p53 (10 page)

In those cases in his data set with no family history of retinoblastoma, faults had obviously developed in both copies of the gene quite by chance over time – explaining why victims tended
to be older than those with familial retinoblastoma and typically to have no more than one tumour.

All this suggested to Knudson that one faulty copy of the gene involved in the development of the retina is not enough on its own to cause tumour growth at that site. The other copy has to
develop a fault also. But – and this was the revolutionary suggestion –
until
the normal copy develops a fault, it seems to keep the already faulty copy in check, so that the
cell functions perfectly normally. Only when both copies are faulty do the cells start to behave erratically and to develop into a tumour. Put another way, in all the cases Knudson observed,
something seemed to be breaking down and the cells involved to be losing their ability to function properly. He proposed that what had happened in these cases was that some kind of brake on
proliferation of cells was lost. It was such a simple suggestion, but it swam against the tide of cancer research which, in 1971 when Knudson published his ‘two-hit hypothesis’, was
intensely preoccupied with the driving forces of cancer, the newly discovered oncogenes, which implied a
gain
of function. He was proposing the existence of an equal and opposite force, an
‘anti-oncogene’ (soon to be renamed a ‘tumour suppressor’) that allows cancer to develop when it is knocked out – implying a
loss
of function.

Here, the car offers a useful analogy for visualising the forces at work inside cancerous cells, as proposed by Knudson. Think of oncogenes as the accelerator pedals, and tumour suppressors as
the brakes: a defective accelerator cable might stick, forcing the car to speed up uncontrollably (a gain of function); brake failure (loss of function) will have a similar effect in that the car
won’t be able to stop. But the analogy can be taken still further, since in most cars there are two brake systems. If one system fails, you still have the other and generally both sets of
brakes have to fail for catastrophe to occur. Likewise, for tumour-suppressor genes to be involved in cancer, it usually requires both copies of the gene, both alleles, to be disabled. This is the
central lesson of retinoblastoma, and the essence of Knudson’s ‘two-hit hypothesis’.

HENRY HARRIS HAS THE SAME IDEA

Knudson had arrived at his theory through mathematical modelling of the data before him. There was, however, more direct experimental evidence to back it up. It came from the
laboratory of cancer geneticist Henry Harris, an obstinate (by his own admission) and independent-minded Australian who had been recruited in 1952 by fellow countryman Howard Florey – famous
for his collaboration with Alexander Fleming in the discovery of penicillin, for which they won a Nobel Prize – to work with him at Oxford University.

Talking of his move from Sydney to Oxford many years later, Harris told an interviewer, ‘I got a telephone call from Hugh Ward, the Professor of Bacteriology [at Sydney], to say that he
had Florey in his office and would I like to meet him. I said, “I must be dreaming, you mean
the
Florey?” He said, “Yes, come on over.” So I dropped what I was
doing and I went over and there was Florey. He looked very much like a moderately successful businessman, but his speech was very laconic, very direct and he said, “Ward tells me that you
like doing experiments, Harris, is that right?” I said, “Yes, I quite enjoy myself, break a bit of glassware, make a noise.” He said, “Well, how would you like to come to
Oxford?” I said, “That’s like asking a man in the desert whether he would like a drink.”’

This was the era of the virologists whose flood of insights into the deepest workings of the cell so excited the cancer community that their theories of how tumours start – typically
through the acquisition of aggressive new powers in the cell – overshadowed all else. ‘My reaction to this unanimity of opinion was intransigent disbelief,’ wrote Harris in a
review for the journal of the Federation of American Societies of Experimental Biology, FASEB. He figured that, with or without the agency of a virus, the rate at which mutations occur naturally in
our cells is such that if malign mutations were always dominant – that is, able to override any other operating instructions in the cell – hardly a child would be born without a tumour
already forming.

In experiments run by fellow cell biologists Georges Barski and Francine Cornefert that seemed to confirm the virologists’ theories of the driving forces in cancer, Harris was struck by
something the two scientists had dismissed in the interpretation of their results. Their experiment involved fusing malignant mouse cells with normal mouse cells to see which set of instructions
prevailed. When, in due course, a tumour developed, they concluded that the genetic material from the malignant cell had dominated that of the normal cell. The fact that the resulting tumour cells
had a depleted number of chromosomes they thought was of no consequence. Harris did not agree. Could it be, he wondered, that in becoming malignant the cells had
lost
genes that might have
suppressed cancer, rather than
gained
genes that encouraged malignancy? It was exactly the question posed a few years later in Texas by Alfred Knudson, looking at the evidence from
retinoblastoma cases.

Over the next few years, Harris and his colleagues at Oxford – in collaboration with a lab in Stockholm that had the best materials to play with – explored this question by fusing
malignant cells with normal cells of various different types. They demonstrated conclusively that for the hybrid cells to produce tumours, something in the DNA had to be lost – something that
presumably was suppressing the malignant growth while it was still present. They published their findings in
Nature
in 1969, two years before Knudson’s retinoblastoma studies –
and well before it was possible to home in on the individual gene or genes that might be responsible.

But Harris and Knudson were up against the limits of technology in proving their theories; they were ahead of their time and their ideas caused barely a ripple in the cancer community.

THE FIRST TUMOUR-SUPPRESSOR GENE IS FOUND

That began to change in the late 1970s when cytologists – scientists who study the structure and function of cells – noticed that in the tumour cells of children
with retinoblastoma, chromosome 13 was unusually short: it seemed to be missing a large chunk of DNA. What is more, in those children with a family history of retinoblastoma, all the cells in their
bodies had a truncated chromosome 13. It gave researchers a place to look for the offending gene, and suddenly a hotly competitive race was on to find it and clone it. This promised to be the
novelty everyone was seeking – something that might explain the many anomalies that were thrown up by their pursuit of oncogenes.

