THEY are the genetic equivalent of the mischievous creatures in the film Gremlins: small, harmless-looking pieces of DNA that lurk within genes, whose destiny is to grow bigger and wreak havoc. Unfortunately, malevolent DNA of this kind 鈥 known as 鈥渦nstable鈥 or 鈥渢rinucleotide repeat鈥 DNA 鈥 is all too real. It first came to light in 1991 as the cause of fragile X syndrome, an inherited form of mental retardation that mainly affects males. Soon afterwards, geneticists found strands of unstable DNA in people with myotonic dystrophy and Huntington鈥檚 disease. Now it is recognised as a key player in the world of genetic disease. Answering questions like 鈥淲here does unstable DNA come from?鈥 and 鈥淲hich genes does it affect?鈥 has become one of the latest goals in the fast-moving science of human molecular genetics.
And not without reason. A Cambridge team examining the evolution of unstable DNA has reached a worrying conclusion: the amount of unstable DNA may be gradually, almost imperceptibly, increasing, driven by its intrinsic ability to expand as it is passed from one generation to the next. In other words, some of our DNA is trying to kill us; and for the moment, at least, there is nothing we can do to stop its advance, or correct its harmful effects (see Diagram).
The rate of spread cannot be measured directly. But judging from comparisons of unstable DNA sequences in different ethnic groups, unstable DNA has been gaining ground for tens of thousands of years. Evolution has yet to equip us with mechanisms that could protect genes from the effects of unstable DNA. Nor is there likely to be much selection pressure to eliminate the faulty DNA from the gene pool because many of the diseases it causes disable or kill people only after they have reached child-bearing age.
Advertisement
Genes that contain a trinucleotide repeat in humans often do so in other animals too. But the number of repeats is usually different from one species to another, and with humans can differ between ethnic groups. The Huntington鈥檚 gene is a case in point. David Rubinsztein and his colleagues at the University of Cambridge have examined the Huntington鈥檚 repeats in people from different ethnic groups, and also in nonhuman primates, especially our closest relatives, chimpanzees. In humans, the repeat tends to be longer in populations where the disease has a high incidence, such as northwest Europeans, than in populations where there is little or no Huntington鈥檚 disease.
This is true even when people with the disease are excluded from the analysis. And herein lies the key implication. It suggests that the repeats which cause Huntington鈥檚 disease originate in the healthy population. In other words, the repeats are 鈥渆volving鈥, gradually increasing in length over many generations until they cross the threshold into the disease range. Not all human populations have reached this point. For example, the incidence of Huntington鈥檚 disease is still vanishingly small among Japanese people and black Africans. Moreover, Rubinsztein鈥檚 results show that chimpanzees are well behind humans in this respect, since their repeats are much smaller than ours and have a long way to go before they approach the disease threshold.
A decade ago none of this could have been predicted. At that time, insights into the molecular causes of genetic illnesses were exceedingly rare. Researchers had linked a handful of diseases, such as sickle-cell anaemia and phenylketonuria, to defective proteins, the products of mutated genes. But geneticists such as myself had only just begun to take seriously the idea of finding human disease genes without knowing what proteins they encode, or how the genes act, or what they look like. We reasoned that by studying the inheritance patterns of a genetic disease in an affected family, and analysing DNA samples of family members using rapidly improving molecular methods, it should be possible to find the gene responsible for any inherited disorder. In a spirit of optimism, we set out to track down the genes for illnesses such as cystic fibrosis, muscular dystrophy and Huntington鈥檚 disease.
Huntington鈥檚 would prove a difficult nut to crack, but by the late 1980s researchers had identified the defective genes that cause cystic fibrosis and muscular dystrophy. It was clear that the gene-hunting approach worked. And yet, at a conceptual level, these early successes produced no great surprises.
The mutations that made the genes harmful involved the kind of corruption of DNA code seen many times before in mice, bacteria and fruit flies. Some mutations affected the gene in just one place, changing the recipe of the protein it encoded by substituting a single amino acid in the normal molecule for another. Others deleted large segments of the DNA code. Still others made the code 鈥渦nreadable鈥 by scrambling its 鈥減unctuation鈥, turning sections of the code into gibberish. Every mutation, however, followed the classical laws of genetics and was faithfully passed on, unaltered, to the next generation, giving rise to the same problem in the offspring as in the parent.
But some genetic diseases had always appeared a bit odd. They didn鈥檛 obey the classical laws. For example, parents with only the mildest of symptoms might produce offspring with much more severe disease. Or the disease might suddenly appear in a family, as if from nowhere, but subsequently be found to have been passed on through the mother on the X chromosome. Among thesewas fragile X syndrome, so called because patients typically have X chromosomes with fragile-looking tips.
