The complex functions of the gastrointestinal tract, which include food propagation and mixing, supply of digestive enzymes, resorption and secretion, as well as appropriate blood flow rely upon an intensive coordination by intrinsic autonomous neural networks (Surprenant,1994; Furness et al.,1995; Knutson et al.,1995; Cooke,2000). These networks are embedded in the gut wall and are based on the interconnection of ganglionic and aganglionic plexus-like structures and also upon highly sophisticated neuronal wiring and polysynaptic circuits (Furness et al.,2004). The ENS forms by far the major part of the autonomic nervous system and, given to its inner structure, organization and chemical coding, is much more similar to the central nervous system than other parts of the peripheral nervous system (PNS); therefore, it is very often referred to as “the second brain” or “brain within the gut” (Gershon,1999; Wood,2008). The number of neurons in the ENS equals the number found in the spinal cord, some hundred millions (Furness and Costa,1987). A second cell type to be found in the ENS is the enteric glial cell. It is a distinct group of glial cells which is different from the Schwann cell in the PNS (Gershon and Rothman,1991) and which contains, at least partially and depending on its functional situation or involvement in an intestinal inflammatory response varying amounts of the astrocytic protein GFAP (Jessen and Mirsky,1985; von Boyen et al.,2002b,2004,2006a,b; Savidge et al.,2007a,b). Roughly, the ENS is organized in two major ganglionated networks, the myenteric and the submucous plexus, and also in several aganglionated plexus within the mucous and muscular layer, and underneath the serosal layer. The number, morphology, and neurochemical characteristics of these networks differ depending on localization along the gastrointestinal axis, as well as from species to species (Timmermans et al.,1997). Although there is a relatively low amount of independent networks in mice and rat, the number of individual distinct submucous plexus can vary. In the human gastrointestinal tract, we find three different ganglionated submucous plexus (Timmermans et al.,1992,1997,2001; Wedel et al.,1999). Obviously, the number and complexity of different networks in different species does not interfere with the capacity of the gut to deliver a sufficient nutritional support. Whether these species-dependent differences are based on different food intakes is a matter of speculation.
During development, where neural crest derived stem cells have to migrate all along the gut axis, but also postnatally, the ENS has to adapt to new challenges, facing microenvironmental and nutritional changes such as inflammation or loss of resorption capacity. The aim of this review is to highlight how the ENS reacts upon these challenges and to raise questions whether and how the ENS plasticity correlates with its intrinsic neural stem cell potential.
PLASTICITY OF THE ENS
Plasticity is the ability of the nervous system to rewire its connections, or to adapt its functions to the actual microenvironmental situation. In the central nervous system, the ability of the brain to adapt to changing physiological conditions or traumatic lesions is crucial, due to a lack of a sufficient regeneration potential. Growth-associated protein-43 is a protein, which is correlated to neuronal growth and regeneration is typically expressed at high levels in the nervous system during development. In adult animals, its expression is lower, but still observable in brain areas showing structural or functional plasticity, such as the hippocampus (Simmons et al.,2008). In the autonomic nervous system, and especially in the ENS, GAP-43 is strongly expressed in the myenteric and submucous ganglia at all ages (Stewart et al.,1992), thus giving evidence for a lifelong capability of the ENS to adapt to new challenges (Giaroni et al.,1999).
The ENS undergoes significant changes in perinatal development. During the first two postnatal weeks, the rat myenteric plexus changes its morphology and its neuronal number depending on time and localization. Neuronal number per ganglionic area increases from duodenum to colon, while their density and total number decrease with the age of the animal (El-Salhy et al.,1999; Schäfer et al.,1999; Saffrey,2004; Wade and Cowen,2004). Neuronal densities decrease further with increasing age (Gabella,1989,2001). This age related decline is correlated with changes in motility, mucosal function and changes in specific neuronal phenotypes, such as the intrinsic sensory neurones, which appear to be the most age-sensitive neurones in the ENS. Additionally, the expression of nitric oxide synthase and the density of peptidergic (Substance P, Vasointestinal peptide, Somatostatin) fibers is reduced in aged rats (Feher and Penzes,1987; Takahashi et al.,2000). In porcine small intestine, similar observations have been made, so the density of nitrergic neurons declines from the late prenatal period onward (Van Ginneken et al.,2001).
