SEARCH

SEARCH BY CITATION

Keywords:

  • neural crest;
  • sox10 transcription factor;
  • enteric nervous system;
  • bacterial artificial chromosome (BAC);
  • live imaging;
  • lineage restriction

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

To facilitate dynamic imaging of neural crest (NC) lineages and discrimination of individual cells in the enteric nervous system (ENS) where close juxtaposition often complicates viewing, we generated a mouse BAC transgenic line that drives a Histone2BVenus (H2BVenus) reporter from Sox10 regulatory regions. This strategy does not alter the endogenous Sox10 locus and thus facilitates analysis of normal NC development. Our Sox10-H2BVenus BAC transgene exhibits temporal, spatial, and cell-type specific expression that reflects endogenous Sox10 patterns. Individual cells exhibiting nuclear-localized fluorescence of the H2BVenus reporter are readily visualized in both fixed and living tissue and are amenable to isolation by fluorescence activated cell sorting (FACS). FACS-isolated H2BVenus+ enteric NC-derived progenitors (ENPs) exhibit multipotency, readily form neurospheres, self-renew in vitro and express a variety of stem cell genes. Dynamic live imaging as H2BVenus+ ENPs migrate down the fetal gut reveals cell fragmentation suggesting that apoptosis occurs at a low frequency during normal development of the ENS. Confocal imaging both during population of the fetal intestine and in postnatal gut muscle strips revealed differential expression between individual cells consistent with down-regulation of the transgene as progression towards non-glial fates occurs. The expression of the Sox10-H2BVenus transgene in multiple regions of the peripheral nervous system will facilitate future studies of NC lineage segregation as this tool is expressed in early NC progenitors and maintained in enteric glia. genesis 49:599–618, 2011. © 2011 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Imaging migration of neural crest (NC) has been an area of intense investigation because of the diversity of cell lineages that derive their origins from this transient population, as well as the desire to understand cell-to-cell communication and the responses of these multipotent progenitors to regulatory cues. Time lapse imaging of developmental processes is most readily achieved in embryos that develop externally like zebrafish (Andersen et al.,2010), chick (Kulesa and Fraser,1998, 2000), and frog (Kieserman et al.,2010). Indeed induction and early stages of NC migration have been particularly amenable to live imaging at the dorsal aspect of these embryos (Ahlstrom and Erickson,2009; Krull et al.,1995; McLennan and Kulesa,2007). However, live imaging of NC migration in the mammalian embryo has proven more challenging due both to placental development as well as the size and thickness of the embryo. Despite these challenges, analysis of NC development is particularly appealing in murine systems because molecular genetic tools comprehensively label migrating NC populations.

Dynamic imaging of enteric neural crest derivatives during migration in the fetal gut is one area where live cell microscopy has been successfully implemented in the mouse. This success has arisen primarily due to developments in mounting and maintaining the fetal intestine in catenary cultures that are amenable to microscopy (Abud et al.,2008; Druckenbrod and Epstein,2005; Young et al.,2004). As a result, migratory behaviors of NC-derived progenitors in the fetal gut have been visualized and the responses of progenitors to specific manipulations have identified key molecules that are essential for migratory behaviors (Anderson et al.,2006b; Breau et al.,2009; Druckenbrod and Epstein,2009). These imaging studies have relied primarily on constructs that either introduce fluorescent reporters by knock-in gene targeting into the coding regions of enteric progenitor genes (Ret; Young et al.,2004), transgenic reporters that reflect behaviors of early neurons (Tyrosine Hydroxylase; Hao et al.,2009) or bi-genic neural crest lineage reporters (Wnt1Cre x RosaYFP; Breau et al.,2009; Druckenbrod and Epstein,2005). Such molecular genetic approaches are advantageous because they label the entire NC population in contrast to cell labeling techniques in chick (electroporation or DiI) that only tag subsets of the NC. These successes have established fundamental baselines on migratory patterns, cell speeds, connectivity, and trajectories.

However, to advance live cell microscopy of mammalian NC further, several challenges must be met. Single locus reporters that do not alter genes that participate in development are needed to avoid the possibility that knock-in alleles could alter cell behaviors. Moreover, bright stable fluorescent reporters expressed from the earliest phases of NC delamination and that are maintained in mature cell types would significantly expand opportunities to investigate NC lineage divergence. A recently reported Sox10-Venus transgenic that provides cytoplasmic fluorescence addresses some of these issues (Shibata et al.,2010). But, as for all previously described NC reporters, this allele relies on imaging cytoplasmic fluorescent expression that does not always permit discrimination of individual cells, particularly those migrating immediately adjacent to one another as streams of early migrating NC or enteric NC-derived progenitors (ENPs) do.

To image NC lineages throughout mammalian development concurrent with visualization of the cell nucleus that often exhibits cell type specific morphology in mature tissues, we developed a transgenic line that drives expression of a Histone2BVenus reporter from the regulatory regions of Sox10. Venus is a bright stable variant of yellow fluorescent protein that matures rapidly in cells (Nagai et al.,2002) and its fusion with the Histone2B moiety labels chromatin of both interphase and dividing cells (Kanda et al.,1998). The expression of the resulting transgene, Tg(Sox10-HIST2H2BE/YFP*)1Sout hereafter referred to as Sox10-H2BVenus, recapitulates expression of Sox10 at multiple stages of NC development including restriction to mature glial cells in the enteric nervous system (ENS). Not only were all Sox10+ cells migrating in the ENS labeled by transgene expression, but mitotic cells were readily visible, offering a means to quantify proliferative rates in situ directly. The transgene facilitates visualization of multipotent NC-derived progenitors in fixed tissue, isolation of these cells by flow cytometric analysis, and dynamic live cell imaging of these progenitors as they populate peripheral tissues. Confocal microscopy of H2BVenus+ cells in the fetal gut revealed differential expression of Sox10 between individual migrating ENPs that likely reflects altered expression of the transgene as lineage segregation proceeds. Moreover, live confocal microscopy identified rare and infrequent nuclear fragmentation that appears to be indicative of apoptosis, a developmental mechanism that has not previously been observed by live imaging of the fetal gut during ENS ontogeny. Because expression of Sox10-H2BVenus offers one of the earliest possible entry points into development of NC-derived lineages and is maintained in enteric glia as well as other types of peripheral glia, this reporter should significantly advance efforts to image the complete continuum of events in NC development including capture of cells undergoing cell fate restrictions during lineage divergence.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Generation and Analysis of Sox10-H2BVenus Transgenic Mice

The Sox10-H2BVenus bacterial artificial chromosome (BAC) construct was generated by integrating a fluorescent fusion protein, Histone2B-Venus (H2BVenus), into a Sox10 BAC derived from the CHORI RP-23 C57BL/6J (B6) genomic library (Fig. 1a). In vitro homologous recombination methods were used to fuse the coding sequences of H2BVenus in frame with the ATG of the Sox10 coding region in Exon 1 (Lee et al.,2001) so that no sequences from the Sox10 locus were deleted. As a consequence of the polyadenylation sequence at the end of the H2BVenus cassette, no Sox10 protein is produced by the BAC construct. The 28O11 BAC spans an interval of 218 kb at the Sox10 locus and has previously been used to drive expression of a βGal reporter and recapitulates patterns and cell-specific expression of the endogenous Sox10 gene(Deal et al.,2006). We demonstrated the functionality of the Sox10-H2BVenus BAC construct by in vitro transfection into enteric glial cells in culture (Fig. 1b). Transfected cells exhibit bright nuclear fluorescence and distinct labeling of mitotic spindles as a consequence of the Histone2B fusion reporter remaining associated with chromatin throughout the cell cycle (Fraser et al.,2005; Hadjantonakis and Papaioannou,2004; Kanda et al.,1998).

