DAY and night, an unseen guardian watches over you. All-knowing and omnipresent, it holds your life in its hands. Gifted with the power to deal out death when it sees fit, it is charged with defending you against the forces of disorder and decay. But it has a dark side too, dictating how likely you are to get cancer and how fast you age.
This guardian is no fabled supernatural being. It is a small protein with the rather unassuming name of p53, and it is probably the most important molecule in cancer. Faults in the protein itself or the processes it oversees are likely to be involved in the development of nearly all tumours. Now, after 25 years of research, we are finally getting to grips with how p53 works. At last we have the exciting possibility of giving our guardian a helping hand to prevent or cure cancer, and perhaps even slow the ageing process.
It is thanks to p53’s ceaseless vigilance that cancer is so rare. This may sound nonsensical at first; after all, 1 in 3 of us will develop some form of cancer at some point in our lives. But it only takes a single mistake in a single cell to start the development of a tumour, and considering the billions of cell divisions that take place over a lifetime, it is a wonder that most of us survive for as long as we do. People with the genetic disease Li-Fraumeni syndrome, which is caused by inherited mutations in p53, show what would happen if it weren’t for our invisible guardian. They are extremely prone to developing cancer, sometimes as young as 2 or 3 years old.
Advertisement
Several viruses associated with cancer are now known to target and shut down p53. Back in 1979, when I discovered p53, such viruses were a hot topic in cancer research. I joined a lab at the Imperial Cancer Research Fund in London, where there was a big project under way on a monkey virus called SV40. It makes just one main protein called T antigen, which dramatically transforms normal cells into cancer cells. But when I tried to purify T antigen from the transformed cells, I found it was always stuck to another protein from the host cell. The protein had a molecular weight of 53,000, so I named it p53 and suggested that T antigen stuck to it and and altered its function in some way (Nature, vol 278, p 261).
Today, we know that a number of viruses make proteins that gag p53. Importantly, they include the human papilloma virus, which can cause genital warts and is associated with cervical cancer. At the time, however, people were reluctant to believe such a radical suggestion. Colleagues were sniffy, suggesting that there was something wrong with my reagents or that p53 was some sort of artefact. Nature initially rejected my paper.
Gratifyingly, confirmation from other labs soon began to roll in. Arnold Levine, then at Princeton University, and his colleagues came out with a similar conclusion shortly afterwards (Cell, vol 17, p 43), as did a couple of other labs. But it was not until 1989 that we really understood what p53 was doing.
In that year, Ed Harlow at Cold Spring Harbor Laboratory in New York and David Livingston at the Dana Farber Cancer Institute in Boston and their colleagues discovered that T antigen also attacks another host-cell protein called the retinoblastoma protein, or Rb. People with faulty versions of the gene develop retinoblastoma, a cancer of the retina, in early childhood. We already knew that Rb keeps the process of cell division in check and so suppresses cancerous growth. If Rb is blocked, the cell cannot stop dividing. We soon realised that p53 was doing something similar. When researchers added p53 genes to cultures of cells, the p53 genes that contained mutations or genetic errors made the cells cancerous, while the normal version of the gene suppressed cell division. Bert Vogelstein at the Johns Hopkins Oncology Center in Baltimore discovered that many of the tumours he was studying had abnormal or missing p53 genes, making the vital link between the loss of normal p53 protein and human cancer.
Then my colleagues and I decided to test a collection of preserved tumour samples. We screened hundreds, and found that about half had altered p53 expression. So by 1990, we had a protein that suppressed tumour development, was targeted by cancer-causing viruses and was mutated in half of all tumours.
Today, more than 15,000 papers have been published about this one small protein. And the picture they paint is one of a molecule that has god-like properties within our bodies, deciding whether individual cells should live or die. If a cell is damaged beyond repair or starts to behave abnormally, p53 will force it to self-destruct. This power to hand down a death sentence to damaged cells is how p53 protects us against cancer.
