Lgr5+ ductal cells of von Ebner’s glands: Candidate stem cells for turnover of posterior tongue taste buds
Theresa A. Harrison, Anthony M. Downs, Alexandria J. Slepian, Johan H. van Es, Hans Clevers, Dennis M. Defoe
TL;DR
This study identifies Lgr5+ cells in von Ebner’s glands as potential stem cells for replacing taste cells in the posterior tongue.
Contribution
The study identifies Lgr5+ ductal cells in von Ebner’s glands as candidate stem cells for posterior tongue taste bud turnover.
Findings
Lgr5+ cells are concentrated in ducts of von Ebner’s glands during postnatal development.
Lgr5+ ductal cells express markers of stem/progenitor cells and give rise to taste bud cells.
Lineage tracing shows Lgr5+ cells contribute to taste cell renewal in posterior tongue papillae.
Abstract
Taste bud cells have a limited lifespan and must be continuously replaced along with the papilla epithelium in which they reside. Previous work has shown that expression of leucine-rich G protein-coupled receptor 5 (Lgr5), a Wnt pathway agonist, serves as a marker of adult stem/progenitor cells for taste buds located in posterior tongue (circumvallate and foliate), but not anterior tongue (fungiform), taste papillae. However, the specific location/niche of the Lgr5-expressing cells supporting renewal and their phenotypic properties have not been fully explored. To address this, the genesis and fate of Lgr5+ cells were examined in developing and adult mice using genetic reporter strains. Evidence from Lgr5-lacZ and Lgr5-GFP mice shows that, while Lgr5 is broadly expressed in the epithelium of nascent circumvallate papillae and their trenches during embryonic development, it becomes…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig 1
Fig 2
Fig 3
Fig 4
Fig 5- —NSF REU
- —NSF REU
- —http://dx.doi.org/10.13039/100000097National Center for Research Resources
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsBiochemical Analysis and Sensing Techniques · Olfactory and Sensory Function Studies · Advanced Chemical Sensor Technologies
Introduction
Mammalian taste buds are complex sensory organs consisting of taste receptor cells that interact with each other via local neurochemical circuits and ultimately communicate with sensory nerves via both conventional synaptic and non-synaptic interactions [1–3]. Along with supporting cells and immature precursors, these receptor cells are arranged as compact clusters (taste buds) embedded within the stratified surface epithelium of the tongue, palate and epiglottis. Lingual taste buds, along with surrounding epithelium and connective tissue, are organized into structures called papillae that are distributed non-uniformly over the dorsal tongue surface. Mouse tongues have three major regions that contain taste papillae: the anterior tongue, across which fungiform papillae are distributed in a patterned array, two sets of foliate papillae located on opposite lateral edges of the posterior tongue, and a single circumvallate papilla that lies in the midline near the back of the tongue.
While taste buds in different tongue regions are structurally very similar, their arrangement within papillae differs markedly. Fungiform papillae are relatively simple dome-shaped structures consisting of a surface epithelium, each with a single taste bud, that envelopes a connective tissue core. By contrast, in circumvallate and foliate papillae several hundred buds occupy the walls of trenches, infoldings formed by epithelial invaginations from the tongue surface. Posterior papillae are also distinguished from anterior papillae by their connection to serous salivary glands, the von Ebner’s glands, which are located deeper in the tongue. These glands empty their contents into the luminal spaces of papilla trenches via excretory ducts. In addition to their structural differences, anterior and posterior papillae have distinct embryological origins. While fungiform papillae are presumed to arise from ectoderm, both circumvallate and foliate papillae are endodermal derivatives [4] that develop in tandem with their associated salivary glands [5–9].
The peripheral taste system of vertebrates is unusual among sensory pathways in that its cellular components are continually turned over in adult animals [10–12]. The molecular signaling mechanisms regulating homeostasis in adults include several that are shared with taste bud development [13,14]. Prominent among these is the canonical Wnt pathway, which plays a key role in both formation and maintenance of taste papillae [15–17]. Wnt-driven transcription, initiated through interaction of nuclear β-catenin with Tcf/Lef transcriptional cofactors, is part of an integrated program for renewal and regeneration in many tissues [18]. One of the targets under Wnt control, leucine-rich G protein-coupled receptor 5 (Lgr5), has been shown to be a marker for stem cells in several different areas of the body [19–23], including the posterior tongue [24,25]. In lineage tracing experiments, Lgr5^+^ cells were shown to be responsible for generation of all taste cell types during normal homeostasis and for replacement of taste buds following neurectomy-induced degeneration [24,25]. In these studies, however, characterization of the long-term progenitors giving rise to taste cells and the niche in which they reside was not undertaken.
In the present report, we have examined Lgr5 expression and lineage of Lgr5^+^ cells in the tongue of developing and adult mice. Our studies establish excretory ducts of von Ebner’s glands, associated with circumvallate and foliate papillae, as major lingual sites of Lgr5 expression in adult animals. Found only in the outer layer of the double-layered excretory ducts and situated at their confluence with papilla epithelium, these Lgr5^+^ cells are unique in exhibiting a set of phenotypic marker proteins that are typically differentially expressed between ductal and papillary epithelial tissues. Lineage analysis shows that ductal Lgr5^+^ cells undergo long-term renewal and give rise to differentiated cells of both taste buds and of their surrounding stratified epithelium; thus, they may represent an important source of stem/progenitor cells for posterior taste buds of adults.