thumbnail image

Figure 1. Sox10-H2BVenus BAC transgene construct and expression. (a) Schematic diagram of Sox10-H2BVenus targeting vector and homologous recombination into the wild-type Sox10 BAC depicts Sox10 exons (black rectangles), initiator ATG (green), and Histone2BVenus (H2B) fusion reporter (grey/yellow). Sox10 homology arms (each ∼500 bp) are shown on either side of the Venus sequence. Excision of the TetR cassette after integration to derive the final Sox10-H2BVenus BAC transgene is shown. (b) Expression of the H2BVenus reporter produces bright nuclear fluorescence in transfected enteric glial cells that is associated with mitotic chromosomes (insets b′ and b″) in a fraction of the transfected cells (×400). (c) Sox10-H2BVenus transgene expression at 10.5 dpc labels multiple neural crest derivatives (tg, trigeminal ganglia; ov, otic vesicle; v vagus; drg, dorsal root ganglion) in the PNS. (d) Sox10-H2BVenus transgene in 14.5 dpc fetal mouse gut labels enteric progenitors. (e) Colocalization of Sox10 protein and the Sox10-H2BVenus transgene expression exhibits complete coincidence in fetal gut at 14.5 dpc (×400). High magnification inset of a Sox10+/H2BVenus+ cell (white arrow) demonstrates the localization of the reporter protein with chromatin in mitotic spindles of dividing cells.

Download figure to PowerPoint

The Sox10-H2BVenus construct was microinjected into fertilized oocytes, and three founders were identified that carried H2BVenus transgene sequences from a total of 147 pups. Two of the three founders passed the transgene to their offspring and were bred to establish lines. Additional genotyping established that both of these lines (designated A and C) carried sequences for the flanking BAC T7 and Sp6 arms suggesting that the Sox10-H2BVenus construct was intact.

To assess transgene integrity and rule out the possibility of small deletions that can remove regulatory elements and impact transgene expression (Chandler et al.,2007; Deal et al.,2006), these two lines were out-crossed to C3HeB/FeJ (C3Fe) mice, and third generation progeny were screened with simple tandem repeat markers (STRs) that are polymorphic between C3Fe and B6 strains as previously described(Deal et al.,2006). The presence of B6 alleles detected by these STRs in genomic DNA from third generation outcross mice for lines A and C indicated that the complete genomic interval derived from the B6-origin Sox10 BAC had integrated into the genome (data not shown). Subsequent analyses were carried forward with both lines A and C because all molecular tests indicated the transgene was intact in these lines and patterns of expression in fetal and adult tissues appeared comparable. These lines were back-crossed to C3HeB/FeJ to establish the transgene on a uniform genetic background and all experiments were performed with C3Fe.Sox10-H2BVenus lines at greater than 10 backcross generations.

Sox10-H2BVenus BAC Exhibits Appropriate Spatial, Temporal, and Cell-Type Specific Expression

To establish that the Sox10-H2BVenus transgene conferred appropriate expression on the H2BVenus reporter, we evaluated patterns of Venus fluorescence in whole-mount embryos at 10.5 days post coitus (dpc) and in 14.5 dpc fetal gut. Both Sox10-H2BVenus transgene lines drive expression that is consistent with previously documented Sox10 expression in cranial ganglia, vagal NC streams migrating through brachial arches, otic vesicle, dorsal root ganglia, and cervical ganglia (Fig. 1c) (Anderson et al.,2006a; Britsch et al.,2001; Herbarth et al.,1998). Robust H2BVenus reporter expression was also readily apparent in 14.5 dpc whole-mount gut (Fig. 1d) and the distribution of fluorescence was consistent with the distribution of ENS progenitors in the intestinal wall.

To establish that the Sox10-H2BVenus transgene reflects expression of Sox10 in migrating ENPs of the ENS, we examined co-localization of H2BVenus with immunohistochemical (IHC) labeling for Sox10 protein. Whole-mount IHC of fetal gut demonstrated complete concordance of transgene expression with enteric progenitors expressing Sox10 at 14.5 dpc (Fig. 1e). Incidentally, we observed that the distinct distribution of cytoplasmic Sox10 and H2BVenus associated with mitotic spindles in dividing cells made proliferating ENPs readily apparent. While Sox10 is present primarily in the nuclei of migrating ENPs, the breakdown of the nuclear membrane during mitosis resulted in dispersal of Sox10 protein to the cytoplasm while the H2BVenus remained associated with the mitotic chromosomes. As a result, mitotic cells were revealed by the presence of a cytoplasmic halo of Sox10 protein surrounding H2BVenus+ condensed chromosomes during mitosis (Fig. 1e, inset).

Migration pathways of enteric progenitors that populate the intestine to generate the ENS have been described (Druckenbrod and Epstein,2005, 2007; Young and Newgreen,2001; Young et al.,1998, 1999, 2000, 2004). To establish that the Sox10-H2BVenus transgene confers appropriate temporal and spatial expression of H2BVenus in migrating ENPs we examined expression in 9.0, 10.0, and 12.0 dpc fetal gut. The earliest migratory NC is labeled by Sox10-H2BVenus expression at 9.0 dpc as cells leave the neural tube and vagal streams of cells can be seen trafficking through the brachial arches toward the foregut (Fig. 2a). Transgenic embryos display appropriate regional expression with individual cells being clearly visible in the foregut of sub-dissected fetal intestine at 9.0 dpc (Fig. 2b). At 10.0 dpc, H2BVenus+ ENPs have progressed through the foregut into the midgut and are visible just rostral to the cecal bulge (Fig. 2c). Notably, high magnification images of H2BVenus+ ENPs at the 10.0 dpc wave-front reveal heterogeneity of reporter expression between individual cells. Some cells are clearly brighter than other adjacent dimmer cells. This differential expression within the population is analogous to heterogeneous Phox2b expression that occurs at this same stage (Corpening et al.,2008). By 12.0 dpc, the H2BVenus+ ENPs have migrated from the midgut around the cecal bulge into the hindgut with distributions similar to Ret+ ENPs (Druckenbrod and Epstein,2005). We also observed that sacral Sox10+ populations were labeled by transgene expression (Fig. 2c and data not shown). Thus Sox10-H2BVenus transgene expression recapitulates multiple aspects of vagal and sacral ENP development.

thumbnail image

Figure 2. Appropriate temporal and spatial patterns of Sox10-H2BVenus transgene expression in migrating enteric NC-derived progenitors. (a) Streams of H2BVenus+ cells are seen in whole-mount 9 dpc Sox10-H2BVenus transgenic embryos imaged by confocal microscopy caudal to the otic vesicle in the glossopharyngeal (IX), vagal (X), and sympathetic chain pathways (sc) (×200). (b) Subdissected 9 dpc Sox10-H2BVenus transgenic gut reveals labeled ENP migrating into and down the developing foregut (×200). (c) ENP progress into the midgut region of a whole-mount 10 dpc gut (×200). Higher magnification inset shows heterogeneity of transgene expression at wavefront of migrating progenitors nearing the future cecal region (×400). Sacral NC labeled by Sox10-H2BVenus expression is indicated (arrow). (d) Extent of H2BVenus+ ENP progress through the hindgut in a 12 dpc sub-dissected transgenic embryo (×200). (ov, otic vesicle; ba1, brachial arch 1; ba2, brachial arch 2; ce, cecal region; hg, hindgut region; mg, midgut region).

