Many recent investigations have generated excitement about the use of either embryonic pluripotent stem cell lines or multipotent stem cell lines derived from fetal or adult tissue to replace cells or rebuild tissue lost or destroyed by disease. There is considerable debate in the literature on the relative merits of the two broad cell types; however, there appears to be a growing consensus that research needs to be done along both lines for the next several years before the true therapeutic potential, if any, of either approach can be established (Gage,2000).
ES Cells and ES Cell-Derived Neural Precursors
Pluripotent ES cell lines possess several distinguishing attributes that may represent potential advantages, depending upon the proposed application. To begin with, they are capable of giving rise to more diverse offspring than multipotent cells. The second important difference relates to the potentials of ES cells to be expanded in culture indefinitely without losing differentiation potential. By contrast, it has been difficult if not impossible to expand adult multipotent stem cells in culture without losing their undifferentiated state. However, although attractive because of their pluripotentiality, there are some obstacles and risks associated with the use of ES cells for therapy, even setting aside the so-called ethical controversies. Spontaneously, only 0.2% of ES cells generate neurospheres (Tropepe et al.,2001), and the number and complexity of signals and transformations required to achieve a neuronal subtype may be formidable (Temple,2001). Recent studies do show, however, that such differentiation can be achieved in vitro (Zhang et al.,2001); whether the neuronal precursors derived from ES cells behave differently than those derived from the CNS remains to be investigated. In addition, there exists at least one identifiable risk that appears to be particularly associated with ES cells: human ES cells injected into mice can produce teratoma-like growths (Reubinoff et al.,2000,2001).
Our preliminary data suggest that grafted ES cells do not differentiate into mature neurons or glia (Sasselli et al.,2006) and that, for an ES-based transplant strategy to successfully repopulate the ENS, the cells will either need pretransplantation “nudging” or further manipulation of the post-transplant environment to acquire an enteric neuronal phenotype, a strategy that we are currently pursuing. The idea that ES cells can be used as a reservoir of cells from which differentiated cells can be isolated has been validated in both rodent and human models (Gorba and Allsopp,2003; Bibel et al.,2004). The advantage of using neurons, glia, or precursors derived from ES cells is that they may not be regionally specified (by contrast with local tissue-specific neural progenitors) and, thus, may be more capable of site-specific integration. Several investigators have shown that cell surface markers, in vitro manipulation of culture conditions, or using tissue-specific promoters can be used to isolate neural stem cells or more restricted precursors. When transplanted in the rodent CNS or spinal cord, ES-derived neural precursors implant and give rise to neurons and glia and are functionally active. The generation of peripheral neurons and neural crest derivatives from ES cells has also been reported (Mizuseki et al.,2003; Pomp et al.,2005). Moreover, and of relevance to this review, our preliminary results suggest that enteric neurons can be generated from ES cells in vitro (Sasselli et al.,2006).
Neural stem cell-like cells have also been observed or induced to arise from a diverse population of other stem cell populations, including those resident in bone marrow (Jiang et al.,2003), and adipose tissue and skin (Safford et al.,2002,2004; Joannides et al.,2004). Whether these derived precursors can truly give rise to sustained and functional neuronal cells after transplantation remains to be established.
NSC have been identified in both the developing and adult CNS and PNS of rodents and humans. These cells can be grown in culture and display the ability to self-renew, although they are more restricted than ES cells and give rise predominantly to neurons and glia (astrocytes and oligodendrocytes). In the CNS, NSC have been identified and isolated from various sites within the rodent brain, including the cerebellum, cerebral cortex, hippocampus, and subventricular zone (the richest source along with the hippocampus), and spinal cord. However, NSC from different sites are not necessarily identical, displaying different growth characteristics, trophic factor requirements, and specific patterns of differentiation (Mayer-Proschel,1997; Rao,1999; Morrison,2001).
The use of multipotent NSC for treatment has several advantages with respect to ES cells. First, cells of a given lineage (e.g., NSC to produce neurons) are “naturally poised” to generate a particular tissue (National Academy Press,2001) and may require little or no manipulation either in vitro or in vivo. Second, such stem cells are able to migrate to sites of injury, or other “homing” sites within an organ (Aboody et al.,2000). Third, they can produce biological factors that may have additional beneficial effects for their “native” tissue (Noble,2000). Thus they can deliver missing neurotransmitters, provide a tissue scaffold for host axonal regeneration, and provide neurotrophic and other active molecules that promote endogenous growth and repair (so-called “bystander” effects; Ourednik and Ourednik,2004,2005; Martino and Pluchino,2006). A potential problem with NSC is that, by contrast with pluripotent ES cells, multipotent human adult stem cells undergo replicative senescence and crisis after approximately 50 cell divisions, related to the lack of telomerase activity and progressive shortening of the telomeres (Cheng et al.,2005). Fetal multipotent cells may have an advantage in this respect, but recent studies also indicate that the challenge of growing even adult stem cells in large numbers may be overcome in the near future (Gottlieb,2002).
