In the news

A common mutation in the alpha2-microglobulin (A2M) gene on chromosome 12 makes carriers 3.5-4 fold more susceptible to the common, late-onset form of Alzheimer's disease [Blacker, D. et al. (1998) Nature Genet. 19, 357-360]. A2M was investigated as a suspect because the protein is known to interact with other proteins implicated in the pathogenesis of Alzheimer's, including beta-amyloid protein and apolipoprotein E. The association between A2M and Alzheimer's risk was uncovered by comparing the frequency of the A2M mutation in Alzheimer's patients with its frequency in their unaffected siblings. The mutation, which is a small deletion, is found in 30% of the population, and increases Alzheimer's risk to the same extent regardless of the APOE genotype (see Alzheimer's disease summary in the Information database for information on the association between the APOE4 allele and Alzheimer's disease risk). In a separate study also reported in Nature Genetics, another group of researchers have found that, within a large group of elderly people, those who developed Alzheimer's disease did so earlier if they carried an APOE4 allele [Meyer, M. et al. (1998) Nature Genet. 19, 321-322]. The suggestion has been made that the A2M mutation may be a strong factor in determining whether someone will get Alzheimer's disease, while the presence or absence of an APOE4 allele determines when (see News and Views article by Jean Marx in the 24 July issue of Science for more details).

Comment: Assuming it can be confirmed by studies in other populations, the A2M association with Alzheimer's makes a major contribution to the understanding of the causes of Alzheimer's disease. Like APOE4 genotype, however, A2M genotype is not a definite predictor of Alzheimer's and there are no grounds for genetic testing or screening for the mutation, except in a research setting. The most important implication of the finding may be the potential for developing new drugs to combat the disease. 

Genetic tests have confirmed that Dolly the sheep was indeed cloned from an adult cell [Ashworth, D. et al. (1998) Nature 394, 329; Signer, E.N. et al. (1998) Nature 394, 329-330], a claim that had been contested by some since the original paper was published last year. And now cloning from an adult cell has also been achieved in a different mammal, the mouse [Wakayama, T. et al. (1998) Nature 394, 369-373; see also News and Views article by Davor Solter on page 315 of the same issue]. Although the mice were cloned by a different method from the one used to produce Dolly, the conclusion is the same: a nucleus from a differentiated cell (that is, a cell whose genetic material has been programmed during development so that it expresses only a specific subset of genes and divides to produce cells of the same type) can be reprogrammed so that it is capable of giving rise to all the different cells that make up the organism.

Comment: The successful cloning of an adult human cell, somewhere in the world, cannot now be far off. Although such an event would spark intense ethical debate, its impact in public health terms, at least for the foreseeable future, is probably negligible; cloning of a complete human being is likely to remain or become illegal in most if not all western countries. However, human cloning may find medical applications in the future if technological and ethical barriers were overcome to the extent that it became possible to clone human tissues for use in cell and tissue transplantation. In the meantime, cloning in the mouse, the best characterised model system for studying mammalian biology, may help to answer many fundamental questions about how an organism's genome is deployed during development. Cloning of other animals, if its success rate can be improved, has many applications, for example in biotechnological drug production and preservation of endangered species.

There has been a recent flurry of papers about the effects of an individual's genetic make-up on their risk of addiction to smoking and of developing smoking-related illness. One important culprit is the gene CYP2A6 encoding the drug-metabolising enzyme cytochrome P-450 2A6. CYP2A6 catalyses an important step in the metabolism of nicotine. About 20% of the white population carries at least one copy of the gene that is non-functional (encoded by a 'null allele' of the gene), and around 1% have two non-functional copies. These percentages vary in different populations, for example they are much lower in African Americans, only 2.5% of whom carry a non-functional allele. A recent paper in the Scientific Correspondence section of Nature [Pianezza, M.L., Sellers, E.M. and Tyndale, R.F. (19980 Nature 393, 750] reports that 'poor metabolisers' (people with at least one null allele) are less likely to become addicted smokers and, if they do smoke, tend to smoke fewer cigarettes. CYP2A6 is also involved in formation of an activated carcinogen from one of the chemicals in tobacco smoke, so fast metabolisers may also be at increased risk of lung cancer, as is observed for African American smokers [see Editorial by Sellers, E.M. (1998) JAMA 280, 179-180].

Comment: No single gene can account for the 'pharmacogenetics' of smoking. Nevertheless, the time may not be far off when it will be possible to identify sets of gene variants that make individuals particularly susceptible to nicotine addiction and/or smoking-related illness. It seems clear that the most sensible public-health policy will always be to discourage anyone from smoking, but if further research confirms the differential susceptibility of different racial groups, for example, there may be a case for targeting part of the anti-smoking campaign preferentially at vulnerable subsets of the population. 