Download figure to PowerPoint

Sox10 is expressed in the initial migratory NC that delaminate from the neural tube and is maintained in mature glia including enteric glia (Kuhlbrodt et al.,1998; Paratore et al.,2001; Young et al.,2003). To establish that the Sox10-H2BVenus transgene is appropriately restricted to specific cell types in the mature ENS, we examined co-localization of H2BVenus with lineage specific markers. Cell-type specific expression of Sox10-H2BVenus was examined by IHC for both neuronal and glial markers in gut muscle strips (GMS) of adult transgenic mice. Expression of the H2BVenus reporter was evident in the nuclei of enteric glial cells, small spindle-shaped cells within and between enteric ganglia, that express Sox10 and S100 (see Fig. 3). Moreover, IHC of transgenic gut muscle strips with neuronal markers PGP9.5 and Hu, revealed that H2BVenus expression is excluded from the enteric neurons, which are evident as larger rounded nuclei within the center of myenteric ganglia (see Fig. 3).

thumbnail image

Figure 3. Sox10-H2BVenus transgene expression is restricted to NC-derived enteric glia in the mature ENS. Gut muscle strips from adult Sox10-H2BVenus transgenic mice stained with antibodies to Sox10, S-100, PGP9.5, and Hu reveal expression of the transgene in mature enteric glia based on colocalization of the H2BVenus reporter with glial markers, Sox10 and S100. Sox10-H2BVenus expression is excluded from neurons labeled by PGP9.5 and Hu. All images ×400.

Download figure to PowerPoint

Sox10-H2BVenus Expression Facilitates Isolation and Analysis of Gene Expression in Multipotent ENPs

The neural crest-derived progenitors that populate the intestine exhibit multipotency and self-renewal characteristic of stem cells (Bixby et al.,2002; Bondurand et al.,2003; Iwashita et al.,2003; Kruger et al.,2002). To determine if expression of the Sox10-H2BVenus transgene would facilitate isolation of enteric neural crest stem cells (NCSC), we investigated the multipotency and self-renewal of H2BVenus+ cells purified from fetal gut by flow cytometry. In dissociates generated from 14.5 dpc fetal gut H2BVenus+ cells accounted for 8.4% ± 1.9% of total events and 16.0% ± 2.7% of viable single cells sorted (n = 18). When Venus+ enteric cells were sorted directly into low-density non-adherent cultures, large homogeneously bright Venus+ neurospheres formed after seven days at a frequency of 0.4% of total cells (Fig. 4a). Cells within these primary neurospheres were able to self-renew yielding an average of 12 secondary neurospheres for each dissociated primary neurosphere. This self-renewal capacity was consistently present in secondary, tertiary, and quaternary neurosphere cultures (data not shown). When single cell suspensions of 14.5 dpc transgene fetal gut were dissociated and immediately subjected to IHC, all H2BVenus+ cells were immunoreactive for p75 and nestin consistent with prior reports of these antigens in NCSC(Kim et al.,2003; Suarez-Rodriguez and Belkind-Gerson,2004) (Fig. 4b,c). When Venus+ gut cells were sorted directly into adherent culture wells at low density in standard conditions (Bixby et al.,2002; Kruger et al.,2002; Walters et al.,2010), 4.5% of all cells formed spatially distinct adherent colonies. Immunostaining of these colonies for cell type lineage markers of neurons (peripherin), glia (glial fibrillary acidic protein, GFAP), and myofibroblasts (smooth muscle actin, SMA) identified multipotent (Fig. 4d) and restricted colony types (data not shown). Of all cells that grew to form colonies 11.6% derived from mutipotent progenitors based on detection of multiple lineages within a single colony, consistent with the frequency of neurosphere formation. Within these colonies we observed that the expression of the Sox10-H2BVenus transgene was extinguished in neuronal and myofibroblast cells as the cultures differentiated, as would be expected with down-regulation of Sox10 expression in these lineages during NC lineage segregation in vivo (Fig. 4d, peripherin and SMA panels).

thumbnail image

Figure 4. ENPs labeled by Sox10-H2BVenus expression exhibit stem cell characteristics. Sox10-H2BVenus+ cells from 14 dpc fetal gut give rise to neurospheres in culture (a) and express the neural crest stem cell markers p75LNTR (b) and nestin (c) detected by Cy3 (red) IHC labeling (×200). Sox10-H2BVenus+ cells isolated from 14 dpc gut grown in low density cultures give rise to multipotent colonies (d) that contain neurons (peripherin+), glia (GFAP+) and myofibroblasts (SMA+) labeled by triple immunofluorescence, ×100 magnification. Detection of myofibroblasts with SMA-FITC resulted in co-visualization of SMA+ cells with H2BVenus+ nuclei in multipotent (SMA panel) colonies. RT-PCR detects expression (e) of stem cell genes (Abcg2, Bmi1, Dll1, Klf4, Lgr5, Msi1, Myc, Nes, Ngfr, Notch1, Pou5f1, Sox2), neural crest genes (FoxD3, Snai1, Snai2, Sox9, Sox10, Twist1), neuronal progenitor marker genes (Ascl1, Neurod1, Neurog1, Phox2b, Prph1, Uchl1), and peripheral glia markers (Erbb3, Fabp7, Gfap, Mpz, Nrg1, Nrtn, Olig2, S100b) in Sox10-H2BVenus+ enteric populations at discrete developmental stages. Housekeeping control genes (Actb, Ipo8, Ubc) are shown for comparison. Lanes include 100 bp molecular weight marker (M), no template control (H2O), total 14.5 dpc fetal mouse RNA (Total E14), flow-sorted 14.5 dpc gut H2BVenus+ ENPs (E14 Sox10+), cultured neurospheres from 14.5 dpc H2BVenus+ gut (E14 NS), postnatal day 6 H2BVenus+ cells isolated from gut muscle strips (P6 GMS).

Download figure to PowerPoint

The ability to purify the Sox10-H2BVenus+ population from the intestine offers an avenue to investigate expression of stem cell-related genes, neural crest cell genes, and neuronal and glial markers at discrete stages of development. We implemented semi-quantitative PCR (Supporting Information Table 1) to allow a comparison of gene expression in H2BVenus+ cells from freshly sorted 14.5 dpc fetal gut suspensions, from cultured neurospheres, and from cells sorted from dissociates of postnatal day 6 (P6) gut muscle strips. We found that stem cell genes Abcg2, Bmi1, Dll1, Klf4, Lgr5, Msi1, Myc, Nes, Ngfr, Notch1, and Sox2 were all detected in 14.5 dpc isolated enteric progenitors, cultured neurospheres, and P6 GMS (Fig. 4e). Pou5f1 (Oct4) was the only stem cell gene not detected in P6 GMS RNA indicating that while this gene is expressed in enteric NCSC, it is down-regulated postnatally. Key neural crest regulatory genes FoxD3, Snai2, Sox9, Sox10, and Twist1 were all detected in 14.5 dpc enteric progenitors and in RNA isolated from cultured primary neurospheres, but Sox9 and Twist1 were down-regulated and undetectable in multiple preparations of P6 GMS. Snai1 was not detected in any of the cell populations evaluated in repeated trials although the positive signal in total RNA from whole 14.5 dpc embryo indicated that the assay for this gene was functional. Neuronal markers Ascl1, Neurod1, Neurog1, Phox2b, Prph1, and UchL1 were all expressed in the 14.5 dpc Venus+ population with only Neurod1 down-regulated in P6 GMS. Expression of glial lineage genes Erbb2, Fabp7, Gfap, Mpz, Nrg1, and S100β were all expressed in freshly sorted 14.5 dpc Venus+ cells. Nrg1 was down-regulated in P6 GMS. Interestingly, Olig2 was the only glial gene absent from 14.5 dpc Venus+ enteric cells, but was clearly upregulated in neurospheres and RNA from freshly isolated P6 GMS. The expression of neural crest genes in 14.5 dpc Venus+ progenitors is consistent with the origin of these cells, while the expression of multiple stem cell genes is consistent with the ability of a subset of these cells to self-renew. Our observation of both neuronal and glial lineage markers in Venus +14.5 dpc population is consistent with ongoing neuro- and gliogenesis that has been described during population of the gut by enteric progenitors.