However, plasticity is not limited to changes during development or aging. The high level of GAP-43 expression leads to the conclusion, that there must be an enormous regeneration potential, which allows the ENS to react upon microenvironmental changes, such as varying microflora, changing dietary habits, mechanical stenosis, or diseases. Nevertheless, these GAP-43 levels do not increase further during inflammation, while other peptides or neurotrophic factors do change in the amount of expression, that is, peptidergic innervation, which increases in colitis (Vento et al.,2001).
Enteric disturbances such as inflammatory bowel diseases in general can lead to a remodeling of the enteric nervous system and to cytokine-induced changes in neurotrophin and -transmitter content and release (Collins et al.,1992; Stead,1992; Sharkey et al.,1999; Reinshagen et al.,2002; von Boyen et al.,2002b,2004,2006a,b; De Giorgio et al.,2004; Vasina et al.,2006; Di Nardo et al.,2008) and to a remodeling of synaptic connections (Krauter et al.,2007). There are also responses to changes in dietary habits. In the stomach myenteric neurones respond immediately to enteral feeding by expression of immediate early genes (Dimaline et al.,1995; Sharkey et al.,1999) and more small intestinal myenteric neurons express nNOS upon the installation of enteral feeding (Oste et al.,2005). In addition to enteral food-evoked mechanisms such as postprandial tissue hypoxia or peristalsis (Pawlik et al.,1995; Bredt,1999), influences of single or complex nutritional components on the enteric nervous system need to be taken into account as well. Myenteric neurones in vitro, which were exposed to butyrate, changed their membrane potential, ionic currents, and intracellular Ca2+ concentration (Haschke et al.,2002), which clearly demonstrates the immediate interaction of nutritional components and the ENS. Especially during the early postnatal period, where the development of the gut and also the ENS is still ongoing, appropriate nutrition will be crucial for the correct formation of an “adult” gut. During the immediate postnatal period, macromolecular transport across the intestinal epithelium is possible, facilitating contact between nutritional components and the ENS. In this context, we analyzed the content of glial cell line derived neurotrophic factor (GDNF) of various breast milk samples and their impact upon enteric neurons and glial cells in vitro. We could show that breast milk samples from individual time points and mothers show varying GDNF concentrations. “Feeding” myenteric plexus cultures with breast milk preparations led to “healthier” neuronal networks with more surviving neurons and higher neuron densities (Fig. 1). However, in vivo experiments could not show an increased neuron density when colostral milk was fed to piglets when compared with milk replacer. In contrast, feeding a milk replacer instead of breast milk gave rise to a higher myenteric VIP and GFAP content in the inner submucous plexus of the porcine small intestine (van Haver et al.,2008). Probably this plasticity of the postnatal enteric nervous system can be explained by indirect effects of nutritional compounds via cytokines (Neunlist et al.,2003; Tjwa et al.,2003; von Boyen et al.,2004; Van Haver et al.,2008), toxins produced by the bacterial flora (Neunlist et al.,2003; Arciszewski et al.,2008; Van Haver et al.,2008), or the interaction with the epithelial layer (Moriez et al.,2009) and has to be studied in detail.
As is the case via nutrition, besides neurones, other cell types of the ENS can be affected in enteric diseases. In Crohn's disease, enteric glia in the submucous plexus undergo transdifferentiation characterized by an increase in GFAP expression (Cabarrocas et al.,2003; von Boyen et al.,2006a; Neunlist et al.,2007; Savidge et al.,2007a,b) as well as in the expression of GDNF, a key trophic factor in ENS development (Moore et al.,1996). Similar findings have been reported from tapeworm infected rat small intestine, where GDNF was found in macrophages (Starke-Buzetti et al., 2008).
Although the increase of GFAP correlates with GDNF increases in the inflamed gut, it might be interesting to have a closer look at reactive astrocytes in the central nervous system, which display increased GFAP expression correlated with increased expressions of the neuroepithelial stem cell marker nestin (Clarke et al.,1994; Lin et al.,1995). Nestin, which is a class IV intermediate filament, can also be found in the pre- and postnatal enteric nervous system (Rauch et al.,2006a,b; Silva et al.,2008), and nestin levels do obviously correspond to inflammatory changes in the gut. A higher rate of nestin expression could be found in enteric glia when exposed to the toxic agent ethanol (in vitro investigation), and also in the inflamed human appendix (personal observations). These observations are consistent with the findings of Coskun et al. (2002), who demonstrated an increase in the number of enteric glia and an increase in nerve growth factor expression in acute appendicitis. So where do these new cells come from? Because of the fact that there is an abundant nestin expression in the adult gut, there might be a strong link between increasing GFAP levels and the recruitment of cells within the enteric nervous system, which do have at least some neural stem cell characteristics, such as the expression of nestin.