Materials and methods
Ethics statement
Experimental procedures involving animals were approved by the institutional review committees at East Tennessee State University (ETSU) and the Hubrecht Institute. They also complied with the National Institutes of Health Guidelines for Care and Use of Animals in Research and with the recommendations of ARRIVE (Animal Research: Reporting of Experiments; https://arriveguidelines.org).
Mice
Lgr5-IRES-lacZ mice were resident at the Hubrecht Institute. These animals possess an IRES-lacZ cassette targeted to the 5′ end of the last exon, removing all transmembrane regions of the encoded Lgr5 protein and placing the β-galactosidase enzyme (lacZ) under the Lgr5 control region [19]. Non-transgenic mice (C57Bl/J6), as well as Lgr5-EGFP-IRES-CreERT2 and Rosa26-td-Tomato, − EGFP (mTmG) strains, were obtained from the Jackson Labs (Bar Harbor, ME) and housed and bred at facilities in the Division of Laboratory Animal Resources at ETSU. Lgr5-EGFP-IRES-CreERT2 mice carry a knock-in allele that both abolishes Lgr5 gene function and expresses EGFP and CreERT2 fusion proteins from Lgr5 promoter/enhancer elements. The mTmG line, a two-color fluorescent Cre-reporter, possesses loxP sites on either side of an expression cassette (inserted into the ROSA 26 locus) targeting tdTomato (mT) to cell membranes, causing strong red fluorescence labeling in all tissues and cell types [26]. When bred to mice with a tamoxifen-activable Cre recombinase, the resulting offspring have the mT cassette deleted in activated cells and the lineages derived from them, allowing expression of membrane-targeted EGFP (mG) via an expression cassette located immediately downstream. Maintenance and genotyping of mice has been described previously [19,26].
For most expression studies, the Lgr5-EGFP-IRES-CreERT2 and Lgr5-IRES-lacZ lines were bred to C57BL/6J mice to produce heterozygotes (referred to hereafter as Lgr5-GFP and Lgr5-lacZ). In the case of lineage tracing, Lgr5-EGFP-IRES-CreERT2 heterozygotes were crossed with homozygous Rosa26-td-Tomato, − EGFP (mTmG) animals and progeny heterozygous for both transgenes were selected for experiments. The latter double heterozygotes were injected with tamoxifen to induce mG in all lineage tracing studies. Uninduced animals with the same genotype (Lgr5^EGFP-IRES-CreERT2/+;R26R-mTmG/+^) were also used in some Lgr5 expression experiments.
In developmental studies, the age of an embryonic mouse was calculated based on the number of days following fertilization, with noon of the day of vaginal plug formation designated as E0.5. Animals were considered to be adults if they were older that 8 weeks. No apparent differences were noted between female and male mice in pilot studies. Consequently, both sexes were pooled in all experiments.
To initiate Cre induction, mice were injected intraperitoneally with a single 200 μl dose of tamoxifen (20 mg ml^-1^; Sigma Chemical Co., St. Louis, MO) in corn oil [27]. For long-term studies (1–2 months), animals received two consecutive tamoxifen injections on days 0 and 1. Control mice received corn oil alone.
Tissue collection, preparation and histological staining
Neonatal and adult mice, including timed-pregnant dams, were euthanized by CO_2_ asphyxiation followed by decapitation. Tongues were then surgically removed prior to fixation, except in the case of embryos, in which the tongues were fixed in situ.
All animals were monitored daily and no procedures involving survival surgery were performed. Because tissue collection occurred immediately following euthanasia, no anesthesia or analgesia was required or used. Euthanasia procedures complied with institutional IACUC protocols and AVMA guidelines and all efforts were made to minimize pain and distress.
For fluorescent protein and immunocytochemical visualization studies, tissues were immersed overnight at 4°C in 2% paraformaldehyde in 0.1 M sodium acetate buffer, pH 6.0 [28,29]. Following brief rinsing in phosphate-buffered saline, pH 7.3 (PBS), the tongues were either left intact for documentation of GFP fluorescence in whole-mounts (see below) or further dissected into anterior, posterior and lateral pieces containing fungiform, circumvallate and foliate papillae, respectively, and prepared for sectioning (see below). In the case of lacZ staining, tissue fixation took place at 4°C in 4% paraformaldehyde in PBS before being processed for paraffin embedding and sectioning (see below).
For experiments on isolated circumvallate papillae, the extirpated tongues from non-transgenic and Lgr5-EGFP-IRES-CreERT2;R26-mTmG reporter mice were injected with approximately 100–200 μl 1% Dispase II (neutral protease; Roche, Indianapolis, IN) dissolved in Balanced Salt Solution (BSS) [30,31]. Injections were localized to the intradermal space immediately beneath the papilla. Following incubation for 15 minutes at room temperature in BSS, the epidermal layer, including that of the papilla, was peeled away and extraneous tissue trimmed. The papilla epidermis was then fixed at room temperature in 4% paraformaldehyde in PBS, rinsed and examined by fluorescence stereomicroscopy (see below).
