Everyone gives Science, Cell and Nature way too much credit. They are flashy, but tend to have less substance than other journals. Their retraction rate is also much higher.
I'm not even going to go into why the correlation between influence and retraction rate exists, that should be blatantly obvious to any with a background in any of the sciences. Also, we want retractions, it means science works.
What I find surprising about this is that the work was apparently conducted by separate groups at Riken, Japan, and Harvard, USA. It would take a fair degree of negligence/incompetence/subterfuge for a basic error such as that proposed to occur (i.e. there were no stem cells produced by this method).
Apparently the paper was in review for 9 months, which suggests at least two back-and-forth discussions between reviewers and authors. If there was that degree of misbehavior suggested, I would probably give the editors and reviewers a pass. Their job is not to detect outright fraud.