Most investigators isolate NSC by dissecting out a region of the fetal or adult brain, typically the subventricular zone (SVZ). The tissue is then disaggregated, and the dissociated cells are grown in culture medium containing a high concentration of fibroblast growth factor-2 (FGF-2) and/or epidermal growth factor (EGF) and allowed to proliferate. It appears that FGF is mitogenic at the beginning of neurogenesis and EGF-responsive cells emerge later (Santa-Olalla and Covarrubias,1999; Martens et al.,2000). In vitro, FGF but not EGF, promotes self-renewal of NSC, whereas together the two growth factors permit the initial commitment of NSC into neuronal and glial phenotypes (Maric et al.,2003).
In culture, NSC can be characteristically made to grow and proliferate in colonies called neurospheres, which are floating spheroid structures containing a mixture of NSC and more differentiated (progenitor) cells suspended together in an extracellular matrix. These structures have served as convenient assays to test both the multipotentiality of NSC (by characterizing the types of cells present within) as well as their ability for self-renewal (with the assumption that isolated NSC should themselves give rise to secondary neurospheres; Campos,2004). Only 3% to 4% of the cells within neurospheres are actually true stem cells, in that they can give rise to all three neural lineages either in vivo or in vitro when induced to differentiate upon withdrawal of growth factors (EGF and FGF) or by specific inductive signals (trophic factors such as glia-derived growth factor [GDNF], brain-derived neurotrophic factor, or nerve growth factor [NGF]; Gritti et al.,1996).
In the past, it has been difficult, if not impossible, to distinguish tissue-specific stem cells from progenitor cells, which can also be found in the same tissue and represent more differentiated cells that divide and give rise to fully differentiated cell types (National Institutes of Health,2001). Rietze et al. have recently provided a potential means to distinguish the NSC phenotype by virtue of nestin positivity, and low to absent expression of both peanut agglutinin and heat stable antigen (nestin+, PNAlo, HASlo; Rietze et al.,2001). Nestin is an intermediate filament protein whose expression is widely used to identify mammalian neuronal precursor cells or stem cells (Lendahl et al.,1990; Zimmerman et al.,1994). The stem cell is initially positive for nestin, but secondary progenitor cells become negative for nestin. 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, similar to what has been described for another putative stem cell marker, Msi-1 (Musashi-1; Sakakibara and Okano,1997; Sakakibara et al.,2001). Furthermore, Uchida et al. have shown that CD133, a marker used to enrich for hematopoietic stem cells, can also be used to isolate human NSC (Uchida et al.,2000). If these results are validated, we may be able to purify NSC to a much greater degree. For the present, however, most investigators rely on neurosphere generation and nestin positivity as surrogate markers for NSC. Traditionally, it has been easier to work with fetal NSC because of their relative abundance; however, it should be noted that it is not yet established with complete certainty that they behave in the same way as their adult counterparts. Interest in the use of fetal NSC as a therapeutic tool has also been fueled 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 (Kuhn and Svendsen,1999; Svendsen et al.,1999).
CNS-Derived NSC Can Improve Function in a Mouse Model of Gastroparesis
Until very recently, CNS-derived NSC (CNS-NSC) were the most well-characterized and experimentally studied neuronal precursor population. Hence, we used these cells for our initial experiments a few years ago, even though we were well aware of the fact that enteric neurons may represent a different lineage. The potentials of CNS-NSC for the treatment of disorders of the ENS have recently been explored by us. NSC isolated from the SVZ of fetal rodent brain express RET and proliferate in response to GDNF (RET is the tyrosine-kinase receptor for GDNF, a trophic factor that plays a critical role during the development of the ENS) in vitro (Micci et al.,2001). Moreover, they express the enzyme responsible for the production of NO (nNOS), the major neurotransmitter in the ENS. When co-cultured with colonic smooth muscle cells CNS-NSC differentiate into neurons and form functional contact with smooth muscle cells (Fig. 2).