In 1991, the odd pattern of inheritance of fragile X syndrome began to make more sense as groups of geneticists in Europe, Australia and the US identified the mutated gene responsible for the disease. The mutation was like nothing they had seen before. It consisted of a piece of DNA that grew in length when it passed from parent to child, and the bigger it became, the more severe the disease it caused.
At around the same time, my colleagues and I at the University of Wales College of Medicine in Cardiff, together with many collaborators and competitors worldwide, were closing in on two other disease genes 鈥 those causing myotonic dystrophy, the most common form of muscular dystrophy in adults, and Huntington鈥檚 disease.
Grip of disease
The quirky behaviour of the myotonic dystrophy gene had long puzzled medical geneticists. In most families with the disease, the affected grandparent has only the mildest of symptoms, usually cataracts. But those affected in the next generation go on to develop the classical form of the disease in adulthood, suffering muscle weakening and an inability to relax their grip after they contract their muscles. If these individuals have children, there is a risk that their offspring will be even more severely affected, suffering respiratory problems, 鈥渇loppiness鈥 and severe mental retardation, the results of which may prove fatal. This progressive increase in severity, with earlier age of onset in each successive generation, is known as 鈥渁nticipation鈥.
In 1988, we teamed up with Keith Johnson鈥檚 group at the Charing Cross and Westminster Medical School, and David Housman鈥檚 team at the Massachusetts Institute of Technology in Boston. Three years on and we had narrowed our search to a handful of genes on the long arm of chromosome 19. We then examined each of these genes in turn, comparing their structures in patients with myotonic dystrophy with those in healthy people. This was a long-winded, repetitive task and most of the DNA comparisons revealed nothing of interest.
But six months later things began to change as we focused on a fragment of a gene known prosaically as either pM1OM6 or cDNA25. Armed with a 鈥渉ealthy鈥 version of this fragment, we probed DNA samples from a family with myotonic dystrophy. The result was startling. As expected, each family member carried the DNA fragment, but the versions carried by affected family members were longer than those carried by their unaffected relations. It was as though the piece of DNA in question had somehow expanded. And the more severely affected the patient, the more the DNA had expanded.
We realised that the expanding DNA seen in our myotonic dystrophy patients was similar to the kind seen in the gene for fragile X syndrome. In both cases the DNA had a distinctive, repetitive appearance. Spelt out in single letter code, base by base, there were sequences such as 鈥渃tgctgctgctgctgctg鈥 or 鈥渃ggcggcggcggcggcgg鈥, compared with the random and featureless strings that one usually sees (such as 鈥渁gctgctattg-cacaaatgcgtagtcgat鈥). The similarity suggested that myotonic dystrophy was caused by the same kind of mutation as fragile X syndrome, a nonclassical one in which a piece of DNA disrupts a gene by growing bigger.
Such mutations, we reckoned, might explain certain other genetic diseases, and a year later we were proved right. A group of laboratories, including ours, found the long sought-after gene for Huntington鈥檚 disease. It, too, contained one of the repetitive DNA sequences. Since then another six examples of such sequences have been reported and the list is growing all the time.
Most of the sequences are linked with a particular genetic disease, usually one that affects the nervous system in some way. It is not yet clear why the nervous system should be especially vulnerable. Possibly it is because nerve cells are more sensitive to the effects of the DNA expansion since they do not go on dividing and would therefore accumulate toxic products of the expanded DNA. But whatever the answer, human geneticists have seized on unstable DNA with glee. Seldom can they claim a new phenomenon that wasn鈥檛 discovered first by their colleagues working on bacteria, mice or fruit flies. Now the race is on to understand why these DNA sequences are such bad news for genes. Only then can we hope to understand 鈥 and perhaps one day halt 鈥 the biological mechanisms that cause diseases like myotonic dystrophy.
The detail of these sequences is instructive. As with all DNA, the sequences are made of strings of nucleotides that carry four different bases, the four chemical 鈥渓etters鈥 that make up the genetic alphabet. But unlike most DNA, the expanding sequences consist of the same three-letter words 鈥 鈥渃tg鈥 or 鈥渃gg鈥 for example 鈥 spelt out over and over again (hence the term 鈥渢rinucleotide repeat鈥). It is this repetition that creates the problems, because it may make it difficult for cells to replicate the DNA accurately. Instead of faithfully reproducing just one piece of trinucleotide repeat DNA, cells will sometimes repeat sequences unnecessarily, causing the DNA to lengthen as it is copied. No organism that commits such copying errors can be sure of passing on reliable, working copies of genes. In effect, some of their genes become unstable.
This instability holds the key to understanding how and why trinucleotide repeats first appeared in the human genome. They probably evolved by chance, by mutations in sequences that happened to be nearly a trinucleotide repeat. The sequence ctgttgctg, for example, could mutate into ctgctgctg.
