Can genes reliably predict risk for an Alzheimer's brain (left)? Credit: A. PASIEKA/SPL

Genetic testing for Alzheimer's disease tells a cautionary tale about the legal, medical and ethical complications of personal genomics, as the story of a Pennsylvania company shows.

Smart Genetics, based in Philadelphia, has stopped offering its controversial 'Alzheimer's Mirror' genetic test just eight months after introducing it. The test checked for variants in a gene, called APOE, that bestow as much as a 15-fold increased risk of developing Alzheimer's. Soon after launching the test, though, Smart Genetics chief executive Julian Awad found himself in a controversy over whether it violated intellectual-property agreements covering APOE testing.

Smart Genetics' tests were performed by Athena Diagnostics, based in Worcester, Massachusetts. Athena had, in turn, licensed the patents from Duke University in Durham, North Carolina, where researcher Allen Roses discovered the APOE link to Alzheimer's in the early 1990s. Roses and Duke argue that Athena's licence covers APOE testing only in people who already have symptoms of dementia.

"The test was never intended to be used for wholesale screening of non-cognitively impaired individuals," adds Alan Herosian, director of corporate alliances for Duke University. He says he has contacted Athena many times in recent months to press this point.

Michael Henry, Athena's vice-president of business development, wouldn't comment on whether the company agreed with this interpretation of its licence. But Smart Genetics is no longer taking new orders for Alzheimer's Mirror. Its website says the test is currently unavailable because of "high demand". The company's phone lines have been disconnected, and a Philadelphia newspaper, the Philadelphia Business Journal, reported earlier this month that the company has closed.

Smart Genetics co-founder Richard Watson would not comment on the newspaper article, and Awad did not respond to e-mails or phone messages. But one member of the company's scientific advisory board, Andrew Faucett of Emory University School of Medicine in Atlanta, Georgia, notes that the firm faced another roadblock: Smart Genetics was charging hundreds of dollars for one test, whereas other gene-scanning firms offer Alzheimer's risk assessments along with other tests. "The financial model was hard to support," he says.

The tests offered by other firms bring issues of their own. For instance, Navigenics of Redwood Shores, California, provides Alzheimer's risk assessments by testing variants of a gene called APOC1, which sits next to APOE on chromosome 19. Navigenics uses APOC1 variants to predict APOE status on the basis of published reports that certain variants of the two genes are often inherited together. But APOC1 is not a perfect proxy for APOE variants. Navigenics chief operating officer Sean George says the company is switching to testing APOE variants directly.

Roses thinks genetic testing for the risk of Alzheimer's will only get more complex. He claims to have unpublished data suggesting that variants in another gene can be used with APOE variants to predict, to within 5–10 years, the age at which a person will develop Alzheimer's disease.

But he also hopes that ethical aspects of risk assessments will change if clinical trials identify drugs to treat the disease. Currently, knowing one's risk of developing the disease may simply cause needless worry, as there is no prevention or treatment. But Roses notes that the firm for which he worked previously, GlaxoSmithKline based near London, has reported preliminary data that its drug rosiglitazone benefits patients with Alzheimer's if they don't carry a high-risk APOE variant (A. D. Roses Alzheimer's Dementia 4, 164–166; 2008). The company is scheduled to finish clinical trials next year to test whether the findings hold up in larger numbers of patients with Alzheimer's.

If they do, Roses notes, it will mean APOE tests might be more useful, as they could help identify patients who will benefit most from treatment: "That's where we are approaching very, very rapidly," he says.