Figure 2. Top: β-Tubulin immunoreactivity in rat central nervous system-derived neural stem cells (CNS-NSC) in co-culture with human colonic smooth muscle cells (visible in the background) indicating that CNS-NSC differentiate into neurons. Bottom: Intracellular calcium measurements in human colonic smooth muscle cells in co-culture with rat CNS-NSC in response to the nicotinic agonist DMPP. Cells were loaded with the calcium-sensitive fluorescent indicator Fura-2 and changes in intracellular calcium were measured in colonic smooth muscle cells in response to bath application of the nicotinic agonist DMPP (20 μM, indicated by the white arrows). A: DMPP induces an increase in intracellular calcium in colonic smooth muscle cells when in co-culture with rat CNS-NSC. B: This effect is not observed when colonic smooth muscle cells are cultured without rat CNS-NSC. C: The DMPP-induced response in colonic smooth muscle cells in co-culture with rat CNS-NSC was blocked by the nicotinic receptor antagonist hexamethonium, demonstrating that DMPP is acting on nicotinic receptor. D: Moreover, the response was blocked by the neurokinin A receptor antagonist (CP-099994-01), suggesting that it is mediated by substance P. These data show that rat CNS-NSC stimulated with DMPP induce an increase in smooth muscle calcium, demonstrating that they make functional contact with smooth muscle cells in vitro.
Download figure to PowerPoint
When CNS-NSC are transplanted in the stomach of a transgenic mouse model of gastroparesis (the nNOS−/− mouse, characterized by delayed gastric emptying and the inability of the pyloric sphincter to relax in response to nerve stimulation), CNS-NSC graft and differentiate into neurons and glia, with the majority of neurons expressing nNOS. Moreover, 1 week after implantation of CNS-NSC, significant improvement of gastric function (known to be impaired in nNOS−/− mice) is observed, thus providing the first functional evidence that transplantation of CNS-NSC in the gut might be beneficial in the treatment of motility disorders (Fig. 3; Micci et al.,2005a). Although promising, these data are limited in their short-term nature and further investigation are necessary to assess long-term survival of CNS-NSC in the gut and their functional effect on motility.
Figure 3. Left: Transplanted neural stem cells (NSC) survive in the pyloric wall and differentiate. A,B: NSC grafts in cross-sections of nNOS−/− mouse pylorus, 1 week post-transplantation showing immunoreactivity for the neuronal marker PGP9.5 and the glial marker GFAP. LM, longitudinal muscle; CM, circular muscle. Scale bar = 20 μm. Center: NSC transplantation results in increased relaxation of the pyloric muscle. A: Electrical field stimulation of pyloric circular muscle under NANC conditions in nNOS−/− mice. B: Quantification of NANC-induced relaxations in response to EFS in nNOS−/− mice as described in A. C: L-NAME (100 μM) blocks relaxations, indicating that the relaxations are NO mediated. Data are means ± SEM. *P < 0.05 and **P = 0.001. Right: Transplanted NSC improve gastric emptying in nNOS−/− mice. Gastric emptying of liquids in nNOS−/− mice 1 week after NSC transplantation or vehicle. *P < 0.01; n = 6 mice for vehicle and 8 mice for NSC. (Reprinted from Micci et al.,2005a.)
Download figure to PowerPoint
Neural Crest Stem Cells and ENP
Within the developing CNS, a group of cells arise from the neural crest, the neural crest stem cells (NCSC). NCSC have been isolated and characterized in multiple species from the neural crest. NCSC generate the PNS and also migrate throughout the gut to form the ENS (Gershon,1997; Heuckeroth and Pachnis,2006). Like CNS-NSC, NCSC are multipotent and generate differentiated progeny by means of the production of more restricted progenitor cells, but in addition to neurons and glia they can also generate non-neuronal cells (i.e., smooth muscle cells, pigment cells, bone and cartilage of the face and skull). NCSC, therefore, differ from CNS-NSC because of their extensive ability to migrate, their epithelial to mesenchymal transition, and the phenotypes that can arise from them. Evidence from chick embryo experiments suggest that NCSC and differentiated CNS cells share a common progenitor (Bronner-Fraser and Fraser,1989; Mujtaba et al.,1998). Spinal cord neuroepithelial cells can generate p75+/nestin+ cells that are morphologically and antigenically similar to previously characterized NCSC (Mujtaba et al.,1998). CNS-NSC–derived p75+ cells differentiate into peripheral neurons, smooth muscle, and Schwann cells in both mass and clonal culture (p75 is the low-affinity receptor for NGF, whose function during development is not completely understood). Moreover, clonal analysis indicated that individual CNS-NSC could generate both CNS and PNS derivatives, providing evidence for the first time of a direct lineage relationship between these two distinct cell types. However, it is unclear whether NCSC can generate CNS derivatives. When NCSC were transplanted into the spinal cord, they gave rise to Schwann cells but not to CNS derivatives, suggesting that the NCSC is a more restricted precursor. These concepts are illustrated in Figure 4.
