by Nick Lane
Weindruch and his colleagues argued that the shift in gene expression was caused by a rise in oxidative stress from damaged mitochondria.
Although the changes measured were correlative — so a causal relationship was not proved — the notion that mitochondria were damaged was supported by the declining activity of respiratory genes (presumably they are not transcribed if they are not needed) and by high levels of oxidative damage to mitochondrial DNA, proteins and lipids. The level of oxidative damage correlated with the activity of pro-inflammatory and stress genes.
Again, a causal relationship was not proved; but the most reasonable conclusion is that failing mitochondria brought about a rise in oxidative stress, which altered the spectrum of genes switched on. If so, then oxidative stress does indeed shift the genetic balance in old age towards stress-resistance and inflammation, as predicted.
Thus, our argument — the double-agent theory — is as follows. Infectious diseases cause a rise in oxidative stress, which is largely responsible 8 Lest these numbers seem small, we should remember that a change in activity of this magnitude in a single gene, such as that for haem oxygenase, can mean the difference between life and death while fighting off an infectious disease. Changes of this magnitude in 450
genes, therefore, are presumably serious.
The Double-Agent Theory of Ageing and Disease • 301
for coordinating our genetic response to the infection. As we age, mitochondrial respiration also causes a rise in oxidative stress, which activates essentially the same genes through a common mechanism that involves transcription factors like NFB. Unlike infections, however, ageing is not easily reversed: mitochondrial damage accumulates continuously. The stress response and inflammation therefore persist, and this creates a harsh environment for the expression of ‘normal’ genes. The expression of normal genes in an oxidized environment is the basis of their negative pleiotropic effects in old age (Figure 12).
We are faced with two components to oxidative stress in the cell: mitochondrial leakage and non-mitochondrial factors such as infection.
Let us assume that the overall degree of oxidative stress is additive. If the stress from old mitochondria is added to the stress from infection, the combination might easily be overwhelming — perhaps this helps to explain why old people are more prone to die from infectious diseases such as influenza and pneumonia. Earlier in life, the stress from mitochondrial leakage is less serious, but other factors may ‘top up’ the overall stress to a higher level. In the case of infections, the rise in oxidative stress is normally constrained and reversible, but other factors, such as smoking or high blood glucose, may be more persistent. Any factor that brings about a general rise in oxidative stress should exert similar effects through the same genes. If oxidative stress is persistent, it might have the effect of simulating old age, at least in some tissues or organs. Factors that cause a ‘premature’ rise in oxidative stress would thus be expected to cause premature ageing, and a greater risk of other diseases.
In the rest of this chapter we will look at how this works out in practice, using the example of Alzheimer’s disease. Dementia illustrates many of the points we have discussed, such as misleadingly complex genetics.
There are several well-established genes that increase the risk of Alzheimer’s disease, which seem to have no association with oxidative stress.
This disease is therefore a good example of how susceptibility genes that seem unrelated to oxygen are in fact influenced by mitochondria, oxidative stress, and inflammation.
Why do some people with no genetic susceptibility still get Alzheimer’s disease? This question should be as revealing as the opposing question —
why do some genetic mutations increase the risk of dementia? After all,
302 • BEYOND GENES AND DESTINY
more than half of the people who develop Alzheimer’s disease have no known genetic risk factors. Following the simplistic nature–nurture interpretation, if the reason is not genetic, it must be environmental. For 30
years, researchers have looked in vain for evidence that aluminium or mercury provoke Alzheimer’s disease, but unequivocal links have never been proved; if there are links, they must be very weak. What about a virus? Many people are infected with the herpes simplex virus (the virus that causes cold sores), which typically localizes in the brain to the same regions that degenerate in Alzheimer’s disease. There is a link, as we shall see; but still, only half the people who are infected with herpes simplex get dementia, so this cannot be the only answer. In recent years, most serious research interest has centred on genetic factors, if only because the clues seem to lead somewhere. The fact that the clues do not point to most people who get Alzheimer’s disease is held in abeyance until our understanding improves. Surely they must be linked in some way!
