When it comes to the issue of transplanting cells for the restoration of the neuronal network in the gut, the central question is: what is the source of the cells to be used? There are legions of approaches for all kinds of diseases using various types of stem cells, and the right choice is not clear. Some of these sources, such as embryonic or haematopoietic stem cells not only have great potential but also may have severe drawbacks. Thus, there are several theoretical advantages of embryonic stem cell lines for ENS restoration. Firstly, they can be maintained and expanded in culture without losing their ‘stemness’. Secondly, they can potentially give rise to many more cell types (e.g. interstitial cells of Cajal) than just neurons or glia, an attribute that could be particularly useful in certain diseases states where multiple lineages are affected. In an appropriate environment, embryonic stem-derived neural precursors have been shown to generate successfully both central and peripheral neurons, glia, enteric neurons and other neural crest derivatives.16–19 However, the use of embryonic stem cells is ethically restricted, can produce teratoma-like growths,20,21 and further, as with cells from other lineages (bone marrow, skin, adipose tissue, etc.), may require additional and perhaps intense reprogramming by as yet poorly defined protocols to produce an enteric neuronal phenotype. A more feasible and perhaps suitable type of cell therefore may be one that is already programmed for a neuronal fate, i.e. a NSC. These can be derived from the CNS, neural crest or postmigratory enteric neural progenitor population. However, before discussing the merits of each of these sources, it is important to first clarify that most of the literature consists of studies on neurospheres and not a pure stem cell population.
Neurospheres vs stem cells When putative NSC are isolated in culture from their source organs, they characteristically grow and proliferate in floating spheroid colonies called neurospheres. Several investigators have successfully isolated neurospheres from both the rodent and human gut which appear to be similar to their CNS-derived counterparts.14,22,23 As in the CNS, neurospheres continue to form the starting material for most studies published to date in the ENS field. However, although this feature has been used as a surrogate marker for ‘stemness’,24 the neurosphere is actually a very heterogeneous entity, consisting of progenitor cells at various levels of commitment as well as neurons, glia and other terminally differentiated cell types. Only 3–4% of the cells within neurospheres are actually true stem cells in that they are self-renewing and can give rise to all three neural lineages.25 Although it is clearly desirable to start with a relatively homogenous population of NSC, this goal has been difficult to attain because of the lack of a suitable marker. Amongst the studies shown in Table 1, only two have started with a relatively homogenous population based on cell sorting by expression of either the Ret receptor9 or the neurotrophin receptor p75,26 and only one of these performed clonal analysis to prove the true stem cell nature of these cells.26 It is not clear whether these receptors represent universal markers for ENS-NSC that can be applied to different species or at different stages of development, as most investigators appear to have limited themselves to using neurospheres. It may therefore be necessary to identify additional and perhaps more specific markers in the self-renewing stem cell population that will allow us to rigorously prove their stemness and consistently and reproducibly isolate them from the gut.
In the CNS, the NSC phenotype has been distinguished by the expression of nestin,27,28 and the low to absent expression of both peanut agglutinin and heat stable antigen (nestin+, PNAlo,HASlo).29 Nestin is an intermediate filament protein whose expression is widely used to identify mammalian neuronal precursor cells or stem cells. The stem cell is initially positive for nestin, but secondary progenitor cells lose this property. However, it is not completely specific: in the adult brain, nestin is expressed not only in NSC in the subependymal zone but also in reactive astrocytes, and in other organs, nestin also stains endothelial or glandular cells.30 A further problem with nestin is that it is not expressed on the surface, making it difficult to isolate these cells based on cell sorting techniques. Despite these limitations, nestin positivity, along with neurosphere generation, has become almost synonymous with stemness in the hands of most investigators. In rodents, neuronal precursors have also been isolated from the embryonic and postnatal guts using antibodies to specific markers known to be expressed by enteric neural crest-derived cells: Ret28 and p75 (the low-affinity receptor for nerve growth factor).26,31 However, it is not known that an approach using these techniques to isolate a more uniform population of precursor cells actually leads to better engraftment and functional restoration as compared with neurospheres alone.
There is therefore obviously a need for investigators to go ‘beyond the neurosphere’ and put greater efforts towards the identification of markers with both high specificity and selectivity, that will allow to harvest stem cells in larger quantities, using minimally invasive techniques, and without other ‘contaminating’ cell types. Such strategies will allow the delivery of a pure source of neuronal tissue, eliminating several confounding variables affecting the post-transplantation outcome.
Heterologous vs autologous sources Although neurospheres can be obtained relatively easily from a variety of sources, there are both ethical and immunological problems associated with their origin. Heterologous transplantation of stem cells into the ENS works relatively well in animal models without immunosuppression,32 but it is not clear whether this will also be true for long-term survival and functional benefit in clinical situations. Further, ethical issues, while surmountable, will continue to present challenges for the use of stem cells from tissues obtained from dead or aborted donors. Clearly, therefore, the best source to be used will be cells isolated from the patient itself, preferably from the same organ as the intended target. This may be feasible even in disorders such as Hirschsprung’s disease, in which the failure to develop ganglia may be caused by defects either in the NCSC or in the environment they need to inhabit [e.g. with endothelin receptor or glial cell line-derived neurotrophic factor (GDNF) mutations]. It is presumed that in the latter group of patients, the autologous source of stem cells will be the ganglionic segment but it remains to be seen whether these cells are actually effective in repairing the ENS. In this regard, one of the authors (Karl-Herbert Schäfer) has been successful in isolating neurospheres from the ganglionic or the transitional zone of Hirschsprung’s disease patients (Fig. 2, Karl-Herbert Schäfer, unpublished data).
