formula. This is scientifically sound: try a treatment, see what happens, withdraw it, see what happens, start it up again. The part that is shocking to a modern reader is that they deliberately concealed this plan from the child’s mother. They wanted her to be an unbiased observer of what happened after the phenylalanine was re-introduced. What happened was instant regression: within a day, the girl had lost nearly all the gains of the preceding ten months. Remarkably, the child’s mother then agreed (perhaps in reality she was given no choice) to a repeat of the experiment. Again the girl was given the phenylalanine-free formula and learned new skills, again (this time during a hospital admission) the toxic (to her) amino acid was added in, again the skills were lost. The treatment worked. Although Sheila undoubtedly had some permanent brain damage, in the long run, she must have been far better off with the treatment than she would have been without it.
Within a short time, this new way of treating PKU was being widely adopted. Treated children made significant gains, and the younger they were when the treatment was started, the better the outcome. This was because there was always some damage that could not be reversed, and the older the child, the further the damage had progressed. A grim clock was ticking from the moment of birth, counting out the time before it was too late to start treatment. Decades later, in the 1990s, I met Frank soon after his diagnosis was made, and we started treatment. There was some improvement in his behaviour, but that was all: the chance for meaningful benefit had long passed.
Back in the 1950s, the best results of all were in babies diagnosed soon after birth not because of symptoms in themselves but because they had an affected older sibling. From those early trials, the early indication — since proved correct — was that the fate of an affected baby could be utterly changed, with a devastating neurological injury completely avoided.
PKU had become a treatable disease,6 and the stage was set for Bob Guthrie’s rise to greatness.
[6 The treatment is very effective, but can be a real challenge for families to manage. The diet is extremely strict, the essential supplements are becoming more palatable but still have a very distinctive taste, and regular blood tests are needed. Especially strict management is needed when a woman with PKU has a pregnancy, because even moderately raised levels can be very harmful to the developing baby. As a side note, if you drink diet soft drinks, have a look at your next can: it may well say ‘Phenylketonurics — contains phenylalanine’ on the side. The artificial sweetener aspartame contains phenylalanine; even the small amount in a can of Diet Coke could be a problem for someone with PKU.]
Guthrie seems to have been something of a difficult character. The tributes written by colleagues after his death are carefully worded, in a way that tells its own story. I asked Bridget Wilcken about Guthrie. Bridget herself is something of a legend in the world of screening — she steered newborn screening in New South Wales through its first 50 years, and has been a world leader in the field for decades. She spoke of Guthrie’s great contributions to medicine, but she also told me that, many years ago, Guthrie was a guest at her house. He omitted to tell her, until dinner was already served on the first night, that he was a vegetarian. It just hadn’t occurred to him to mention it. This was at a time when it was uncommon to be vegetarian, and as a host you wouldn’t think to ask. Others speak of Guthrie calling collaborators at all hours of the night to discuss some new idea, or the details of an ongoing project. Later in the story of screening, he apparently resisted the addition of some other conditions to newborn screening, perhaps because they didn’t form part of his vision for the test.
Perhaps, though, that one-eyed focus on the task at hand was the secret of his success. Guthrie was a father of six; his second child, John, had intellectual disability and as a result, Guthrie and his wife became very active in the Buffalo Chapter of the New York State Association for Retarded Children. Through meetings of this group, Guthrie became aware of the existence of PKU and its treatment. One of the difficulties of managing PKU was measuring the levels of phenylalanine in the blood. Guthrie had been working in cancer research for over a decade and realised that it would be possible to adapt a simple test he had been using in his work to do this.7 He moved to the Buffalo children’s hospital and began developing the test. One of his greatest achievements, and one of which he was most proud, was the development of filter-paper cards onto which blood from the baby’s heel is dropped. Once the blood has dried, the cards can easily be mailed back to the laboratory that does the testing. If you’ve had a baby, it’s almost certain that he or she was screened, using tests that are only possible because of the existence of these cards, which are still known as Guthrie cards.
[7 The test is simple but elegant. Bacteria are grown on agar jelly that contains a substance that stops the bacteria from using phenylalanine. This inability to use the amino acid prevents them from growing — effectively they are being starved. A circle of filter paper soaked with the baby’s blood is placed on the agar. If there is a lot of phenylalanine in the blood, it overcomes the effect of the blocking substance, and the bacteria can grow. The more phenylalanine, the better the growth. You put spots of blood from lots of children onto a tray of the jelly, keeping track of which spot came from which baby, of course; incubate; then see which spots (if any) have made the bacteria grow.]