There are several kinds of faults that set p53’s alarm bells ringing (see Diagram). The one we understand best at the moment is DNA damage. Left unattended, DNA damage makes the genome unstable and prone to mutation and disintegration. Gangs of proteins patrol the genome looking out for faulty bases or breaks in the double helix. If they find any they signal to p53, which leaps into action with dramatic effect. The molecules team up into groups of four and slam the brakes on the process of cell division, giving the cell precious time to repair the damage. If the genome is beyond repair, p53 unleashes the cell’s self-destruct program. The system is so sensitive that just one snap in both sides of the cell’s double helix will trigger the p53 response. Considering that our DNA is some 3 billion letters long, this is no mean feat: p53 has truly earned the title “guardian of the genome”.
“There is the fascinating possibility that we may one day manipulate p53 to control both cancer and ageing”
Complex, sensitive and deadly efficient: so how does this omnipotent guardian go wrong? Things in our environment, such as cigarette smoke and ultraviolet light, can damage the p53 gene, and many cause such specific changes to the protein that we can often tell which carcinogen a cancer patient has been exposed to. Tantalisingly, there are hints that we may soon be able to repair this damage and restore the protein’s function.
Now we know that about half of all cancers have mutations in p53, the obvious thing to do is to try to restore p53 in tumour cells. In principle, the simplest approach is to add normal p53 back into these cells by using gene therapy – typically injecting the body with viruses primed to reinsert the healthy gene. Indeed, this is already being done: a p53 gene therapy developed by Chinese biotech company Shenzhen SiBiono GenTech was approved for clinical use in China earlier this year. When combined with radiotherapy, the gene treatment can eliminate tumours in 64 per cent of patients with head and neck cancer. A similar therapy developed by biopharmaceuticals company Introgen Therapeutics is currently undergoing clinical trials in the US and Europe.
Another possibility is to find ways of patching up the faulty protein produced by mutated p53 genes. Most of the mutations found in the p53 gene twist the protein out of shape and prevent it from working correctly. Our team showed it is possible to prod the protein back in to shape by sticking protein fragments called peptides onto it. In 2002, Klas Wiman and his team at the Karolinska Cancer Center (CCK) in Stockholm, Sweden, reported that they had found a small molecule that could do the same thing and restore the function of mutated p53 in cultures of cancer cells. (Nature Medicine, vol 8, p 282). And in February this year, Steven Dowdy at the Howard Hughes Medical Institute in Maryland and his colleagues used a peptide to restore p53 function in mice with advanced ovarian cancer, extending their survival time sixfold.
But there are other, subtler ways of harnessing p53’s power. Perhaps we could manipulate the mechanisms that control it, or bypass p53 completely and switch on the processes it normally activates. Such an approach would be ideal for the 50 per cent of patients whose tumours still contain intact p53, but whose overall p53 response is failing.
Understanding how the response works is key. We now know that p53 sits at the hub of a bewilderingly complex and delicately balanced network of biochemical pathways (see Diagram). This network is normally dormant, and only becomes active if things start to go wrong. When it is activated, the levels of p53 in the cell rise rapidly. This is not because the cell makes more of the protein: p53 is made at a constant rate. Rather, the processes that break it down are blocked. The p53 molecules then stick together in their groups of four and activate the genes needed to halt cell division or set off the self-destruct program.
Many existing cancer treatments work by activating the p53 response in tumour cells, mainly by deliberately damaging DNA. But what if we could boost p53 levels some other way, without having to wreck the patient’s DNA and potentially cause yet more mutations? There is encouraging evidence that we will soon be able to do so, by blocking the processes that normally break down p53.