A panel convened by the US Centres for Disease Control and Prevention and the US National Human Genome Research Institute has come out against population screening for hereditary haemochromatosis [Burke, W. et al. (1998) JAMA 280, 172-178 (Summary)] Two different mutations in a gene called HFE account for at least 90% of people with hereditary haemochromatosis, and the carrier rate in white populations is high (about 10%). Affected individuals can be offered treatment (regular phlebotomy) to reduce iron overload. This report argues against  population genetic screening for HFE mutations, arguing that  too little is know about their penetrance (i.e. the chances that people who are homozygous for the mutations will develop the disease), about the burden of disease caused by the mutations, or about the possible adverse social effects of identifying large numbers of mutation carriers.

Comment: Since the identification of the HFE gene in 1996, it has often been argued that hereditary haemochromatosis represents the ideal condition for population genetic screening: a high mutation frequency, an accurate genetic test, and a simple, cheap and effective preventive treatment. The US statement sounds a welcome note of caution. Although hereditary haemochromatosis may be underdiagnosed, the introduction of screening could run the risk of labelling as "ill" - and subjecting to treatment - many people who may actually never show symptoms of the disease.  

An interaction between the genetic make-up of mother and foetus can affect both the birth weight of the child and its susceptibility to non-insulin-dependent diabetes in later life [Hattersley, A.T. et al. (1998) Nature Genet. 19, 268-270 (Medline Abstract). See also the News And Views article by Mark McCarthy on p.209]. Babies who carry a mutation in the glucokinase gene, but whose mothers are normal for this gene, have significantly lower birth weights than babies whose glucokinase genotype is the same as their mother's (i.e. both normal or both mutant). Hattersley et al. suggest that it is an interaction between the circulating maternal glucose level (determined by the mother's genotype) and the ability of the foetus to regulate insulin secretion in response (an ability determined by the foetus's genotype) that leads to a higher or lower growth rate of the foetus.

Comment:: Although glucokinase mutations are rare, the importance of this work is that it demonstrates a genetic explanation for the observation that low birth weight is a risk factor for non-insulin-dependent diabetes later in life. Other more common genetic factors that affect both foetal and adult insulin metabolism might also affect both birth weight and risk of later diabetes and hypertension, though the overall picture is likely to be complex and to involve environmental effects as well. In public health terms, the implication is that concentrating on improving maternal nutrition or behaviour to combat low birth weight and its associated risks might turn out be of limited benefit. 

Prenatal diagnosis of Down's syndrome by analysis of the chromosome complement of foetal cells obtained by amniocentesis usually takes about 15 days because of the time needed to grow the foetal cells in culture. By using the polymerase chain reaction to amplify specific DNA sequences on chromosome 21, and then measuring the number of copies of these sequences  in the foetal cells, the time for diagnosis can be shortened to one day [Verma, L. et al. (1998) Lancet 352, 9-12]. Normal individuals  have two copies of chromosome 21 sequences, while Down's syndrome individuals have three. The method is very accurate (no false positive or negative results so far in more than 2000 analyses) and over 99% informative. 

Comment:  This technique looks very promising. If the authors' estimates of 5000 samples per scientist per year (cf 400 for chromosome analysis) and a total consumable cost of less than £5 per PCR assay are accurate, this may become the method of choice for Down's syndrome diagnosis.  

A prospective population-based study from the Netherlands [Ott, A. et al. (1998) Lancet 351, 1840-1843)] confirms previous suggestions that smokers are at increased risk both of vascular dementia and of Alzheimer's disease, but reports that the increased risk of Alzheimer's does not apply to smokers who carry the E4 allele of the blood protein apolipoprotein E (for an explanation of the relationship between APOE genotype and risk of Alzheimer's, see Alzheimer's disease summary in the Information database). In other words, there appears to be an interaction between an individual's APOE genotype and the biological effects of smoking.

Comment:  This work provides further support for the conclusion that there is no justification for APOE genetic testing to assess genetic risk of Alzheimer's disease. Smoking is just one of many 'environmental' factors that are likely to interact with an individual's genetic make-up in determining whether they will develop disease. 

To the handful of recent consensus statements from the US on genetic testing for susceptibility to Alzheimer's disease can be added a statement from the Stanford Program in Genomics, Ethics and Society, published in the July issue of Nature Medicine [McConnell, L.M. et al. (1998) Nature Med. 4, 757-759]. In addition to discouraging predictive genetic testing for Alzheimer's, the Stanford statement comes out firmly against the use of APOE genotyping as a diagnostic test, arguing that "[t]he small increase in diagnostic confidence provided by APOE genotyping does not justify the burdens of testing". These burdens include the inevitable revelation of information about probable/possible APOE status to the close relatives of the affected person, and the lack of any benefit to the affected person in terms of their treatment or prognosis.

Comment: This statement generally concurs with other recent consensus statements from the US on genetic testing for susceptibility to Alzheimer's disease, but goes further in its opposition to diagnostic genetic testing, which has received some cautious support from other quarters. At present the use of APOE genotyping appears only justifiable in a research setting, in studies to clarify the relationship between APOE allele status and the risk or nature of disease.