Paraffin embedding and sectioning, as well tissue staining for the presence of lacZ activity, have been described [19]. Processing of dissected tongue specimens for frozen sectioning and procedures used for immunolabeling of sections followed previously published protocols [32]. Primary antibodies used included rabbit anti-cytokeratin 14 (Thermo Scientific, Waltham, MA; 1:250), rat anti-cytokeratin 8 (TROMA-I; Developmental Studies Hybridoma Bank; 1:500) and rabbit anti-Sox9 (Millipore, Burlington, MA; 1:750). Detection of antibody binding was with the appropriate species-specific Biotin-SP-conjugated goat anti-antibodies (1:400) followed by AlexaFluor 555-conjugated Streptavidin (1:600) (Life Technologies, Carlsbad, CA).
All studies of Lgr5 expression, through its surrogate reporters, exploited the strong lacZ and GFP signals from the Lgr5-IRES-lacZ and Lgr5-EGFP-IRES-CreERT2 transgenes. No labeled antibodies were used for intensification of the expressed enzyme and fluorescent proteins.
Microscopy
LacZ reaction product in 5 μm paraffin sections was viewed on an Olympus BH-2 light microscope (Olympus, Center Valley, PA) and images captured and digitized using an Insight CCT camera (Model 142 Color Mosaic) and Spot Software System (Diagnostic Instruments, Sterling Heights, MI). Fluorescence images of whole-mounted tongues and isolated papillae were obtained using a Leica MZ16FA stereomicroscope with a Fluo III fluorescence module (Leica, Heidelberg, Germany) and documented with a color digital CCD camera and image capture software (Q Imaging Retiga EXi camera and Q Capture software, Surrey, Canada). Frozen sections of adult (10 μm) and embryonic/neonatal (20 μm) mice were examined in Leica SP2 and SP8 laser scanning confocal microscopes (Leica, Heidelberg, Germany) using 20X and 60X infinity-corrected objectives and viewed through the GFP and RFP fluorescence channels or using manufacturer-provided Alexa 488 and Alexa 555 excitation and emission settings, as appropriate for the labeling paradigm used. The resulting images were processed using Photoshop CS2 (Adobe Systems, San Jose, CA).
Co-expression of GFP with cytokeratin 14, cytokeratin 8, or Sox9 was assessed by qualitative, cell-by-cell examination of single 1 µm optical sections from frozen circumvallate papillae of Lgr5-GFP reporter mice. Analyses were restricted to well-oriented regions of the bi-layered excretory duct epithelium, in which individual GFP^+^ duct cells and their boundaries could be clearly resolved. For each marker, GFP^+^ cells within these regions were identified and the presence or absence of detectable immunolabeling was recorded manually based on consistent labeling patterns and signal localization. Because the number of evaluable ductal regions was limited and derived from archived material, co-expression frequencies are presented as descriptive estimates rather than as inferential statistical measurements.
Results
Lgr5 expression by developing taste papillae
Lgr5 expression was initially examined in tongue whole-mounts from neonatal and adult mice with an Lgr5-EGFP-IRES-CreERT2 allele (Lgr5-GFP reporter mice). Consistent with a previous report [24], in these mice robust GFP labeling demarcates all three types of lingual taste papillae at 1 week after birth, when taste buds have begun forming at all sites (S1A, B Fig; [14]). Expression in circumvallate and foliate papillae is maintained at high levels in 12 week-old animals (S1C Fig), but is undetectable in fungiform papillae at this age (S1D Fig).
To characterize the distribution of Lgr5^+^ cells in developing circumvallate papillae, 20 μm sections of posterior tongue from Lgr5-GFP reporter animals were viewed at selected times from mid-gestation to the initial postnatal period. Early in papilla morphogenesis, GFP fluorescence in the central posterior tongue is restricted to columnar cells of the presumptive taste epithelium, which forms a continuous single layer that covers the surface of the dome-like central portion and lines the invaginations of the developing papilla trenches (Fig 1A). By late gestation (Fig 1B), labeling of the multilayered epithelium appears patchy due to an irregular distribution of Lgr5^+^ cells. At this stage, Lgr5^+^ cells are found more deeply in the mesenchyme, both in the lengthening invaginations of the papilla and in the area directly below it (Fig 1B). At birth (Fig 1C, D), expression is still evident in the medial and lateral walls of the trenches. Extending from the bottom of the trench at this stage are labeled nascent ducts of von Ebner’s salivary glands, characterized by their very narrow epithelium-lined lumens. However, by 1 week postnatally (Fig 1E, F), conspicuous concentrations of GFP fluorescence appear below the trenches, in regions where salivary gland ducts meet the wall of the taste papilla. By this stage, these are the areas of highest Lgr5-GFP expression associated with the papilla.
Lgr5 expression by developing circumvallate papillae of Lgr5-GFP knock-in mice.(A) E14.5. A layer of GFP-labeled epithelial cells is observed covering the papilla dome (bracket) and lining the invaginations of the nascent trenches (bars). (B) E18.5. Labeling within the papilla epithelium is interrupted by non-expressing cells in both the dome and trench areas. Lgr5+ cells are also found deeper in the mesenchyme, at potential sites of salivary duct formation. At birth (C), GFP fluorescence can be seen in the medial and lateral walls of the papilla trenches (e.g., bar) and in the epithelium of the forming ducts (between arrows). The boxed area in (C) is shown at higher magnification in (D). (E and F) P7. Lgr5 labeling in the papilla itself appears mainly in the layer around the base of taste buds arranged in the trench walls. Fluorescence is most intense at the terminations of salivary ducts, three of which are indicated by arrowheads in (F), which is a higher magnification image of the boxed area in (E). Images represent maximal projections of 3-5 single 1 μm confocal scans. Scale bar in (A) also applies to (B) and (C). These images were selected from tongue sections (20 μm) obtained from 3-5 animals at each time point.
