In which we discover that p53 protects us from cancer by stopping potentially dangerous cells in their tracks as they attempt to divide, and sending in the repair team to mend the damaged DNA.
***
I worry about p53 a lot. I’m paid to do it, but perhaps we all should, as the correct functioning of this 393-amino-acid nuclear protein is apparently all that lies between us and an early death from cancer.
David Lane
A critical part of the answer to the question of how p53 suppresses tumours came from a fellow medic of Vogelstein’s at Johns Hopkins Hospital, Michael Kastan, who came to p53 research through his work as a paediatric oncologist. ‘People say it must be depressing to treat kids with cancer,’ commented Kastan when I spoke to him by phone from North Carolina, where he now lives and works. ‘But first of all, kids do much better than adults; we cure 80 per cent of children with cancer, which is amazing. Also, they deal with it much better; and you get to know the families really well, so it was socially a more attractive field for me than other aspects of cancer.’
A single case that seemed to epitomise the combined social appeal and scientific challenge of paediatric oncology for Kastan – and convinced him this was the right field for him – was that of Dora Squires, a little girl with an unusual form of Down’s syndrome. So-called ‘translocation’ Down’s meant that, instead of having a complete extra copy of chromosome 21, as is most often the case, Dora had an extra scrap of chromosome 21 that had translocated and attached itself to another chromosome. When Kastan, as a young doctor on the wards, met her she was three years old and already had a long history of cancer.
Dora had been born with leukaemia – a not unusual occurrence in Down’s children – which resolved itself without treatment, Kastan told me. She did fine until she was two-and-a-half years old, when she developed a tumour on her face, a rapidly growing sarcoma that responded well and melted away with radiation therapy. But the little girl was soon diagnosed with acute myelogenous leukaemia, the condition for which she was being treated when Kastan appeared on the wards.
‘So I come on service and I hear this story, and I say, “Here is a girl with a known chromosomal translocation that in the space of three years has had three different tumours, and no one is saving her blood to try to figure out why.” Now this is before the era of knowing about oncogenes, but I said, “We don’t know what question to ask now, but if we save her blood samples, some day we’ll be able to ask questions about what it is with this translocation that led to this story.” Scientifically I found it extraordinarily interesting, and I felt it was a field ripe for discovery.
‘But the other piece of the puzzle was that Dora was a typical Down’s child: she was very very happy, and she always loved to see us when we came on our rounds. Her father weighed 275kg (600lb) and was too large to sleep on the parent bed in the ward, so he used to sleep on the floor – and she would sleep on his belly. So when we came on our rounds in the morning we’d open the door, she would see us and she would sit up on his belly and put out her arms for hugs from everyone on the team. You can’t not melt . . . So that’s how I decided to go into paediatric oncology – it was because of Dora Squires.’
Kastan had sandwiched his medical training around a PhD in molecular biology, for which he studied how cells respond when their DNA is damaged. And when, after completing his specialist training in paediatrics at Johns Hopkins, he started his own lab – a modest set-up consisting of himself and an inexperienced young assistant – to do research on the side, this was the topic he was intent on pursuing. His daily experience on the wards had convinced him it was central to cancer biology, he says, and nothing has happened since to change that view.
‘We know DNA damage causes cancer, right? We know this from animal models where you can take an animal, treat it with a carcinogen or radiation and cause tumours. We know it from human experience – Hiroshima and Nagasaki, for example, showed us how radiation causes cancer. We know it from exposure to carcinogens in the environment, which is why we have laws about what chemical companies can put in the water. And we know it from familial cancer syndromes, most of which are due to mutations in DNA repair genes.
‘So we know DNA damage causes cancer. But we also use DNA damage to cure cancer: radiotherapy and most of our chemotherapeutic agents target DNA. And most of the side effects of treating cancer – the hair loss, the bone-marrow suppression, the nausea and vomiting – are because DNA damage is killing normal cells. So from a clinical perspective, or a cancer-biology perspective, DNA damage causes the disease; DNA damage is used to treat the disease; DNA damage is responsible for the side effects of treating the disease . . . It makes it a pretty important phenomenon! As an oncologist, understanding and characterising DNA damage signals is important in every aspect of cancer.’