But though the discovery had narrowed the field considerably, finding the retinoblastoma gene remained a Herculean task, for chromosome 13 is a mighty bundle of DNA some 60 million base pairs
long. Furthermore, scientists weren’t even sure whether they were looking for a single gene or a clutch of genes that normally worked in concert to suppress tumours. They got their answers by
an almost impossible stroke of luck. Arriving at Bob Weinberg’s lab in the mid-1980s, a young postdoc named Steven Friend announced he wanted to clone the retinoblastoma gene. As Weinberg
tells it, he met this request from his new recruit with frank astonishment: ‘
What?
How on earth are you going to do that? You don’t know anything about cloning; nobody knows
exactly where it is in chromosome 13.’ But Friend was not deterred. ‘Don’t worry. I’ll do it,’ he said.

With what Weinberg calls ‘irrational enthusiasm – totally irrational and illogical’, Friend went ahead. He struck up a collaboration with a doctor working at the nearby
Massachusetts Eye and Ear Infirmary, Ted Dryja, whose caseload at the hospital included children with retinoblastoma. Driven by concern for his small patients as well as by intellectual curiosity
and the desire to learn something about DNA, Dryja, who had no formal training in molecular biology, had started to do some lab research to try to find out what lay at the root of this dreadful
affliction. Focusing on chromosome 13, he had chopped out and cloned small fragments of the DNA. For him, this was just a means of acquiring some basic skills, but these clones created useful
probes for investigating the chromosome further. Dryja shared his new tools with Steve Friend and, soon after the young scientist began his search, ‘Lo and behold, one of these probes landed
right in the middle of the retinoblastoma gene and allowed it to be cloned out,’ Weinberg told his audience in the lecture hall at MIT. Spreading his arms wide to emphasise the length of the
DNA strand, then stabbing with his finger to indicate the extraordinary landing site of the probe, right on target, Weinberg spoke with a voice slow and deliberate with amazement. ‘Now you
know how many mega-bases each human chromosome is long; and you know how astronomically unlikely this stroke of luck was – or is. But it happened . . . This is what’s termed an
“unearned run”. It was a terrific finding.’

Steve Friend and Ted Dryja published their story in
Nature
in October 1986 and now the world was listening. A scientist called Webster Cavanee, then a postdoc at the University of Utah,
had earlier narrowed the field of search down even further, to a specific region on chromosome 13, and his 1983 paper was the first to confirm that Knudson’s two-hit hypothesis was right. On
hearing the news that the actual gene had been found and cloned, Cavenee commented, ‘I take my hat off to these guys. You can call it luck, but they did the right experiment, an elegant
experiment, and it worked. What more do you have to do before they stop calling you lucky and start calling you a good scientist?’

Alfred Knudson, too, was excited. ‘I’m delighted this has happened,’ he said. ‘Before, we could only concoct theories about what the retinoblastoma gene does. Now that we
have the gene, we can get to work on the facts.’

This was nothing less than a paradigm shift – a whole new way of looking at tumour formation as a battle of competing forces between oncogenes and tumour suppressors, the accelerators and
brakes of our car analogy above. It led also to the recognition, finally, that cancer is altogether an aberration. Oncogenes are not there primarily to drive cancer, and tumour suppressors are not
there primarily to suppress cancer; all these genes have regular work to do, including promoting or controlling the growth of cells as part of the endless cycle of building and maintaining our
bodies. Only when these vital genes become corrupted and start to malfunction do they acquire the ability to cause cancer.

When, very soon afterwards, p53 was finally revealed as being the same kind of gene as the retinoblastoma gene – a tumour suppressor, and a powerful one at that – it had an
electrifying effect on the field. Researchers reacted like a flock of starlings over a winter field, wheeling around to fly in a new direction, and those who had begun to lose faith in p53 and to
consider moving on to other things returned to their work with renewed enthusiasm.

***

The evidence that p53 had been miscast as an oncogene and was in fact a tumour
suppressor
had been accumulating in parallel with the work on the retinoblastoma gene.
At the forefront of this challenge to received wisdom was Bert Vogelstein, a legendary figure in the p53 story, whose lab has been involved in many of the most important discoveries relating to the
gene. On a stunningly hot afternoon in July 2012, I travelled to Baltimore to meet him, climbing the stairs of an elegant office block, all glass and sunlight and potted greenery, overlooking the
original old hospital of Johns Hopkins.

Vogelstein’s lab is famed almost as much for its fun as for its hard work. For years he headed a rock band, a bunch of musicians from his lab who called themselves Wild Type and who played
at scientific conferences and other venues. The band broke up when the drummer’s wife died of leukaemia and, with children to care for alone, he had to drop out.

‘We started the band mainly to develop
esprit de corps
in the lab. Everybody liked it, scientists liked it – we used to
play for scientists. And it
was fun!’ Vogelstein told me later. ‘I think it’s important to have outside activities. I certainly encourage everyone who works here to do so. Most people who have looked at
creativity have recognised that inspiration often comes in the off moments when you’re not focusing on exactly what you’re doing.’

Arriving a little early for our meeting, I waited in the lobby and leafed through a photo album I found lying on a low coffee table – pictures of Vogelstein’s lab
‘family’ at conferences and social gatherings, and of Wild Type in their heyday playing gigs. Beside the albums on the coffee table was a copy of
Grant Making for Dummies.
When
he emerged from his office, hand outstretched, I was struck by how slight a figure Vogelstein is and by the expression of impish fun on his face as he led the way into his large, cool office and
motioned towards a swivel chair. Against the back wall I noticed the keyboard which, he told me, he likes to tinkle on from time to time during the day.