As Neil Ganem spent many hours peering down a microscope in search of dividing cells, his adviser David Pellman at the Dana-Farber Cancer Institute in Boston, Massachusetts, would sometimes joke that he was “looking at his wife's teeth”.

Pellman was referring to philosopher Bertrand Russell's words about Aristotle's assertion that women have fewer teeth than men: “Although he was twice married, it never occurred to him to verify this statement by examining his wives' mouths.” Ganem was testing, by direct observation, a long-held assumption in cell biology — and, similarly to Aristotle's assertion, it turned out to be wrong.

Most cancer cells have extra centrosomes — small structures that control the formation of the mitotic spindle, which organizes and segregates chromosomes during cell division. Dividing cells normally have two centrosomes, one at each end of the cell, which pull chromosomes towards the poles. If everything works, the cell splits into two daughter cells with equal complements of chromosomes. But many cancer cells are chromosomally unstable, meaning that they often 'missegregate' their chromosomes.

Many scientists assumed that the extra centrosomes in cancer cells generated this instability by giving rise to multipolar cell divisions, resulting in three or more viable daughter cells with abnormal chromosome numbers. “But no one had looked to see if it was really happening,” says Ganem.

Using a microscope equipped with an incubator, Ganem watched thousands of cells grow and divide, looking for those with multiple poles and following the fate of their daughter cells — sometimes getting motion-sickness in the process, he laughs.

He found that cells with extra centrosomes rarely produce multipolar divisions. And when they do, the daughter cells typically perish. Having established that extra centrosomes do not lead to chromosome instability through multipolar division, Ganem asked: “Do extra centrosomes contribute to instability at all?”

To address this, he was inspired by the work of his graduate adviser, Duane Compton of Dartmouth Medical School in Hanover, New Hampshire. In normal cell division, one copy of each chromosome attaches to each pole. But sometimes one chromosome copy attaches to both poles — a defect known as merotely. “If not corrected, the chromosome can get stuck in a tug of war, and sometimes ends up going to the wrong daughter,” explains Ganem.

Compton showed that merotely was a major cause of chromosome missegregation in cells with chromosome instability. So Ganem decided to test whether having extra centrosomes leads to merotely. He created two sets of cells, one set with extra centrosomes and the other with two centrosomes as normal. Turning to the microscope again, Ganem and colleagues found that cells with extra centrosomes were much more likely to form merotelic attachments and missegregate their chromosomes (see page 278). “Having extra centrosomes is definitely a defect that promotes missegregation,” says Ganem.

The finding raises several questions to follow up, such as where the extra centrosomes come from. But for now, Ganem, who undertook the study as a side project, will be turning his attention back to his main research — a genome-wide screen to identify novel tumour suppressors.