Kastan’s research had always focused on blood cells and what goes wrong with them to cause leukaemia. It was his preference for this cell type over the ones that cause solid tumours, or carcinomas, that enabled him, serendipitously, to discover the key mechanism by which wild-type p53 protects us from cancer – simply because, unlike carcinoma cells, leukaemia cells almost never have mutant p53. Thus what he observed in his experiments was the activity of normal p53 under a variety of circumstances, whereas he would not have seen anything – because nothing would have happened – if the cells had contained mutant p53. It was serendipity also because p53 was far from his mind, with no place in his research agenda, when he began his experiments to look at DNA damage and repair in cells that are dividing.
The cell cycle, as this dividing process is called, has several phases, and Kastan had been intrigued by a paper he had read from researchers studying yeast – one of nature’s simplest, most pared-down organisms, consisting of a single cell – that described how, if its DNA was damaged by radiation, yeast would stop at a ‘checkpoint’ in its cycle while the DNA was repaired, before carrying on through the cycle. The two researchers had found the gene responsible for this exquisite control of the cell cycle after damage, and this set up a challenge to Kastan: could he identify the genes and proteins that might be doing a similar job in us?
The first thing he did with his leukaemia cells was to bombard them with ionising radiation – which typically causes extreme damage to the DNA by breaking both strands of the double helix – and to take note of the changes in the cell cycle as a consequence. He found that his damaged cells arrested at checkpoints, demonstrating for the first time that what happened to yeast was a general phenomenon, applicable to the human body too. Now he could start asking the questions that really interested him: were there proteins whose level in the cells increased as a result of the damage, indicating that a particular gene or genes had been activated by the event? And if so, which of these genes were responsible for arresting the cells at the checkpoints?
Kastan was pinning his bets on the known oncogenes, and he was surprised to see no changes in these. But he had developed especially sensitive tools for measuring protein levels, and he noticed that the reading for p53 was slightly elevated. This was unusual since, for reasons that will soon become clear, p53 protein is normally present at levels that are barely detectable in cells. Could this slight increase be significant, he wondered?
Indeed it could. As we shall see, Kastan had begun to uncover the mechanism by which p53 suppresses tumours – by halting defective cells in their tracks so that they cannot divide. He was on the brink of a momentous discovery, but it would take time and hard work to tease it out.
FOLLOWING THE CLUES
While Kastan was busy with his initial experiments, a paper came out from Steve Friend – the scientist, you will recall, who had discovered the first-ever tumour suppressor, the retinoblastoma gene, Rb. Friend’s paper showed that if you pushed a dividing cell into producing an over-abundance of p53, it came to a temporary halt at a checkpoint named G1. It was a simple observation; Friend did not know whether this ever happened in real life, nor what might trigger an over-abundance of p53, but it made Kastan sit up. Could this be part of the same picture he had observed in his damaged leukaemia cells? Could the damaging event, the ionising radiation, be what activated p53? And could p53 therefore be the protein that was responsible for the checkpoint arrest?
‘This is when I got Bert Vogelstein involved,’ he says. ‘You know, I was a clinician who happened to have a very small lab and was doing these cell-cycle studies; Bert had this big machine . . . Since we were at the same institution I knew they were able to sequence p53. So I got on the phone to him one day in between seeing patients, and I said, “Bert, I think we know what p53 is doing. Will you sequence these cell lines for us?”’ Kastan described his experiments and his hypothesis to Vogelstein and explained that he wanted to check which of his cells had wild-type p53 and which had mutants, and to compare their activity. ‘Bert didn’t believe a word of my story!’ laughed Kastan. ‘But he said sure, he’d have the cells sequenced. And lo and behold, when we tested them, those that had wild-type p53 arrested at the G1 checkpoint after radiation, and those that had mutant p53 didn’t . . . I immediately had a sense of how important this might be: all of those experiments told us p53 plays a role in DNA damage responses.’
Kastan published his findings in Cancer Research in 1991 hot on the heels of news from Vogelstein’s own lab, working in collaboration with Carol Prives, that p53 was a master switch. All of a sudden, scattered pieces of the jigsaw began to fall into place.
