Science News - March 10, 2012

“What we have is a provocative set of findings
that need to be replicated.” — PAOLA SEBASTIANI
Centenarians
distinguished by
DNA signatures

By Tina Hesman Saey

People who live to be 100 often credit their aging success to particular dietary or lifestyle habits, religious faith or a generally positive outlook. But scientists have long believed extreme longevity is at least partly in the genes.

Now a group of researchers has identified 281 genetic variants that together distinguish people who live to be 105 or more from everyone else with about 85 percent accuracy.

Further analysis reveals several genetic signatures among centenarians, researchers led by Paola Sebastiani and Thomas Perls of Boston University report January 18 in the online journal PLoS ONE. While the findings are drawing some criticism, the results suggest that there is a genetic component to extreme longevity.

Centenarians in the study have just as many disease-associated genetic variants as other people. So the researchers think that the inherited component probably includes versions of genes that protect against age-related diseases.

“What we have is a provocative set of findings that need to be replicated,” Sebastiani says.

In an earlier version of the work that was published online in Science in 2010, she and the other Boston University researchers claimed to have found a set of 150 genetic variants that could correctly predict who would be a centenarian 77 percent of the time. But the paper soon came under fire for technical flaws. The researchers fixed the technical problem and engaged an independent lab at Yale University to analyze the data.

Despite those revisions, the paper was retracted from Science last year because the journal said the results no longer met standards for publication. Science’s reasoning is disingenuous, says Nir Barzilai, director of the Institute for Aging Research at Albert Einstein College of Medicine in New York City. “The results, if anything, are stronger,” he says. “The data are the data, and

No disease difference though 281 genetic variants can be used to distinguish centenarians from younger controls, few of those differences are due to a lack of disease-associated gene variants. a look at 1,214 disease-associated locations in the genomes of 1,054 centenarians and 4,118 controls found similar proportions of high-risk genetic variants in the groups. sOurce: p. sebastiani ET AL/PLOS ONE 2012

Genetic susceptibility to diseases

Lupus High cholesterol Macular degeneration

Condition

Parkinson’s disease
Crohn’s disease
Diabetes
Cardiovascular disease
Controls
Cancer
Centenarians
Alzheimer’s

0.1 0.2 0.7 0.8

it’s very striking.”
But other geneticists have expressed
vague unease with the findings.

“The obvious technical issues have been corrected,” says geneticist Greg Cooper of the HudsonAlpha Institute for Biotechnology in Huntsville, Ala. But longevity “is a messy trait” that may be too complicated to explain with a small number of genetic variants. “I’m not totally sold,” Cooper says.

In the new study, the researchers combed the genetic blueprints of 801 centenarians and 914 healthy younger people for longevity-associated variants. The researchers also replicated the findings with two additional rounds of testing, first with a separate group of 253 people in their 90s and 100s and a control group of 341 younger people, then with a third set of 60 centenarians and 2,863 other people.

The researchers detected only one variant, linked to the Alzheimer’s disease–associated APOE gene, that met statistical standards for separating supercentenarians from people with a more average life span. Many other variants also looked as if they might be tied to longevity, but none passed statistical muster.

So Sebastiani and her team began summing the effects of variants that didn’t quite rise to the statistical threshold to see if those individual differences added up to a genetic signature that could predict longevity. As the researchers added in more and more variants, up to the 281 reported in the study, their power to predict centenarians increased.

Such grouping of genetic variants has been used to study characteristics such as height, body mass and intelligence. That type of analysis may help detect an underlying genetic component to a trait, but it doesn’t indicate which biological processes are important, says Elizabeth Cirulli, a geneticist at Duke University’s Center for Human Genome Variation.