Figure 2. Neurospheres derived from either the ganglionic (A) or transitional (B) areas of children undergoing surgery for Hirschsprung’s disease showing that the transitional zone gives rise to smaller spheres, which is consistent with the finding of Bondurand et al.22 in an animal model. The neurospheres were generated by isolating myenteric plexus or single ganglia from both areas using a technique based on enzymatical (collagenase) digestion and mechanical agitation.60 Briefly, the resected areas were stored on ice and processed within hours from surgery. Muscle and submucous layer were separated and the muscle tissue from the most proximal, ganglionic, as well as from the transitional zone was incubated in a collagenase solution (1 mg mL−1) for up to 5 h. After vortexing, muscle cells and myenteric plexus could be identified in both samples. The plexus tissue was dissociated and plated in a stem cell medium as previously reported14 (Karl-Herbert Schäfer, unpublished data).
Download figure to PowerPoint
Source tissues Central nervous system-derived NSC remain the most well characterized and studied of all NSC and appear to be closest to clinical reality, at least for CNS diseases. Interest in the use of fetal NSC as a therapeutic tool has also been fuelled in part by a report of the isolation of a human fetal NSC that can be expanded in vitro for years and that can be readily differentiated into neurons and glia, thus holding the promise of a renewable, plentiful and standardized source of human neural cells;33,34 recent studies show that such an approach is useful for adult subventricular zone cells as well.35 Indeed, CNS-NSC provided the earliest in vivo proof-of-principle for successful functional engraftment in the gut.7 However, long-term survival is still an issue as more than 90% of the grafted neurons usually die upon grafting, both in animal and in human studies.36,37 A large portion of this cell death occurs as programmed cell death, or apoptosis, and occurs within the first week after transplantation.38,39 Therefore, CNS-NSC treatment for gastrointestinal neuromuscular disorders will require strategies to circumvent or attenuate this phenomenon.
Although NCSC and differentiated CNS cells share a common progenitor,40,41 NCSC are potentially more attractive because in nature, they give rise to both the peripheral nervous system and the ENS, in addition to smooth muscle cells, pigment cells, bone and cartilage in other regions of the body.42,43 However, it can logically be argued that the most appropriate cell type for ENS therapy is the postmigratory ENP. These cells are downstream of the NCSC and appear to be more committed than other neural crest derivatives (such as sciatic nerve stem cells) in terms of their commitment to a neuronal fate.44 Recent developments have added to the enthusiasm for this approach. These include the discovery of endogenous NSC within the immature as well as adult ENS.5,26,44–46 NSC isolated from the small intestine of lactating and adult mice express nestin, vimentin, and the pro-neural transcription factors neurogenin-2 (ngn-2), Sox-10 and Mash-1.45 These cells can differentiate into various cell types, particularly neurons, smooth muscle, and glia with the neurons expressing several characteristic neurotransmitters and receptors including calcitonin gene-related peptide, neuropeptide Y, peptide YY, substance P, vasoactive intestinal polypeptide, galanin and c-KIT. However, in keeping with the previous discussion, it should be pointed out that these experiments used unfractionated primary cell cultures of whole gut and therefore it is difficult to ascribe any of these attributes to those emanating from pure NSC.
It is not known whether CNS-NSC can also express such a profile but these results do suggest that ENP are more likely to respond to gut-specific environmental cues. Indeed, isolation and expansion of precursor cells from the developing and postnatal human ENS have recently been reported using bowel samples from human fetuses and children (from the ninth week of gestation to 5 years postnatal). Such cells can be differentiated and also be transplanted after dissociation into aganglionic bowel in vitro.11–13 Another advantage of ENP, particularly for autologous approaches, lies in the accessibility of the gut by minimally invasive means (unlike the CNS). This has been highlighted by the recent dramatic discovery of potential neuronal progenitors in the mucosa and submucosal region which can be accessed by performing a simple mucosal biopsy. These cells can give rise to neurospheres in vitro which can be differentiated into neurons.47 Although it has still to be proven whether this method can generate all the neurons required to restore fully gastrointestinal function, this is a very promising and exciting development for the field.
Another attractive source for ENP is the appendix which harbours a fully developed ENS48 and can easily be removed by minimal invasive surgery. Thus, enteric nervous tissue can be isolated from surgically removed appendices and neurospheres generated and neuronal and glial cells cultivated (Karl-Herbert Schäfer, unpublished data).
Finally, endoscopic techniques are being developed that can provide access to the muscular layer and associated ganglia in a relatively non-invasive manner.49,50 It is quite possible that these layers may provide an alternative source for stem cells in the future.
Although these locations are very promising autologous sources for enteric NSC, further studies and more effort have to be directed at characterizing the amount and quality of neurospheres that can be obtained from individual biopsies or appendices.