THE JIGSAW BEGINS TO TAKE SHAPE
Another of those jigsaw pieces came from Kastan’s clinical casebook and involved a rare, inherited, neurodegenerative condition called ataxia telangiectasia, or AT. This devastating disease affects between one in 40,000 and one in 100,000 people worldwide. Children typically start to show signs of AT as toddlers, as it kills brain cells and progressively disrupts their motor co-ordination, affecting everything from walking and balancing to speaking, swallowing and moving the eyes; it generally sees them wheelchair-bound by the time they are in their teens. Treating his young patients with the condition, Kastan knew their risk of cancer was exceptionally high – in fact, 37 to 100 times that of the general population. He knew also that they were especially sensitive to ionising radiation, so medical procedures such as X-rays and CT scans were to be avoided if at all possible. Now, he began to wonder if both phenomena might have something to do with p53. Were the cells of patients with AT able to halt the cycle at the GI checkpoint and activate a DNA repair programme as they should, or was this mechanism defective?
‘We had cell lines from patients with AT and it became clear very early on that p53 did not get induced normally,’ Kastan says. ‘We had absolutely no idea what the gene was that was missing in these patients. But whatever it was, we realised it was somehow required for the induction of p53 after radiation.’
At the same time as investigating the AT connection, Kastan was collaborating with a scientist at MIT, Tyler Jacks, who had created experimental mice with no p53. Sure enough, thymocytes – important components of the immune system – in Jacks’ mice failed to arrest at the G1 checkpoint when bombarded with radiation. Together with Vogelstein, Kastan was also collaborating with a third group, at the National Cancer Institute near Washington DC, who had discovered a collection of genes called GADDs that are directly responsible for arresting growth of cells with damaged DNA (indeed, their name is derived, imaginatively, from Growth Arrest and DNA Damage). The three teams found that GADD 45 was controlled by p53, and was one of the genes switched on by the tumour suppressor to cause arrest at the GI checkpoint that Kastan had first uncovered. Very soon, Vogelstein found another gene, p21, involved in the same event and also controlled directly by p53.
The picture that emerged of p53 from these disparate bits of research was of a master switch at the hub of a communication network within cells. Its job is to respond to incoming signals indicating DNA damage by recruiting the relevant genes ‘downstream’ to halt growth of the cell pending future decisions about its fate. In this way cells with scrambled DNA that might threaten the organism are disabled.
It was this picture that the researchers described in a paper they published together in Cell in 1992 and that Kastan says was ‘the most fun thing I ever did in my scientific career’. Just before the paper came out, he attended his first big p53 meeting, hosted that year by Moshe Oren and Varda Rotter in Israel. Oren had seen Kastan’s original paper on checkpoint arrest following radiation and been sufficiently excited to invite the American to speak at the plenary session – to the full, august gathering of the p53 community. ‘What was so much fun was that I was a total unknown in the p53 field,’ said Kastan. ‘I go to this meeting; I get up to the podium and give this talk about this whole signal transduction pathway the day the paper was published in Cell.’ No one had seen the data before, and it had a powerful effect on the audience.
‘I was a nobody with a no-technology lab,’ he continued, ‘but I just happened to ask an important question because I read the literature carefully. And I asked it at the right time, with the right techniques and in the right cell type.’
RARE DEGENERATIVE DISEASE HOLDS THE KEY
Not everyone was ready to accept Kastan’s model entirely. The fuzziest part of the picture at that stage, in 1992, was the connection with ataxia telangiectasia. No one knew what the missing element was in these patients that made them so sensitive to radiation; they knew only that, in normal circumstances, it was essential for signalling to p53 that the DNA was dangerously damaged and for turning the whole damage-response system on. Things became clearer when, after a Herculean effort by 30 international scientists and hot competition between the labs to find the gene or genes responsible, a team led by Yossi Shiloh at Tel Aviv University announced success in 1995.