As described earlier (von Boyen et al.,2004), the source of these GFAP increases is to be found in the enteric glia. The amount of GFAP can vary, as seen in Crohn's disease patients, from patient to patient, but also within the gut, as seen in dilated and non dilated areas of colon from chagasic patients (Da Silveira, et al., 2009), and also in Hirschsprung's disease (HCSR).
We have investigated the ratio between the enteric glial marker S100 and the astrocyte marker GFAP in 13 patients with HCSR. The results demonstrate a significant difference between ganglionated and hypoganglionated regions in the affected gut. Within the transitional zone, where a reduced amount of enteric neuronal tissue can be found, the amount of GFAP increases relative to the S100 expression. Although in the ganglionated gut segments, the GFAP positive glia represent only a minor part of the whole ganglion, at the end of the transition zone the major part of the individual myenteric glial cells is GFAP positive. In the same samples, nestin expression was evaluated in the transitional zone. Here we could find a strong nestin expression in the myenteric plexus, while only a very reduced nestin expression could be seen within the submucous plexus (unpublished data, Fig. 2). Although it is not yet clear what exactly and how GFAP levels are changed, the mechanical stenosis presented in HRSC can obviously lead also to changes in the structure of myenteric ganglia. So an artificially induced stenosis can induce neuronal hypertrophy in a subset of myenteric neurones oral to the stenotic region (Brehmer et al.,2000; Brehmer et al.,2001). Motility defects caused by long-term obstruction are at least partially reversible. A gradual recovery of the enteric neurotransmission can be seen after the removal of the obstruction (Bertoni et al.,2006).
During the last years, another player in plasticity has come into the focus of ENS research: the bacterial microflora. At birth, bacterial colonization of the gut starts and depends on various factors. Because of stress, antibiotical treatments, or changes and variations in food intake, the microbiological flora in the gut can undergo substantial changes, which may affect the enteric nervous system and deliver a prerequisite for inflammatory bowel diseases (Wood,2007). The ENS can be activated by lipopolysaccharides released from luminal microflora via Toll-like-receptors (Rumio et al.,2006).
As such, the enteric nervous system plays an important role in the response to noxious influences on the gastrointestinal tract and obviously at the same time provides a source of regeneration and remodeling. The enteric glia is most probably the home to intrinsic neural stem cells, which can be activated and recruited whenever needed.
In contrast to the CNS, the neural stem cell source of the enteric nervous system is based upon the neural crest where neural stem cells derive from three distinct spatial sources: Neural crest cells derived from vagal and truncal regions colonize the whole embryonic gut via the foregut mesenchyme, while sacral derived neural crest cells exclusively contribute to the hindgut colonization (Heanue and Pachnis,2007). The migration routes of these two groups are complementary, although the majority of these cells derive from the vagal neural crest (Le Douarin and Teillet,1973). The colonization of the gut is completed after 7 weeks in human or at E 15 in mice (Fu et al.,2003; Druckenbrod and Epstein,2005), although the ENS is far from being completely developed (Rauch et al.,2006b). Vagal truncal derived neural crest cells migrate in a rostrocaudal direction, while sacral derived cells migrate the opposite way, after the rostrocaudal migration has been completed (Burns and Le Douarin,1999).
The migration of neural crest cells has recently been analyzed using time-lapsed based live-cell-imaging techniques, where neural crest cells were genetically labeled with GFP (green fluorescent protein). Here it could be demonstrated that the leading cells in the wave front of migration were organized in chains or strands with varying grades of progress (Young et al.,2004; Druckenbrod and Epstein,2005). Obviously, the microenvironment within the embryonic gut is changing with site and also with progress of development. Guiding molecules such as GDNF and EDN3 are expressed at varying points in time in the stomach, in the caecum, and in the distal hindgut (Young et al.,2001; Burns and Thapar,2006; Heanue and Pachnis,2007). This means that the “stemcellness,” the proliferation or differentiation capacity of the neural crest cells vary and change along the migrating route and movement patterns of migrating cells within or behind the wavefront are manifold (Druckenbrod and Epstein,2007). This influences the fate of the neural crest cells while migrating. We do have neural crest derived stem cells, with the capacity to give birth to the complete lineage of ENS neurons and glial cells from the beginning of migration onwards. These cells will be found especially in the wave front of the migrating neural crest derived cells, while subgroups of these cells do deliver committed progenitors for individual subpopulations of enteric neurons at each individual site in the gut. How this local differentiation takes place and how ganglia are organized are still matter of intense investigations (Faure et al.,2007). The coexistence of neural crest derived stem cells, committed neural progenitors, and fully differentiated neurons and glial cells needs a complex communication between the individual cells and their microenvironment to establish a functional network of enteric ganglia.
