DNA ii. clinical applications

Summary
DNA testing is one part of the investigation of families with genetic diseases. Tests can investigate disorders of the 3 main modes of Mendelian inheritance. Clinical genetic services can provide advice on the most appropriate tests for a particular family.

Autosomal recessive conditions
Among the several hundred genetic conditions for which prenatal diagnosis is available, more than three quarters are autosomal recessive disorders. Most are rare inborn errors of metabolism and are still diagnosed prenatally by traditional biochemical techniques because these are applicable to all families. In most conditions, there are many known mutations and most patients have two different mutations rather than two copies of the same mutation, making diagnosis by DNA methods particularly complex. There are some autosomal recessive conditions in which DNA techniques are preferred because the biochemical defect is unknown or is not expressed in the tissues accessible for prenatal testing. The way DNA tests are used in autosomal recessive disorders is described in the section on cystic fibrosis.

Although linkage is relatively simple because only the previously affected child and the two parents are needed for testing, the risk of recombination varies widely. This risk is generally low if markers within the gene are available, but may be high if only markers outside the gene are available (see the section on spinal muscular atrophy). It is absolutely essential for families to be investigated before starting a pregnancy to ensure that a test can be offered. The correct clinical diagnosis is essential as confusion of conditions with similar clinical effects could lead to testing for the wrong gene. Some conditions can be caused by faults in any one of several different genes making testing difficult or even impossible.

Sicklecell disease
Worldwide, this is probably the most common of all autosomal recessive conditions. Every patient is always homozygous for the same single base substitution which results in the formation of haemoglobin S. There are 5 or 6 different elegant and simple DNA methods available to detect the mutation.

Thalassaemia
This rivals sicklecell disease in frequency, but is genetically more complex. Beta-thalassaemias are the result of mutations (usually point mutations) which reduce the production of betaglobin chains and alpha-thalassaemias result from a deficient production of alpha chains (usually due to deletions). In each part of the world where one of these diseases is prevalent, different mutations predominate. The number of frequent mutations in each region is small enough to make it possible to test for these common mutations in each family and then to offer mutation detection for prenatal testing. However, in some families, it is still necessary to use linkage analysis.

Cystic fibrosis
Cystic fibrosis is the most frequent autosomal recessive condition in Caucasian populations. Over 400 different mutations have been detected, but, fortunately, one mutation ( F508) accounts for 70% or more of mutant genes in most population groups. If tests for another 4 or 5 relatively common mutations are added, then 80-85% of mutations can be detected. Many intragenic microsatellite markers are available for linkage testing.

First, the affected child is tested. If both mutations are identified, then no other family samples are needed before prenatal diagnosis can be offered. If one mutation is detected, it will be found in one parent and linkage can be used to track the mutant gene from the other parent. Occasionally, neither mutation is found and linkage alone is used.

Other family members who may be carrying the gene often want to know if they are also at risk of having children with cystic fibrosis. The key question in this situation is whether the partner of the relative is also a carrier. As only about 80% of mutations can be recognised, the unrelated partner cannot be advised with absolute certainty that they are not a carrier.

There has been much discussion about communitywide screening for the carrier state, but the fact that 20% of carriers will be missed has deterred most communities from introducing such programs.

Spinal muscular atrophy (SMA)
The severe and intermediate forms of SMA appear to be caused by mutations in the same gene. The gene has been localised, but not isolated, so linkage analysis is appropriate. There are many linkage probes available, but this disease illustrates the magnitude of the errors that can be introduced by genetic recombination.

If just one linked probe is used and there is a 5% risk of recombination between the marker and a mutation in the gene, then the risk of an erroneous result is 20%. This is because a recombination could cause the probe to fail to track the passage of the gene from either parent to either the previous affected child or the new fetus. If two markers, one on either side of the gene (flanking), can be used, then the results are much more accurate. A single recombination between the markers can be identified, but one can no longer track the gene mutation, so no test can be offered. If a single recombination has not occurred, the accuracy of the test is >= 99%, and the only remaining risk is of a double recombination which would have a probability of (0.05)2 = 0.25%.

Autosomal dominant disorders
There are many important autosomal dominant diseases which emerge in adult life. Myotonic dystrophy and Huntington's disease are two well known examples which always cause great anxiety to families and geneticists alike. Each produces a progressive degenerative disease which incapacitates and kills. The defective gene may be inadvertently passed on to the next generation as the patient's symptoms usually develop after they have had children. In both diseases, linkage tests were available before the genes were isolated. These tests were tedious for both the laboratories and the families. An immense amount of work was involved and often key family members were dead or uncooperative. In our laboratory, one scientist worked fulltime on Huntington's disease and was able to provide presymptomatic diagnosis for only 20-25 families a year. Even then, the result was only a probability statement such as a 98% chance of having the disease or a less than 5% chance of having it.

When the genes for both these diseases were isolated, it turned out that all families have the same type of mutation ­ expansion of a 3 base pair (triplet) repeat which is present in multiple copies.¹ This simplified the testing procedure so that our one scientist can now test 20 families in about two weeks.