The single-minded search for the AT gene took more than 15 years of his life, Shiloh told me when I spoke to him over the phone from New York, where he was on sabbatical in 2012. It began when his mentor at university, Professor Maimon Cohen, suggested that the young scientist join him on a field trip to a small village in southern Israel; there they would meet a family of Moroccan Jewish origin afflicted with ataxia telangiectasia. Shiloh had recently completed his Masters degree and was casting around for a topic for his PhD thesis. ‘Professor Cohen had a hidden agenda – to interest me in AT,’ he said. ‘It worked very well because when I saw those patients I decided almost on the spot that this was an important problem to work on. First, because it’s an extreme human tragedy and at that time it was an “orphan disease” – no one cared much about these rare diseases with long names. And second, it was clear that understanding AT would have broad ramifications in many areas of medicine – neurology, immunology, genetic predisposition to cancer and whatnot – because AT is like a microcosm of medicine, it involves so many systems in the human body.’
Shiloh had no illusions about how difficult it would be to find a common cause for such diverse symptoms – and for many years the consensus among AT researchers was that there were four distinct types of the disease and probably at least four different genes responsible. The first breakthrough – what Shiloh identifies as the starting gun for the race to find the genes – came from Richard Gatti at the University of California in Los Angeles, whose study population was the Amish people of Ohio. In 1988, Gatti had managed to localise the gene responsible for AT to a region on chromosome 11, homing in on this stretch of DNA through a technique called linkage analysis, which looks for genetic markers – small strips of DNA with unusual ‘spelling’ dotted along the genome that are consistently present in people with a particular genetic disease, and never found in healthy individuals. The researchers then use statistics to suggest which marker or markers is closest to the target gene. This narrows the search area, but finding the actual gene is still akin to looking for a person’s house when you have only the name of the city in which they live to go on, and it was another eight years before Shiloh and his team managed to achieve their aim.
‘When I look back I’m surprised yet again that for eight years the entire lab was working on that one project . . .’ he said. ‘You know, scientists are very individual . . . Today we still work on AT, but every student in the lab has his or her own project. At that time the entire lab, several generations of students and postdocs, was focused on just fishing out genes from that region of chromosome 11, analysing them, cloning them.’
Today, thanks to the Human Genome Project and the wealth of data about genes and sequences available at the click of a computer mouse, such an exercise is relatively straightforward. But in the mid-1990s it was slow and labour-intensive, and relied on close co-operation with the AT-affected families whose personal DNA was the lifeblood of the research. Among the hundreds of genes Shiloh’s team cloned was one that specially caught their attention because it was unusually long – so long in fact that they had to repeat the cloning exercise a number of times to convince themselves it was real. Clearly this was the recipe for a huge protein – and one, they soon discovered, that had the hallmarks of a ‘signalling’ protein responsible for sending messages within the cell.
Shiloh remembers the day they realised this was what everyone had been looking for. ‘I had been teaching and when I came back to the lab from my class my student was holding a Southern blot7 in her hand. She said to me, and I remember her words clearly, “There is something odd about this gene in this family.” This was one of our Palestinian Arab families. I looked at the blot and it was clear that a big portion of that specific gene was deleted in that family. It was a very dramatic result. Of course my heart skipped several beats, but I said to her as calmly and quietly as possible, “This indeed looks interesting, there might be something here. Why don’t you repeat the experiment with DNA samples from the entire family and additional controls?”’
She did so and the conclusion was inescapable: here was the gene whose corruption was the cause of the disease Shiloh’s team were seeing in all their AT patients. It was a time of high tension, recalls Shiloh. The race to find the gene was at its peak, with frequent rumours in the air that someone or other had succeeded, and the temptation to publish his lab’s results immediately was heavy. But he had a hunch that there might in the end be just one gene – not the four that everyone supposed – responsible for the different manifestations of ataxia telangiectasia, and it would take time to prove it. Someone else might get there first, but after intense discussion among themselves everyone in his lab agreed to hold off announcing their results until they had tested their hypothesis. It was a nail-biting time, but the gamble paid off: AT is indeed caused by defects in a single gene, which the international consortium named ATM, short for ataxia telangiectasia mutated.