Neural crest cells from different sources do depend on different regulation. The truncal crest derived neural stem cells depend on Mash-1, which is an essential transcription factor for the autonomic nervous system. Mash-1 can be found in sympathetic precursors and in invading foregut neural crest derived stem cells. Mash-1 deficient mice do not develop enteric neurones within the esophagus, while the ENS is intact in the rest of the gastrointestinal tract. Most of the vagal derived neural crest cells are Phox2b and Ret-dependent. Phox2b is a homeodomain-containing transcription factor, and its absence leads to embryonic death in utero, although migration analysis performed during development could demonstrate NCSC's within the foregut, they do not migrate further. A similar phenotype can be seen in case of a deficient Ret tyrosine kinase receptor, or other components of the GDNF receptor complex. This receptor is a multicomponent receptor, which is formed by two Ret-molecules in combination with two GFRalpha 1-4 molecules; GDNF, neurturin, artemin, or persephin are the ligands for the individual GFRalpha components. Mutations in either Ret, GDNF-family members, or the glycosylphosphatidylinositol-anchored binding protein (GFRalpha 1-4) will lead to severe deficits in the innervation of the gastrointestinal tract. An additional signaling pathway which plays an important role in ENS development is the endothelin3 (EDN3)/endothelin3-receptor (EDNRB) pathway (Druckenbrod et al.,2008).
HIRSCHSPRUNG's DISEASE: AETIOLOGY AND FUTURE TREATMENT STRATEGIES BY NEURAL STEM CELLS FROM THE GUT
In humans, an undisturbed function of the enteric nervous system is needed to fulfil the needs of a well working gastro-intestinal tract. Because of the developmental deficits, a broad range of so called neurocristopathies, which often lead to severe dysganglionic disorders of the gut, can be seen. These might turn out as slight hypoganglionosis, aganglionosis of a small segment, or a merely complete lack of the ENS. In the latter, most severe cases, the children cannot survive unless depending on total parenteral nutrition. The most common situation is Hirschsprung's disease (HSCR) with a frequency of 1 in 5,000 living births (Brooks et al.,2005). HSCR is characterized by a lack of enteric innervation within the distal colonic region. The amount of aganglionosis can vary from ultrashort segments to most of the colon and even parts of the small intestine. So far several candidate genes could be identified which are involved in the aetiology of HSCR.
Besides the classical Hirschsprung's disease, there is a broad field of allied disorders, which are much rarer and often associated with other syndromes, such as Waardenburg or MEN2b syndrome. Total aganglionosis is the most severe situation and can be found in about 5% of the HSCR cases leading to the Zuelzer-Wilson-Syndrome (Amiel et al.,2008). Although there are similar clinical manifestations, the underlying mechanisms may differ and involve disturbances in colonization, survival, or differentiation of the migrating precursor cells (Milla,1999). In more than 40% of the HSCR cases, germline mutations in genes coding for the glial cell line derived neurotrophic factor GDNF and neurturin or its receptor, the Ret tyrosine kinase receptor (Fitze et al.,2003; de Pelet et al.,2006), the endothelin 3 receptor (Milla,1998) or others such as SOX10, Phox2B, Smad interacting protein 1, or PMX2B (Pingault et al.,1998; Pattyn et al.,1999) are involved. A combination of at least two different mechanisms can cause focal neuronal cell losses as seen in HSCR:END3 and Ret/GDNF (Barlow et al.,2003). Mutations affecting either EDN3 or EDNRB are associated with Waardenburg syndrome, characterized by colonic aganglionosis, pigmentation defects, and deafness (Brooks et al.,2005). The current treatment for Hirschsprung's disease involves surgical resection of the aganglionic colon and reanastomosis of the healthy colon with the anal sphincter. However, obstructive symptoms recur after surgical correction in up to more than 40% of the cases, mostly because of insufficient resection of the dysfunctional gut. This might be partially due to a defective innervation involving apparently “healthy” gut regions. In animal models for Hirschsprung's disease, the innervation of the ganglionated parts of the gut was also altered when compared with normal controls (Sandgren et al.,2002; von Boyen et al.,2002a). Recurrent symptoms in these patients range from mild constipation to severe episodes of abdominal distension and enterocolitis. In many cases, this means a second operation or other surgical and nonsurgical measures such as sphincter myectomy, botulinum toxin injections, or the massive use of laxatives or subsequent irrigations. In many cases, a long-lasting and painful dilation of the external sphincter is part of the postoperative treatment. Taking into account that a complete restoration of gut and sphincter functions is possible only in a subpopulation of HRSC patients, there is a need for new therapeutical approaches, either replacing the resecting surgical procedure or at least as a second step to reduce or improve the postoperative complications.
