|Home | About | Journals | Submit | Contact Us | Français|
The Ascl3 transcription factor marks a subset of salivary gland duct cells present in the three major salivary glands of the mouse. In vivo, these cells generate both duct and secretory acinar cell descendants. Here, we have analyzed whether Ascl3-expressing cells retain this multipotent lineage potential in adult glands. Cells isolated from mouse salivary glands were cultured in vitro as non-adherent spheres. Lineage tracing of the Ascl3-expressing cells within the spheres demonstrates that Ascl3+ cells isolated from adult glands remain multipotent, generating both duct and acinar cell types in vitro. Furthermore, we demonstrate that the progenitor cells characterized by Keratin 5 expression are an independent population from Ascl3+ progenitor cells. We conclude that the Ascl3+ cells are intermediate lineage-restricted progenitor cells of the adult salivary glands.
Saliva produced by the three major pairs of salivary glands is critical for many facets of oral function and health. Radiation treatment of head and neck cancers, drug therapies, and autoimmune diseases, such as Sjögren’s syndrome, all inflict damage on the salivary glands, causing a significant decrease in saliva production. Although there is currently no satisfactory treatment to reverse salivary gland damage, intense interest is now focused on bio-therapeutic approaches such as cell replacement, or the development of artificial salivary glands. A high priority for both approaches is the identification of suitable donor cells with the ability to repopulate damaged or artificial glands, in order to restore salivary function.
In a search for cells responsible for salivary gland maintenance and renewal, several candidate progenitor cell populations have been identified in the mouse [1–5]. For the progenitor populations expressing the markers cKit, α6β1 integrin, Ascl3 or Keratin 5, there is direct evidence that these cells have either multi-lineage or regenerative potential within the salivary glands [1, 2, 4, 5]. It is not yet clear how these cell groups are related, but the prevailing hypothesis is that multiple progenitor cell types are involved in salivary gland development, maintenance and repair [6, 7].
There are three identified members of the mammalian achaete scute homolog (Ascl) gene family. All encode transcription factors expressed by tissue-specific progenitor cells, and are implicated in cell fate determination and differentiation events [8, 9]. Progenitor cells characterized by expression of Ascl3 are present in all three major salivary glands of the mouse . During prenatal development of the glands, only a small number of epithelial duct cells express the Ascl3 transcription factor . Lineage tracing showed that these Ascl3-expressing (Ascl3+) cells are progenitors of both duct and acinar cells in the mature glands . These data suggested that Ascl3+ progenitor cells are involved in normal maintenance and renewal of the salivary glands. However, the lineage tracing was based on a constitutively active Cre recombinase. Thus, it was not clear if acinar and duct cells are generated from Ascl3+ progenitors only during gland development, or also in the adult gland. In order to directly test whether the Ascl3+ cells are progenitor cells in the adult gland, we have used an in vitro culture system to generate salivary gland spheres .
In vitro formation and culture of non-adherent spheres from single-cell suspensions has been reported for many tissues, including the salivary glands [5, 11–13]. The cultivation of non-adherent spheres in serum-free media permits experimental characterization of cell proliferation and differentiation, and can be used to demonstrate the presence and viability of undifferentiated stem or progenitor cells within adult tissues (reviewed in ). Salivary gland spheres have been shown to include stem cells able to restore secretion and tissue regeneration in radiation-damaged salivary glands of mice . Using the sphere assay, we have analyzed the differentiation capacity of Ascl3+ cells isolated from adult salivary glands, and have investigated the relationship of Ascl3+ progenitors to the Keratin 5 progenitor cell population.
Spheres are formed after culturing dissociated cell suspensions from adult mouse salivary glands for two to three days in serum-free media . While a majority of the cells undergo cell death, we obtained approximately 3×103 spheres from 1.5×106 dissociated cells of an adult submandibular gland. Cell-mixing experiments were conducted to ascertain that sphere expansion is not due to cell aggregation. Spheres were generated using dissociated salivary gland cells from wild type C57Bl/6 and from CAG-EGFP mice. The latter express EGFP in a ubiquitous manner, driven by the chicken beta-actin promoter . Cell suspensions of submandibular glands were separately prepared from same-sex animals, to rule out sexually dimorphic differences present in rodent salivary glands . Cells were mixed and plated in a 1:1 ratio of each genotype. At 3, 5 and 8 days after plating, the spheres were scored based on their composition: all wild type cells, all GFP+ cells, or mixed wild type and GFP+ (data not shown). The majority of the spheres (>80%; 884/1056 counted; P=0.0001) were exclusively either wild type or GFP+. We conclude that salivary gland spheres form predominantly from proliferation of component cells, and only minimally from random aggregation of cell clumps.