Expression of Lgr5 in adult posterior papillae
Lgr5 expression in adult animals was investigated by comparing posterior tongue sections from mouse strains harboring either the Lgr5-IRES-lacZ or Lgr5-EGFP-IRES-CreERT2 alleles. The pattern of labeling visualized in each of these lines is very similar (Figs 2 and S2). The most consistent and intense labeling is seen at junctures of ducts of von Ebner’s salivary glands with the walls of the circumvallate (Figs 2A and S2A, B) and foliate (Fig 2C) papillae, where it is co-extensive with the distal, excretory segments of the ducts (Figs 2B, D and S2A, C, F). Labeling often extends a small distance into the extragemmal taste epithelium immediately surrounding the “entry zones” where ducts merge with the papilla wall to open into the trench (Fig 2B, D). At increasing distances from these entry zones, Lgr5 expression, as indicated by β-galactosidase reaction product or GFP fluorescence, gradually decreases. Labeling appears to be absent in taste buds themselves.
Expression of Lgr5 in posterior tongue of adult Lgr5-LacZ reporter mice.(A and B) Circumvallate papilla. In the low power photomicrograph (A), lacZ reaction product is evident in the taste bud-containing epithelium lining the trenches and extending into extrapapillary tissue near the base of the papilla. In the boxed area, shown at 2X magnification in (B), three salivary ducts (brackets) can be seen near the base of the trench. LacZ staining is intense where von Ebner’s gland ducts intersect the papilla (arrowheads) and around the base of taste buds (asterisks) in the adjacent extragemmal taste epithelium. (C and D) Foliate papillae. (C) Lgr5 expression is shown at low power for three foliate papillae. At higher power (D), two salivary ducts (brackets) can be seen to merge as they contact the base of one papilla. Lgr5 labeling is evident in the ducts (arrowheads) and around taste buds (asterisks), as in the circumvallate papilla in (A) and (B). Scale bars in (A) and (B) pertain to (C) and (D), respectively. Images are representative of processed tissues from 6 mice.
While the most prominent labeling observed in Lgr5 reporter mice is associated with identifiable excretory ducts, it is also present in some areas near the base of taste buds when no duct is evident (Fig 2B, D). Examination of serial sections reveals that these areas are usually found to extend from entry zones of excretory ducts that are identifiable in adjacent sections (Fig 3E, Ei-Eviii; see description below). Multiple duct entry zones are often present in the same section (e.g., Figs 2A, B and S2F, G). While the majority of ducts intersect the papilla wall at or near the base of the papilla, entry zones of labeled ducts can be seen distributed along medial and lateral papilla walls at loci throughout the deeper one-half of the of the papilla, but generally not more superficially. No GFP labeling appears below the two-layered excretory region in any of the papilla-directed salivary ducts (e.g., see Fig 3A) or in glands from which they arise.
Characterization of Lgr5+ cells in circumvallate papillae of Lgr5-GFP reporter mice co-labeled for epithelial and ductal markers.(A-D) Immunolabeling for Krt14. In the low power confocal overlay image (A), expression of this epithelial cell marker (red) is demonstrated in von Ebner’s salivary glands (vEG), their ducts (arrows), the epithelium of the papilla and the basal layer of the surface epithelium of the tongue. A zone of high GFP fluorescence corresponding to Lgr5 expression (green) is evident in the terminal portion of the salivary duct on the left, where it contacts the base of the papilla and its lumen opens into the papilla trench. In higher magnification images of the boxed region in A, Lgr5 expression (B) and Krt14 immunoreactivity (C) are seen to partially overlap in the distal, excretory portion of the duct (D). (E) Immunolabeling for Krt8. Anti-Krt8 reactivity (red) is located throughout the single-layered epithelium of the salivary duct (arrow) and in the internal (luminal) layer of the bi-layered excretory segment of the duct proximal to the base of the papilla. Staining for Krt8 is also seen within taste buds (asterisks). GFP fluorescence is limited to the external (basal) layer of the excretory segment of the duct (arrowheads) surrounding the Krt8-immunoreactive cells of the internal layer. Lgr5 expression appears to extend into basal portions of the adjacent papilla epithelium, but fluorescence falls off in intensity with distance from the duct. (Ei-Eviii) Images of subsequent serial sections through this circumvallate papilla further illustrate the consistent association of high Lgr5-GFP expression with segments of at least 5 excretory ducts (numbered arrowheads) and not with the distal, single-layered portions of the ducts (red labeling only). (F) Immunolabeling for the nuclear marker Sox9. Co-localization of Lgr5 labeling (green) and Sox9 immunoreactivity (red) is seen in the excretory segment of the salivary duct where the two overlap (yellow nuclei, arrowheads). The insets show the respective single channel images for the overlay. (A-D): Single optical sections. (E and Ei-Eviii) Maximum projection of 3 optical sections. (F): Single optical sections. Images were selected from immunoreacted tissues of 5 (anti-Krt14), 6 (anti-Krt8) and 3 (anti-Sox9) animals.