Once the test had been developed, it became possible to diagnose babies with PKU in the first weeks of life, before irreversible damage had been done to their brains, and get treatment started. Guthrie made it his mission to promote screening and to make it possible for others to get started. The first trials of screening using this method8 began in 1960, and, by 1963, 400,000 babies from 29 US states had been screened, with 39 babies with PKU identified — and saved.
[8 A less effective urine-based test had already been introduced in some areas.]
One reason for the rapid early progress of screening was that US president John F. Kennedy had a sister with intellectual disability, and had ensured that the Children’s Bureau, w
hich had oversight of this kind of program, was well funded. There are other examples of politicians with a connection to an issue making a difference of this kind; particularly relevant for our purposes is former British prime minister David Cameron, whose son had a rare, severe form of epilepsy, Ohtahara syndrome. Cameron’s interest in genetic conditions had a great deal to do with the subsequent developments in genomic medicine in the United Kingdom, which has a world-leading diagnostic and research program.
Over more than half a century, newborn screening has improved and prospered. Most services in developed countries screen for 40 or more different conditions. Not all have such clear-cut benefits from early diagnosis as PKU, but there is no doubt that many tens of thousands of children have had their lives saved, or profoundly improved, by that simple heel prick. Because it seems like such a simple thing, it’s easy to take newborn screening for granted. Unless someone in your family is affected by one of the conditions, it’s something that barely touches your life. A few minutes of time lost in the flood of events and emotions soon after the birth of a new baby — many people quickly forget the test was even done. But make no mistake, this is one of the great triumphs of medicine.
Newborn screening, wonderful though it is, has the disadvantage that it can only ever be done on a child who has already been born. For many conditions, however, there are no effective treatments yet, and many parents would prefer to have a choice about whether they will have a child with a significant genetic condition. For this reason, screening for genetic conditions during and even before pregnancy has been developed.
Screening for genetic conditions during pregnancy also goes back a long way. As early as 1955, a method had been developed for collecting a sample of amniotic fluid during pregnancy (amniocentesis), and this was initially used for determining the sex of the fetus by examining the cells found in the amniotic fluid. By 1966, the technique had been improved to the point that cells from the amniotic fluid could be grown in the laboratory and their chromosomes analysed. In 1968, a 29-year-old woman who was known to have an inherited rearrangement of her chromosomes9 that gave her a high chance of having children with Down syndrome attended the Downstate Hospital in Brooklyn, New York. She was 16 weeks into her third pregnancy, having already had a healthy daughter and a son with Down syndrome. An amniocentesis was done and showed that this fetus was also affected, and the woman chose to have a termination of pregnancy. At this time, amniocentesis had already been used to diagnose several different conditions in the fetus with biochemical tests. Now, for the first time, a genetic test had been used to make a diagnosis during pregnancy. The next few years saw rapid development in the field of prenatal diagnosis. Once the technical skills to perform the procedure and the laboratory capacity to do the analysis became widely available, screening of large numbers of pregnant women also became a possibility.
[9 A Robertsonian translocation, to be exact.]
Over the decades that followed, screening for Down syndrome (and, later, other chromosomal conditions) has been performed using ever more sophisticated methods. The goal is to identify pregnancies with a higher chance of being affected, so that an invasive test such as amniocentesis10 need only be offered to a subset of women to definitively answer the question as to whether the baby is affected. The first screening test was simply to ask the woman her date of birth. The chance of having a baby with Down syndrome, and some other chromosomal conditions, rises with the mother’s age.11 Women are born with all the eggs they will ever have, held in a kind of suspended animation just short of maturity. If you have just turned 35, your eggs are a little older than that … and their capacity to complete that last step in their development without errors has been declining slowly for years. There’s no hard age cut-off at which the chance of having an affected baby suddenly jumps up; the probability rises slowly and smoothly, although the curve does steepen in the mid-30s. For a 20-year-old mother, the chance of having a baby with Down syndrome is about 1 in 1,500; for a 35-year-old, it’s 1 in 340; and for a 45-year-old, it’s 1 in 32. When I started in genetics, one of the main reasons for having a prenatal test for chromosomal conditions was still ‘AMA’ — advanced maternal age.12
[10 Chorionic villus sampling (CVS) is the other main method in use.]
[11 Men aren’t completely off the hook here, because some other types of genetic condition become more likely as fathers age.]
[12 I once put my foot in it very badly by wishing one of our genetic counsellors ‘happy AMA day’ on her 35th birthday. The term ‘advanced maternal age’ is unfortunate, considering that 35 years is not really an advanced age in most other contexts. There’s an even worse term in obstetrics — an ‘elderly primigravida’ is a woman who becomes pregnant for the first time at 35 or older. ‘Elderly’? What were they thinking?]
