Michael K. Richardson
Leiden University, the Netherlands
A developmental biologist highlights potential pitfalls of using stem cells that can 'remember' their origins.
For me, embryos are beautiful and their development is endlessly fascinating. They are experts at making new tissues, and accomplish this by using stem cells. Stem cells can develop into mature tissues such as bone or muscle; but, cleverly, some of their progeny remain in an undeveloped state, forming reserve supplies that remain in our bodies into adulthood.
Adult stem cells are found in tissues where cell populations are constantly being renewed, such as the testes, hair follicles and bones. We replace our entire skeleton every decade or so, and rely on stem cells in our bones to do this. Stem cells also have an important role in repair, swinging into action to deal with broken bones and other mishaps.
A recent study in mice yielded remarkable evidence that some of these adult stem cells remember where in the embryo they came from. Jill Helms and her colleagues at Stanford University in California grafted stem cells from one bone into another to see whether they would help repair fractures in the 'wrong' location. Stem cells transplanted from leg bones into fractured jaws failed to produce new bone (P. Leucht et al. Development 135, 2845–2854; 2008). Interestingly, the uncooperative stem cells continued to express a gene, Hoxa11, that acts as a kind of embryonic 'postcode' for the leg.
These findings have broad implications. They validate the concept of non-equivalence — that seemingly identical cells differ if they come from different places in the embryo — first enunciated by Julian Lewis and Lewis Wolpert in the 1970s, and show that it holds in the adult. More pragmatically, if some stem cells also have positional memory, doctors may need to make sure that they take stem cells from the right location to heal damaged tissues.
Comments
There are, however, equivalence groups among progenitor populations. Transplanting cells derived from the calvarium can and does repair/reconstruct the jaw (both neural crest-derived and populations originally expressing few if any Hox gene members); from the tibia, one might in theory expect the cells to be able to repair the femur, but also perhaps the more anterior, mesodermally-derived bone in the clavicle.
However, ribs and iliac bone are regularly used for mandibular reconstruction. And apparently, vascularized grafts are becoming all the rage because of a more successful outcome. This is not simply cartilage in the place of the jaw (eg. the femur used in Benlidayi 2009 PMID: 19446201).
The other observation I would like to make is that Hox-expressing trunk-level neural crest cells, after a while in culture, will discontinue Hox expression. While normally unable to make cartilage and bone, they can then do so, like their cranial counterparts (Abzhanov et al. 2003 PMID: 12925584). Leucht et al. were rather dismissive of both the paper and concept, which is surprising, given that Abzhanov has published with Helms in the past. And Couly et al. 1998 (PMID: 9693148) had shown a decade earlier that neural crest cells within a Hox+ or Hox- equivalence group can form appropriate skeletal structures, and outside of that equivalence group, keep their Hox expression status when grafted heterotopically. However, one could imagine that it is possible to apply exogenous influences not available in the normal tissue environment to alter Hox (or other master transcriptional regulator) gene expression status. And thereby, cell potential and eventual fate.
Posted by: Heather Etchevers | May 28, 2009 10:38 PM