In which we learn that besides lung cancer, other types of cancer, including liver and skin, frequently have mutant p53 that carries the unique fingerprint of the agent that caused the disease.
***
The most exciting phrase to hear in science, the one that heralds the most discoveries, is not ‘Eureka!’ (I found it!) but ‘That’s funny . . .’
Isaac Asimov
Liver cancer is the seventh most common cancer worldwide, but in South East Asia and sub-Saharan Africa, where the great majority of cases occur, it kills more people every year than any other tumour type. In these regions Hepatitis B, which is a major risk factor for liver cancer everywhere, is extremely widespread – passed between sexually active adults and from mother to child, much like HIV. And like the AIDS virus, too, it can wreak havoc in a person’s body without them being aware of the infection, and become endemic in communities. Hep B generally takes many years to cause liver cancer, but in the high-incidence countries of Asia and Africa people’s risk of getting the disease is compounded by exposure also to aflatoxin, a poison produced by the fungus Aspergillus that flourishes on peanuts and grains stored in warm, damp conditions without adequate ventilation.
Aflatoxin is a known carcinogen and was one of the chemicals investigated by Curt Harris’s lab in the late 1980s and early ’90s for its mechanism of action in human cells. As with BaP in tobacco tar, Harris knew that aflatoxin is metabolised and transformed in cells into a substance that sticks to DNA and causes mutations. But it was work he did with colleagues in China, analysing the genetic mutations in liver tumours in Qidong county, on the north side of the Yangtze River opposite Shanghai, that showed the poison at work in the real world and pointed the finger at p53 as being the target for mutation. Rates of liver cancer in the county were exceptionally high; so too was people’s exposure to aflatoxin from mouldy grains and beans in their diet, and the researchers were struck by the frequency of an unusual mutation in p53 at codon 249. This resulted in the building blocks of the gene being swapped, a G to a T – the same as with tobacco tar, but in a different hot spot on the gene. Could this be the fingerprint of aflatoxin?
As the paper describing Harris and his colleagues’ findings and suggesting such a possibility was about to go to press, a visiting scientist to Harris’s lab at the National Cancer Institute mentioned casually that another group, working in South Africa, had also discovered an unusual p53 mutation in liver tumours, but were unsure of its significance or what to do with their findings. Realising that this strengthened their case for a direct link between aflatoxin and p53 in liver cancer, Harris pushed the other group, led by Mehmet Ozturk, to write up their research in double-quick time so that the two papers could be published together, and they came out back to back in Nature in April 1991.
The coincidence of aflatoxin exposure and a distinctive p53 mutation in liver-cancer patients soon became apparent in many other warm, humid places with poor storage for crops. But what was going on in the machinery of their cells? Pierre Hainaut joined the quest to find out. Despite his initial unease at being drawn into the tobacco and cancer controversy, Hainaut is a natural-born sleuth, never happier than when he is doing research on the front line, where faulty tumour-suppressor genes are affecting the lives of real people. He has followed the fingerprints of p53 from China to Brazil, Iran to West Africa and South East Asia, and many other countries. With aflatoxin, his research has focused mainly on Mali, The Gambia and Thailand – three countries where liver cancer is a huge problem. Over the years he and others working on this issue, including Harris and Pfeifer, have revealed a devilish relationship between aflatoxin, the ubiquitous Hep B virus and p53, as they co-operate to cause liver cancer.
First the scientists worked out how Hep B virus on its own can lead to liver cancer. A viral gene – known simply as ‘x’ because for a long time no one had a clue what it did – codes for a protein that has a dual function: one end of the protein encourages proliferation of the liver cells it’s infecting; the other end promotes apoptosis, cell death. In this way the virus tries to maintain some kind of balance in the population of infected liver cells. But in so doing, it causes cycles of inflammation, damage and repair to the liver that result in cirrhosis – a liver greatly enlarged and distorted by scar tissue and lumps, or nodules, of regenerated cells.
‘These cycles of destruction and regeneration can go on for a while, and sometimes they can kill people – they can die from cirrhosis of the liver without getting cancer,’ explained Hainaut. ‘But at some point, in the absence of mutant p53, what happens to people with chronic liver disease and cirrhosis is that HBx becomes accidentally integrated into the genome of the liver cells. At that point it loses it pro-apoptotic part, and what remains is just the part that activates proliferation: the cells then escape destruction and are on their way to cancer. This is why cancer develops as a sequel to cirrhosis in the context of wild-type p53.’