Lgr5+ cells exhibit properties of ductal cells
To better define the phenotypic properties of Lgr5^+^ cells in the ducts, we compared GFP fluorescence with immunolabeling for cell type-specific markers using Lgr5-GFP reporter mice. In our experiments, keratin 14 (Krt14)-labeled cells include Lgr5^+^ cells of von Ebner’s gland excretory ducts, as well as Lgr5^–^ cells of gland ducts and acini and of the epithelium covering the tongue and lining papilla trenches (Fig 3A). As shown in Fig 3B–D, Lgr5^+^ cells co-label with anti-Krt14 in the terminal region of salivary gland ducts. On the other hand, immunolabeling for keratin 8 (Krt8) localizes to taste buds and salivary gland ducts, both simple and stratified, but shows minimal overlap with GFP fluorescence (Fig 3Ei-Eviii). The majority of strongly GFP fluorescent cells are located in a layer surrounding the single layer of Krt8-immunoreactive epithelial cells forming the excretory duct luminal wall (Fig 3E). This labeling pattern is reproduced in multiple salivary ducts interfacing with the trench that can be observed in serial sections (Fig 3Ei-Eviii). Expression of Sox9 (Fig 3F), a transcription factor generally considered a marker for ductal cells, characteristically appears in von Ebner’s salivary gland ducts, including in some Lgr5^+^ cells of excretory zones (Fig 3F). Sox9 immunolabeling, however, is not seen in those cells with lower levels of GFP fluorescence found in basal or other regions of the circumvallate papilla epithelium beyond duct entry points, or in any other taste epithelium-associated cells.
Across markers, qualitative cell-by-cell assessment consistently indicates that the majority of GFP^+^ duct cells co-expressed Krt14, showed only rare overlap with Krt8 and that approximately half expressed Sox9. These co-expression frequencies are reported as descriptive estimates derived from assessment of the best-preserved archived sections and are intended to summarize consistent patterns observed across samples rather than to provide inferential statistical measurements; no formal statistical testing was performed due to the limited amount of evaluable material.
Lgr5+ ductal cells are progenitors for taste bud cells
For lineage tracing studies, Lgr5-EGFP-IRES-CreERT2;R26-mTmG reporter mice were used. In these animals, induction of Cre recombinase following tamoxifen injection causes a switch from expression of the red fluorescent protein tdTomato to that of GFP in the cell membranes of activated cells. This allows fate mapping of cells indelibly marked by surface labeling, while at the same time monitoring continuous Lgr5 expression, visualized as GFP targeted to the cytoplasmic compartment. While the same fluorophore is used to follow labeling in both cell compartments, these are readily distinguished by relative intensity; membrane-localized GFP is driven by the strong ROSA promoter, while cytoplasmic GFP is regulated by much weaker Lgr5 control elements.
When tissues are examined at 1 day following injection, the earliest time point sampled, cells with GFP-labeled membranes are primarily Lgr5^+^ cells located in the excretory portion of salivary ducts (Fig 4A, B). Induced cells are also found at the base of taste buds and in perigemmal areas (Fig 4C, D). Some of these cells appear to lack or have low cytoplasmic GFP labeling. This is consistent with their identity as precursor cells which, having become membrane labeled as ductal Lgr5^+^ cells, migrate to the perigemmal area and display reduced Lgr5 expression. However, because we also find low expressing cells in areas of taste epithelium of animals not treated with tamoxifen, it is also possible that these cells were induced to become membrane-labeled as papillary basal and perigemmal cells. By 2 days post-injection, cells with membrane GFP first appear within taste buds (Fig 4E). Along with perigemmal and surface epithelial cells, marked taste bud cells continue to accumulate over the next several days (Fig 4F) and, by 1 week post-injection, can be seen to comprise multiple cells in some taste buds (Fig 4G). Importantly, membrane labeling persists in Lgr5^+^ excretory duct cells as late as two months following tamoxifen injection, providing evidence that this population comprises progenitor cells that are undergoing self-renewal (S3C Fig).
Lineage tracing of Lgr5-expressing cells in circumvallate papillae.Lgr5-GFP-IRES-CreERT2;mTmG mice were injected with a single dose of tamoxifen (day 0) and examined at the indicated time intervals for induction of Cre-mediated expression of membrane GFP (mG). (A-D) 1 day post-injection. (A) is a low power image of one side of a papilla (dashed boundary line) with associated ducts of von Ebner’s glands (arrows). Boxed areas are seen at 2X magnification in (B), (C) and (D). Cells with induced membrane labeling are prominent at the terminations of salivary ducts, where they can be identified by the distinct circular mG profile (open arrowheads in B) surrounding their less intense cytoplasmic Lgr5-GFP expression. These cells, as well as uninduced Lgr5+ cells displaying cytoplasmic labeling only (closed arrowheads), are restricted to the basal layer of the bi-layered ductal epithelium (dots indicate luminal layer cells). Induced cells are also present in the perigemmal region in (C) and (D) (open arrowheads). In (C), asterisks indicate two longitudinally sectioned taste buds while in (D), taste buds are cut in cross section. (E) 2 days post-injection. The progeny of GFP-expressing cells are first seen in taste buds at this time point, illustrated by the surface-marked cell at upper left (open arrowhead). Cells with induced labeling appear at all levels within the taste epithelium, including at the tissue surface. At 3 days post-injection (F), membrane-labeled cells continue to accumulate within the epithelium. A continuous line of surface-marked cells (open arrowheads) can be seen extending from the basal layer of the duct (arrow) to the basal region of the papilla below an adjacent taste bud. (G) 7 days post-injection. Duct cells with induced labeling remain (open arrowheads), with their progeny occupying taste buds as well as the surrounding papilla epithelium. Representative images were selected from processed tissues of 5 (1 Day), 3 (2 Days), 5 (3 Days) and 6 (7 Days) tamoxifen-injected mice.