Sox10-H2BVenus Can be Used to Assess Effect of Mutant Alleles on NC Migration, and Proliferation

Investigating the effect of mutant alleles on NC development has greatly expanded our understanding of individual gene effects. To demonstrate the ready ability of the Sox10-H2BVenus line to facilitate analysis of mutant alleles, we crossed Sox10-H2BVenus transgenics with C3Fe.Sox10Dom/+ mutants (Cantrell et al.,2004) to investigate developmental processes as ENPs are populating the intestine. Migration of ENPs in wild-type and Sox10Dom/+ littermate embryos was evaluated based on H2BVenus reporter fluorescence. The brightness of the H2BVenus reporter allowed ready visualization of ENPs in viable or fixed whole-mount gut samples. We observed that the Sox10Dom allele dramatically delays ENP migration within the gut. In whole-mount 13.5 dpc fetal guts ENP migration in Sox10Dom/+ fetal gut samples was markedly delayed behind that of wild-type littermates in multiple independent litters (Fig. 5a). Within wild-type fetal guts H2BVenus+ ENPs had progressed midway through the hindgut while ENPs in the majority of Sox10Dom/+ embryos had only progressed to the mid-gut with lead migratory cells having reached the cecum in only a few samples. Our observation is consistent with prior reports of delayed ENP migration in Sox10Dom/+ embryos based on histological sections (Lane and Liu,1984; Paratore et al.,2002) or LacZ staining of neuronal transgenes (Kapur et al.,1996).

thumbnail image

Figure 5. Sox10-H2BVenus transgene visualizes effects of Sox10Dom mutation on enteric development in situ. Enteric progenitors labeled by H2BVenus expression exhibit normal extent of migration in wild type 13.5 dpc fetal guts in contrast to delayed migration in Sox10Dom/+ mutant littermates (a, ×100). High magnification confocal images (b, ×400) taken from wild type and Sox10Dom/+ mutant samples in gut regions labeled one to four in Panel a at top to quantify proliferating progenitors. Proliferation of H2BVenus+ ENPs (white circles) is revealed by co-labeling with Sox10 antibody such that a red halo of Sox10 protein in the cytoplasm surrounds condensing H2BVenu+ mitotic chromosomes.

Download figure to PowerPoint

Some models of ENS ontogeny include a prominent role for proliferation at the wave front of the invading cell population and suggest that the proliferation is a driving factor in migration of ENPs down the developing intestine (Landman et al.,2007; Simpson et al.,2006, 2007). Our studies however have indicated that the Sox10Dom/+ mutation does not impact proliferation of ENPs in the fetal gut (Walters et al.,2010), but relied upon dissociation, plating, and staining of cells with subsequent loss of spatial information relative to migratory position within the gut segments examined. The ability to view mitotic cells expressing the Sox10-H2BVenus transgene allowed us to re-examine proliferation of ENPs in Sox10Dom/+ animals in situ based on the co-localization of the H2BVenus reporter with mitotic chromosomes. Cytoplasmic localization of Sox10 detected by IHC concurrent with Venus localization in mitotic chromosomes provides ready visibility in situ of dividing cells from prometaphase through late telophase as shown in Figure 1 and Supporting Information Figure 1. We imaged Sox10-H2BVenus+ cells at four positions in the fetal gut: midgut (Region 1), ileocecal border (Region 2), just after the cecum (Region 3), and mid-hindgut (Region 4; Fig. 5b). Fewer mitotic cells were observed in Sox10Dom/+ fetal guts (n = 10) compared to wild type littermates (n = 9); however, many fewer total ENPs are present in samples carrying the Sox10Dom/+ mutant allele. The numbers of mitotic cells were normalized relative to the total numbers of cells present in each image field. In wild-type fetal gut the percentage of cells exhibiting mitotic figures among all Sox10-H2BVenus+ cells ranged from 1.73 to 2.49% of total cells. In the three most proximal gut regions of Sox10Dom/+ mutants the percentage of cells exhibiting mitotic figures ranged from 0.88 to 2.42% of total cells which was not significantly different from wild-type values. No H2BVenus+ cells were evident in the distal most segments (Region 4) among any of the Sox10Dom/+ mutant samples imaged. These values are similar to those reported in wild-type animals based on phosphoshistone3 immunohistochemical detection of proliferative ENPs (Wallace et al.,2010). Our findings confirm that delayed migration seen in Sox10Dom/+ mutants is not due to reduced proliferation and demonstrates the attractiveness of this transgene reporter for proliferative studies in situ.

Live Imaging of ENP Migration by Sox10-H2BVenus Expression

To view dynamic migration and proliferation of ENPs as they migrate through the developing gut, catenary cultures of fetal intestine from Sox10-H2BVenus transgenic embryos were established. Segments of fetal intestine from just proximal to the cecum to the distal tip of the hindgut were mounted on supporting filter paper and imaged by confocal microscopy. Short time lapse sequences were taken every 15 to 20 min with a 20× lens using 1% laser power to observe behaviors of H2BVenus+ cells (Fig. 6; Movies 1 and 2). Images were captured such that the wave-front of migrating ENPs entered the field of view and progressed across the field as H2BVenus+ cells migrated caudally during the culture period. Although cells most often migrated immediately adjacent to one another, nuclei of individual cells were clearly visible (see Fig. 6). ENPs migrated in chains with leading cells taking paths that were subsequently followed by more proximal cells filling in behind the wave-front. Actively migrating cells exhibited elongated nuclei. When cells divided, their nuclei transiently rounded and brightened as the H2BVenus reporter associated with mitotic chromosomes and then dispersed back into a more diffuse nuclear localization (Fig. 6a, Movie 1). Dividing cells were observed all along the length of the imaged segment: at the wave-front, just behind the wave-front, and well behind the wave-front at the proximal border of the viewing field.

thumbnail image

Figure 6. Proliferation of Sox10-H2BVenus+ ENPs observed by dynamic imaging of fetal gut. Nuclei of individual cells labeled by H2BVenus expression are apparent in selected frames of the hindgut from a 12.5 dpc transgenic mouse. A single arrow highlights an ENP that undergoes cell division between 80 and 100 min of the recording with daughter cells indicated by two arrows. Other mitotic spindles are highlighted (asterisk). Scale bar, 20 μm. A complete frame series (Movie 1) of this preparation is available at the Genesis website.