The pathology of Alzheimer’s disease is characterized by two striking features, which were originally described by Alois Alzheimer himself in 1906: tangles and plaques. The tangles consist of twisted fibrils of a protein called tau. These fibrils are remnants of the extensive network of tubules that normally maintains the structure and function of neurons.
As the tangles form, the neurons die off around them, finally exposing the tangled fibrils like bones in an excavated graveyard. In contrast, the plaques are formed outside the neurons. They consist of dense deposits of a protein fragment called amyloid, mixed up with inflammatory cells (some types of glial cell and invading white blood cells) and detritus. In the venerable tradition of scientific controversy, the plaques and tangles have each attracted a dedicated tribe of researchers, convinced that their preferred pathology is the primary cause of dementia; and ne’er the twain shall meet. Most researchers, though, are open-minded enough to admit that there are problems relating the two features together. The premise that one of them must come first, and somehow produce the other, is difficult to verify. Biopsies of brain tissue can only be taken at autopsy, so most pathological data relate to the late stages of the disease. Animal models offer a potential escape from this problem, but so far no animal model of Alzheimer’s disease parallels the human syndrome in all its details.
The theory that amyloid toxicity is central to Alzheimer’s disease gained ground in the mid 1990s as virtually all the mutations that make us prone to the disease promote the deposition of amyloid. Amyloid is a
The Double-Agent Theory of Ageing and Disease • 303
fragment of a larger protein called the amyloid precursor protein (APP), which straddles the external membrane of neurons. In healthy brains, APP is cleaved to produce the soluble fragment, amyloid, which then circulates in the cerebrospinal fluid. What it does there is uncertain, but some evidence suggests that it is necessary for normal neuronal function.
Amyloid only becomes toxic when precipitated into dense clumps. The first clue to this came from people who had inherited very rare mutations in the APP gene, which cause dementia in early middle age (familial Alzheimer’s disease). The mutations shift the cleavage point of the amyloid precursor protein, so that the amyloid fragments are longer and
‘stickier’. The sticky fragments clump together to form plaques more easily. In 1995, two further genes were discovered, presenilin 1 and presenilin 2, which also cause dementia in middle age when mutated. The presenilin genes code for proteins that are thought to help in processing APP. Again, people with these mutations generate long, sticky amyloid fragments, which precipitate readily. Apolipoprotein E4, the protein product of the ApoE4 allele, also exacerbates amyloid deposition, but exactly how is not clear — we will return to this. Regardless of the exact mechanisms involved, all established genetic factors point to amyloid deposition as the primary pathology in Alzheimer’s disease.
There are two problems with this interpretation of the disease. First, tangles are often formed before the amyloid plaques, and indeed some people with classic symptoms of Alzheimer’s disease never develop amyloid plaques. In general, the onset of dementia corresponds to t
he loss of neurons, rather than the quantity of amyloid in the brain. Second, transgenic mice with mutations in the APP or the presenilin genes produce amyloid plaques, but only form tangles in old animals. Once the tangles have formed, the senile mice begin to lose neurons, and show signs of dementia, insofar as that can be measured in mice. Something similar is true of rhesus monkeys injected with amyloid — they develop tangles and lose neurons only in old age. Amyloid alone, it seems, is insufficient to produce Alzheimer’s disease, despite the fact that all known genetic mutations point to it as the prime suspect. Perhaps this should not be entirely surprising. After all, even people with APP or presenilin mutations do not develop Alzheimer’s disease until middle age — well past their childhood, and therefore unlike other single-gene disorders such as haemophilia. Something seems to be missing from the equation: might this be the same something that produces dementia in people with no genetic susceptibility?