This approach reduces the number of women who might receive more-invasive screening, but the problem remains that, if there is a 1 in 200 chance that an older woman is carrying a baby with a chromosome condition of some kind, the follow-up invasive testing will return 199 negative results for every one affected pregnancy that you identify. When you consider that amniocentesis carries a roughly 1 in 200 chance of causing a miscarriage, you can see that there’s a trade-off happening. Are those 199 tests, with their costs and the anxiety they cause, plus one miscarriage, worth identifying one affected pregnancy?
On top of that, screening only women aged 35 and above will miss most affected pregnancies, because, even today, the great majority of babies are born to younger women.
There have been various efforts to address these problems. The first worked by measuring combinations of several substances in the mother’s blood that change during pregnancy, with levels that tend to be higher or lower if the baby has a chromosomal problem. The results of these tests, combined with the mother’s age, were used to refine the risk assessment. Later, measurement of the thickness of skin at the back of the neck was added into the calculation. Many different medical conditions, including Down syndrome, can lead to a build-up of fluid in this area early in pregnancy, so the measurement improved the accuracy of the screen.
Every version of these tests still suffered from the same basic problem. Even as they became more sensitive, with lower false-positive rates, that trade-off remained; for most women who were flagged as being at higher risk, it was still much more likely that, if they chose to have an amniocentesis, the result would be normal — yet they still were exposed to the risk of miscarriage from the procedure. In that sense, having a screening test carries risks — the risk that you might have an invasive test you don’t need, and possibly even lose the pregnancy.
Recently, a far better screen has become available.13 Non-invasive prenatal screening (NIPS) uses the new genetic sequencing technology in a nifty way: it treats DNA simply as something to be counted. Usually, when we extract DNA from a blood sample, we take it from the white blood cells, each of which has a nucleus (red blood cells have lost theirs, and their mitochondria, so they don’t contain DNA). However, there is a small amount of DNA in the plasma, the fluid that makes up the half or so of your blood that isn’t blood cells. In the late 1990s, it was discovered that, if you take blood from a pregnant woman, discard the cells, and extract DNA from the plasma, some of that DNA will come from the placenta. It’s often called the fetal fraction, but the fact that the DNA is from the placenta, not directly from the fetus, is important. We’ll see why shortly.
[13 It’s a bit harder to pin down who the key people were for some of the prenatal screening methods, because of contributions from many different people over a long period of time. But Dennis Lo from Hong Kong was the first to develop NIPS, and Kypros Nicolaides from London led the use of ultrasound for Down syndrome screening.]
There are different ways that NIPS can be done, with several companies having come up with their own approaches. The most common approach involves sequencing DNA using one of the new sequencers t
hat reads many individual strands of DNA at once (the same ones discussed in chapter 5). Sequences are read from regions spread across the genome. Each continuous stretch of sequence, representing one molecule of DNA, is a ‘read’. The fetal fraction varies between blood samples, but suppose for example that it is 9 per cent. On average, you expect that, at any one place, there will be 100 reads from the mother’s DNA and ten from the placenta (representing the fetus); ten out of 110 is 9 per cent. You don’t necessarily even have to distinguish which are which: all you need to do is count. If you have, on average, 110 reads for every chromosome except 21, but 115 reads14 across chromosome 21 — then there are half again as many from the placenta as you expect: there must be three copies of that chromosome instead of two. The baby has Down syndrome.
[14 You’ll likely need to do more sequencing than 110 reads for the statistics to work, but the principle remains the same.]
Except if it doesn’t.
There are various ways you can be fooled into thinking the baby has a chromosome abnormality when really it doesn’t. One of the most important reasons15 this can happen is that, sometimes, there are chromosomal changes in the placenta that are not in the fetus. The early embryo is a ball of cells that splits into two: part of it goes on to become the fetus, and eventually the baby, and the rest becomes the placenta, membranes, and so on. If a mistake in cell division happens after that split, the placenta can have a chromosome change it doesn’t share with the fetus … and since the fetal fraction is really a placental fraction, the test can be fooled. It really is correctly counting an extra copy of chromosome 21 — it’s just that the extra copy is in a tissue where it doesn’t matter.
[15 Most of the other reasons are technical, but a rare cause is that the mother has cancer, the cancer cells have abnormal chromosomes, and they in turn are spilling DNA into her blood that has abnormal chromosomes. There have been a number of women around the world who have unexpectedly had a cancer diagnosis made this way.]