But not all trinucleotide repeats expand during replication. We all have hundreds 鈥 maybe thousands 鈥 of trinucleotide repeats in our genomes, and only a minority of these cause genetic instability. The essential difference between those that can expand and cause disease, and those that don鈥檛, seems to be solely to do with their length. DNA sequences containing more than about 50 repeats of a trinucleotide become unstable and expand during replication; sequences that contain fewer repeats are relatively stable and behave like any normal piece of DNA, although they too may sometimes expand during the processes of cell division that produce sperm and egg cells.
In effect, there is an instability threshold, and it is this threshold that explains the unusual inheritance pattern of diseases like myotonic dystrophy and fragile X syndrome. In families with these diseases, unaffected members have DNA with less than the magic 50 repeat threshold, while affected members have DNA with repeats above that threshold. In myotonic dystrophy and fragile X syndrome, the symptoms of people just above the repeat threshold are very mild. They only cause serious problems when they expand as they are passed to children.
Why do the sequences behave as they do? Most changes in an organism鈥檚 DNA sequence (mutations) occur when the DNA is replicated just before cell division. All cells, from bacteria to those of humans, have an elaborate proofreading mechanism that compares newly synthesised strands of DNA with their 鈥減arent鈥 strands, and corrects any mistakes that have crept in. This is no small task, given that the human genome contains some three billion nucleotides. Unstable DNA somehow escapes the correction process.
To understand why, imagine being interrupted in the middle of proofreading a page of normal text. Provided you could remember a few words from the bit you had just read, you could easily find your place again. But now imagine proofreading a page consisting of a thousand copies of the word 鈥渃at鈥. If you were disturbed in the middle of checking, you would have to go right back to the beginning.
The behaviour of unstable DNA suggests that cells have similar problems. Their replication and proofreading machinery deals with DNA in lengths of a few hundred nucleotides at a time. Once the length of a trinucleotide repeat becomes greater than this, the accuracy of the process breaks down. The cell loses count of how many copies of the repeat it has made and cannot tell if it has added a few extra ones.
And this kind of proofreading error is not the only reason why unstable DNA expands during replication. Unlike normal DNA, trinucleotide repeats seem to have an inbuilt tendency to get bigger. It is almost as though they 鈥渨ant鈥 to grow. A simple experiment illustrates this. If a strand of trinucleotide repeat DNA is incubated in a test tube with the enzyme that catalyses DNA replication, new strands are formed that are bigger than the original. This does not happen with 鈥渙rdinary鈥 DNA. It seems that trinucleotide DNA can form unusual molecular structures, which allow the two entwined strands of the DNA double helix to slip against each other, leaving 鈥渓oose ends鈥 to which the enzyme will add extra DNA bases.
So, the genetic instability that causes diseases such as myotonic dystrophy and fragile X syndrome is the product of at least two things: the 鈥済o for growth鈥 tendency of trinucleotide repeats combined with the failure of the error correction apparatus to put things right.
In fact, the story may be even more complicated. One of the most puzzling aspects of unstable DNA is that it is not always unstable. Cells isolated from patients can be grown in culture for many generations without further changes in the length of the trinucleotide DNA sequence. And studying diseased tissues taken from aborted fetuses or postmortems shows that most of the expansion seems to occur during the first few cell divisions of embryonic development. After that trinucleotide repeats are relatively stable.
What is it about these first few cell divisions that makes them so vulnerable to replication errors? One explanation is that this is the time when cells are dividing most rapidly. Charles Laird and his colleagues at the University of Washington in Seattle have shown that the DNA sequence in fragile X syndrome is among the slowest to be replicated during early cell division. Once the DNA sequence has expanded, as it does in the disease state, replication becomes positively snail-like. If the cells are trying to divide faster than normal, this could create extra problems. Parts of the genome could still be replicating while the chromosomes are being pulled apart into the two new cells. This behaviour may cause the characteristic, broken appearance of fragile X chromosomes under the microscope.
However, most trinucleotide repeats, including those for myotonic dystrophy and Huntington鈥檚 disease, don鈥檛 form these 鈥渇ragile sites鈥. A second explanation could be connected with the fact that the newly fertilised egg relies on proteins and other cell components from the ovum until it can start making its own. Maybe a molecule required for maintaining DNA stability is lacking in the ovum and must be synthesised in the embryo. If so, then any delay in producing it might be responsible for the trinucleotide instability in the very early stages.
Fascinating though the molecular biology involved in DNA instability is, from the medical viewpoint it is just as important to understand the effects of unstable DNA on human cells and how it causes the symptoms of disease. This is no easy task. While the mechanism of DNA instability might be the same for each disease, the genes involved and the ways they are affected appear to be quite diverse.
