Although ES cells can differentiate into neural precursors in the absence of any external influence, only a small percentage of neural cells survive and differentiate under these conditions [29
]. In this report we have therefore developed a modified protocol to enrich our culture in ES-derived neural progenitors. Using this approach we obtained multi-cellular aggregates that share many properties with neurospheres generated from CNS-derived neural stem cells, namely, sphere morphology, nestin expression and the ability to produce neurons. Moreover, ES-NS express p75, a marker found in neural crest-derived stem cells [31
]. Having established a neuronally biased ES population, we next examined the effects of an enteric neuromuscular environment on its further differentiation. It has been shown that co-culture of progenitor cells with mature cells or tissues can drive their differentiation toward desired phenotypes [32
]. We therefore established a controlled in vitro
system where ES-NS are co-cultured with strips of small intestine longitudinal muscle and the adherent myenteric plexus (LM-MP). Although the presence of the myenteric plexus in the LM-MP preparation might suppress ES-NS differentiation, our results show that intestinal LM-MP exert a considerable neurogenic effect on ES-NS, as evidenced by the enhanced expression of several neuronal markers, including peripherin, the intermediate filament which is a well established marker of peripheral neurons [34
] and the enzyme neuronal nitric oxide synthase (nNOS), responsible for the production of the critical enteric neurotransmitter, nitric oxide [35
]. On the other hand, ChAT was not expressed in ES-NS under either co-culture or control conditions. Although our data does not provide an explanation for the lack of ChAT expression, it is interesting to note that nNOS-expressing neurons appear early during ENS development while ChAT-expressing neurons are only detected at later embryonic stages and post-natally [36
The neurogenic effect observed after co-culture of ES-NS with LM-MP could be attributed to either increased proliferation and/or improved differentiation of neural progenitors within ES-NS. Our data show that the number of neurons (PGP9.5 and nNOS positive) is increased after co-culture with LM-MP, but not the number of proliferating cells (Ki67 positive). Moreover, co-culture with LM-MP clearly affects the morphology of neurons produced, which possess longer axons and form a more extensive and organized network as compared to controls. This suggests that soluble factors released from the LM-MP promote survival and differentiation of neural progenitors but not their proliferation. Therefore the main effect in our ES-NS co-culturing protocol is to favor differentiation of neurons.
In addition to inducing neuronal differentiation, we also observed a minor myogenic effect of LM-MP on ES-NS. Previous reports have shown that CNS-derived NS can be induced to express smooth muscle markers in vitro
], on the other hand, neural crest cells (NCC), which give rise to enteric neurons and glia during development, are also able to generate smooth muscle cells in different embryonic compartments [40
]. Therefore our findings could be due to an in vitro
artifact or to a NCC-like potential of ES-NS.
ES-NS failed to express the glial marker GFAP under any of our culture conditions, unlike other studies using ES-derived neural precursors, which have reported the capability of these progenitors to differentiate into both the neuronal and glial lineage in vitro
]. The discrepancy between these reports and our results may be due to differences in culture conditions used to produce ES-NS. Alternatively, the fact that neural progenitors have a stage-dependent potential giving rise first to neurons and then to glia during neural commitment [41
], may suggest that prolonged culture periods are necessary to generate glial cells. Therefore differentiation of ES-NS for longer periods of time (more than 10
days) and analysis of a more widespread panel of glial markers will need to be performed.
It is well known that the environment plays a key role on the differentiation of ES cells into defined phenotypes. We have developed a robust and informative in vitro system where soluble factors released by gut tissues are able to drive neuronal differentiation of ES-derivatives appropriate for the ENS lineage. The nature of these signals has yet to be determined, but may include well-known soluble factors of the gut such as GDNF, NO, TGFβ (respectively produced by glial, neuronal and muscle cells) or other molecules. Further analysis of the LM-MP conditioned medium will help to identify these factors.
Although transplantation of ES-NS in the pyloric muscle layer of wild type mice showed adequate survival of grafted cells at 1
week, no markers of terminal differentiation were noted. This contrasts with many reports that showed successful transplantation of ES-derived neural precursors in the CNS [13
]. This may be due to the presence of inhibitory factors, either secreted or contact-based that were not captured in vitro
using isolated LM-MP co-cultures. This difference also probably reflects the vastly more complex environment of the gut wall, consisting of a mixture of epithelium, neurons, glia, interstitial cells of Cajal and smooth muscle, along with resident and circulation-derived immunocytes. These cellular elements may provide conflicting cues to ES cells, resulting in inhibition or retardation of their differentiation. Further, non-neural factors such as neo-angiogenesis may be required for the optimal differentiation of these cells.
We have previously shown that CNS-derived NS grafted in the pylorus of mice generate neurons and glia and partially restore gastric function [7
]. We here show that ES-derived neurospheres do not compare to CNS-NS after transplantation. Clearly, there is a lot more to be done in order to fully understand the mechanisms for this different behavior. It is possible that the lack of GFAP expression (and glia?) accounts for the poor outcome in vivo
. Further work will need to be done to address this question.
In this regard, our study had several limitations. First, it should be noted that the time frame of these experiments was relatively short. Long-term survival and differentiation of grafted cells needs to be investigated in further studies to fully assess the potential of ES-NS to form enteric neurons in vivo. It is also possible that “priming” of the ES, perhaps by exposure to LM-MP prior to the transplantation may result in a better outcome. Finally, the inability to precisely localize the injection into the muscle layer may be a factor in preventing the appropriate signals from reaching the ES cells (in our study most of them were found in the submucosa). Therefore, this study cannot confidently exclude the therapeutic potential of ES cell for repopulating the ENS. On the other hand, if they truly do not differentiate, they may retain the ability to proliferate (which we did not examine), which is of concern because of the risk of tumorigenesis.