In the American sitcom Friends, Monica bakes mouth-watering cookies for her housemates, but — having been overweight as a teenager (pictured) — she is reluctant to eat them herself. Kirsty Spalding and colleagues now provide scientific evidence for why Monica has every reason to be cautious (K. L. Spalding et al. Nature doi:10.1038/nature06902; 2008).

Two factors contribute to an increase in fat mass: the number of fat cells and how much fat each of them stores (their volume). The authors studied the dynamics of fat-cell number in some 700 adults, both lean and obese, and combined their data with previous observations in children and adolescents.

A clear pattern emerged: irrespective of weight, the number of fat cells seems to rise steadily from birth to the early twenties, but remains constant thereafter. Moreover, in patients observed before and up to two years after surgical treatments that facilitate weight loss by reducing stomach size, no decrease in fat-cell numbers was detected — although their volume did drop.

So, are fat cells that are generated in early life doomed to remain with us till death us do part? In animal studies, this question can be addressed by labelling DNA nucleotides with radioactive isotopes such as 14C. Differentiated fat cells do not divide, and so radioisotopes, incorporated in their DNA in the last round of division before differentiation, remain there throughout the cells' life. The time of radiolabel incorporation, which is worked out from its half-life, is therefore the 'birth date' of these cells. But the potential toxicity of radioisotopes means that such studies cannot be performed in humans.

Spalding et al. cleverly thought of the next-best option. Atmospheric levels of 14C have remained relatively constant for centuries, with the only major increase occurring between 1955 and 1963, when nuclear bombs were being tested above ground. A chain of reactions ensures that, at any given time, the radioisotope content of human DNA matches that of the atmosphere. The authors could thus follow fat-cell dynamics in individuals born around 1955–63.

As Spalding and colleagues' results show, fat cells have a high turnover: new cells are continually being born to replace their dead predecessors. The average age of a fat cell seems to be about 10 years in both lean and obese individuals, and the number of fat cells as a proportion of all cells remains constant in each weight group. But the total number of new fat cells was higher in obese subjects, suggesting that they are replenishing an existing larger pool.

So do the lean among us need to worry about our diet if we have fewer fat cells? Yes, we do: our fewer fat cells can still store large amounts of fat. Also, can obese people do anything about their weight? After all, they've already accumulated a large pool of fat cells in childhood and adolescence? Again, the answer is yes. As Monica seems to have guessed, they can still reduce the volume, if not the number, of their fat cells. A further corollary of the paper is that researchers seeking drugs to cure obesity should consider targeting the mechanisms underlying fat-cell turnover.