Harris’s group found also that HBx protein sticks to p53 protein, forming a complex in much the same way as SV40 does with p53. They assumed that in so doing HBx had a similar effect of crippling the tumour-suppressor function of p53, and that this was one of the driving forces towards cancer. However, very recent research by Hainaut and his colleagues in West Africa suggests this assumption is wrong; it has the relationship between the two proteins back to front, for what really seems to be happening is that p53 is blocking the ability of the virus protein, HBx, to trigger apoptosis, while leaving its ability to drive proliferation of cells intact. The crucial point here is that, in real life, the bond between p53 and HBx that transforms the viral protein is only really strong when p53 has the aflatoxin-induced mutation, at codon 249. Then the brakes are off and the liver is especially vulnerable to cancer. ‘The risk of having liver cancer for someone who is a chronic carrier of Hep B is about 5-7 times compared to a non-chronic carrier,’ Hainaut told me. ‘The risk of getting liver cancer with just aflatoxin is very difficult to measure, but is probably no more than twofold. However, the risk of having liver cancer if you have the two is at least 20 times, and some measures suggest it is up to 60 times greater than usual. So it’s truly multiplicative.’
The revelation that mutant p53 transforms the function of the virus rather than the other way round has also helped to explain an abiding mystery in African liver-cancer patients. When someone infected with Hep B virus finally succumbs to liver cancer, he or she generally has signs of advanced cirrhosis from years of damage and repair. ‘This is the rule in the Western world,’ said Hainaut. ‘The patient who doesn’t develop cirrhosis before liver cancer is really the exception.’ But this is not what they have found among patients with liver cancer in Africa, despite chronic infection with Hep B. ‘I would say that maybe 15 per cent have cirrhosis beforehand, and then a number of them develop cirrhosis during the proliferation of cancer, as a sort of secondary response of the liver to the inflammatory state, but it does not precede cancer.’
Hainaut’s theory is that by blocking the virus’s killer function, the mutant p53 prevents the regular cycles of inflammation, damage and repair that cause the scars and nodules of cirrhosis. Thus, paradoxically, aflatoxin exposure can be protective of people with chronic Hep B infection, often for years, until other events in the ordinary course of living render them vulnerable to cancer. ‘We could never understand why we have so little liver cirrhosis in these populations. It’s something I observed about 15 years ago – we had very few patients with cirrhosis. The common response was, “Ah, you’re not looking for them . . . They’re not reporting to doctors, so detection is not good . . . The diagnosis is not accurate,” and so on. But since then we’ve done a few cohort studies (which follow a group of people who share common characteristics and lifestyles) and still we find most patients presenting with liver cancer without any trace of cirrhosis beforehand.’
In their studies among liver-cancer patients in Thailand, Hainaut and his team found the same phenomenon – those who had the aflatoxin mutation as well as Hep B infection had no signs of cirrhosis. But though aflatoxin-mutated p53 may be protective of livers in the short term, the case for controlling the offending fungus to reduce the burden of liver cancer is overwhelming, as events in Mali have shown serendipitously. Poring over the cancer register in the capital city, Bamako, very recently, Hainaut and colleagues faced another mystery – liver-cancer cases seemed to be plummeting. In the 15 years since the late 1990s, the rate of new cases had declined by about 75 per cent. They looked for flaws and biases in the records, but could find nothing obvious to explain away the dramatic figures.
On further investigation they discovered that in the mid-1990s, the agriculture ministry had started a programme to prevent aflatoxin contamination of the country’s crops. The primary motivation was not public health but economics: Mali wanted to export its crops for animal feed and needed to comply with international regulations. But this has had far-reaching consequences for the man and woman in the street, said Hainaut. ‘The first thing is that the contamination of food has decreased; and second, most of the crop production has been diverted towards export, so the diet has changed.’ Today, the exposure to aflatoxin of people in Mali is only a tiny fraction of the exposure of people in The Gambia, where rates of liver cancer remain as high as ever.
But there is a downside to this story: as Hainaut and his colleagues predicted, doctors are beginning to see more people with liver cirrhosis in Mali than ever before, as Hep B is still widespread but the factor that keeps the offending viral gene under control – aflatoxin-mutated p53 – is no longer so common.