304 • BEYOND GENES AND DESTINY
The secret of plaques and tangles is not to be found in genetics, but in chemistry. Let us think about amyloid first. People with no genetic susceptibility to amyloid deposition still deposit amyloid in their brains. This is because the deposition of normal amyloid (and perhaps even the sticky variety) depends on its oxidation: it clumps together when oxidized. In the dense amyloid plaques, amyloid is invariably oxidized. Amyloid is more likely to become oxidized later in life, as oxidative stress in the brain rises. As we have seen, oxidative stress rises in everyone in the end, regardless of their genetic make-up, because mitochondrial respiration inevitably damages neurons. But does oxidative stress rise first and then cause amyloid deposition? We do not know for sure, but the question can be recast more practically: if oxidative stress occurs earlier in life for some reason, does it bring about a correspondingly early deposition of amyloid and onset of Alzheimer’s disease?
One clue to the relationship between oxidative stress and dementia is provided by people with Down syndrome, who often get Alzheimer’s disease in early middle age. Again, the disease almost seems to have ‘moved forward’ to an earlier time slot. We saw in Chapter 10 that people with Down syndrome suffer from oxidative stress, as a result of an imbalance in antioxidant enzymes.9 Could a rise in oxidative stress underlie the early onset of dementia in people with Down syndrome? Quite probably, according to pathologists George Perry and Mark Smith, and their team at Case Western Reserve University, Cleveland, in a study published in 2000.
This team measured the oxidation of proteins and DNA in people with Down syndrome, and found that a marked rise in oxidative stress always preceded the deposition of amyloid. Oxidized proteins and DNA began to build up in their late teens and 20s, with amyloid deposition occurring by their 30s. Thus, it seems that a rise in oxidative stress does foreshadow an increased risk of Alzheimer’s disease, regardless of age.
What about the tau protein, the chief component of the tangles, the other pathological trait of Alzheimer’s disease? In a 1995 study, Olaf Schweers and his colleagues at the Max Planck Unit for Structural Molecular Biology in Hamburg showed that the tau protein only coagulates when 9 People with Down syndrome inherit an extra copy of chromosome 21, which includes the gene for SOD among others. SOD eliminates superoxide radicals, but in doing this produces hydrogen peroxide. Unless the hydrogen peroxide is eliminated by catalase, the extra SOD
increases oxidative stress. Other genes on chromosome 21 include the APP gene, so people with Down syndrome also overexpress the amyloid precursor protein and presumably amyloid; however, the amyloid fragment is normal, not the long, sticky version.
The Double-Agent Theory of Ageing and Disease • 305
oxidized. Conversely, if oxidation is prevented by antioxidants, tau will not coagulate.10 In other words, tangles, like plaques, usually form under conditions of oxidative stress. This is presumably why tangles do not form in transgenic mice or rhesus monkeys, despite extensive amyloid deposition: the pathology must ‘wait’ for a more general rise in oxidative stress.
Thus, for most people with Alzheimer’s disease, oxidative stress is the earliest pathological change, and is responsible for producing both main features, the tangles and the plaques. This idea is borne out by the effects of apolipoprotein E. Recall that the ApoE gene is polymorphic, which means that there are several different versions of the same gene. None of these versions is a mutant, in the sense that none is broken: all have been maintained by evolution, so all must have positive benefits. One of these benefits seems to be some degree of antioxidant activity. In old age, though, we have seen that ApoE4 raises our susceptibility to Alzheimer’s disease. The phrase ‘raises susceptibility’ is misleading. If ApoE4 is beneficial earlier in life, then it makes more sense to see it as beneficial later on too — but perhaps less so than its cousins. In other words, rather than raising risk, it might be less effective at suppressing risk. Why this should be is not known, but the ApoE4 protein is known to be more sensitive to free-radical attack than its cousins: it is therefore plausible that the ApoE4
protein is ‘lost’ preferentially with age. Early in life, when selection pressure is high, this difference is not felt because oxidative stress is low. Later in life, as oxidative stress rises, the benefits of ApoE4 (whether antioxidant effects or more general effects on cholesterol transport) are gradually lost, as more and more ApoE4 proteins are disabled by free radicals.
