Joshua Schiffman is a paediatric oncologist in Salt Lake City, and a good part of his clinic is children with Li-Fraumeni syndrome. They are born with one broken copy of a gene called TP53, and the numbers that follow them are brutal. Half of the women with the syndrome have had cancer by 31. Half the men by 46. By 70, nearly everyone.9 Schiffman spends his days watching what happens to a body when its main defence against runaway cells is running at half strength.
A few years ago he was at a conference and heard an evolutionary biologist, Carlo Maley, describe a puzzle. Elephants have far more cells than we do and live about as long, so by the ordinary arithmetic of cancer they should be riddled with tumours. They aren't. And elephants, Maley said, seem to have around 40 copies of the TP53 gene, where humans have two. Schiffman, who had spent his career on patients with one, remembers thinking: what if we could get elephant blood and test it next to the blood from our patients.11
Not long after, he took his family to the elephant show at Utah's Hogle Zoo. A keeper mentioned, as a piece of trivia for the crowd, that the elephants gave blood every week for routine health checks. Schiffman waited until the show ended and asked how he could get some. It took months of applications to the zoo's review committee. Then, each week, someone from his lab drove over, picked up a tube of elephant blood still warm from the animal, and brought it back to be irradiated.11
The arithmetic that doesn't work
The puzzle Maley described has a name, Peto's paradox, after the Oxford epidemiologist Richard Peto. In the mid-1970s Peto was studying how cancer rates climb with age in mice and in people, and he noticed that the two species follow the same mathematical curve on completely different timescales, as if a mouse packed a human lifetime of cancer risk into three years.2 In a 1977 essay he drew out the awkward implication. If every cell carries roughly the same chance of turning malignant each time it divides, then an animal with a thousand times more cells, living thirty times longer, should face a cancer risk tens of thousands of times higher. Humans should be far more cancer-prone than mice. Whales should barely survive childhood.1
They aren't. Within a species, size and lifespan do predict cancer: tall people and big dogs get more of it. Between species, the relationship mostly vanishes. The largest test of this came in 2022, when researchers pooled veterinary records on more than 110,000 zoo mammals across 191 species. Cancer mortality varied enormously, and carnivores suffered the most, but it was essentially unrelated to body mass or life expectancy.4
I should be honest that the paradox is softer than its name suggests. A 2025 analysis of amphibians, birds, mammals and reptiles found that larger species do have somewhat higher cancer prevalence, just far less than the cell-count arithmetic predicts, and that lineages which grew large quickly tend to have lower cancer than expected, as though rapid growth forced them to evolve better controls.5 Other statisticians have pointed out that the zoo datasets are small and noisy.18 So the right way to say it is not that size doesn't matter, but that big, long-lived animals have clearly evolved something that lets them get away with it. A blue whale has around a thousand times as many cells as you do.3 Something is protecting it.
The elephant's answer
The irradiated elephant blood told Schiffman's team what that something was, at least for elephants. The results came out in JAMA in October 2015, alongside an independent study from Vincent Lynch's lab in Chicago that reached the same conclusion from the genome alone.67
First, the numbers confirmed the paradox. From necropsy records of 644 elephants, cancer accounted for about 5 percent of deaths. In humans the figure is somewhere between 11 and 25 percent, depending on the population.6 An elephant has perhaps a hundred times as many cells as a person and lives into its sixties or seventies. It should be worse off than we are, and it is better.
Second, Maley's gene count was right. Humans carry one TP53 gene, two copies in each cell. African elephants carry at least 20 genes, 40 copies. One is the original. The other 19 are retrogenes, copies that were pasted back into the genome from messenger RNA at some point in the elephant lineage and, unusually, kept working rather than decaying into junk.6 Lynch's group showed that the copy number grew in step with body size as the elephant lineage got larger over millions of years, and that elephant cells have what they called a hyperactive TP53 response.7
The third finding is the one I think about most. TP53 is often called the guardian of the genome, and when a cell's DNA is damaged the protein it makes has two options. It can pause the cell and try to repair the damage, or it can order the cell to kill itself. Schiffman's team exposed white blood cells to radiation and counted how many chose death. In cells from Li-Fraumeni patients, with one working copy, about 3 percent did. In healthy human cells, about 7 percent. In elephant cells, about 15 percent.6 Elephant cells did not repair DNA faster than ours. They were simply far more willing to quit. As Schiffman put it when the paper came out: "If you kill the damaged cell, it's gone, and it can't turn into cancer."10
In 2018, Lynch's lab added a twist. Elephants carry a gene called LIF6 that in most mammals is a dead pseudogene, a fossil. In elephants it has been brought back to life. When TP53 detects DNA damage, it switches LIF6 on, and the protein travels to the mitochondria and punches holes in them, which triggers the cell's self-destruct sequence. The DNA switch that lets TP53 control it evolved at about the time the ancestors of elephants became large.8 A zombie gene, resurrected to serve as an executioner.