Download figure to PowerPoint

Previously, apoptosis of ENPs has not been reported in dynamic imaging studies. Multiple efforts have been made to document apoptosis in the development of the ENS using static approaches that capture a snapshot of cell death by immunohistochemistry for activated caspases (Gianino et al.,2003; Kruger et al.,2003; Wallace et al.,2009, 2010). In sectional analysis that imaged cross sections through the gut (Gianino et al.,2003; Kruger et al.,2003) no caspase3+ cells were identified. However when the fetal gut was imaged in whole-mount so that the entire length could be imaged, rare caspase3+ cells (∼0.2% of total Sox10+ cells) were observed (Wallace et al.,2010). The rarity of apoptotic events among ENPs migrating in the fetal gut poses a substantial challenge for detailed analysis of this process. Imaging of nuclear-localized reporters may circumvent this difficulty. We observed nuclear fragmentation and dispersal of H2BVenus+ fragments by time-lapse imaging that are highly reminiscent of apoptosing progenitors observed in zebrafish (Lyons et al.,2005; van Ham et al.,2010). A small proportion of the migrating H2BVenus+ ENPs would suddenly and dramatically fragment their nuclei while other cells around them continued to divide and migrate normally (Fig. 7a,b, Movies 1 and 2). This behavior was observed at very low frequency, 0.7–3.5% of total Sox10-H2BVenus+ cells, within the viewing field imaged every 15 to 20 min and was evident at a variety of locations along the imaged segment. The residual nuclear fragments labeled by H2BVenus persisted in the viewing field over variable periods of time from 80 to 200 min. When we counted the number of initiating nuclear fragmentation events relative to the total lapsed time of the imaging session, the average rate of nuclear fragmentation events was found to be 0.23 ± 0.12 events per hour. Nuclear fragmentation occurred both in cells that were at the wave front and in those that had just moved into the viewing field (Movie 1) as well as in interphase cells and those that had just completed mitosis. To exclude the possibility that the nuclear fragments were due to phototoxicity, we collected H2BVenus+ fetal gut samples, fixed them, and then mounted them for confocal imaging. Even in samples that had not previously been illuminated, we observed rare nuclear fragmentation events. Nuclei were evident that had already fragmented or that were just beginning to coalesce into nuclear fragments while nearby cells appeared normal (Fig. 7c). This behavior strongly suggests that cell elimination does occur among migrating ENPs populating the intestine. Given the role of programmed cell death in neurons and proliferating neural precursors (Enomoto,2009) and a recent report that that inhibition of cell death leads to hyperganglionosis of the foregut (Wallace et al.,2009), this elimination process likely plays a critical role in normal development of the ENS.

thumbnail image

Figure 7. Elimination of Sox10-H2BVenus+ ENPs observed during dynamic imaging. Rapid fragmentation of cell nuclei that occur both at the wavefront (a) and behind the wavefront (b) are shown in selected frames of two different imaging sessions. Circles indicate the position of cell nuclei before (a, 100′; b, 280′), during, and after fragmentation. A complete frame series of the imaging session shown in Panel b (Movie 2) is available at the Genesis website. Nuclear fragmentation events were also identified in multiple independent fetal gut samples (d) that had been not illuminated prior to fixation and mounting as indicated by arrows. Scale bar Panels a and b, 20 μm; scale bar panels shown in c, 10 μm.

Download figure to PowerPoint

Sox10-H2BVenus Expression Labels NC in Multiple NCSC Locations

NC-derived progenitors with stem cell characteristics similar to those of enteric NCSC have been reported in the cornea (Yoshida et al.,2006), skin (Fernandes et al.,2004; Sieber-Blum et al.,2004; Wong et al.,2006), and sciatic nerve (Bixby et al.,2002; Morrison et al.,1999). Certainly NC derivatives contribute to a wide variety of developing structures in the embryo. To determine if the Sox10-H2BVenus transgene drives expression of the H2BVenus reporter in locations consistent with known locations of NCSC, we surveyed tissues of transgenic mouse embryos at 14.5 dpc. Sox10-H2BVenus expression was readily apparent in multiple aspects of the PNS including cranial ganglia, dorsal root ganglia, superior cervical ganglia, oligodendrocyte precursors in the spinal cord, and nerve tracts in the digital branches of the radial and ulnar nerves in the forelimb as well as in the developing diaphragm. (Figs. 1 and 8 and data not shown). H2BVenus reporter signal was evident in the cornea where it is known that NCSC reside (Yoshida et al.,2006) and play an essential role in corneal fibril organization (Takatsuka et al.,2008). Moreover, sections through 14.5 dpc eye revealed expression in nearby lacrimal glands as well, which are thought to be of neuroectodermal origin (Tonks et al.,2003; Tripathi and Tripathi,1990). We also observed expression surrounding embryonic whisker follicles and at the base of hair follicles in adult ear punch biopsies (data not shown). These expression patterns are consistent with known roles of Sox10 in PNS and melanocyte development.

thumbnail image

Figure 8. Sox10-H2BVenus transgene labels NC-derived progenitors in multiple aspects of the PNS. Expression of the Sox10-H2BVenus transgene is apparent at 14.5 dpc in other regions where NC contribute to developing structures including the cornea (a), follicles in the whisker pad (b), oligodendrocytes in the spine and glia of the dorsal root ganglia (c), nerve tracts in the forelimb (d), and innervation of the diaphragm (e).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Live imaging studies are often challenged by technical difficulties due to low fluorescence of reporter constructs or close apposition of migrating cells that are difficult to discern when cytoplasmic fluorophores are used. We present the characterization of a new transgenic allele of Sox10 that drives bright, stable, nuclear fluorescence of Venus, Sox10-H2BVenus. We have fully validated the spatial, temporal, and cell-type specific expression of this transgene during ENS ontogeny in the mouse. We demonstrate the usefulness of the Sox10-H2BVenus allele for isolation of neural crest stem cells and analysis of gene expression in enteric lineages at discrete stages of development. Given the ubiquitous expression of Sox10 in NC progenitors, this tool opens avenues for live cell imaging and studies of NC-derivatives in multiple tissues. In our own dynamic imaging studies of ENP migration, we demonstrate the value of being able to track individual nuclei and have identified unexpected nuclear fragmentation of ENPs that suggests this population does experience natural pruning through processes involving cell death.

Sox10 is an exciting entry point to target for directing reporters of NC migration. The gene initiates expression just as neural crest progenitors delaminate off the neural tube, is maintained in migrating progenitors throughout the embryo, and continues expression in peripheral glia. Several alleles that target the Sox10 locus have been derived to facilitate analyses of NC (Britsch et al.,2001; Ludwig et al.,2004), however these rely upon gene targeting of the locus to integrate reporter genes and result in subsequent haploinsufficiency of Sox10 protein. Because NC development is exquisitely sensitive to levels of Sox10, haploinsufficiency results in deficits across multiple NC lineages and is not ideal for monitoring normal NC processes. Transgenic approaches attempting to label NC populations have been implemented using isolated Sox10 enhancer elements that drive Cre (Stine et al.,2009), but a bright stable fluorescent reporter that facilitates imaging of discrete individual cells from Sox10 regulatory regions has not been achieved previously. The allele we describe here contains regulatory elements needed to recapitulate the complete expression profile of Sox10, enables live imaging of cells in dynamic studies, and has proven genetically stable by passage through multiple generations of transgenic animals.