If this reading is correct, then the loss of ApoE4 proteins with rising oxidative stress accounts for two findings that otherwise resist interpretation. First, we noted that infection with herpes simplex virus increases the risk of Alzheimer’s disease. This effect is marked in people with two ApoE4
genes, but barely noticeable in people with ApoE3 or ApoE2. Activation of herpes simplex in the brain produces oxidative stress and inflammation.
As the ApoE4 protein is sensitive to raised oxidative stress, its beneficial 10The most obvious alteration in tau is its reaction with phosphate: it becomes abnormally highly phosphorylated in Alzheimer’s disease. Researchers had tacitly assumed that phosphorylation was necessary for the tau fibrils to coagulate into tangles. Schweers and his colleagues contradicted this view: phosphorylation is not necessary for tau deposition, and tau fibrils can form even in the complete absence of phosphate. More recent studies have confirmed their findings.
306 • BEYOND GENES AND DESTINY
effects are lost preferentially in people who have both ApoE4 and herpes simplex. In other words, people with ApoE4 are sensitive to oxidative stress anyway, and if they happen to harbour the herpes simplex virus, they are more likely to develop oxidative stress. They are therefore more likely to succumb to Alzheimer’s disease.
Second, and on the brighter side, people with two ApoE4 genes gain most from antioxidant therapies. This is demonstrably the case in all conditions for which ApoE4 is a risk factor, including dementia, heart disease and stroke. In one sense, this is a puzzle: as we have seen, antioxidants have little effect on mitochondrial respiration, and by suppressing the genetic stress response might even exacerbate oxidative stress inside cells.
However, they can help protect the ApoE4 protein from free-radical attack outside cells, as the external fluids are at once more accessible to antioxidants and less tightly controlled by gene activity than the insides of cells. Antioxidants may therefore shield the ApoE4 protein from oxidation, or supplement its own failing antioxidant actions. This in turn postpones the onset, or slows the progression, of dementia. One antioxidant proved to delay the onset of Alzheimer’s disease is vitamin E. If you know you have two ApoE4 genes ask your doctor about taking vitamin E supplements. If you have other ApoE genes, you may not gain much from taking extra vitamin E; but neither will you lose much (as long as you don’t overdo it).
Once formed, the tangles and plaques exacerbate oxidative stress both directly and indirectly. Direct amyloid toxicity is dependent on the binding of metal ions, such as iron and copper, which can catalyse the formation of free radicals. Such
metals undoubtedly bind to amyloid plaques in the brains of people with Alzheimer’s disease. If amyloid is added to cells grown in culture, its toxicity depends on free-radical formation in this way.
Conversely, amyloid toxicity is abolished, in the same simple system, by free-radical scavengers or metal chelators (which block the action of iron or copper). Thus, amyloid plaques are formed through the action of free radicals on amyloid, and then exert their toxic effects by producing more free radicals. They are free-radical amplifiers. In the brain, it seems likely that amyloid damages the neurons surrounding the plaques in this way, though it is unlikely to injure more distant neurons.
The indirect toxicity of both tangles and plaques almost certainly results from inflammation. Plaques and tangles are recognized as alien by the brain’s resident inflammatory cells, the microglial cells, which attempt
The Double-Agent Theory of Ageing and Disease • 307
to engulf the ‘invaders’ and attack them with chemicals, including free radicals. The plaques, especially, are indigestible and fester away. The jumpy microglial cells pump out inflammatory messengers that recruit and activate immune cells from elsewhere in the brain, as well as the blood stream. The entire brain is put on perpetual red alert. The chemical balance of the brain shifts inexorably towards oxidative stress, and susceptible neurons begin to die off. The most vulnerable neurones often die of ‘excito-toxicity’, in which they are provoked into a frenzy of electrical firing, and finally sink exhausted into an early grave. Thus, brain inflammation promotes the formation of more and more tangles and plaques, and finally the loss of neurons on a huge scale. By the time that Alzheimer’s disease can be diagnosed by standard clinical criteria, a quarter of the brain’s neurons — 25 billion of them — are dead. This vicious circle of inflammation is so important that the Canadian researchers Patrick and Edith McGeer have described Alzheimer’s disease as arthritis of the brain.