Figure 4. During development, early neuro-ectodermal stem cells generate two subpopulations of multipotent neural stem cells: Central nervous system-derived neural stem cells (CNS-NSC) that give rise to neurons and glia of the CNS and neural crest stem cells that generate the neurons and glia of the peripheral nervous system (PNS), including the enteric nervous system (ENS), as well as other non-neural lineages. Recent reports suggest that individual CNS-NSC could generate both CNS and PNS derivatives, providing evidence for the first time of a direct lineage relationship between these two distinct cell types. Specific markers expressed by each population are indicated in red.
Download figure to PowerPoint
NSC have also been isolated from the immature as well as adult ENS (Kruger et al.,2002; Sidebotham et al.,2002a,b; Bixby et al.,2002; Suarez-Rodriguez and Belkind-Gerson,2004; Burns,2005), a discovery that raises the possibility of an intrinsic mechanism for enteric neuronal replacement within the gut, even though we are as yet unaware of what it takes to harness this potential. Whether these cells have the same potential of NCSC is not clear but are probably more restricted to a gut phenotype and more appropriately termed enteric neuronal precursors, or ENP. Such precursor cells have been isolated from the embryonic and postnatal gut of rats using antibodies to specific markers known to be expressed by enteric neural crest-derived cells: RET (the tyrosine-kinase receptor for GDNF; Lo and Anderson,1995) and p75 (Kruger et al.,2002).
ENP have also been isolated from the gut without using these markers. When the embryonic or postnatal gut is dissociated and the cells cultured in high concentrations of EGF and/or FGF, neurospheres are generated that contain nestin+ proliferating progenitor cells, neurons, and glia (Bondurand et al.,2003; Schafer et al.,2003; Suarez-Rodriguez and Belkind-Gerson,2004). These cells partially recapitulate the genetic program of ENS development from NCSC in that they express Sox10 but lack detectable levels of RET or lineage specification markers, such as MASH1 (the transcription factor Sox10 plays a critical role in development of the PNS including sensory, autonomic, and enteric ganglia [Britsch et al.,2001] while Mash1 encodes a basic helix–loop–helix pro-neural factor that is required for specific subsets of enteric neurons; Blaugrund et al.,1996).
ENP isolated from the small intestine of lactating and adult mice can differentiate into various cell types, particularly neurons, smooth muscle, and glia with the neurons expressing several sensory and motor neurotransmitters present in the CNS and ENS, including calcitonin gene-related peptide, neuropeptideY, peptideYY, substance P, vasoactive intestinal polypeptide, and galanin; along with glia, these neurons form elaborate intercellular connections. Both endothelin 3 and SOX10 are important factors in maintaining these cells in an uncommitted state (Bondurand et al.,2006). GDNF promotes ENS precursor survival through PI-3 kinase (Srinivasan et al.,2005), but its effects on proliferation are opposed by Sonic hedgehog (Shh), which also prevents their differentiation (Fu et al.,2004).
ENP differ from their neural crest counterparts in the sciatic nerve—the cells from the gut are more responsive to neurogenic factors, while sciatic nerve cells are more responsive to gliogenic factors (Bixby et al.,2002). When transplantated into peripheral nerves in vivo, sciatic NSC generate only glia, while gut NSC give rise primarily to neurons. Thus ENP, by virtue of their genetic lineage, may have quite different potentials than their CNS counterparts and may be more suitable for therapeutic purposes as they are more likely to respond to gut-specific environmental cues.
There are several compelling reasons to think that ENP may become the preferred route for clinical use. First, it has been shown by Dr. Pachnis' laboratory that multipotential progenitors can be derived from biopsy-sized gut segments, allowing minimally invasive harvesting of tissue (Bondurand et al.,2003). The gut is readily accessible endoscopically in humans, thus rendering this an attractive feature. The same team has also shown that ENP can also be isolated from unaffected segments of the gut of Ret mutant animals. Furthermore, isolation and expansion of precursor cells from the developing and postnatal human ENS has recently been reported using bowel samples from human fetuses and children (from the ninth week of gestation to 5 years postnatal). Neurospheres obtained by such methods could be differentiated and also be transplanted after dissociation into aganglionic bowel in vitro (Rauch et al.,2005), thus confirming earlier experimental results with murine ENP (Bondurand et al.,2003). These are exciting results and highlight the potential for the use of ENP in postnatal motility disorders.