Background information. Substantial evidence indicates the existence of NCSCs (neural crest-derived stem cells) in embryonic mandibular processes; however, they have not been fully investigated or isolated. The aim of the present study was to isolate stem cells from mandibular process during embryonic development by MACS (magnetic-activated cell sorting). The findings show that the cells are multipotent and self-renewing.
Results. LNGFR (low-affinity nerve-growth-factor receptor)+ cells were isolated from rat embryonic mandibular processes by MACS. The cells were grown in clonal culture by limiting dilution to assess their developmental potential. Clone analysis indicated that, first, LNGFR+ cells are multipotent, being able to generate at least neurons and Schwann cells, similar to peripheral neural crest stem cells. Secondly, multipotent LNGFR+ cells generate multipotent progenies, indicating that they are capable of self-renewal and therefore are stem cells. Thirdly, manipulation of the medium supplementation alters the fate of the isolated LNGFR+ cells.
Conclusions. These results indicate that LNGFR antibodies label NCSCs with high specificity and purity, and suggest that positive selection using these antibodies may become the method of choice for obtaining multipotent cells from rat embryonic mandibular processes for tissue engineering or regenerative therapeutic use.
NCSCs (neural crest-derived stem cells) have aroused a great deal of interest (reviewed in Cameron and McKay, 1998; Gage, 1998; Temple and Alvarez-Buylla, 1999), not only because of their developmental importance, but also for their therapeutic potential (Gage et al., 1995). They spread over the embryo and settle in various tissues where they differentiate into a large variety of cell types, such as neurons, glial cells of the peripheral nervous system, melanocytes, and certain endocrine and paraendocrine cells. In addition, the cranial neural crest cells that have migrated to mandibular process yield mesectodermal derivatives, which are also known as ectomesenchymal stem cells, which further develop to most of the oral and dental tissues (Clouthier and Williams, 2000; Chai et al., 2000). In chicken, multipotent neural crest progenitors have been detected in post-migratory trunk crest derivatives, including the skin (Richardson and Sieber-Blum, 1993), dorsal root ganglion (Duff et al., 1991; Sextier-Sainte-Claire Deville et al., 1992), sympathetic ganglion (Duff et al., 1991) and gut (Sextier-Sainte-Claire Deville et al., 1992). For many years, the development of NCSCs was addressed primarily at the cellular level using avian embryos as an experimental system (Noden, 1993). Recently, it has become possible to observe mammalian NCSCs, by using mouse and rat models (reviewed in Stemple and Anderson, 1993; Anderson, 1997).
A major limitation in the study of NCSCs has been the inability to identify them prospectively. This is because there have been no markers to isolate the stem cells or to distinguish them from restricted progenitors in vivo. Thus multipotent self-renewing neural stem cells have all been isolated after a period of growth in culture that could change their properties (Stemple and Anderson, 1992; Kilpatrick and Bartlett, 1993; Davis and Temple, 1994; Gritti and Parati, 1996; Johe and Hazel, 1996; Kalyani et al., 1997; Palmer et al., 1997). It is therefore not yet clear whether such cells are derived from cells with similar properties in vivo. The expression of certain neural lineage markers confirms that the mesenchyme examined is of crest origin, i.e. ectomesenchyme. Among markers investigated, LNGFR (low-affinity nerve-growth-factor receptor) is a cell surface marker that has been used to prospectively isolate mammalian NCSCs from fetal sciatic nerve, and is suitable for antibody-based cell sorting (Morrison et al., 1999). Thus we isolated LNGFR+ subpopulations by MACS (magnetic-activated cell sorting) and cultured them in clonal culture by limiting dilution to assess their neural developmental potential and self-renewal capacity. Their mesenchymal lineage genetic capacities were also observed in culture study. Taken together, stem cells with dual neural and mesenchymal lineage potentials still persist in rat mandibular tissues until at least a week after the onset of neural crest migration by self-renewing cell divisions. This persistence may explain the origin of dental pulp stem cells found in adult dental pulp, and therefore may be critical for late development or regeneration of oral and dental structures. It also provides a blueprint for the isolation of other crest-derived stem cells for potential tissue engineering and therapeutic use.