A new perspective lies in the use of neural stem cells to be transplanted into the affected regions of the gut (Natarajan et al.,1999; Sandgren et al.,2000; Schäfer,2000). Neural stem cells from the central nervous system could successfully be used for the restoration of gastric dysfunctions (Micci et al.,2005a). They could easily be isolated from E13 mouse embryonic brain, expanded and transplanted. Unfortunately, the long time survival, though improved by using apoptosis inhibitors (Micci et al.,2005b) was not satisfying. As already mentioned, neural crest derived stem cells migrate from several sources and critical numbers of neural crest derived stem cells are required to colonize the whole gut (Barlow et al.,2008). The isolation of these cells and their propagation in vitro could be a more convenient and also suitable alternative. The wave front of these migrating cells, the true neural stem cells from the gut, consists of only a few cells. In mice this group has on average only about 70 cells, which nevertheless do have the potency to colonize the whole colon. This could be shown by Emma Sidebotham and coworkers in an impressive way, using embryonic colon explants before colonization (Sidebotham et al.,2002). In practice, these cells would have to be isolated using an appropriate marker. P75, the low affinity neurotrophin receptor was already identified as a suitable candidate (Young et al.,1998), but interestingly, PGP 9.5, which is a common marker for enteric neurones, does stain for the same set of cells (Sidebotham et al.,2001). Appropriate markers for the identification of neural crest derived stem cells are an absolute prerequisite for the practical use of these cells for cell repair, especially because the cells would have to be isolated from the individual patient who is affected with the specific dysganglionic disease. Therefore, these markers would have to be identified also in the human gut. Rauch et al. could show that several neuronal precursor markers were present in the early formation of the enteric nervous system in the human fetus (Rauch et al.,2006b). They followed the expression of PGP 9.5, nestin, the neuroepithelial stem cell marker, as well as for peripherin, internexin, and tyrosine-hydroxylase (TH). With the exception of TH, all markers could be found in the early fetal gut, but only PGP 9.5, peripherin and internexin were present in the earliest samples. Interestingly, PGP 9.5 was coexpressed with nestin in the early stages of pregnancy, while this coexpression switched after the 14th week of gestation. At this point of time as well as later, PGP 9.5 stained only for neurons, whereas nestin clearly stained the glial fraction of the myenteric ganglia. Nestin expression could also be found in samples from adults up to elderly patients, which means that there is a persistent subgroup of cells with a certain neural stem cell potential. This was consistent with findings from other groups (Kruger et al.,2002; Suarez-Rodriguez and Belkind-Gerson,2004). These cells are the targets for the isolation from gut in postnatal patients. Postnatal neurospheres can be cultured from rodents and humans (Bondurand et al.,2003; Schäfer et al.,2003; Rauch et al.,2006a; Lindley et al.,2008) but show less proliferative activity. Depending on the amount of NCSC to be found in the postnatal gut from patients with dysganglionic disorders, a sufficient amount of enteric neurospheres could be harvested and used for transplantation. The enteric neurospheres have a great potential to generate dense neurite networks in vitro (Schäfer et al.,2003). Using extracellular matrix gels as a model (Schäfer and Mestres,2000) for transplantation in vivo, it could be demonstrated that single neurospheres cultured at sufficient distances from each other, express significant neurite outgrowth within 24 hr, which is not observed to a similar extent when neurospheres generated from the subventricular zone of the same animal are used (Fig. 3).