The great appeal of molecular epidemiology to those involved is that it is often swiftly and directly applicable to real life, and this is the case with liver cancer and mouldy grains. Having worked out the relationship between aflatoxin, p53 and Hep B virus, the scientists find they can read much of what is going on in a person’s liver with a blood test. ‘The point is that when material is being cleared from the liver it goes either into the bile or the blood. It can’t go anywhere else – there’s no direct route to the outside world like in the digestive tract or the lungs,’ explained Hainaut. ‘That means that every bit of DNA from liver cells ends up in the bloodstream. And the liver is such a massive organ that a large part of the free-circulating DNA that you find in the blood comes from the liver.’
The scientists have worked out a method for retrieving that DNA and screening it for aflatoxin-mutated p53. They are also able to monitor the components of the viral genome and to look at what’s happening with HBx. Unfortunately, however, no such simple test exists for skin cancer, where the carcinogen, sunlight, leaves an equally clear fingerprint on p53. People just have to be on the lookout themselves for the signs and symptoms of disease.
THE SUN’S FINGERPRINT
I was seven years old when my father was posted to Borneo to run a TB clinic and a general hospital serving the indigenous Dyak tribespeople. The family set sail from Liverpool, and I vividly remember the days on deck, far out in the ocean under a blank blue sky and scorching sun. We all got sunburnt, coming out in big blisters on our shoulders that were too sore to touch and meant we had to sleep, spread-eagled, on our stomachs. Eventually the skin peeled off in long strips like wallpaper, but my younger sister’s nose never seemed to heal completely and family photos of our Borneo days show her with a patch of sticking plaster across it most of the time. In those days, the 1950s, and indeed for several more decades to come, we had no idea of the risks we were taking in not protecting ourselves from the sun.
We know now that ultraviolet light (UV) is the main cause of skin cancer, and in the early 1990s Douglas Brash and colleagues at Yale University discovered that it too damages p53 and leaves a characteristic fingerprint mutation. When Brash first started investigating the effect of UV radiation, a known carcinogen, on skin cells in the late 1980s there were three main theories about how it causes cancer. One was that it disrupts the immune system so that it fails to remove damaged cells from the surface of the skin as normal; another that sunlight directly stimulates cell growth; and the third that it damages DNA, knocking out a vital gene or genes.
Several groups in Europe and North America working independently on skin cancer had already discovered that UVB rays are absorbed most readily by the squamous cells, flat disc-shaped cells just below the surface, and slightly less readily by the basal cells deeper in the skin that nevertheless account for the majority of skin-cancer cases (at that time, the effect of the sun’s rays on the melanocytes, the cells involved in melanoma – the least common but most deadly form of skin cancer – was still unclear). The researchers had discovered also that UVB radiation damages DNA – and that it does so in a very specific way: it hits always at the point where the two bases, cytosine (C) and thymine (T), are adjacent to each other on a strand of DNA, swivelling them round so that a C is replaced by T, and sometimes two Cs by two Ts. This causes slight but crucial changes in the recipe of the protein the strand produces. This exact mutation isn’t seen in tumours anywhere else in the body, where sunlight cannot reach, and is thus considered to be a fingerprint of UVB.
In us, as in all living things, DNA damage and mutation happen all the time as we pass through the mill of life. Mutation is, of course, what drives evolution and adaptation to the environment, so it can be a force for good as well as bad. Brash knew therefore that the fact that UVB causes mutation did not automatically point to this as being the culprit in skin cancer. However, he was persuaded by the general pattern of the disease – and particularly by some evidence from Australia, which together with New Zealand has by far the highest rates of skin cancer in the world – that his best bet was to explore the idea of damaged genes.
Typically, skin cancer develops in middle age and beyond. The Australian researchers had noticed that rates among pale-skinned immigrants – generally from the UK and other parts of northern Europe – who had arrived in the country as adults were lower than among those who had arrived as children. Among the more recent immigrants, skin-cancer rates tended to reflect the rates in their home countries, while those who had been in Australia since childhood had rates similar to other white Australians. This suggested that the insult to a person’s skin from UV radiation that had set him or her off on the path to cancer had occurred years earlier, with those who had been in the country longest obviously at greater risk from the powerful Australian sun than those who had spent their childhoods in the clouds and rain of northern Europe.
Brash reasoned that if UV radiation affected the immune system or triggered runaway growth of cells directly, the effect would be much more immediate and transient, and the age at immigration would make no difference to the risk. The fact that there was a difference pointed to a mutated gene, which has a lasting effect, as the most likely cause. His task therefore was to find out which gene was affected. Once again it was like searching for a needle in a haystack: the human genome had not yet been sequenced and everyone still believed it contained at least 100,000 genes, not fewer than 30,000, as we now know to be the case.
