symptom of Alzheimer's is the appearance in brain cells of 'plaques' of insoluble protein whose growth damages the cell. A viral infection was once suspected to be the cause, as was a history of frequent blows to the head. The presence of aluminium in the plaques threw suspicion on aluminium cooking pots for a while. The conventional wisdom was that genetics had little or nothing to do with the disease. 'It is not inherited,' said one textbook firmly.
But as Paul Berg, co-inventor of genetic engineering, has said, 'all disease is genetic' even when it is also something else. Pedigrees in which Alzheimer's disease appeared with high frequency were 2 0 2 G E N O M E
eventually discovered among the American descendants of some Volga Germans and by the early 1990s at least three genes had been associated with early-onset Alzheimer's disease, one on chromosome 21 and two on chromosome 14. But a far more significant discovery in 1993 was that a gene on chromosome 19 seemed to be associated with the disease in old people and that Alzheimer's in the elderly might also have a partial genetic basis. Quite soon the culprit gene was discovered to be none other than APOE itself.5
The association of a blood-lipid gene with a brain disease should not have come as such a surprise as it did. After all, it had been noticed for some time that Alzheimer's victims quite often had high cholesterol. None the less, the scale of the effect came as a shock. Once again, the 'bad' version of the gene is E4. In families that are especially prone to Alzheimer's disease, the chances of getting Alzheimer's are twenty per cent for those with no E4 gene and the mean age of onset is eighty-four. For those with one E4
gene, the probability rises to forty-seven per cent and the mean age of onset drops to seventy-five. For those with two E4 genes, the probability is ninety-one per cent and the mean age of onset sixty-eight years. In other words, if you carry two E4 genes (and seven per cent of Europeans do), your chances of eventually getting Alzheimer's disease are much greater than those of the population at large. There will still be some who escape either fate - indeed, one study found an eighty-six-year-old E4/E4 man with all his wits. In many people who show no symptoms of memory loss, the classic plaques of Alzheimer's are none the less present, and they are usually worse in E4 carriers than E3. Those with at least one E2 version of the gene are even less likely to get Alzheimer's than those with E3 genes, though the difference is small. This is no accidental side-effect or statistical coincidence: this looks like something central to the mechanism of the disease.6
Recall that E4 is rare among Oriental people, commoner among whites, commoner still among Africans and commonest in New Guinean Melanesians. It should follow that Alzheimer's obeys the same gradient, but it is not quite so simple. The relative risk of P R E V E N T I O N 263
getting Alzheimer's is much higher for white E4/E4S than for black or Hispanic E4/E4S - compared with the risk for E3/E3S. Presumably, susceptibility to Alzheimer's is affected by other genes, which vary between different races. Also, E4's effects seem to be more severe among women than men. Not only do more women than men get Alzheimer's, but females who are E 4 / E 3 ate just as much at risk as those who are E4/E4. Among men, having one E3 gene reduces risk.7
You may be wondering why E4 exists at all, let alone at such high frequencies. If it exacerbates both heart disease and Alzheimer's, it should surely have been driven extinct by the more benign E3 and E2 long ago. I'm tempted to answer the question by saying that high-fat diets were until recently so rare that the coronary side-effects were of little importance, while Alzheimer's disease is all but irrelevant to natural selection, since it not only happens to people who have long ago reared their own children to independence, but strikes at an age when most Stone-Age folk were long dead anyway. But I am not sure that is a good enough answer, because meaty and even cheesy diets have been around a long time in some parts of the world — long enough for natural selection to go to work. I suspect that E4 plays yet another role in the body, which we do not know about, and at which it is better than E3 . Remember: G E N E S A R E
N O T T H E R E T O C A U S E D I S E A S E S .
The difference between E4 and the commoner E3 is that the 334th 'letter' in the gene is G instead of A. The difference between E3 and E2 is that the 472nd 'letter' is a G instead of an A. The effect is to give E2's protein two extra cysteines and E4's two extra arginines compared with each other, E3 being intermediate. These tiny changes in a gene that is 897 'letters' long are sufficient to alter the way APOE's protein does its job. Quite what that job is remains obscure, but one theory is that it is to stabilise another protein called tau, which is supposed in turn to keep in shape the tubular 'skeleton'
of a neuron. Tau has an addiction to phosphate, which prevents it doing its j ob; APOE's job is to keep tau off the phosphate. Another theory is that APOE's job in the brain is not unlike its job in the 2 6 4 G E N O M E
blood. It carries cholesterol between and within brain cells so they can build and repair their fat-insulated cell membranes. A third and more direct theory is that, whatever APOE's job, the E4 version has a special affinity for something called amyloid beta peptide, which is the substance that builds up inside neurons of Alzheimer's sufferers. Somehow, it assists the growth of these destructive plaques.
The details will matter one day, but for now the important fact is that we are suddenly in possession of a means of making predictions. We can test the genes of individuals and make very good forecasts about whether they will get Alzheimer's disease. The geneticist Eric Lander recently raised an alarming possibility. We now know that Ronald Reagan has Alzheimer's, and it seems likely in retrospect that he had the early stages of the disease when he was in the White House. Suppose that some enterprising but biased journalist, anxious to find some way of discrediting Reagan as a presidential candidate in 1979, had snatched a napkin on which Reagan had wiped his mouth and tested the D N A on it (gloss over the fact that the test was not then invented). Suppose he had discovered that this second-oldest-ever presidential candidate was very likely to develop the disease in his term of office and had printed this finding in his newspaper.
The story illustrates the dangers for civil liberties that genetic testing brings with it. When asked if we should offer APOE tests to individuals curious to know if they will get Alzheimer's, most in the medical profession say no. After cogitating on the issue recently, the Nuffield Council on Bioethics, Britain's leading think-tank on such matters, reached the same conclusion. To test somebody for a disease that is incurable is dubious at best. It can buy reassurance for those who find themselves with no E4 gene, but at a terrible price: the almost-certain sentence to an incurable dementia for those with two E4 genes. If the diagnosis were absolutely certain, then (as Nancy Wexler argued in the case of Huntington's - see the chapter on chromosome 4), the test could be even more devastating.
On the other hand, it would at least not be misleading. But in cases P R E V E N T I O N 2 6 5