This unusual type of mutation was first identified in the fragileX mental retardation syndrome and has now been found in a total of 7 conditions (Table 1). Doubtless, others will be added.

The discovery of this type of mutation also resolved the question of anticipation which had been debated for many years, especially in relation to myotonic dystrophy. Anticipation is a progressive increase in the severity of the disease and a reduction in the age of onset from generation to generation. This has now been explained by a progressive increase in the size of triplet repeat from one generation to the next.

Familial cancer syndromes e.g. bowel cancer²
Geneticists have always known of rare cancers which showed autosomal dominant inheritance. Now we know that there are dominantly inherited forms of most common cancers. Probably 5-15% of common cancers will prove to be inherited with 5 or 6 genes involved in each type of cancer. The mutant genes which have been discovered cause cancer in almost all the people who inherit them.

Three different genes responsible for inherited colon cancer have now been isolated. The first gene to be isolated was also the least frequent: the APC gene which causes familial adenomatous polyposis coli, a dominantly inherited precancerous condition. A relatively simple test which can identify 95% of mutations has been discovered, so that the investigation of the offspring of an affected individual can begin with a DNA test which will classify each individual as affected or unaffected. Only those individuals with the gene need repeated colonic examinations to choose the right time for prophylactic colectomy.

More recently, genes on chromosome 2 (hMSH2) and chromosome 3 (hMLH1) have been found culpable in more frequent forms of inherited colon cancer. They probably account for about 3% and 6% respectively of all colon cancer. It is too early to know what form of testing will be possible for these genes. If there are only a small number of mutations in hMSH2 and hMLH1, and if simple tests to detect these mutations can be developed, it will be desirable to test all new cases of colon cancer. Investigations can then be offered to the relatives of those individuals who have the mutations. If, however, there are many mutations and/or no simple mutation detection tests, the situation is less satisfactory. Only if families have a number of affected individuals can linkage be used to estimate the risks of carrying the bowel cancer causing gene to other family members.

One can predict that DNA tests will have quite a large impact on the early treatment of cancer during the next decade. The family based testing that will be possible is much more efficient than mass community screening.

Xlinked disorders
In Xlinked conditions, males inheriting the mutation are affected. Females inheriting the mutation are carriers and are usually unaffected. However, some carriers may show partial effects e.g. in fragileX mental retardation syndrome. Carrier females are usually difficult to detect clinically, but identification is desirable because half of their sons will develop the disease. It is particularly difficult to deal with the families in which there is a single affected male because this individual may have inherited a mutation present in all his mother's cells (in which case, half of his sisters will be carriers), a mutation present in just one egg cell (in which case, neither his sister's nor his mother's future children are at risk), or a mutation present in a proportion of the mother's egg cells (germinal mosaicism with a risk of further affected sons or carrier daughters).

Duchenne/Becker muscular dystrophy
All cases of these conditions are caused by mutations in a very large gene, the dystrophin gene. Mutations are generally different in each family. However, about two thirds of the Duchenne mutations are deletions and these deletions cluster in two regions of the gene, making it possible to design sets of polymerase chain reactions (multiplex PCR) which can pick up most of the deletion cases. By specifically amplifying segments of the gene, a segment's presence (normal result) or absence (deletion) can be determined. This technique can only be applied to males. Complex methods have to be used to try to work out whether related females are carriers, but these may not be sufficiently reliable to exclude the carrier status completely. One approach is to identify the sex of the fetus in the pregnancies of an affected boy's mother, sisters and aunts and offer prenatal testing for the deletion if the fetus is male.

In the one-third of patients who do not have deletions, the situation is even more difficult. Linkage used to involve considerable errors because the dystrophin gene is so large that the probability of a recombination between one end of the gene and the other is over 10%. Fortunately, multiple microsatellite probes within the gene are becoming available which yield more accurate results.

FragileX mental retardation³
This is the most common inherited cause of mental retardation and shows some very unusual features for an Xlinked condition. Most males carrying the mutation are severely retarded and more than 30% of female carriers show some intellectual disability, or even moderate to severe mental retardation. The physical stigmata are of large testes in males, and joint laxity, protruding jaw, large nose and large ears in affected males and females. The mutation is nearly always a marked expansion of a triplet repeat sequence present in the part of the gene which is transcribed to RNA, but not translated to protein. In males, there is a clear relationship between the size of the expanded triplet repeat and the severity of clinical effects. This relationship is much less precise in females because of the variable inactivation of the mutant or normal X chromosome. Some males have a modest increase in size of the triplet repeat which causes no clinical symptoms in them or their daughters, but the repeat may expand considerably and cause symptoms when passed on by these daughters to their children. Now that the gene has been isolated and the type of mutation identified, counselling is more straightforward than previously. Prenatal diagnosis is available, but it remains difficult to predict the clinical outcome for a female with a moderate to large expansion of the triplet repeat.

Conclusion
DNA testing is just one part of the process of counselling a family with a genetic disease. Knowledge of genetic diseases is advancing very rapidly, and consultation with a clinical geneticist is recommended before giving advice or arranging DNA tests.