A series of headline-grabbing results in the late 1990s and early this decade suggested that specific adult stem cells such as those from the blood seemed able to exceed expectations and transform themselves into other organs and tissues. When engineered to glow green and injected into mice, these studies revealed glowing cells in unexpected organs such as the brain, heart and liver. The discoverers dubbed the process transdifferentiation.

Within a few years, however, laboratories that tried to replicate the findings either failed or came up with simple explanations for them such as fusion between the injected cells and those in the identified organs. A September 2002 follow-up study found no evidence for widespread transdifferentiation of blood-forming stem cells in brain, liver, kidney, gut and other tissues. Transdifferentiation may occur, but if so it must be a rare event, says stem cell and cardiovascular researcher Kenneth Chien of the Harvard Stem Cell Institute.

Scientists were even more stunned in July 2002 when researchers led by stem cell biologist Catherine Verfaillie at the University of Minnesota reported that bone marrow–derived cells they had injected into young embryos contributed to all three embryonic layers, just as embryonic stem cells would do.

Verfaillie and colleagues were trying to grow stem cells extracted from rodent mesenchyme, a component of bone marrow that contributes to fat, skeleton and muscle. Bone marrow cells do not normally grow in the lab, but the team reported that by carefully controlling the amount of oxygen the cells received along with other growth conditions, they could keep the cells alive for at least a year. In the process, the cells seemed to turn into something else, the experiments indicated. Rather than showing up only in the mesoderm, or middle layer of the embryo where future mesenchymal cells reside, the cells had spread to the two adjacent layers, which form other tissues such as skin, brain and gut, according to the group's report in Nature.

The researchers speculated that their cells, which they called multipotent adult precursor cells (MAPCs), might have regressed to a more primitive state in culture or were left over from embryonic development. "That started a lot of excitement that there might be these adult stem cells that had the potential of embryonic stem cells,'' says stem cell researcher Amy Wagers of the Joslin Diabetes Center in Boston.

Hoping to confirm the result by replicating it, other laboratories tried unsuccessfully to grow the cells. Wagers, then a postdoctoral fellow in the Stanford University lab of stem cell biologist Irving Weissman, spent a week in Verfaillie's lab in 2002 trying to learn to culture MAPCs. "They're very fussy," she says. Even Verfaillie, who announced she would teach a course on growing the cells in lab dishes, had problems culturing them at times, according to Weissman.

Difficulties growing stem cells are not unheard of. "The culture conditions needed to grow various kinds of stem cells can be very tricky," down to the source of the water, says neural stem cell biologist David Anderson of the California Institute of Technology in Pasadena. But the longer a study goes unreplicated, despite researchers' best efforts, the less sound a result seems.

Against that backdrop, the London-based magazine New Scientist reported in February that data presented in Verfaillie's July Nature paper appeared under a different label but was otherwise identical in a second paper published in Experimental Hematology in August of 2002. The data referred to cell surface proteins that supposedly defined the cells. That month, an expert panel convened by the University of Minnesota deemed the data "significantly flawed" and "potentially incorrect" for further inconsistencies in the data, but did not accuse Verfaillie or anyone in her lab of deliberately manipulating it.

The flawed data did not contradict the paper's main claim that the injected cells have a potential similar to that of embryonic stem cells, researchers say, because they could simply have had other cell surface proteins not included in the data. [see endnote]

Comments