BrdU incorporation for 24 hours has been used to show that cells within the spheres are undergoing proliferation . Because BrdU is retained in labeled cells through several divisions, we repeated this experiment, shortening the labeling time to 3 hours. In spheres treated at days 3, 5 and 7 of culture, cells located on the outer periphery of the spheres were rarely labeled with BrdU after the 3-hour labeling period. However, BrdU was incorporated into proliferating cells located at the center of the spheres (Figures 1A and B). This is consistent with earlier reports [5, 17] and suggests that proliferating cells are located within the spheres.
We examined the cellular composition of spheres formed from dissociated submandibular gland by immunostaining with lineage-specific markers for differentiated cell types. Mist1, a transcription factor, is expressed exclusively in differentiated serous acinar cells of the salivary glands [10, 18]. Antibody to Mist1 was used to stain submandibular gland spheres after 7 days of culture, and detects Mist1-positive cells localized at the peripheral surface of the submandibular gland spheres (Figures 1C and D). Acinar cells of the salivary glands also express the water channel, aquaporin 5 (AQP5) . Antibody to AQP5 also labels cells at the periphery of the spheres (Figures 1E and F), indicating that differentiated acinar cells are localized to the outside of the spheres. Antibody to Keratin 8 (Krt8), a cytoskeletal component of duct cells in the submandibular and parotid glands [20, 21], also consistently labeled cells that were on the outer periphery of the spheres (Figures G and H). These data suggest that the SG spheres display distinct cellular organization, with proliferating cells inside and differentiated cells at the periphery.
Salivary gland spheres have been shown to contain undifferentiated stem or progenitor cells . We used RT-PCR to monitor the expression of several putative stem cell markers in spheres at days 0, 2, 3, 4 and 7 of culture. Prior to day 3, spheres are not distinguishable from cell aggregates, and RNA samples most likely include non-sphere cells. Primers to LE32, a 60S ribosomal protein, served as controls. The expression level of four genes known as stem or progenitor cell markers, cKit, Sca1 , Keratin 5 (Krt5, ), and Notch1 , remained relatively constant within the spheres over time in culture (Figure 1I). Unexpectedly, Ascl2, a member of the achaete scute-like gene family and a stem cell-associated factor in the intestine , is not expressed in such a constant manner. Instead, Ascl2 is activated in spheres only after more than 3 days of in vitro sphere culture. Ascl2 expression is found predominantly in the salivary gland duct cells of developing embryos and is significantly upregulated in adult glands (http://sgmap.nidcr.nih.gov/sgmap/sgexp.html; our unpublished observations). The increase in Ascl2 expression within the spheres over time could reflect differentiation to duct cells, and suggests that Ascl2 is not a progenitor cell marker in the salivary gland. We also note an increase in the expression of enhanced yellow fluorescent protein (EYFP), a lineage reporter for descendants of Ascl3+ cells (see Figures 3D and E).
To determine if Ascl3+ progenitor cells are included in the salivary gland spheres, the Ascl3 EGFP-Cre mouse strain was used for sphere formation. In this strain, Ascl3+ cells express EGFP driven by the endogenous Ascl3 promoter . At day 7 of culture, EGFP-positive cells were present in spheres derived from submandibular (Figure 2A, arrowheads), sublingual, and parotid glands (not shown). The outer cells of the sphere are stained with antibody to the differentiated acinar cell marker salivary androgen-binding protein (Sabpa; also known as Scgb1b), a secretoglobin protein . Ascl3 expression within the spheres was analyzed by RT-PCR to confirm that the EGFP expression reflects that of endogenous Ascl3, using RNA isolated from spheres at days 0, 2, 3, 4 and 7 of culture. At day 0, the RNA sample is derived from the total cell suspension and includes many differentiated cells. Primers were designed to detect Ascl3 or LE32, a ribosomal protein, used as an internal control. Expression of Ascl3 mRNA was clearly detectable in salivary gland spheres (Figure 2B), as well as in secondary spheres after 7 days of culture (data not shown).