Discussion
Collectively, our data support a role for Lgr5-expressing cells located in lingual salivary gland ducts as long-term progenitors/stem cells for the genesis and maintenance of posterior tongue taste buds. First, during the period of initial taste bud formation in early postnatal animals a gradual accumulation of Lgr5^+^ cells appears in nascent ducts while their numbers decrease within the papilla. Second, Lgr5 expression persists at this location in adult animals, restricted to cells in the basal (outer) layer of terminal excretory ducts associated with von Ebner’s glands. Third, lineage tracing studies demonstrate that this specific population of cells undergoes long-term self-renewal and gives rise to differentiated cells both within and outside of taste buds in the papillae, implicating these Lgr5^+^ cells in the turnover of mature taste epithelium.
In previous expression and lineage tracing studies with Lgr5 reporter mice, conflicting suggestions have been offered as to the location of taste stem cells. The presence of Lgr5^+^ cells surrounding the basal aspect of taste buds in adult papillae lead Takeda et al. [24] to propose that these cells were themselves the long-term progenitors for taste cell turnover during homeostasis and in regeneration following denervation injury. On the other hand, in fate-mapping studies Yee et al. [25] stated that “the strongest GFP signal is at the bottom of the trench area below the CV papilla and adjacent to the opening of the ducts of von Ebner’s glands”, implicating this tissue region as the source of stem cells maintaining taste bud renewal. However, in neither case did the authors identify phenotypically the specific cell type(s) expressing Lgr5 in these varying locations. These differing observations can be reconciled by our findings that Lgr5 labeling achieves its highest density in excretory ducts of von Ebner’s salivary glands, as well as in papillae cells directly adjoining duct entry points. Thus, the very strong Lgr5 expression noted by Yee et al. [25] near the trench base is consistent with the fact that this is the site where the largest number of duct entry points are found. Similarly, areas of elevated labeling adjacent to the base of taste buds elsewhere in the papilla, including in its lateral walls [24], may correspond to additional locations where ducts enter. Importantly, in the present study we observed that, at the earliest time point following tamoxifen injection, induced membrane labeling was apparent in the Lgr5^+^ population of cells localized to excretory ducts, as well as in extragemmal cells in immediate continuity with this ductal population. This supports the view that Lgr5^+^ cells may move into the taste papillae at numerous locations, from progenitor niches distributed in multiple salivary ducts (Fig 5). Immunocytochemical studies showing that the Lgr5^+^ cells in excretory ducts can co-express both Krt14 and the duct marker protein Sox9, but are devoid of Krt8, suggests a possible signature that identifies a population of taste progenitor/stem cells.
Diagram illustrating the niche and cell relationships of putative stem/progenitor cells mediating renewal of taste bud cells in the circumvallate papilla.The upper drawing represents the junction between terminal portions of salivary gland ducts and taste epithelium at the base of the papillary trench (see schematic below). Stem cells (Lgr5+high) occupy primarily the basal layer of the stratified columnar epithelium of excretory ducts. Arrows indicate the proposed pathway of cell transfer of descendent cells from the ductal niche, via Lgr5+low cells, to the basal layer of taste epithelium.
An extra-papillary source of progenitor cells capable of generating taste cells has been suggested in previous work. Using grafts of posterior tongue taste papillae into the anterior chamber of rat eyes, Zalewski [33] found that, while taste buds disappeared in the walls of circumvallate papillae, they sometimes appeared in ductal remnants attached to the graft. The author interpreted this as suggesting a potentially important role for duct epithelium in taste bud generation. More recently, Yu et al. [34] used a Sox10-Cre reporter mouse strain in cell fate mapping experiments to demonstrate Sox10-Cre-labeled cells within taste buds in all three types of taste papillae. These authors found that Sox10 transcripts were absent from the taste epithelium, but present in subepithelial connective tissue and in secretory (acinar) and duct cells of von Ebner’s glands. They, therefore, concluded that labeled taste bud cells were derived from precursor cells in one or more of the latter non-taste tissue compartments. Subsequent work with an inducible Sox10-Cre reporter strain and single-cell RNA sequencing (scRNA-seq) has led the same group to propose that the Sox10-Cre-labeled taste cells were derived from von Ebner’s gland ducts rather than the connective tissue core of taste papillae [35].