For the clinical approach, neural crest derived stem cells will have to be isolated from the patient at the given age, that is, there will be significant differences concerning the quality of the isolated neural stem cells, which is mirrored in the growth potential of early postnatal and adult enteric nervous tissue in human (Schäfer and Mestres,1997). The needs for the optimal growth of neural crest derived stem cells of different ages are still to be investigated, and will most probably differ significantly.
Looking at the changes in GFAP expression during inflammation, there is much evidence for a neural stem cell induction correlated with nestin upregulation as seen in the CNS (Lin et al.,1995), which might be used clinically by inducing higher stem cell rates before sampling.
A further key question in the cell therapeutical strategy is whether the microenvironment of the diseased gut does allow a survival of the transplanted cells for infinite periods. To address this question, homogenates from ganglionic and aganglionic HRSC smooth muscle were analyzed concerning GDNF content and also used in culture to support the growth of myenteric and dorsal root neurones. It could be demonstrated that all homogenates support neurite outgrowth, independently of a reduced GDNF content, which supports a general susceptibility of the aganglionic gut for cell transplantation (Schäfer et al.,2008). Dissociated neurospheres differentiated under the influence of the HRSC homogenates into morphologically distinct neurons (Fig. 4), which display electrophysiologically neuronal properties. The morphologies found in theses cultures could not be distinguished from those seen in cultures from postnatal dissociated myenteric plexus (Schäfer and Mestres,1999; Schäfer et al.,1999).
STRATEGIES TO USE ENS DERIVED CELLS FOR CNS REPAIR
Using enteric neural stem cells for the repair of neuronal disorders of the gut is a logic alternative, but could the gut also be used as an autologous stem cell source for the repair of central nervous system disorders? First attempts to use the ENS derived neurons for transplantation into the CNS were already performed in the eighties. Isolated myenteric plexus was transplanted into the striatum or the hippocampus to substitute cholinergic losses (Lawrence et al.,1991; Tew et al.,1996) or used to restore spinal cord or blood-brain-barrier function (Jaeger et al.,1993; Jiang et al.,2003a,b,2005). In the latter, adult myenteric ganglia were transplanted into injured spinal chord and did survive for at least a short time period. But only half of the transplants could be recovered, due to ongoing immune responses. Immunological consequences could be avoided by using NCSC's from the same animals, respectively, later on, in a clinical setting, from the same patient.
To use the ENS as an autologous source for central nervous system repair, it is absolutly necessary to investigate the stem cell potentials all over the life span. We, therefore, started looking at nestin expression in a specific GFP-nestin transgenic mouse model. These transgenic mice, which express green fluorescent protein (GFP) under the neural enhancer element of the nestin promotor, have been proven an appropriate tool to demonstrate the location of neural stem cells throughout the central nervous system (Yamaguchi et al.,2000; Mignone et al.,2004). During embryogenesis, nestin-GFP expression in the brain is confined to the two neural stem cell types in the ventrical zone: neuro-endothelial cells and radial glia cells; in the radial glia cells, it is colocalized with GFAP and other specific radial glia markers (Götz and Huttner,2005). In the adult brain of the transgenic mice, nestin-GFP expression is found in the neural stem cells in the two neurogenic areas of the adult brain, in the dentate gyrus of the hippocampus and also in the subventricular zone. In both cell types, descendents of embryonic radial glia (Ihrie and Avarez-Buylla,2008), nestin is coexpressed with GFAP; in addition, nestin-GFP appears to be expressed in PSA-NCAM-positive neuroblasts formed in the adult subventricular zone that migrate along the rostral migratory stream to the olfactory bulb, where they differentiate into periglomerular and granular interneurons. Despite their differentiation into mature interneurons these cells maintain GFP expression in these transgenic mice even after incorporation into the olfactory bulb circuitry. Moreover, nestin-GFP expression in the adult brain can be found in reactive astrocytes in response to various types of brain injury, and in oligodendrocyte precursor cells that dedifferentiate into neural stem cell like cells under demyelinating conditions.
In the GFP-nestin transgenic mice, we observed that nestin-GFP cells were located along the gastrointestinal tract, in the myenteric and submucous plexus. The amount of nestin-GFP varies according to location and also decreases with age. Nevertheless, there is an abundant nestin-GFP expression in the adult gut. Although the nestin expression is mainly found in glial cells, there is a small number of cells, which do not show a S-100 staining, but display a neuronal morphology within a myenteric ganglion.