By crossing the Ascl3 EGFP-Cre mice with a Cre-activated reporter strain (Rosa26LacZ), progeny derived from Ascl3+ cells can be detected and traced, even if the Ascl3 gene expression is down regulated . Ascl3 EGFP-Cre/Rosa26LacZ mice were used for the preparation of spheres from all three major salivary gland cell types. At day 4 of culture, the spheres were fixed and stained for β-galactosidase (LacZ) expression. The expression of LacZ is dependent on activation by Cre recombinase driven by the Ascl3 promoter, and indicates that the cell has either expressed Ascl3 or is descended from an Ascl3+ cell. Spheres formed from submandibular, sublingual and parotid salivary gland cell preparations all included LacZ+ descendants (Figures 3A–C).
The Rosa26YFP reporter mouse was also crossed to Ascl3 EGFP-Cre for lineage tracing. The EYFP reporter allowed us to maintain the spheres in culture and follow the generation of descendant cells over time. A small number of peripheral cells expressing EYFP were present in spheres at day 4 of culture (Figure 3D), and this number increased over time, indicating that the Ascl3+ cells continue to generate progeny (Figure 3E). Furthermore, the EYFP-positive progeny are generally located at the periphery of the cultured spheres (Figure 3F). RT-PCR analysis confirmed an increase in EYFP expression (Figure 1I). These results demonstrate that adult Ascl3+ progenitor cells retain the capacity to generate progeny in vitro.
In order to identify the cell types generated from adult Ascl3+ progenitors, spheres from Ascl3 EGFP-Cre; Rosa26EYFP mice were analyzed with antibodies to cell type-specific markers. Antibody to Sabpa detects seromucous acinar cells . Labeling of a day 4 sphere reveals significant co-localization of EYFP expression with Sabpa, indicating that adult Ascl3+ progenitor cells can generate seromucous acinar cells (Figures 4A–C). In contrast, no co-localization was found in numerous spheres labeled with antibody to Mist1, a transcription factor expressed in serous acinar cells  (Figure 4D–F). The absence of overlap between Ascl3+ descendants and Mist1-expressing cells suggests that Ascl3+ progenitors do not generate serous acinar cells. This result is consistent with earlier findings that few Ascl3+ progenitor cells can be lineage traced in the parotid gland, which is predominantly comprised of serous acinar cells .
Additional co-localization experiments were performed to test whether other cell types are also derived from Ascl3+ progenitors. Although Sca1 has been suggested to delineate salivary gland progenitor cells , it is broadly expressed in salivary gland duct cells . Sca1 antibody labels cells that are located at the periphery of spheres (Figure 4H), and many of these are also EYFP-positive (Figures 4G–I), leading us to conclude that the Sca1+/EYFP+ cells are differentiated duct cells derived from Ascl3+ progenitors. Antibody to Keratin 8 also labels cells located exclusively at the periphery of the spheres (Figure 4K). The Keratin 8+ duct cells are co-localized with EYFP-labeled Ascl3+ descendants (Figures 4J–L). Taken together, these data demonstrate that adult Ascl3+ cells retain multilineage potential, and can act as adult progenitor cells, to generate both duct and acinar cell types.
The relationship between the populations of putative progenitor cells so far identified in the salivary glands [1–5, 24, 25] is not yet clear. Ascl3+ progenitor cells are located in the ducts in vivo, as are the basal cell Krt5+ progenitors . We analyzed spheres by lineage tracing to determine whether the Krt5+ cells are derived from the Ascl3+ progenitor lineage. In spheres, YFP-labeled descendants of Ascl3+ progenitors do not co-localize with Krt5+ cells (Figures 5A–C). In fact, the staining patterns demonstrate that Krt5+ and YFP+ cells are clearly separate populations. Thus, Ascl3+ cells are not precursors of the Krt5+ progenitors. Lineage tracing of the Krt5+ cells will be required to determine whether they might be upstream precursors of Ascl3+ progenitor cells.
Non-adherent spheres formed in vitro from dissociated salivary gland tissue have been shown to include progenitor and/or stem cells that contribute to the restoration of secretory function in damaged glands . Here we demonstrate that cells expressing the transcription factor Ascl3 are adult progenitor cells present in all three major salivary glands.
