The belief that Carrel’s strain stood for, that cells taken out of the body can go on dividing for ever, lasted until 1961. What ended it was not a repetition of Carrel’s experiment but a different one, designed so that it could fail. This piece is about that experiment, told as far as possible in the words of the man who ran it.

A side excursion

In 1958 Leonard Hayflick, newly arrived at the Wistar Institute in Philadelphia, set out to look for viruses that might cause human cancer. His plan was to grow human tumour cells and normal human cells side by side and see whether anything from the tumour cultures changed the normal ones. For the normal cells he wanted fetal tissue, which was less likely than adult tissue to carry stray viruses of its own, and he grew fibroblasts, the same cell type as Carrel’s, from lung and other tissues of a number of fetuses over many months. Then he noticed something that the dogma said he should not have seen. “Of the dozen or so human fetal fibroblast cell populations that were set at random times, only those that had been subcultivated longest had stopped dividing. Cultures only stopped dividing after about fifty population doublings regardless of when in time they were initiated.” The same technician used the same batches of medium and the same glassware for all of them; young cultures in the next rack were flourishing while old ones, grown under identical conditions, slowed and died. He named the stages: Phase I for the first outgrowth, Phase II for the months of luxuriant growth, Phase III for the decline and the stop. “What I thought would be a little side excursion lasted for thirty years. I never did return to the cancer project.”

The experiment that could fail

He had been taught, as everyone had, that cultures died because nobody yet knew how to feed them. With Paul Moorhead, a cytogeneticist, he set out to test whether the dying was an artefact. Moorhead showed, by counting chromosomes, that the strains were normal, which no one had done before for a culture that went on so long. Then they ran the experiment that separated the cells from their surroundings. “We mixed the same number of male cells at the fortieth doubling with an equal number of female cells at the twentieth doubling. After twenty more doublings of the mixture we found that the only cells present were female. Unmixed control cultures stopped dividing at the anticipated times. The older male cells in the mixture stopped dividing and were overgrown by the younger and still flourishing female cells. Clearly, any virus, toxic factor, nutrient deficiency or cultural condition that might have eliminated the male cells could not be expected to spare the younger female cells. Nor were the young female cells capable of rescuing the older male cells.” Whatever stopped a cell was counting inside it, not in the dish.

They ruled out the obvious alternative with arithmetic. If cells stopped because they ran out of some essential molecule they could not make for themselves, diluted at each division like petrol in a tank, then a cell at the first doubling would have had to carry so much of it that, even if it were hydrogen, the cell would have weighed three times what cells weigh. They inoculated their strains into the cheek pouches of hamsters, where cancer cell lines grow and normal cells do not, and the strains did not grow. And they reported one more thing that mattered later: cells frozen at any doubling and thawed years afterwards picked up the count where they had left it, so that the doublings before and after freezing always added up to about fifty. Something in the cell was keeping a tally.

The rejection

“So intimidated were we by the dogma,” Hayflick wrote, “that we were still fearful of publishing the results.” When they did, in 1960, the Journal of Experimental Medicine, the journal in which Carrel and Ebeling had published every report on the strain, turned the paper down. The editor was Peyton Rous, who had worked at the Institute alongside Carrel for thirty years and would win a Nobel Prize in 1966 for the chicken tumour virus that bears his name. He rejected it with the statement that “The largest fact to have come out from tissue culture in the last fifty years is that cells inherently capable of multiplying will do so indefinitely if supplied with the right milieu in vitro.” Hayflick’s comment, nearly forty years later: “That belief was tantamount to the belief that, given the right milieu in vivo, human beings also will live forever.” Their chief at the Wistar, Hilary Koprowski, knew an editor of Experimental Cell Research, George Klein, and the paper was accepted there within two months without change and published in December 1961. By 1982 it had been cited 1,560 times, one of the two hundred most cited papers of those twenty-one years; the count is now in the thousands. The number has a name, the Hayflick limit. “Fifty population doublings,” he liked to point out, “will result in the production of twenty million metric tons of cells.” A limit is not a small thing.