Obviously, there are at least two different types of nestin-positive cells to be distinguished in the enteric nervous system (Fig. 5). A similar situation is to be found in the CNS, where also two types of nestin-positive cells can be found, one being stainable with the astrocytic marker GFAP, the other stainable for neural cell-adhesion-molecules (N-CAM), and also displaying neuronal electrophysiological characteristics (Fukuda et al.,2003). The presence of nestin expressing glial cells in the myenteric and submucous plexus in the postnatal and adult gut gives hope for a successful approach to isolate and expand neural stem cells from the adult gastrointestinal tract, though nestin might not be the optimal marker for screening strategies based on, that is, FACS sorting. Nestin can be expressed in glial cells or neurons, as well as in non-neuronal cells, such as endothelia (Vanderwinden et al.,2002). The ideal marker should be able to identify individual cells, which fulfill all criteria of “real neural stem cells,” which includes self-renewing, clonal growth and differentiation into neurons and glial cells. The number of theses cells will be limited and following actual isolation strategies, they will be mixed up with many cells of different developmental properties. In the developing gut, we find all kinds of cells with a potential “stemcellness.” Migrating along the gut axis, NCSC do start to form ganglia, while a subpopulation is still heading toward the aganglionic areas. Also most of these cells can express nestin, individual cell will express individual markers concerning their actual role or fate in migration, differentiation, or ganglion formation. There are markers available (P75, ret), which can be used to isolate a majority of these cells, without discriminating “true stem cells” from already committed cells. For the future, appropriate isolating, sorting, and differentiation strategies have to be developed, for both the use in ENS and CNS transplantation. Although using neural crest derived stem cells for the ENS will have to focus on the isolation and expansion of a relatively low number of cells which is feasible by minimal invasive surgery, that is, appendices or mucosal biopsies (Metzger et al.,2008; Schäfer et al.,2009), their potential use for the CNS will additionally depend on techniques to transdifferentiate the NCSC into cell types needed for the individual disease: oligodendrocytes in multiple sclerosis, motorneurones in amyotrophic lateral sclerosic, etc. To avoid severe side effects such as viral infections or the generation of tumors (Amariglio et al.,2009), autologous transplantations where NCSC's have been epigenetically modified will have to be developed. Before starting to think about clinical trials, there must be an intense investigation concerning the integration of NCSC's into the brain or spinal chord. Basic experiments will have to show how NCSC's will be influenced by the altered microenvironment in the diseased or traumatic brain, modified cells will have to be transplanted and followed for significant periods of time to investigate whether the ENS cells transdifferentiate permanently into appropriate CNS cells. The access to a sufficient amount of suitable ENS tissue will be a major bottleneck. Although larger amounts of ENS tissue could be harvested using transmural biopsies (including submucous and myenteric plexus), the question has to be raised how far the disease which has to be treated does influence the quality of the ENS. In traumatic lesions, the ENS will most probably not be affected, while neurodegenerative diseases can also hit and alter the ENS.
The most important question will be: how intensively the ENS participates in CNS disorders (Sakakibara et al.,2008). In Parkinson's disease, the ENS is also affected (Chaumette et al.,2009) and Lewis bodies and Synuclein can be found also in the gut (Wakabayashi and Takahashi,1997; Braak et al.,2006). The amount of dopaminergic cells does decrease in at least a subgroup of theses patients. So the majority of the neurons survive. There is not much known about the glial population in these patients, so the major focus in future approaches has to be put onto the characterization of the glial population and so the presumptive neural stem cell niche.
Plasticity and neural stem cells are two distinct principles in the ENS, which might be much more interconnected than we understand yet.
The enteric nervous system harbours a potential source of neural crest derived stem cells which may be responsible for a great deal of the plasticity in the ENS which is to be seen all over its life span. Developing strategies to identify, isolate and use these neural stem cells for cell therapeutical approaches are currently under investigation in many groups and will lead to their use in clinical trials for the treatment of dysganglionic disorders in the gastrointestinal tract. Moreover, due to its similarities with the central nervous system, the ENS could also be a potent and easily accessible autologous source for central nervous system repair. Using this approach, strategies to adapt neural crest derived stem cells to the needs for their use in the CNS have to be developed. These will include a “chemical” reprogramming by cytokines and transcription factors as well as the use of transfection technologies, which can integrate appropriate genes into the neural crest genome.
We thank M. Gross for proofreading the manuscript.