In the adult salivary gland, Ascl3+ cells are found exclusively in the ducts, from the smallest intercalated, to the striated, granular, and large excretory ducts, but represent less than 5% of total duct cells . In developing glands, Ascl3+ cells are mitotically active , and lineage tracing has established that they are precursors of both duct and acinar cells . To establish that Ascl3+ cells continue to act as progenitors in the adult gland, we have cultured spheres from adult gland tissue and demonstrated that they generate differentiated progeny over time in culture. Furthermore, the Ascl3+ cells isolated from adult glands retain their multilineage potential, generating both duct and acinar cell types in vitro within the spheres. Notably, no sphere was found in which all cells were generated from Ascl3+ progenitors. Furthermore, the co-localization experiments indicate that Ascl3+ cells are not capable of generating all cell types in the salivary gland, and there was no evidence of self-renewal within the spheres. Thus, the Ascl3+ cells are clearly not stem cells, but represent an intermediate lineage-restricted progenitor population of the adult glands.
In this report, co-localization studies using lineage tracing and immunohistochemistry revealed that Ascl3+ progenitors in the spheres are precursors of duct cells expressing the differentiation marker Keratin 8. In addition, some Ascl3+ descendants express the stem/progenitor cell marker Sca1. Although Sca1 was identified as a progenitor cell marker in the salivary gland, it is widely expressed in differentiated epithelia of several tissues . We conclude that the Sca1+/EYFP+ cells are likely differentiated duct cells derived from Ascl3+ progenitors. Co-localization was also found between Ascl3+ descendants and secretory cells expressing salivary androgen binding protein (Sabpa), a member of the secretoglobin protein family. Sabpa, for which at least six paralogues have been detected in the mouse genome , is expressed by a subpopulation of acinar cells in the submandibular gland, which are presumed to be seromucous cells . In contrast, we saw no co-localization of the Mist1 transcription factor with EYFP+ progeny of Ascl3+ cells. In the salivary glands, Mist1 expression is found in serous-secreting acinar cells, but not in serous demilune or mucous-secreting cells . We have previously demonstrated that Ascl3+ cells are progenitors of serous demilune cells in the sublingual glands . Thus, it is clear that secretory acinar cells are a heterogeneous population, and may arise from separate progenitors. We propose that Ascl3+ progenitors generate ductal, mucous-secreting, and possibly, seromucous type cells, but that they are not precursors of serous acinar cells. Expression of a closely related transcription factor, Ascl1, was recently shown to define one of several competence-restricted progenitor populations in the retina . In an analogous manner, Ascl3 expression may define a restricted set of progenitor cells in the salivary gland.
A number of progenitor cell populations have been identified in the salivary gland [1, 2, 4, 5, 24, 25]. In addition to Ascl3, c-Kit and Keratin 5 also mark progenitor cells [4, 5]. To begin to establish the relationship between these cell populations, we analyzed Ascl3+ descendants in the spheres for Krt5 expression. We found no co-localization of EYFP-labeled progeny with cells expressing Krt5. Thus, Krt5+ cells are not lineally derived from Ascl3+ progenitors. We recently generated a targeted cell ablation model in the salivary glands in order to establish the role of the Ascl3+ progenitor cells. In the absence of Ascl3+ cells, the salivary glands appeared to develop normally, and were able to efficiently regenerate after ductal ligation . Consistent with the lineage tracing results reported here, the Krt5+ basal cells are still present in the Ascl3-cell ablation model . These results clearly indicate that adult salivary glands harbor more than a single population with progenitor cell capacity. Observations in other tissues suggest that rather than being redundant, these cell populations may contribute differently to maintenance and repair [29–32].
We observed a distinct spatial organization of cell types within the spheres. Distinct cellular organization has also been reported in spheres derived from mammary gland  and tracheal basal cells . In the salivary gland spheres, cells expressing differentiation-specific markers, including the acinar cell markers Mist1 and Sabpa, as well as the duct cell-specific marker Krt8, were located on the outer periphery. In contrast, proliferating cells incorporating BrdU were localized in the center of the spheres. Furthermore Ascl3+ cells were found within the spheres, while differentiated descendants marked by EYFP were predominantly found at the periphery. Absence of Ascl3 expression in peripheral cells is consistent with our earlier report that Ascl3 expression is down regulated in differentiated progeny . However, in contrast to the internal localization of Ascl3+ progenitor cells, immunostaining demonstrated that Krt5+ progenitor cells are localized at the outside of the spheres. This was also observed of Sca1+ cells, another proposed marker of progenitor cells . Both Krt5+ and Sca1+ are expressed by ductal basal cells in vivo. Thus their external localization on the spheres may reflect a similarity in interaction of their cell membranes with neighboring cells. The biological relevance of cellular organization within the spheres is not yet clear, and further characterization of each progenitor type is required to resolve their individual roles.