Our IHC results demonstrate that the Sox10-H2BVenus transgene becomes appropriately restricted to mature enteric glia in the ENS, a level of rigorous characterization that has not previously been shown for other Sox10 alleles. In addition it facilitates identification of cells in the ENS based on nuclear morphology. Glial cells in the mature ENS express Sox10, S100, GFAP, and BFABP (Hoff et al.,2008; Young et al.,2003). While intimately associated with enteric neurons, enteric glia exhibit smaller, spindle-shaped nuclei that clearly contrast with the large rounded nuclei of adjacent neurons. In our analysis of postnatal GMS containing myenteric ganglia, we observed two distinct types of enteric glia that are readily discriminated by nuclear morphology and location. H2BVenus+ enteric glia that are present within enteric ganglia exhibit smaller somewhat oval nuclei. H2BVenus+ enteric glia outside the ganglia that were associated with small interconnecting fibers exhibit extremely elongated, thread-like nuclei. This morphological distinction is consistent with IHC and electrophysiological reports that suggest heterogeneity among enteric glia (Cabarrocas et al.,2003; von Boyen et al.,2004). Ready visual localization of these distinct cell types based on H2BVenus labeling of their nuclei will facilitate future developmental, functional, and pharmacological studies of enteric glia.

Expression of the Sox10-H2BVenus transgene allows imaging of all NC-derived progenitors in the gut that are expressing or have recently expressed Sox10. This is particularly advantageous because it enables earlier isolation of ENPs instead of limiting efforts to those that have progressed to expression of known enteric neuronal markers like Tyrosine Hydroxylase and Dopamine β-Hydroxylase. Moreover the retention of H2BVenus label for a time in chromatin of differentiating ENPs offers an avenue for capture of distinct populations as they undergo lineage commitment and begin differentiating. We observed both bright and dimly labeled H2BVenus+ ENPs in the 10.0 dpc fetal mouse gut. Such differential expression has been observed for another critical transcription factor in ENS development, Phox2b (Corpening et al.,2008). Prior reports indicated variable expression of Ret-TGM (Young et al.,2004) and Ret antibody staining (Schiltz et al.,1999; Young et al.,2003) among ENPs, however it was not evident at the time that this differential expression might reflect changes in gene regulation as enteric NC lineages diverge. The opportunity to capture and define gene regulatory networks that control processes of enteric glial cell differentiation is particularly exciting as relatively little is known about the genes that participate in controlling development of this lineage. Indeed our initial results identified expression of Olig2 in P6 H2BVenus+ cells sorted from the gut suggesting a role for this oligodendrocyte transcription factor in postnatal maturation of enteric glia.

Most dramatically our dynamic imaging of the Sox10-H2BVenus expression in the fetal gut revealed the advantage of nuclear localized reporters for visualizing cell elimination in the ENS. Cells exhibiting nuclear fragmentation were observed at a relatively low frequency, 0.7–3% of total Sox10-H2BVenus+ cells in each viewing field exhibited fragmenting nuclei in images taken every 15 to 20 min. Multiple efforts have been made previously to identify apoptosis of ENPs in the developing intestine using static methods (Gianino et al.,2003; Kruger et al.,2003; Maka et al.,2005; Wallace et al.,2009, 2010). Only a single study has previously identified apoptotic cells within the walls of the fetal gut using static methods and established that ∼0.2% of all Sox10+ cells were caspase3+ throughout the length of the fetal gut at 11.5 dpc (Wallace et al.,2010). The frequency we determined is also very low. By counting the number of times new nuclear fragmentation events appeared during imaging sessions relative to total cells visible in the imaging field, we determined that 1.2% ± 0.6% of total Sox10-H2BVenus+ cells were being eliminated, consistent with the frequency reported by Wallace et al. The dramatic dispersal and then persistence of H2BVenus+ nuclear fragments facilitates the identification of cells being eliminated. Use of such nuclear-localized reporters is likely to facilitate future analyses of the cell behaviors in the vicinity of the fragmenting progenitors.

Initially we were concerned that this nuclear fragmentation phenomenon could be attributed to localization of the H2BVenus label in the cell nucleus that would make cells more susceptible to cell death, although detrimental effects of chromatin fluorescence have not been reported previously in dynamic imaging studies (Fraser et al.,2005; Hadjantonakis and Papaioannou,2004; Nowotschin et al.,2009). However, over the course of numerous imaging sessions we noted that fragmenting cells were observed at a variety of locations along the length of the image field and not just in cells at the wave-front that had spent the longest period of time under illumination. We did not observe an increase in rate of fragmentation over the course of imaging sessions. We noted nuclear fragmentation occurred both in cells that had recently divided as well as those that had not divided and was not more frequent in those cells that had recently concentrated the H2BVenus label into mitotic chromosomes. These nuclear fragments were present only transiently but were readily apparent in our preparations due to the continued association of H2BVenus with chromatin. Finally, we confirmed that nuclear fragmentation events occur in gut preparations that have not been exposed to the illumination that is required for live imaging by identifying nuclear fragmentation in freshly fixed preparations (Fig. 7c). This cell elimination process could have been missed in prior studies that visualized migration of ENPs due to imaging with cytoplasmic fluorophores (Druckenbrod and Epstein,2005; Young et al.,2004) that might not stay associated with cellular debris. The ability to visualize cell elimination during population of the ENS is tantalizing and raises multiple questions about the signaling pathways that mediate this process. It has been established that cell death is an important process in the ENS using dominant negative forms of caspase-9 in the chick (Wallace et al.,2009). However, other efforts to identify apoptotic cells in gut cross-sections from normal fetal mice did not find evidence of caspase-3 mediated programmed cell death, nor were enteric neuron numbers reduced in Bax-/- or Bid-/- mutants (Gianino et al.,2003). When the entire length of the fetal gut was examined in whole-mount, infrequent caspase-3+ cells in the ENS were identified at very low frequencies (Wallace et al.,2010). Our live imaging studies are consistent with cell elimination occurring during ENS ontogeny and based on nuclear fragmentation identified slightly higher frequencies than those reported by Wallace et al. The slightly higher incidence we report may be due to the means of detection (visualizing nuclear fragmentation independent of any immunohistochemical labels) or may indicate that multiple signaling pathways participate in cell elimination to sculpt the ENS. Future studies should discriminate between these two possibilities. Moreover, the ability to intrinsically label cells being eliminated during ENS development offers the opportunity to determine how Ret receptor signaling plays a role in controlling enteric neuron numbers since it has been suggested that cell death is an event that contributes to Hirschsprung disease etiology (Uesaka et al.,2008).

The Sox10-H2BVenus transgene offers the advantage of capturing NC progenitors during normal development as these cells undergo lineage restriction and other developmental processes, but our studies also make apparent the need for additional tools to be able to track lineages visibly after down-regulation of Sox10. Cre fate mapping can be temporally controlled with tamoxifen, but many studies would benefit from the ability to map descendents of cell populations that can be discriminated based upon location in a tissue or threshold levels of gene expression. Laser excitation using photoconvertible fluorescent proteins, like Kikume-GR (Kulesa et al.,2008; Nowotschin and Hadjantonakis,2009), allows groups of cells to be tracked following exposure to discrete wavelengths, but the cytoplasmic fluorescence is rapidly diluted in rapidly dividing progenitors populations like NC derivatives. Photoactivatible forms of Cre that could indelibly mark cells subsequent to excitation (Edwards et al.,2009) would greatly facilitate lineage tracing in the ENS and other NC derivatives where cells are closely juxtaposed. In the meantime, straightforward use of bright stable fluorescent alleles, like the Sox10-H2BVenus we describe, offer access to initial processes of lineage segregation in the ENS and other NC derivatives.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Animals

All animal protocols were approved by the Institutional Animal Care and Use Committee at Vanderbilt University. Timed matings were set to obtain staged mouse embryos, designating the morning of plug formation as 0.5 days post coitus (dpc). At the desired time point, intact embryos were dissected, fixed overnight in NBF at 4°C, and stored for up to a week in the dark at 4°C without significant loss of H2BVenus fluorescent signal.