Brash and his group followed several fruitless lines of enquiry, looking at known oncogenes, before their luck changed. One day, Arnie Levine appeared at Yale to give a talk in which he mentioned that p53 was found to be mutated in many cancers. Brash heard the talk and suddenly the pieces fell into place: a tumour-suppressor gene was a much more likely candidate for skin cancer than an oncogene because, to cause malignancy, a tumour-suppressor gene requires both alleles to be damaged – or both sets of brakes to fail, to return to our car analogy of Chapter 7. These events could be years apart, thus accounting for the typically slow development of skin cancer often many years after the victim first suffered a bad dose of sunburn.
And there was another intriguing clue that p53 might be involved. People suffering from an extremely rare skin disease called Lewandowsky-Lutz dysplasia develop warty growths, particularly on their hands and feet, that can be profuse and that readily turn malignant when exposed to the sun. Lewandowsky-Lutz disease is caused by infection with certain strains of the human papilloma virus, HPV, which is known to target and destroy p53 protein and lead to cancer in other organs, notably the cervix.
To start the investigation, Brash’s group pulled from the medical archive blocks of tissue taken from non-melanoma skin tumours of patients in New York City and in Uppsala, Sweden, where Jan Pontén of the University Hospital had become interested and joined the research effort. All the tumour samples came from sites on the patients’ bodies exposed to the sun, such as face and hands. The researchers extracted DNA from each and homed in on the p53 gene, looking for mutations. They found them in 90 per cent of the samples – the great majority bearing the fingerprint of UV radiation and thus capable of producing an active protein with the characteristic modifications to the recipe. Brash and his fellow researchers wrote up their findings in PNAS in 1991. But this was before anyone knew how normal p53 worked, and it took another few years – and research by many other groups as well as theirs – for a clear picture to emerge of what happens in the normal course of events when we sit out in the sun, and what can go wrong to cause skin cancer. In essence the picture looks like this: at some point in their lives, most probably during childhood, people who develop skin cancer will have suffered an episode of sunburn which caused a cell or cells to sustain mutation to the p53. We now know that UV radiation causes extensive damage to DNA, but our bodies have an efficient mechanism for repairing it if it’s not too serious: enzymes in our cells snip out the damaged stretch of DNA and replace it with a healthy copy. However, the mechanism can fail, and the mutation hot spots are the sites in the genes where, for some reason, the repair process is least efficient.
Peter Hall and David Lane’s maverick experiment with the sun lamp on Hall’s arm showed that p53 is activated in our skin when we sit out in the sun. Other researchers have since found that in the normal course of events, this activated p53 protein will trigger apoptosis in cells that cannot repair their sun-damaged DNA. But a cell in which p53 itself is damaged will resist apoptosis and will sit around, reproducing the fateful mutation from one generation to the next, until further insults to the skin, often decades later, turn it cancerous.
At this point, says Brash, sunlight delivers a double whammy. A rogue cell with mutant p53 is surrounded by normal cells that, when damaged by sunlight, will respond, as they should, by committing suicide. This gives more elbow room for the rogue cell to spread. ‘By inducing healthy cells to kill themselves off, sunlight favours the proliferation of p53-mutated cells,’ he explained in an article for Scientific American with fellow skin-cancer specialist David Leffell. ‘In effect, sunlight acts twice to cause cancer: once to mutate the p53 gene and then afterwards to set up conditions for the unrestrained growth of the altered cell line.’ This is the crux of the matter, for it is the expansion of clones of that single mutated cell that precipitates cancer.
‘What most people don’t realise,’ says Brash, ‘is that clonal expansion is numerically more important to cancer than making the initial mutation. Exposing yourself to the sun five times will make five times as many mutations. But favouring cell division of the p53 mutant five times will make many more mutant cells. Like compound interest at the bank, this exponential increase soon leads to very large numbers.’ The pre-malignant lesion thus formed provides an increasingly easy target for the mutagen – in this case UVB radiation – to inflict the crucial ‘second hit’ on a p53-mutant cell that will turn it cancerous.
***
What we’ve been discussing here are somatic mutations – ones that occur by chance in individual cells of the body at some point in a person’s life. Sometimes the cell that receives the hit is a sperm or an egg, and the mutant gene can then be passed on to future generations. This is called a germline mutation and it can be very bad news for those who inherit it, because every cell in their bodies will carry the mutant gene.