After two years of steady advances that have turned cell-reprogramming into one of the most promising areas of biology, researchers in Worcester, Mass., have discovered a major glitch that could delay the march toward a more intimate understanding of disease and even toward possible treatments.

Scientists at a company called Advanced Cell Technology compared cells grown from embryonic stem cells with others grown from the reprogrammed equivalent known as iPS cells. They found that many of the early blood, blood vessel and eye cells grown through reprogramming died or aged prematurely, a severe blow to the company, which had been hoping to go through the approval process to translate these techniques into treatments for patients.

"There's no way the FDA or any other regulatory agency is going to allow cells with those kind of problems into the clinic," said Robert Lanza, chief scientific officer at Advanced Cell Technology and one of the scientists reporting these finding in a paper published online Thursday in the journal Stem Cells Express.

"No one wants to hear the bad news, and we didn't want to either."

Lanza said, however, that there are early indications that scientists may be able to fix the problem by using a different reprogramming method. In a separate set of experiments, Lanza's team did not use the original technique of rewinding cells back to the embryonic state by inserting genes into them, but instead used proteins to accomplish the feat. The protein method appeared to produce cells that perform as they should.

"It made a huge difference," he said.

Still, Lanza said, "We have a very serious problem here," and work remains before it can be considered solved.

"It takes time to digest essentially revolutionary ideas," said Timothy Kamp, a stem cell scientists who did not work on this paper, but directs the University of Wisconsin's Stem Cell and Regenerative Medicine Center.

"It's like any new discovery. It takes time for the scientific community to work through a new discovery and fully understand it. . . . This isn't the end of the story."

Both Lanza and Kamp said the new results underscore the caution of stem cell scientists who have insisted that the reprogramming breakthroughs should not end research on human embryonic stem cells.

Cell reprogramming, the ability to take a mature cell such as a skin cell and return it to its embryonic origin, was achieved with the cells of mice in 2006 by the Japanese scientist Shinya Yamanaka, and then with human cells the following year by both Yamanaka and James Thomson at University of Wisconsin-Madison.

The discoveries triggered great excitement, though there have been cautions all along. The Yamanaka and Thomson labs both started out using viruses to deliver outside genes that set reprogramming in motion, techniques that tinkered with the genome and carried a risk of causing cancer.

From early on, scientists stressed that these were only technical difficulties, and within months a string of new papers presented alternative, safer ways to change cells.

Still, additional concerns have nagged researchers. Although the reprogrammed cells appear very similar to embryonic stem cells, scientists have found differences in the degree to which they express various genes. The full significance of these differences remains a mystery.

Moreover, reprogramming has proved highly inefficient, delivering just 1% or fewer of the cells back to the embryonic state. Many of the cells end up in varying degrees of partial reprogramming.

Lanza said he believes the problem his team discovered has to do with the outside genes being inserted into the cells.

"You can't just cut into the DNA all over the place and not have consequences," he said.

In this case, the consequences not only render the cells unusable for patients, but also leave them so abnormal that they cannot be used to gain an accurate picture of how diseases progress, Lanza said. Since 2007, Thomson and other scientists have stressed that long before the reprogrammed cells can be safely transplanted into patients, they should provide models for various diseases and targets for discovery of new drugs. However, if they do not accurately reflect living cells in the body, "There's no way we can use these cells, even for modeling diseases," Lanza said.

Beverly Torok-Storb, a stem cell researcher in Seattle, said it may turn out that changing cells directly from one type to another, for example skin to blood, may be more practical than taking a detour by first sending cells all the way back to the embryonic state.

Kamp further cautioned that the problems revealed by the new paper may occur only when the reprogrammed cells grow in the artificial environment of a lab dish rather than the natural environment of a body. He suggested that a more rigorous test would be to insert the cells into living animals to see how they function.

The new study, Kamp said, "suggests we don't really understand iPS cells very well yet."