Generation of the Sox10-Histone2BVenus BAC Transgene

The Histone2BVenus fusion protein coding sequence was introduced into a well-characterized BAC clone, 28O11, using established in vitro recombineering methods (Lee et al.,2001). The ATG initiator codon of H2BVenus was fused in frame with the ATG initiator methionine of the Sox10 gene similarly to introduction of prior reporter modules (Deal et al.,2006) and was implemented so as not to delete any Sox10 gene sequences. Appropriate integration of the H2BVenus reporter was confirmed by direct sequencing of the BAC clone in conjunction with restriction enzyme finger-printing. Full details of the modification steps are provided in Supporting Information.

In Vitro Transfection of the Sox10-H2BVenus BAC

Functionality of the BAC construct was tested by transient transfection into cultured enteric glial cells (EGC; Ruhl et al.,2004). EGCs were transfected in 24-well plates at ∼50% confluency using Lipofectamine (Invitrogen) as described (Montigny et al.,2003). At 36-h post transfection, cells were then fixed in neutral buffered formalin (NBF) on ice for 10 min, washed and imaged by fluorescence microscopy.

Generation of Sox10-H2BVenus BAC Transgenic Lines

Sox10-H2BVenusFP BAC DNA was prepared by CsCl banding and subsequent dialysis into microinjection buffer (10 mM Tris, pH 7.5; 0.1 mM EDTA, 10 mM NaCl, 30 mM spermine, and 70 mM spermidine). The modified BAC was injected into fertilized mouse eggs obtained from the mating of C57BL/6J females with (C57BL/6J X SJL)F1 males using standard protocols (Camper et al., 1995) and founder transgenic pups were identified by PCR amplification of three independent sequences as previously described (Deal et al.,2006): Primers complementary to the Histone2B-Venus moiety of each construct were utilized to amplify the “internal” BAC sequence (5′-CTGGTCGAG CTCGACGGCGACGTA-3′ and 5′-AGTCGCGGCCGCTT TACTTG-3′), while primers spanning the vector backbone into the BAC insert were used to detect the SP6 (5′-GGCACTTTCATGTTATCTGAGG-3′ and 5′-GTTTTTTGCGATCTGCCGTTTC-3′) and T7 (5′-TCGAGCTTGACATTGTAGGAC-3′ and 5′-AAGAGCAAGCCTTGGAACTG-3′) construct arms. Transgene copy numbers were determined in offspring by using a semi-quantitative PCR assay (data not shown) in comparison to dilution standards containing 1, 2, 4, and 8 copies of each BAC/genome equivalent using previously described methods (Deal et al.,2006). Transgene copy numbers ranged from two to eight copies per genome. Simple tandem repeat (STR) markers were used to screen for integrity of the Sox10 WT28O11 BAC backbone as previously described (Deal et al.,2006).

Immunohistochemistry and Imaging of H2BVenus+ Cell Types

For analysis of postnatal gut muscle strips gut pieces were flushed with PBS and fixed for 20 min in NBF with 0.5% Triton X-100 the processed for immunohistochemical detection as described (Walters et al.,2010). Antibodies used included: Sox10 (N20; Santa Cruz, 1:25); S-100 (DAKO, 1:500), PGP9.5 (Biogenesis, 1:4,000); and Hu (Fairman et al., 1995, 1:800). Secondary immunoreagents consisted of Cy3-conjugated donkey anti-rabbit (Jackson Immuno Research; 1:1,000) and donkey anti-human Texas Red (Jackson Immuno Research; 1:100). After immunolabeling muscle strips were flat-mounted in Aqua Poly/Mount (PolySciences) and coverslipped for imaging on a Zeiss Scanning Microscope LSM510 using a YFP long-pass (LP) 530 filter to visualize the fluorescent Venus reporter present in this transgene, as well as a rhodamine (Cy3, Texas Red) LP 560 filter to visualize endogenous protein stained with antibodies in gut muscle strips.

Images of sub-dissected 14.5 dpc organs were captured on a Zeiss Stereo Lumar.V12 fluorescence stereomicroscope equipped with a Q-Imaging 4000R digital camera and software. Additional images were obtained on an Olympus BX41 equipped with a DP70 camera and software.

Flow Cytometric Isolation of H2BVenus+ ENPs

C3Fe.Sox10-H2BVenus mice were mated to C3HeB/FeJ wild-type mice to generate transgenic embryos at 14.5 dpc. Litters were screened for expression of the transgene under a fluorescent stereo-dissecting microscope and segregated into wild-type and transgenic pools. Fetal guts were dissected, from the proximal stomach to the anus, and dissociated into a single cell suspensions as previously described (Walters et al.,2010). Viability dye was added to cell suspensions in staining media (7-AAD, 1:1,000 dilution, Life Technologies) prior to flow cytometric isolation to exclude dead cells.

Cell sorts and analyses were performed on a BD FACSAria flow cytometer as previously described (Walters et al.,2010) using cell suspensions from non-transgenic littermate embryos to set compensation controls. YFP+ cells from transgenic fetal guts were sorted at low density (500 cells per well) into self-renewal medium within six-well tissue culture plates coated with poly-D lysine and fibronectin using established methods (Bixby et al.,2002; Kruger et al.,2002; Walters et al.,2010). NCSC were cultured in self-renewal media at 37°C in gas-tight Billups-Rothenberg chambers to achieve near-physiological levels of oxygen (Morrison et al. 2000), replenishing with a mixture of 1% O2/5% CO2/balance N2 every 3–4 days to an O2 level of 3–4%. After 7 days, the medium was changed to differentiation medium containing decreased concentrations of growth factors and chick embryo extract (Bixby et al.2002). Following 7 days in differentiation medium, immunohistochemistry (IHC) was performed to determine the composition of resulting colonies as previously described (Walters et al.,2010). Cultures were washed, fixed, and stained with antibodies to Peripherin (Chemicon, 1:1,000), GFAP-Cy3 (Sigma, 1:800) and αSMA-FITC (Sigma, 1:800). Donkey anti-rabbit Cy5-conjugated secondary antibody (Jackson Immuno Research, 1:1,000) was used to visualize peripherin labeling. Plates were post-fixed and counterstained with DAPI to label all nuclei. Colony phenotypes were assigned based on IHC detection of peripherin (neurons), GFAP (glia), and myofibroblasts (αSMA). Multipotent colonies were identified by the presence of all three lineage markers within a single isolated colony while bi-potent colonies contained only two cell types and in the case of uni-potent colonies only a single cell type was present.

Neurosphere Culture

Neurospheres were cultured from H2BVenus+ cells by sorting 1,000 cells into self-renewal media (Bixby et al., 2000) in uncoated nonadherent tissue culture plates (Costar). Cultures were maintained in Billups-Rothenberg chambers at 3–4% O2 levels as described (Morrison et al.,1999). Growth factors (bFGF and IGF-1) were replenished to the initial concentration (10 and 20 ng ml−1, respectively) every 3 days for a total culture period of 14 days.