Results and discussion
Immunohistochemical characterization of rat embryonic mandibular process tissues
E12.5 (embryonic day 12.5)-E17.5 rat embryonic mandibular process tissues were fixed and sectioned as described in the Materials and methods section. Immunohistochemical/histological analyses indicated that they contained different cell populations (Table 1), similar to their avian counterpart (Baroffio and Bolt, 1991). Most of the cells underlying the epithelium of these tissues expressed S100 and vimentin, suggesting they were of neural crest origin (Mahendra and Anderson, 1997). Some cells expressed differentiated markers such as α-SMA (α-smooth muscle actin), indicating they were precursors committed to smooth muscle cells. No mature markers, such as desmin and peripherin expression, were found at the time point from E12.5 to E17.5, suggesting the cells were not fully differentiated at this stage. Importantly, expression of LNGFR, nestin and HNK-1 (human natural killer 1), markers of NCSCs, were also observed, but with more restricted distribution (Figure 1), indicating a small subpopulation of high multipotency, which was purified by MACS for further analysis.
Table 1. Molecular markers expressed in rat embryonic mandibular tissues−, no expression; ±, weak expression; +, the majority of cells were positive.
Mature muscle cells
Neural progenitor cells
Neural stem cells, immature glial cells
Neural stem cells, immature glial cells
Characterization of potentiality of isolated LNGFR+ cells
Ectomesenchymal cells were enzymatically isolated from rat embryonic mandibular process from E12.5 to E17.5. LNGFR+ cells were purified by MACS (see the Materials and methods section). Cells harvested from embryos at different embryonic days gave similar positive rates of approx. 1%. It was slightly less than the rate the other article reported in migrating neural crest cells (Baroffio and Bolt, 1991). The colonies of LNGFR+ cells containing neurons, Schwann cells and myofibroblasts were designated multipotent clones (Figure 2). By triple labelling with antibodies to peripherin, GFAP (glial fibrillar acid protein) and α-SMA, we identified five types of colonies in clonal cultures of LNGFR+ cells from different ages. They could also be recognized by their heterogeneous morphology and high cell population (on average 1.0×105 cells) after cultured for 14 days. They represented almost 60% of the colonies verified, and their frequency did not significantly decline with the stage of embryonic development until E17.5 (Table 2). In addition to these multipotent colonies, the remainder contained only one or two cell types, and exhibited lower cell population ranging from less than ten to more than several thousand. Further studies are needed in vivo to investigate whether all crest derivatives are produced by one multipotent population or by precursors that are channelled into different lineages. The isolated LNGFR+ cells were then investigated under conditions which favoured muscle, bone and melanocyte development.
Table 2. Frequencies of different progenitor types from freshly isolated LNGFR+ NCSCs based on the types of colonies that formed in clonal cultureN, S and M indicate the presence of neurons, Schwann cells and myofibroblasts respectively in colonies. For example, N+S+M colonies contained neurons, Schwann cells and myofibroblasts. Plating efficiency expresses the percentage of cells added to culture that formed colonies after 2 weeks of culture.
Frequency of colony types (%±S.D.)
After culture of the cells with TGF-β (transforming growth factor β) for 2 days, their morphology changed. Cells became larger and flatter (Figure 3). Approx. 70% of these cells expressed α-SMA, a well-characterized smooth muscle marker. The remaining cells displayed a similar smooth muscle cell-like morphology, despite the lack of expression of α-SMA. An increasing proportion of cells adopted this morphology over the culture time, although the proportion of immunoreactive cells did not change. No GFAP+ or NF (neural filament)+ cells were observed under these conditions. In the control group, 35.7% cells were observed to be α-SMA+, although none were GFAP+ or NF+. These data indicate that most of the flat cells differentiating in the presence of TGF-β have the characteristics of smooth muscle cells and that there is a tendency for the cells to differentiate spontaneously to this phenotype.