Other animals, other answers
If the elephant were the only case, it would be a curiosity. What makes this a field is that evolution solved the same problem several different ways, and almost none of them look like the elephant's.
The naked mole rat weighs about as much as a mouse and lives more than 30 years, eight times longer. In decades of observing large colonies, researchers have struggled to find a single case of cancer. In 2013, Vera Gorbunova and Andrei Seluanov's lab at Rochester found that mole rat cells secrete a form of hyaluronan, the gel-like sugar that fills the space between cells, that is more than five times heavier than the human or mouse version. It makes their tissues so viscous that cells can't crowd into the dense masses a tumour needs. Strip the heavy hyaluronan out and mole rat cells become as easy to transform as a mouse's, and readily grow tumours.12 The same lab found that a different burrowing rodent, the blind mole rat, does something stranger. Let its cells divide past a certain point and they release interferon and die all at once, the whole dish collapsing within three days.13
Then there is the bowhead whale, which can live more than 200 years and weigh over 80 tonnes. The surprising result, published in Nature in late 2025, is that bowhead cells are not harder to turn cancerous than ours. They actually need fewer genetic hits. What they have instead is better repair. Bowhead cells fix double-strand breaks in DNA more efficiently and more accurately than cells from any other mammal tested, and accumulate fewer mutations. The team traced much of this to a protein called CIRBP, which is present at very high levels in whale tissues and which, when added to human cells, improves their repair as well.14 The elephant kills damaged cells. The whale mends them. Both work.
Why we can't just copy it
The obvious question is whether any of this could be given to people. Schiffman thinks so. He co-founded a company, Peel Therapeutics, that packages elephant p53 protein into nanoparticles, and has reported that the particles kill several kinds of human and dog cancer cells in a dish.16 As far as I can tell from public reports, the elephant protein has not yet been tested in a human trial. That is not a criticism. Protein drugs that have to get inside cells are hard, and the company's first clinical programme is a more conventional chemotherapy.
There is also a reason for caution that comes from the biology itself. In 2002, researchers made mice whose p53 was hyperactive, closer to the elephant setting. The mice were strikingly resistant to tumours. They also aged early, with osteoporosis, shrunken organs and shortened lives.15 A cell that quits at the first sign of trouble is a cell you have to replace, and a body that does this too often runs out of replacements. Elephants may get away with it because they have so many cells to spare, or because they have other adaptations we haven't found. A human given elephant-strength TP53 might buy cancer resistance and pay for it in years.
And the paradox itself is not resolved by one gene. A 2025 survey of gene copy numbers across 94 mammals found no general pattern in which long-lived species stock up on extra tumour suppressors. The elephant did it, but most large animals found another route.17 Mathematical modelling suggests how little it would take: cutting the mutation rate or the rate of stem-cell division by two- or threefold would be enough to bring a whale's cancer risk down to a human's.3 Evolution had many small knobs to turn, and different lineages turned different ones.
The question, fifty years on
I keep coming back to the gap between Peto's essay and the elephant paper. He posed the question in 1977. The first real answer took thirty-eight years and a zookeeper's offhand remark. In the decade since, the answers have multiplied, and they point in a direction that is easy to miss when you're inside a hospital. We tend to think of cancer as a disease to be cured, one treatment at a time. The animals suggest it is better thought of as a cost of being large and long-lived, a cost that every lineage has to pay somehow, and that each has found its own way to pay. The mole rat pays with sticky tissue. The whale pays with meticulous repair. The elephant pays by throwing damaged cells away before they can do harm.
We pay more than any of them, and the Li-Fraumeni children in Schiffman's clinic pay most of all. Whether an elephant's spare copies can ever be lent to them is an open question. But the fact that an animal exists which has already solved the problem we are trying to solve seems, to me, like the most hopeful fact in oncology.