One approach is already being tested. Left to its own devices, p53 is so deadly to cells that they keep it under control by chaining it to a burly security guard, a protein called Mdm2. Mdm2 stops individual p53 proteins from sticking together in the crucial groups of four, and also ensures it is rapidly broken down. Our team here in Dundee has shown that you can block the interaction between p53 and Mdm2 using peptides. Pharmaceutical company Roche has since discovered a small molecule called Nutlin that can do the same thing and can actually halt the growth of human tumours grafted into mice (Science, vol 303, p 844). Nutlin is the subject of intense study by many academics and pharmaceutical companies who are trying to turn it into a viable drug.
So we are now on the brink of having the power to turn our genome guardian on and off at will. As if that were not exciting enough, something else has happened in the past two years that has pushed people’s enthusiasm and excitement about p53 to a new level. And that is the link between p53 and ageing.
The story began in 2002 in Lawrence Donehower’s lab at Baylor College of Medicine in Houston, Texas. His team tried to replace the normal version of the p53 gene in mice with a version containing a specific mutation. But the experiment went wrong, and instead of making the protein the researchers wanted, the mice began to make just a tiny fragment of p53. Donehower’s team noticed that as well as being unusually small, the mice were ageing much more rapidly than normal. Curiously, they also suffered from a lower incidence of cancer than normal mice. But because these mice were making a weird, experimentally induced fragment of p53, nobody could really be sure what to make of the finding.
Then, a few months ago, Heidi Scrable and her team at the University of Virginia in Charlottesville engineered mice with a shortened form of p53 that occurs naturally and lacks the first 44 amino acids. For some reason cells sometimes make this protein alongside the full-length p53. Scrable found that mice making high levels of short p53 aged unnaturally quickly, but only if they also made the normal version of p53. It seems that the short form of p53 is controlling the activity of normal p53, although no one yet knows how.
The price of protection
So we seem to have a gene that can both limit cancer and accelerate ageing. How could this be? In fact these observations do fit in with the way we think about ageing. The body ages because it gradually loses its ability to regenerate damaged tissues. This happens as the limited number of stem cells that replenish these tissues die, often as a result of becoming stressed or damaged and then being forced to commit suicide by the p53 response. So the mechanisms that control p53 have to strike a very delicate balance. If the system is too sensitive – in other words, if p53 kills cells after only a small amount of damage – then you don’t get cancer but you also age much more rapidly. But if the system is not sensitive enough, your stem cells tolerate more damage and you age more slowly, but you have a much higher risk of dying from cancer. So ageing is the price we pay for a cancer-free youth.
Although it is still highly speculative, there is the fascinating possibility that we may one day manipulate p53 to control both cancer and ageing. Perhaps we will develop a treatment to suppress p53 for most of the time to slow ageing, but switch it on every few years to clear out any accumulated tumour cells. Researchers are already beginning to experiment with mice that have been engineered to produce a form of p53 whose activity can be controlled. They hope to be able to selectively kill off tumour cells, whose alarm systems are constantly ringing to activate p53, while sparing stem cells, which only sound the alarm in times of stress.
Lastly, if we can fathom the incredible subtlety of p53’s regulation, we may be able to understand why people have different susceptibilities to ageing and cancer. There are already hints that some people have a slightly different threshold for p53 activation than others. Gareth Bond and his team at the Cancer Institute of New Jersey announced in November they had found variations in the region of DNA that controls the activity of the gene for Mdm2. A particular variant may make people more cancer-prone (Cell, vol 119, p 591).
In the short term, I think there is definitely the potential to use p53 to develop really outstanding cancer therapies and possibly even drugs to prevent it developing at all. But realising the full potential of our guardian is not going to be easy. The more we discover about it, the more we see just how awesomely complex it is. We have found, for example, that p53 tends to activate different sets of genes in different kinds of cells. And it can be decorated with extra molecules like a Christmas tree, adorned with probably hundreds of small molecules in different combinations. There could be as many as a thousand or more different forms of p53 in the cell, each doing different things. It gives us an inkling of how incredibly complex biology is. This shouldn’t surprise us – but it is daunting.