While both we (the present report) and Yu et al. [35] have presented evidence for ductal cells as taste bud progenitors, there are substantial differences between the two studies. Using mouse reporter strains with an Lgr5-EGFP-IRES-CreERT2 transgene, we have localized cells exhibiting high Lgr5-GFP expression and traced these cells with high spatial and temporal resolution to taste epithelium, including taste bud cells. These Lgr5^+high^ cells are components of von Ebner’s glands, identifiable as excretory ducts by virtue of their distinctive histology (stratified columnar epithelium) and position (at duct termini, opening directly into papillary trenches) [36,37]. Restricted to the cuboidal basal cell layer, these Lgr5^+high^ basal cells are contiguous with Lgr5^+low^ basal cells in the adjoining stratified squamous epithelium of taste papillae (Fig 5). Yu et al. [35], on the other hand, have relied primarily on data from scRNA-seq studies to demonstrate that Sox10 transcripts are enriched in cells that express genes characteristic of two ductal cell populations found in other exocrine glands (Slc4a4^+^/Duct-1 and Muc1^+^/Duct-2). Citing a preprint of the present work [38], referenced incorrectly as Defoe et al. (2024), Yu et al. [35] challenge our contention that the Lgr5^+high^ cells we observe are, in fact, components of von Ebner’s gland ducts. Using scRNA-seq data showing that Lgr5 and Krt5 are generally co-expressed they argue that the Lgr5^+high^ cells we see must be papillary basal cells. However, Krt5/Krt14 expression is not exclusive to stratified squamous taste epithelium. As shown here and in our preprint, Lgr5^+high^ basal cells of excretory ducts and Lgr5^+low^ epithelial basal cells of taste epithelium both label with anti-Krt14 antibodies. Interestingly, after prolonged tamoxifen administration to a Sox10-CreER reporter strain, these authors were able to show induced labeling not only of taste bud cells (primarily Type III), but also of cells occupying von Ebner’s gland ducts. While they classify the latter structures as main (or collecting) ducts, a term generally reserved for salivary glands whose total output is conveyed via a single common duct, their position at the opening into papillary trenches would seem to identify them as excretory ducts, of which many have been shown to empty separately into the trench of each circumvallate papilla in the present report.
Recently, very low levels of Lgr5 expression have been reported in Type I and Type IV taste bud cells, as well as in perigemmal cells, of circumvallate papillae using antibody-enhanced visualization of GFP in Lgr5-EGFP-IRES-CreERT2 mice and through mining of scRNA-seq data [39]. The authors also performed Iineage analysis following either a single injection, or multiple injections, of tamoxifen in Lgr5-tdTomato reporter mice. These latter studies demonstrated that Lgr5-expressing cells in the intragemmal regions and in “cryptic base/folds” of papillae (what has been shown in the present report to be multiple entry points of salivary gland excretory ducts at trench bases) retained induced tdTomato label long-term (up to 60 days post-injection). Such results were presented as evidence that both sites harbor self-renewing stem/progenitor cells. However, the authors argue that cells of “cryptic folds”, with their much stronger Lgr5-expression are insufficient to account for turnover of all taste cells. This is based, in part, on data they present showing that the proportion of tdTomato-labeled intragemmal cells seen never increased above approximately 50%, even after multiple continuous tamoxifen injections. Consequently, the authors suggest that taste cells may also originate from a population of non-Lgr5^+^ cells. However, induced expression of similar mouse reporter genes in 100% of target cells has rarely been achieved under any conditions.
In addition to in vivo studies, it has also been possible to examine the generation of taste cells ex vivo in organoids from Lgr5-expressing cells [40,41]. Results from these experiments have shown that a self-replenishing system that generates taste cells can arise from single Lgr5-expressing cells isolated directly from dissociated papillae [40] or from papilla explants maintained initially in culture [41]. As seen in our S4A, B Fig, when the circumvallate papilla epidermis is stripped from the tongue using a technique similar to that employed by Aihara et al. [41] for organoid generation, the resulting preparation includes proximal segments of salivary ducts, which would contain the cells composing the excretory ducts. As we have shown, these ducts contain Lgr5^+^ cells, some of which co-express the progenitor/ductal cell marker Sox9, whose expression is a consistent feature of the organoids generated using explants [41]. Thus, it is plausible that the taste cell generating functions of organoids may derive at least in part from Sox9-expressing Lgr5^+^ cells.
Immunocytochemical results in the present study further characterize ductal Lgr5^+^ cells as being, in addition to Sox9^+^, also Krt14^+^. Previous lineage tracing studies using mice with an inducible Krt14-Cre allele have documented that taste bud cells and surrounding keratinocytes are continuously generated from long-term stem cells expressing Krt14 in both neonatal and adult animals [42]. Based on immunolabeling with antibodies to Krt14 and other markers, these progenitors were inferred to be epithelial basal cells adjacent to taste buds. However, the Lgr5^+^/Krt14^+^ ductal cells that we describe in this study provide a possible alternative population of progenitors for the lineage of Krt14-Cre-labeled cells identified in taste papillae.
The present study has several limitations that should be considered when interpreting these findings. Analyses relied on archived tissue material, which limited the number of ductal regions suitable for detailed cell-by-cell evaluation. In addition, the small size and complex anatomy of lingual papillae, together with the convergence of multiple salivary ducts into shared trench regions, preclude unbiased assignment of labeled progeny to individual Lgr5^+^ duct cells across serial sections. As a result, quantitative lineage tracing metrics such as clone size distributions or the fraction of labeled taste buds over time could not be assessed in a statistically rigorous manner within the scope of this study. Nevertheless, the consistent qualitative patterns observed across samples support the conclusions drawn here regarding ductal Lgr5^+^ cells as long-term progenitors contributing to posterior tongue taste bud turnover. In conclusion, our data point to the ducts of von Ebner’s glands as possible niches for stems cells providing renewal of taste bud and surrounding cells within circumvallate and foliate papillae. Posterior tongue taste papillae and von Ebner’s glands are closely associated with one another structurally throughout their development and, in adults, there is evidence that the two structures collaborate functionally. First, these salivary glands have been shown to secrete enzymes and binding proteins that, in addition to aiding digestive processes, might also participate in perireceptoral events influencing taste transduction [43–47]. Furthermore, at least one brainstem autonomic reflex circuit that controls von Ebner’s gland secretory activity includes gustatory afferent input, and the potential for differential neural control of salivary secretion to modulate taste activity has also been suggested [48]. Finally, the growth factor TGFα, whose actions conceivably could support maintenance of posterior lingual taste buds, is present in von Ebner’s gland secretory vesicles [49]. Based on past studies of these interrelationships, the idea that circumvallate and foliate taste papillae and von Ebner’s salivary glands form a functional unit that could be considered a single complex organ has been proposed [48,50]. Our data extend this notion by demonstrating that collaboration between these two structures may be essential for taste bud turnover in posterior taste papillae and, thus, for maintaining functional integrity of this limb of the sensory system for taste.