After culture of the cells for 4 days in mineralization-promoting medium, the morphology changed from fibroblast-like to a multilateral form. After culture for 14 days, the cells became osteoblast-like in morphology, with a cuboidal shape and a tightly packed arrangement (Figure 4A). Most of the cells (82%) expressed collagen type I and showed increasing alkaline phosphatase activity after 5 and 10 days in culture. After 25 days in culture, the nodules stained intensively with Von Kossa stain (Figure 4C).
α-MEM (α-minimal essential medium) supplemented with 5 ng/ml phorbol esters to promote melanocyte differentiation gave rise to cells of melanocyte morphology and the cells were observed to express melanin. The dopa reaction was performed to detect melanocytes, as described by Hirobe (1992) (Figure 5). In brief, cultures were fixed with 4% paraformaldehyde in PBS for 30 min and then incubated with 0.1% dopa in PBS (pH 6.8) for 4 h at 37°C. The reaction was stopped by rinsing with several changes of PBS and further fixed for 60 min with 4% paraformaldehyde at room temperature.
Self-renewal of multipotent isolated LNGFR+ cells
To characterize the progeny of multipotent LNGFR+ progenitors, we carried out serial subcloning. Primary clones (5- and 10-day-old) were dispersed and replated as single cells. Multipotent colonies were distinguished by their appearance. The cells from the 5- or 10-day-old clones formed subclones with an efficiency of approx. 67% and 51% respectively (Table 3). Many of these subclones gave rise to colonies containing cells expressing two or three phenotypes, indicating that these subclones were multipotent, although the properties of the various subclones differed from those in the clones derived from the freshly isolated LNGFR+ cells. Thus subcloned cells do not necessarily stay true to their original phenotype during enrichment of the cell populations under laboratory conditions.
Table 3. Subcloning of primitive colonies after 5 or 10 days in culture (E14.5)N, S and M indicate the presence of neurons, Schwann cells and myofibroblasts respectively.
Average number of subclones per founder colony
Day of subcloning
Following examination of immunohistochemical characterization of embryonic mandibular tissues, the present study demonstrated that LNGFR+ subpopulation represented a population of NCSCs which had the capacity of producing dual neural and mesenchymal lineage derivatives, and its fate would be influenced by environmental cues. The cells we isolated are very similar to those of pluripotent NCSCs. Such stem cells may be equivalent to NCSCs or NCSC-derived-intermediate stem cells. A similar view has been proposed for the NCSCs in peripheral nervous system (Kruger et al., 2002). On the other hand, appearances of such stem cells including phenotype and multipotency are significantly different from those of known mesenchymal-type stem cells in bone marrow and dental pulp. There is a major distinction in the approaches used to study the two stem cell types (Morrison et al., 1995). HSCs (haematopoietic stem cells) can be identified by surface marker expression, isolated by FACS and transplanted in vivo without being cultured in vitro (Spangrude et al., 1988). In contrast, NCSCs have been isolated from cultures of neural tissue. This is a critical difference, since it is precisely the ability to prospectively identify and isolate HSCs that has facilitated the rapid progress toward understanding their properties. This approach has allowed analyses of the genes expressed by HSCs (for example, see Matthews and Jordan, 1991; Morrison and Prowse, 1996), their lineage relationship with other multipotent progenitor populations (Morrison and Wandycz, 1997) and aspects of their developmental potential (Ikuta and Kina, 1990; Geiger and Sick, 1998). The inability to isolate NCSCs directly from tissue has made it impossible to study many of their properties in vivo.