The use of stem or progenitor cells for therapeutic treatment requires precise knowledge of their differentiation potential. Using lineage tracing to definitively establish the progenitor-progeny relationship, we have shown that Ascl3+ cells are multipotent lineage-restricted progenitors of the adult salivary glands. The characterization of such progenitor cells is a critical step in the process of developing cell replacement therapies.
Generation of the Ascl3 EGFP-Cre mouse strain has been previously described . Expression of Ascl3 is detected through EGFP or Cre recombinase expression driven by the Ascl3 promoter. Lineage tracing was done using progeny from crosses of Ascl3 EGFP-Cre/+ with the ROSA26LacZ (Gt(ROSA)26Sortm1Sor; ) or ROSA26EYFP reporter mouse strains. The CAG-GFP (C57BL/6-Tg(CAG-EGFP) stock # 006567) strain was purchased from Jackson Laboratory. This transgenic line expresses EGFP, driven by the chicken beta-actin promoter, in nearly all tissues of the adult mouse, including the salivary glands . Wild type C57BL/6J mice were purchased from Jackson Laboratory. The University Committee on Animal Resources at the University of Rochester approved all procedures and protocols involving animals.
Single-cell suspensions were prepared from salivary glands isolated from either male or female mice, 2 weeks to 9 months of age as described . All three pairs of major salivary glands, submandibular, sublingual, and parotid, were dissociated individually, in HBSS containing CaCl2 and MgCl2, with 10mg Collagenase type II, Hyaluronidase (100 mg/mL, Applichem) warmed to 37°C. Cell suspensions were filtered thru 0.40 micron mesh, and plated in serum-free media [1:1 DMEM (Cellgro Mediatech), and Ham’s F-12 media (Cellgro Mediatech), with PennStrep (1x, Gibco), Glutamax (1x, Gibco), EGF (20 ng/mL, Sigma-Aldrich), bFGF-2 (20ng/mL, Invitrogen), N2 supplement (1x, Gibco), Insulin (10μg/mL), Dexamethasone (1μM, Sigma-Aldrich)] on non-coated plastic dishes at a concentration of 3 × 105 cells/ml. Cells were incubated at 37°C in 5% CO2. Media was changed every 48 hours by gently pelleting the spheres, and resuspending in fresh media. Spheres became evident between 2–3 days.
Secondary spheres were generated after 6 or 7 days in culture by rinsing gently pelleted spheres in 1xPBS for 10 min, followed by pelleting and resuspension in trypsin for 2 min. Spheres were then mechanically disrupted by rapid trituration, rinsed, and single cells were replated in sphere medium as above.
Single-cell suspensions of submandibular gland cells were prepared independently from wild type C57Bl/6 and from CAG-EGFP mice, as described above. All preparations were made from same sex animals, because of distinct sexually dimorphic differences in rodent salivary glands. Equal numbers of cells of each genotype were mixed in a 1:1 ratio, and the mixtures were incubated in serum-free media for 5 days. Spheres were then scored as non-fluorescent (EGFP−), completely fluorescent (EGFP+), or mixed, by direct visualization using a Nikon Eclipse TS100 microscope with Nikon intensilight C-HGFI illuminator. Three separate experiments were performed. Total number of spheres scored was 1315.
BrdU labeling of dividing cells was performed at days 3, 5 or 7 after culturing in serum-free media. 10 μM BrdU was added directly to the media for 3 hours, and incubation continued at 37°C. Spheres were fixed in Carnoy’s fixative, and immunostained with antibody to BrdU according to the manufacturer’s instructions (Roche).
Spheres were fixed in 4% PFA for 20 minutes at room temperature and then rinsed in detergent buffer [2mM MgCl2, 0.01% sodium deoxycholate, 0.02% IGEPAL (Sigma-Aldrich), in 0.1M PO4 buffer]. Staining for B-galactosidase activity was performed as previously described . Spheres were photographed using an Olympus DX41 microscope and DSL camera, with DSL software.
