The creation of human–animal hybrid embryos — proposed as a way to generate embryonic stem cells without relying on scarce human eggs — has met with legislative hurdles and public outcry. But a paper published this week suggests that the approach has another, more fundamental problem: it may simply not work.

Robert Lanza of Advanced Cell Technology, a stem-cell company based in Los Angeles, California, and his colleagues show that in their labs, early-stage human–cow, human–mouse and human–rabbit hybrid embryos fail to grow beyond 16 cells (Y. Chung et al. Cloning Stem Cells doi:10.1089/clo.2009.0004; 2009). The hybrid embryos also failed to properly express genes thought to be critical for pluripotency — the ability to develop into a wide variety of cell types.

Lanza and his co-workers created their hybrid embryos using a process called somatic-cell nuclear transfer, a technique made famous when it was used to create Dolly the cloned sheep in 1996. This time, the researchers replaced the nuclei of human, cow, mouse and rabbit eggs with nuclei from human non-sex, or somatic, cells.

Human–human embryos developed normally and increased their expression of many genes, including several known to be involved in pluripotency. Hybrid embryos, however, were short-lived, and failed to express known pluripotency genes properly. Lanza says that his team has ploughed through many different protocols and "thousands" of embryos over the years, without success. "At first we thought it would just be a matter of tweaking the culture conditions," says Lanza. But "the problem was far more fundamental".

It's too early to write off interspecies hybrids.

Others in the field argue that all is not lost. "Understanding what has to be done to overcome these problems would help us understand what reprogramming is all about," says reproductive biologist Justin St John of the University of Warwick in Coventry, UK, who is developing mouse–pig hybrid embryos. The paper outlines only one set of conditions used to create the embryos, St John adds, meaning that it is impossible to assess all of the options that Lanza's team tried.

Advanced Cell Technology previously cloned an endangered bull of the gaur species, using eggs from the common cow to create hybrid gaur–cow embryos. However, these two species are closely related. It may be possible to create hybrid embryos using human somatic cells and eggs from non-human primates, but primate eggs are also in short supply, says Lanza. Although Hui Zhen Sheng from the Shanghai Second Medical University in China and her colleagues have reported creating human–rabbit embryos (Y. Chen et al. Cell Res. 13, 251–263; 2003), several labs have been unable to replicate the findings, according to Lanza and others in the field.

There are many processes that might cause hybrid embryos to fail. Embryonic development is initially guided by proteins and RNA found in the egg, with control eventually passing to DNA in the nucleus. This transfer of power occurs in humans when the embryo has reached four to eight cells; but in mice it happens at the two-cell stage, and this mismatch may disrupt development.

Furthermore, the nuclear genome may have difficulty communicating with energy-producing structures called mitochondria — which are inherited directly from the mother, through the egg — from distantly related species. If researchers can find the reason why some hybrid embryos stop developing, they might be able to circumvent those roadblocks by altering the expression of specific genes, says cell biologist Jose Cibelli of Michigan State University in East Lansing.

Meanwhile, there may be other ways to reprogram a cell with a different species' DNA, notes embryologist Anthony Perry at the RIKEN Center for Developmental Biology in Kobe, Japan. "Is there really only one path that will give you a pup? Surely the answer is no," he says. "It's too early to write off interspecies hybrids."