RNA Isolation

Total RNA was isolated using methods modified from Iwashita et al. (2003). Specifically, H2BVenus+ cells were sorted directly into TRIzol-LS (Life Technologies) with carrier glycogen added to a final concentration of 250 μg ml−1. Subsequent processing was performed according to the manufacturer's instructions. RNA was DNAseI-treated (Ambion) to remove contaminating genomic DNA. After DNAse treatment, RNA samples were diluted into RLT buffer with poly-D-inosine (Epicentre) added as carrier and purification performed according to the manufacturer's instructions (Qiagen RNeasy, Qiagen). RNA isolates included: 14.5 dpc total embryo (control); three biological replicates of Sox10-H2BVenus+ cells sorted from 14.5 dpc transgenic gut; three biological replicates of neurosphere RNA cultured from 14.5 dpc flow-sorted Sox10-H2BVenus+ transgenic guts; and three biological replicates of Sox10-H2BVenus+ cells sorted from postnatal day 6 (P6) guts. Integrity of RNA isolates was assessed by electrophoresis on Agilent 2100 Bioanalyzer PicoChips.

Semi-Quantitative Gene Expression Analysis

To evaluate the presence of mRNAs for neural crest, stem cell, neuronal and glial genes in the limiting quantities of RNA obtained from flow-sorted cells, TaqMan Assay-On-Demand (AoD) gene expression assays were implemented using the PreAmp Master Mix System (ABI). cDNA was synthesized from 25 ng of total RNA by reverse transcription in a reaction containing dNTPs (Invitrogen), OligodT23VN (NEB), Random Hexamers (NEB), RT Buffer (ABI), RNasin (Promega), and MultiScribe Reverse Transcriptase (ABI). Before adding Reverse Transcriptase to the reactions, a small aliquot of the total volume was removed to serve as the minus RT control. Reactions were incubated at 37°C for 2 h. TaqMan Assay-on-Demand (AoD) gene expression assays (ABI) to be evaluated were combined such that the final concentration of each assay was 0.2×. TaqMan PreAmp Master Mix (ABI) was added to the pooled assay mixture and this master mix was divided and distributed among the different cDNA samples to perform the preamplification reaction. The preamplification thermocyling was performed as follows: a hold of 95°C for 10 min, 14 two-step cycles of 95°C for 15 s, 60°C for 4 min. Upon completion, reactions were immediately placed on ice and the preamplification product was diluted 1:20 with 1X TE Buffer. Aliquots were stored at −20°C for use in subsequent PCR reactions using TaqMan AoD gene expression assays. Each preamplified cDNA sample (1:20 dilution) was evaluated with 33 experimental and three control TaqMan AoD Gene Expression Assays (see Supporting Information Table 1). The reactions contained 2X TaqMan Universal PCR Master Mix (ABI), 20× AoD gene expression assay, water and the appropriate preamplified cDNA sample. To increase the specificity for three of the AoD gene expression assays (Pou5f1, Foxd3, and Gfap), DMSO (1M final concentration) or Betaine (5% final concentration) was added to the reaction mix. Amplification was achieved with a thermoprofile of: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 15 s, 60°C for 1 min. PCR products were subsequently electrophoresed on 10% acrylamide gels to confirm amplification of the appropriately sized gene product.

In Situ Imaging of Proliferation

Fetal intestines were sub-dissected from 13.5 dpc embryos derived through crosses of Sox10-H2BVenus transgenics with C3Fe.Sox10Dom mutants. H2BVenus+ embryos were collected and processed individually for collection of fetal gut and tail samples for genotyping. Sub-dissected gut tissue was fixed in NBF on ice and stained with anti-Sox10 (N20; Santa Cruz, 1:25) to visualize endogenous Sox10 protein. Samples were mounted and imaged using a Zeiss Scanning Microscope LSM510. The number of cells with visible mitotic spindles was determined in each field and a mitotic ratio calculated based on the total number of cells in that particular field. Mean values for the proliferation index were calculated within genotype classes (n = 10, Sox10Dom; n = 9 wild type) and P values were generated using Welch's T-test to account for unequal sample size and unequal variance. Although the wavefront migration was noticeably delayed in Sox10Dom mutants, all samples were subsequently genotyped by routine methods to confirm the identity of those carrying the Sox10Dom allele (Southard-Smith et al.,1998).

Live Imaging of Sox10-H2BVenus+ Fetal Gut

Fetal gut samples were harvested at 12.5 dpc and subdissected from the ileocecal junction to the anus. These gut segments were mounted on nitrocellulouse membranes (Millipore HABGO47S1) in glass-bottom dishes (MatTek) with 10% FBS, 1% Penicillin/Streptomycin in Phenol Red-free DMEM for optimal imaging. MatTek dishes were placed in a humidified environmental chamber with 5% CO2 on a heated stage at 37°C for imaging on a LSM 510 Meta Confocal. The average step interval was 5 μm. The average z-stack collected was 175-μm thick and 29–49 images were collected in 15 to 20 min intervals then stacked at each time point. Images were processed using LSM Image Browser software (Version 3.5.0.376). Each preparation was imaged for 7–16 h.

Paraffin Embedding of Sox10-H2BVenus+ Tissues

Fixed tissues were processed through a graded alcohol series and then treated with Citrisolve (Fisher Scientific) for two separate periods of 30 and 40 min followed by a 1-h incubation in Paraplast X-tra Paraffin (Shandon). Tissues were paraffin infiltrated by a 2-h incubation in 60°C vacuum oven before embedding. Paraffin sections were mounted on glass slides, deparaffinized in Citrisolve, rehydrated through a graded series of alcohols, rinsed in 1× PBS, and mounted in Aqua Poly/Mount for imaging.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

The authors thank Karen Joyce Dunn and Dr. Chris Wright for encouragement and critical conversations during the conception of these studies. They thank Dr. Neil Copeland for the EL250 E. coli strain and BAC recombineering methods; Dr. Atsushi Miyawaki for providing the Venus coding sequence, Catherine Alford for suggestions on cell dissociation methods, flow cytometry support, and immunohistochemical analysis of NCSC colonies; Kevin Weller, David Flaherty, and Brittany Matlock for support in the Flow Cytometry Shared Resource at Vanderbilt University; Drs. Sean Morrison and Jack Mosher for sharing NCSC isolation and culture methods; Dr. Miles Epstein for generously providing Hu antibody; Dr. Heather Young for technical advice in establishing catenary culture conditions. They also thank Drs. Thomas Saunders and Maggie Van Keuren for outstanding support in generation of BAC transgenics at the University of Michigan Transgenic Animal Model Core. They gratefully acknowledge and thank Dr. Sam Wells and the support staff of the Cell Imaging Shared Resource Core at Vanderbilt for advice and assistance in confocal imaging. The Cell Imaging Shared Resource Core is supported by NIH grants CA68485, DK20593, DK58404, HD15052, DK59637 and EY08126. The VMC Flow Cytometry Shared Resource is supported by the Vanderbilt Ingram Cancer Center (P30 CA68485) and the Vanderbilt Digestive Disease Research Center (DK058404). This work was supported by funding from a US National Institutes of Health grant from NIDDK (DK064251) and March of Dimes Research Award (FY-06-390) to E.M.S2. J.C.C. was supported by a Vanderbilt Brain Institute Scholar's Award.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. METHODS
  7. Acknowledgements
  8. LITERATURE CITED
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVG_20748_sm_suppinfotable1.xls22KSupporting Information Table 1.
DVG_20748_sm_suppinfofig1.tif9368KSupporting Information Figure 1.
DVG_20748_sm_suppinfo.doc38KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.