In the present study, we isolated LNGFR+ cells from rat embryonic mandibular processes by MACS. The cells were cultured in clonal density to assess their developmental potential. Clone analysis indicates that, first, single LNGFR+ cells are multipotent, being able to generate at least neurons and Schwann cells, like peripheral neural crest stem cells. Secondly, multipotent LNGFR+ cells generate multipotent progenies, indicating that they are capable of self-renewal and therefore were stem cells. Thirdly, manipulation of medium supplementation alters the fate of the isolated LNGFR+ cells. The data reported in the present study have important implications for modelling the neural crest development system. The study has demonstrated the feasibility of using a cell phenotype marker of neural crest cell lineage for isolation by MACS of cells population of interest from the heterogeneous population of embryonic tissue. Importantly, it has identified a suitable cell surface marker, LNGFR, for the isolation and prospective identification of such stem cells from these more heterogeneous cell populations. Although a number of candidate markers were investigated, LNGFR was the only one that was consistently expressed throughout the embryonic stages examined and was not an intracellular protein; the latter is a prerequisite for techniques selecting by specific affinity for cell surface molecules.
Prospective identification is an important property for research into in vivo behaviours of target cells. It will now be interesting to extend these investigations to mature dental pulp to see whether this approach can be used for isolation of post-natal stem cells for use in tissue engineering and regenerative strategies for teeth. The approach may also be applicable to other tissues as well.
The LNGFR+ cells isolated from rat mandibular processes showed multipotenticity and thus, if such potentiality is retained in their post-natal equivalents, there may be a broad range of potential applications in medicine. The role of NCSCs in craniofacial development is well recognized (ten Cate, 1995; Chai et al., 2000), although their specific contribution to the development of the various tissues still needs to be fully elucidated. While dental and bone derived post-natal stem cells appear to show a number of similarities, differences also exist. However, it is probable that the ultimate fate of these various cell populations will be determined by their interaction with various signalling and extracellular matrix molecules within their tissue environment in vivo.
Materials and methods
Pregnant Sprague—Dawley rats were obtained from Fourth Military Medical University, Xi'an, People's Republic of China. For timed pregnancies, animals were mated in the afternoon, and the morning on which the plug was observed was designated E0.5. All animal procedures were undertaken in accordance with guidelines of Fourth Military Medical University.
Preparation of NCSCs
The rat mandibular tissues from E12.5 to E17.5 embryos (indicated by a red region in Figure 1) were dissected into ice-cold PBS (pH 7.2) and minced, and then incubated in a 0.025% trypsin solution (Gibco) containing 1 mg/ml type III collagenase (Sigma) at 37°C for 10 min. The enzyme digestion was stopped by neutralization with fresh DMEM (Dulbecco's modified Eagle's medium) containing 1 mg/ml BSA (Gibco), 10 mM Hepes (pH 7.4) and 100 units/ml penicillin/streptomycin (Sigma). Cells were dissociated by gentle stirring and collected by centrifugation, and were then triturated, filtered through a steel mesh screen (75 μm) to remove aggregates of cells and undigested tissues, and then resuspended for MACS.
Isolation of LNGFR+ cells by MACS
Primary LNGFR+ cells were isolated by their affinity to magnetic beads coated with sheep anti-rabbit IgG (Dynal Biotech) according to the manufacturer's instructions. In brief, freshly dissociated cells were suspended in a 1/200 dilution of H-92, a rabbit anti-LNGFR antibody (Santa Cruz), for 30 min on ice. After washing, the mixed cell suspensions were incubated with the beads at a concentration of 1×107 beads/ml of cell suspension. After incubation at 2–8°C for 15 min with gentle mixing in a Dynal mixer, the beads were isolated by placing the suspensions in a Dynal magnetic recovery device for 2 min. The resettled cells were resuspended and separated from the beads with release buffer. The dissociated cells were further examined by fluorescence immunohistochemistry for LNGFR reactivity using H-92 antibody and a sheep anti-rabbit secondary antibody (FITC-labelled; Sigma).