Non-adherent spheres were gently pelleted, resuspended in 800 μl serum-free media (described above), plated onto 4-well Permanox™ plastic chamber slides (LabTek), and allowed to attach. Subsequent steps were performed on attached spheres in chamber slides. Spheres were fixed in 4% PFA for 20 minutes, washed in 1xPBS, followed by 1 mM glycine. For the Mist1 antibody stain only, permeabilization was performed with 0.5% TritonX-100 in PBS, followed by 1xPBS wash. Spheres were then incubated in 10% normal donkey serum in 1xPBS. Antibodies used for immunocytochemistry included: anti-rabbit Mist1 diluted 1:2000 (gift from Dr. Stephen Konieczny, Purdue University); anti-rabbit salivary antigen binding protein alpha subunit (Sabpa) diluted 1:100 (gift from Dr. Art Hand, University of Connecticut); anti-chicken Keratin 8 (Krt8) diluted 1:500 (Sigma-Aldrich); anti-rabbit Sca1 diluted 1:400 (R&D Systems); anti-rabbit Keratin 5 (Krt5) diluted 1:500 (Covance). Antibodies were incubated overnight. Secondary antibodies used were Cy3-labeled anti-rabbit, or anti-chicken diluted 1:500 (Jackson Immunoresearch). Following the last wash, 10 μg/mL DAPI (Invitrogen) in PBS was applied to tissue for 5 minutes at room temperature. The tissue was rinsed 3× 5 minutes in PBS. The plastic chamber scaffold was removed, and the slides were mounted using Immu-mount (Thermo Scientific).
Imaging was performed using an FV1000 Olympus laser scanning confocal microscope with sequential scanning to eliminate bleed-through. The intensity for each fluorophore was optimized using gray scale analysis to insure each fluorophore was imaged in the linear range. Post-processing and 3D analyses were performed using FV1000 post-processing capabilities. Photographs in Figure 5 were made using the Bright Z feature of the FV1000 software.
Total RNA was isolated from aliquots of spheres taken at days 0, 2, 3, 4 and 7 after plating. RNA was isolated according to manufacturer’s instructions, using the RNA-easy Qiagen minikit. DNAse treatment was included in the protocol. cDNA was synthesized from 1 μg of RNA using the iScript cDNA Synthesis Kit (BioRad) according to the manufacturer’s instructions. Primers (Integrated DNA Technologies) used for RT-PCR were the following: Ascl3F: GGTGAAAGGAAACGATGGACACC; Ascl3R: TTCTCCAGGTAGTCCTCCGGA; LE32F: TTCATCAGGCACCAGTCAGACC; LE32R: ACACAAGCCATCTACTCATTTTCTTCG; ckitF: CGGGCCTACAATGGACAGCC; ckitR: CCCAAGCCATGCAAACGGTGC; Krt5F: GCCTGGACCTGGACAGCATCATC; Krt5R: CTGGTACTCCCGCAGCAGCCTGG; Ascl2F: GCCCGTGAAGGTGCAAACGTCCA; Ascl2R: CTCATCGGACGGCGGAGTAGCG; EYFP F: GCGAGGGCGAGGGCGATGCCAC; EYFP R: CAGCTCGTCCATGCCGAGAGTGA; Sca1F: CTCTGAGGATGGACACTTCTCAC; Sca1R: GGTCTGCAGGAGGACTGAGCTCA; Notch1F: GCAAGAGGCTTGAGATGCTCCC; Notch1R: AGCGATAGGAGCCGATCTCATTG. EYFPF: AAAGTCGCTCTGAGTTGTTAT; EYFPR: AAGACCGCGAAGAGTTTGTC. RT-PCR conditions were: 94° 3 minutes, followed by 35 cycles of 94°, 30 sec; 60°, 30 sec; 72°, 30 sec. RT-PCR products were separated on 1% agarose gels and photographed using an Alpha Imager (™).
We are grateful to Linda Callahan for technical advice and skilled assistance with the confocal microscopy. We thank Michael Rogers, Laurie Koek and Yasna Jamarillo for excellent mouse husbandry and genotyping. Sabpa antibody was generously donated from Drs. Art Hand (Univ of Connecticut) and Lily Mirels (UC Berkeley). The antibody to Mist1 was a generous gift from Dr. Stephen Konieczny (Purdue University). We thank Dirk Bohmann for critical reading of the manuscript. This work was supported by grants R01 DE008921 and R01 DE018896 from NIDCR (CEO).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.