What it changed

The 1961 paper made six claims, and the first two were aimed at the half-century behind it: that the cells were normal in every respect, and that their stopping “was not an artifact of culture conditions that had been believed for the previous sixty years.” The third was the interpretation that the field has argued about since: that the stop is “a manifestation of aging at the cell level.” Hayflick’s own summary of what had gone before is the shortest account of Carrel’s legacy: “During the first half of this century it was believed that because cultured normal cells were immortal, aging must be caused by extracellular events.” If cells could not age, the study of ageing had no reason to look inside them, and for fifty years it mostly did not.

The practical consequence came first. One of Hayflick’s strains, WI-38, started in 1962 from the lung of a fetus, had the widest range of susceptibility to human viruses of any cells then known, and he proposed it as a safer substrate for vaccines than the monkey kidney cells then in use. After what he called a ten-year struggle with the regulators it became the first normal human cell strain licensed for vaccine production; polio, rubella, measles, chickenpox and rabies vaccines have been grown in it or in strains like it, and hundreds of millions of people have received them. The strain’s frozen ampoules still remember their doubling level: after thirty-six years in liquid nitrogen, he reported in 1998, the memory was intact.

The mechanism took thirty years. Hayflick had proposed in 1961 that normal cells must contain an intracellular clock or event counter; he later preferred to call it a replicometer, since it counted divisions rather than time. In the 1990s the counter was found. The ends of the chromosomes, the telomeres, lose a little of their length at each replication, and when they are short enough the cell stops dividing. “Normal cells are mortal because telomeres shorten at each division. Immortal cancer cells express the enzyme telomerase that prevents shortening.” In 1998 the telomerase gene, put into normal cells, made them immortal. Carrel’s two classes of life, the mortal body and the immortal cell, had turned out to be the wrong way round: the normal cell is the mortal one, and the cells that go on for ever are the abnormal ones, most of them with the properties of cancer.

The chicken

And the chick? Its cells were measured in the decade after 1961, and they stop sooner than human ones. Hay and Strehler, in 1967, found that fibroblasts from the chick embryo manage about twenty-five doublings in sixty to eighty days and die; Pontén, in 1970, grew sixteen strains for between twenty and twenty-seven doublings each, and not one of them turned spontaneously into the kind of abnormal line that grows without limit, which chick cells almost never do. Hayflick’s verdict on the strain that had stood in the way of his paper was the verdict of the method itself: “Despite many efforts, no one has ever confirmed Carrel’s work even to the extent of keeping normal chick cells proliferating for as short period of time as two or three years. Since confirmation is lacking and the scientific method demands this, Carrel’s studies are invalid.”

What it comes to

Carrel’s strain could only ever say “still alive.” Hayflick’s cultures could say “stopped,” and could say it at a number, and because they could fail they could be compared: old against young, male against female, frozen against fresh, human against chick. The finding that overturned fifty years was not a longer culture but a shorter one, counted. An experiment earns its result by the ways it could have come out otherwise.

Sources

• Leonard Hayflick, “A Brief History of the Mortality and Immortality of Cultured Cells,” Keio Journal of Medicine 47 (1998), 174–182. Every quotation above is from this lecture: the cancer project, the fifty doublings, Phases I to III, the male and female cells, the hamster pouch, the hydrogen argument, Rous’s rejection and his sentence, Koprowski and Klein, the citation count, WI-38 and the vaccines, the replicometer and the telomeres. jstage.jst.go.jp
• Leonard Hayflick and Paul S. Moorhead, “The Serial Cultivation of Human Diploid Cell Strains,” Experimental Cell Research 25 (1961), 585–621; Leonard Hayflick, “The Limited in Vitro Lifetime of Human Diploid Cell Strains,” Experimental Cell Research 37 (1965), 614–636. In copyright and paywalled; described here through the 1998 lecture and Witkowski. doi:10.1016/0014-4827(61)90192-6, doi:10.1016/0014-4827(65)90211-9
• Jan A. Witkowski, “Dr. Carrel’s Immortal Cells,” Medical History 24 (1980), 129–142, section 3, for Hay and Strehler, Pontén, and the summary of the limit. pmc.ncbi.nlm.nih.gov