This was the missing detail in Kastan’s picture of the DNA damage response: in time he and others were able to show how the signals are passed down the line from ATM, which first senses the broken strands of DNA, to p53, which then throws the relevant genetic switches to halt the division of the cell. This was biochemical proof of the mechanism, and it finally convinced the doubters that p53’s response to DNA damage is at the heart of its action as a tumour suppressor.
‘You know, you can’t overstate the importance of what Yossi did in cloning the AT gene,’ commented Kastan. ‘He will be somewhat humble in telling it, but people were searching for that gene for 20 years – including him – and it made such an impact . . . It really opened up the whole DNA damage-signalling field. Yossi is a fastidious scientist and it’s because of that fastidious approach that they got to that point.
‘He flew to Baltimore to tell me he had the gene clone, and I remember very distinctly, he was sitting in my living room, saying, “Okay, we got the gene, and we’re calling it ATM for ‘ataxia telangiectasia mutated’.” I looked at him and I said, “Well, that sounds great, but you know ATM has another meaning in the US?” And I explained to him about these new automated teller machines. His face dropped, and I said, “Don’t worry, Yossi, people will know that’s where the money’s at in the signalling pathway!” And it’s been true. The field just exploded at that point.’
MULTIPLE STRESSES, MULTIPLE RESPONSES
In labs everywhere, researchers began testing the model, and evidence soon mounted that many more insults to the DNA – as well as more subtle stresses on the cellular machinery – can trigger the p53 response to halt the cell cycle. The increasingly long list includes UV radiation from sunlight, chemicals in the environment, and activated oncogenes, as well as natural ageing and dangerously low levels of oxygen and essential nutrients like glucose in the cell. Importantly, each stressor has its own characteristic pathway – from the protein that sends out the first alarm signal that all is not ideal for the division of the cell, thus triggering the response, to the range of genes that p53 switches on. But they all have the same effect of preventing potentially harmful mutations from being passed on from one generation of cells to the next.
The frenzy of activity among researchers was fuelled also by revelations that, just as there are many different stressors that can trigger p53, there is also a variety of outcomes to the response. Besides inducing a temporary halt in a dividing cell while DNA damage is repaired, p53 can induce a state of permanent arrest, called senescence. And under certain circumstances, it will instruct a seriously damaged cell to commit suicide – a process that many people feel is the most important weapon in its armoury.
In July 1992, David Lane, p53’s co-discoverer, pulled all the information together from widely scattered publications in a review for Nature in which he dubbed p53 ‘the guardian of the genome’ – essentially, the policeman in our cells taking action to clear dangerous individuals from the scene. As a reflection of what many people were thinking, it was neat; but as a statement from a scientist it was unusually bold. ‘In a sense it was sticking my neck on the block,’ said Lane with a mischievous chuckle. ‘You write a scientific paper and you say: it’s not unreasonable to speculate . . . But in this I said: this is how it works! Then everyone thinks, well there’s a challenge! But is it true? Is it not true? Not everyone believes it even now, but it provoked debate, which is what it was intended to do. That’s very important to the progression of science.’
AHEAD OF HIS TIME
There is a poignant footnote to this story. It involves a young scientist called Warren Maltzman, who worked briefly as a postdoc in Arnie Levine’s lab in the early 1980s, before moving on to Rutgers, the State University of New Jersey. Maltzman’s doctoral research at Stanford had focused on how cells repair damaged DNA, and when he joined Levine’s team he became involved, naturally, in p53. At Rutgers the two fields came together when Maltzman observed that in normal, non-cancerous cells subjected to UV radiation (as in sunlight), the levels of p53 shot up. He published his findings in the journal Molecular and Cell Biology in 1984. ‘At that stage,’ says Levine, ‘we didn’t know p53 was a tumour suppressor; we didn’t know what it meant that the level went up, and so his paper was roundly ignored. Had everybody picked it up, we’d have known p53 was involved in DNA damage and repair responses right away; we might have found that it transcribes genes . . . But . . . the time was not ready for anybody to make sense of it.’ Despite a good reference from Levine when he applied subsequently for a research post, Maltzman’s academic career faltered and he went into industry. ‘I feel badly about that because this man made a contribution whose time had not come . . . In many ways it’s the human story of science,’ mused Levine.