They come in two basic types. In the first, exemplified by Huntington鈥檚 disease, the trinucleotide repeat sequence is in the part of the gene that is translated into a protein. This means that the repeated DNA sequence gets turned into a repeated sequence of amino acids, the building blocks from which all proteins are made. In Huntington鈥檚 disease and all the other cases of this type, the amino acid which is repeated is glutamine. Usually there are between 10 and 35 glutamines in the normal protein encoded by the gene, but around 40 to 100 in the protein from people with the disease. The cell does not seem to tolerate expansions of the gene that would result in more than 100 glutamines.
The function of 鈥渉untingtin鈥, as the protein encoded by the gene responsible for Huntington鈥檚 has been christened, is still a mystery. But many proteins with glutamine repeats are transcription factors 鈥 molecules that control the expression of other genes in the cell and which are often crucial to the cell鈥檚 survival or development. Take the case of Kennedy鈥檚 disease, a condition that affects the nerves controlling movement, leading to progressive muscle wasting. The cause is a trinucleotide repeat. When this repeat expands, it results in a string of excess glutamines being inserted into a transcription factor known as the androgen receptor. These glutamines reduce the protein鈥檚 ability to switch on other genes and the result is a partial loss of control of the many functions normally influenced by the hormone androgen.
It is a reasonable guess that the huntingtin protein might be similarly affected. The degenerative symptoms of Huntington鈥檚 disease result from the death of certain types of neurons in the brain, but nobody yet knows how this process is linked to the extra glutamines in the huntingtin protein. Obviously, new treatments for the disease depend on solving this puzzle.
Not all trinucleotide diseases follow the gene-to-protein pattern, however. In myotonic dystrophy and fragile X syndrome, for instance, the expanding sequences are in stretches of DNA that lie next to genes and are thought to influence levels of gene activity 鈥 how genes respond to signals within cells. And in these regions, trinucleotide expansions are far less constrained because they are not translated into long strings of glutamine in the protein that the gene encodes, which would probably prove lethal for the cell. So they can increase to several thousand copies of the trinucleotide in severely affected patients without destroying the cells which ensure their replication.
At the moment, the genetic effects of such enormous expansions of DNA are clear only in the case of fragile X syndrome. Here the expansion switches the gene off so no protein is made at all. Exactly what the FraX protein does in cells is still a mystery, but it appears to interact with RNA. Since RNA represents a kind of halfway house between genes and proteins 鈥 it transcribes the genetic code and then acts as a template for building proteins 鈥 it is possible that the FraX protein has some role in controlling the production of proteins in cells. Only time will tell whether this theory is correct. Eventually, researchers may also discover how trinucleotide repeats affect the myotonic dystrophy gene. What is clear is that the gene encodes a type of protein 鈥 a 鈥渒inase鈥 鈥 that could be vital to the biochemical balance of a cell.
At the moment, the prospects for treating diseases caused by unstable DNA are poor. An obvious place to intervene would be in the transmission of trinucleotide genes from parents to offspring. But that could mean tampering with the genes of germline cells (sperms and eggs) in ways that are not 鈥 and indeed may never be 鈥 sanctioned by society. An alternative might be to try to design drugs that prevent trinucleotide repeats from expanding. But here, too, the obstacles would be enormous. Replicating and proofreading DNA are two of the most fundamental biological processes. Attempts to interfere with them could produce unpredictable and dangerous side effects.
Another possibility for treating Huntington鈥檚 disease and other diseases caused by excessive trinucleotide repeats might be gene therapy. So far, this has only been tried for diseases such as cystic fibrosis, where it is clear that the disease is caused by a lack of a particular protein, and the affected tissue (in this case, the lungs) is very accessible. The situation for trinucleotide repeat diseases is more complicated. For a start, only in the case of fragile X syndrome do we know that a protein is missing in the diseased state. In the other diseases, the effects of the repeated DNA are likely to be more subtle 鈥 and therefore more difficult to correct.
Diagnosing trinucleotide diseases, on the other hand, is much easier. An expanded trinucleotide repeat virtually always indicates that the person will be affected (although how severely cannot always be predicted), and it is technically quite easy to perform prenatal and presymptomatic diagnosis for those at risk using a small DNA sample obtained from amniocentesis, blood or a mouthwash. There are complications in relating the degree of expansion to the likely degree of severity of disease (鈥淎 fragile case for screening鈥, 25 December 1993/1 January 1994). So clinical geneticists will continue to insist (correctly) that testing should only be carried out if appropriate counselling and support is available for the individuals and their families.
Many people have questioned the usefulness of the Human Gene Mapping Project, an expensive endeavour that seeks to identify all one hundred thousand or so human genes. But the lesson from trinucleotide repeat DNA is that such basic studies in human biology have their own special value. They may uncover completely new biological phenomena that have far-reaching consequences not just for human health but for the very way we perceive our genetic heritage.