Clonal culture of freshly isolated LNGFR+ cells
Single LNGFR+ cell suspensions were plated at limiting dilution of 1 cell/well in ten 96-well plates coated with poly(d-lysine) and 0.15 mg/ml human fibronectin (Biomedical Technologies), with DMEM basal medium containing insulin, BMP2 (bone morphogenetic protein 2), epidermal growth factor, basic fibroblast growth factor and nerve growth factor as described previously (Stemple and Anderson, 1992). The positions of single cells in the wells were marked under phase-contrast microscopy for verification that clones formed were derived from these single cells. After culture for 6 days in this medium, cells were switched to another supplemented media (with 1% chick embryos extracts and 10 ng/ml basic fibroblast growth factor that favours differentiation) for another 8 days before immunohistochemical characterization of colonies. All cultures were maintained in a humidified atmosphere comprising 95% air/5% CO2 at 37°C.
Serial subcloning of freshly isolated LNGFR+ cells
Primary clones cultured for 5 and 10 days were harvested and replated as following: cells were removed by treatment with 10 μl of 0.05% trypsin at 37°C for 3 min. The enzyme digestion was stopped by addition of 90 μl of medium and vigorous trituration. The cells were resuspended at a density of 1 cell/100 μl of medium. Single cell suspensions were cultured, and cultures were analysed as described below.
Immunohistochemistry of embryonic mandibular tissues and cells
Heads of rat embryos (E12.5-E17.5) were dissected and fixed in freshly prepared 4% paraformaldehyde and, following routine histological processing, were embedded in paraffin. Serial sections were stained with antibodies against known neural crest lineage markers including HNK-1 for mouse anti-CD57 (NeoMarker), PJM50 for mouse anti-peripherin (Novocastra Laboratories), H-92 for rabbit anti-LNGFR/p75 (Santa Cruz), Rat104 for mouse anti-nestin (Boster Co., Wuhan, China) and rabbit anti-S100 (DAKO), as well as E29 for mouse anti-[human EMA (epithelial membrane antigen)] (DAKO) as a negative control. Other antibodies used for cell characterization in tissues included D33 for desmin (Immunologicals Direct), 2F11 for NF160 (DAKO), 6F2 for GFAP (DAKO), α-SMA (Boster Co.), VIM3B4 for vimentin (DAKO) and collagen types I and III (Boster Co.).
Neural crest cells have been reported previously to generate neurons, glial and myofibroblasts (Stemple and Anderson, 1992; Parsek et al., 1998; Morrison et al., 1999). Therefore for clone analysis, we identified these three types in cultures. Neurons were typically identified by detecting expression of peripherin, a marker of mature PNS (peripheral nervous system) neurons, as well as neurofilament 160 and neuron-specific tublin (TuJ1). Schwann (glial) cells were identified by examining expression of GFAP and LNGFR; other cells co-expressed α-SMA and S100 and were defined as myofibroblasts, which do not express peripherin, GFAP or LNGFR. Staining procedures for clone analysis were carried out as described previously (Stemple and Anderson, 1992).
Differentiation of muscle, bone and melanin-secreting cells from LNGFR+ cells
Freshly isolated LNGFR+ cells were plated at 50000 cells/35 mm culture dish (5200 cells/ml) in α-MEM standard medium containing 15% FBS (fetal bovine serum) and antibiotics. After 24 h, the medium was removed, and specific supplemented media for induction of differentiation were added as shown in Table 4. [All supplemented reagents were obtained from Sigma except TGF-β, which was prepared in our laboratory (Si et al., 1998)]. Medium was replaced every 3 days. The morphology of cultured cells was observed by phase-contrast microscopy and immunohistochemical analyses were used to characterize the differentiated cells.
Table 4. Lineage-specific differentiation induced by medium supplementation
Smooth muscle cells
0.1 M NEAA (non-essential amino acids), 60 pM TGF-β
10 mM β-glycerophosphate, 50 μg/ml ascorbic acid, 10 nM dexamethasone