Supporting information
S1 FigVisualization of Lgr5 expression by taste papillae in neonatal and adult tongue whole-mounts from heterozygous Lgr5-GFP mice.(A) and (B) 1 week. In the posterior tongue (A), GFP is localized centrally at the site of the circumvallate papilla (CVP) and laterally in the foliate papillae (FOP). Arrays of punctate label are also seen on both the posterior (A) and anterior (B) tongue at sites where single fungiform papillae (FFP) are located. (C) and (D) 12 weeks. Note that, at this time, Lgr5 labeling is maintained in circumvallate (C) and foliate (C, inset) papillae, but is no longer evident in the fungiform papillae (D). (A) and (B) are taken from a single tongue. The main images in (C) and (D) are of the same tongue, whereas the inset image pictured in (C) is from a different mouse. The scale bar in (B) applies to (A); scale bar in (D) to (C). Tongues from 8 (1 week) and 5 (12 weeks) mice were examined in these experiments.(TIF)
S2 FigLgr5 expression by circumvallate papillae of adult Lgr5-GFP-IRES-CreERT2;R26-mTmG mice.(A) and (B) Low power images of a single papilla. In (A), GFP fluorescence is concentrated at the base of the papilla on both sides. Cell outlines revealed by expression of membrane Tomato (mT), seen in the corresponding image (B), show that Lgr5^+^ cells in both regions are associated with excretory ducts (brackets). The plane of section passes through the center of the duct on the right and tangentially through the epithelial wall of the duct on the left. The boxed areas in (A) and (B) are shown at 2X magnification in (C) and (D), respectively. Highly fluorescent GFP^+^ cuboidal cells (arrowheads) occupy the outer/basal epithelial layer of von Ebner’s gland excretory ducts, while columnar cells in the inner/luminal layer of the ducts (indicated by dots) are unlabeled. (F) and (G) Intersection of excretory ducts with the circumvallate papilla. Intense GFP labeling is observed in the walls of four ducts (indicated by brackets) where they merge with the papilla. While two ducts merge at the base, the others join the papilla at more superficial levels laterally. Dashed lines indicate papilla boundaries. Taste buds are indicated by asterisks. (A), (B), (F) and (G): Maximum projection of 3 optical sections. (C-E): Single optical sections. Sections of circumvallate papillae were obtained from 6 non-tamoxifen treated mice.(TIF)
S3 FigLong-term generation of taste bud cells by Lgr5^+^ duct cells associated with circumvallate papillae.Unlike all other reported experiments, in this case Lgr5-GFP-IRES-CreERT2;R26-mTmG mice were injected with two consecutive doses of tamoxifen (days 0 and 1), and then examined at the indicated times for induced membrane GFP expression. At 1 week (A), 4 weeks (B) and 8 weeks (C), numerous membrane-labeled cells are present in and around taste buds, including in the superficial epithelial layers lining the trench. Duct-associated Lgr5^+^ cells, both with and without tamoxifen-induced labeling (open and filled arrowheads, respectively), are also present throughout the 8-week period, the longest time examined. Asterisks indicate individual taste buds. Dashed lines demarcate the borders of the papilla. Tissues were examined from 4 (1 week) and 3 (4 and 8 weeks) tamoxifen treated mice.(TIF)
S4 FigCircumvallate papilla taste epithelia isolated by combined protease treatment and mechanical stripping of epidermis from mouse tongue.(A) Frozen tissue section from a non-transgenic animal labeled with anti-Krt8. Background tissue fluorescence (green) overlaid with Krt8 immunoreactivity (red) produces yellow-appearing taste buds (asterisks). Associated with the papilla are proximal segments of salivary ducts (brackets), which remain attached after the isolation procedure. (B) Papilla epithelium whole mount from an adult Lgr5-GFP;R26-mTmG mouse; all cells express membrane tomato (mT, red). Lgr5^+^ cells with cytoplasmic GFP (green) are present in the proximal excretory segments (arrowheads) of duct fragments (brackets) that partition with the excised tissue. Tissues from 3 mice were used in each experiment.(TIF)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Finger TE, Danilova V, Barrows J, Bartel DL, Vigers AJ, Stone L, et al. ATP signaling is crucial for communication from taste buds to gustatory nerves. Science. 2005;310(5753):1495–9. doi: 10.1126/science.1118435 16322458 · doi ↗ · pubmed ↗
- 2Ma Z, Taruno A, Ohmoto M, Jyotaki M, Lim JC, Miyazaki H, et al. CALHM 3 Is Essential for Rapid Ion Channel-Mediated Purinergic Neurotransmission of GPCR-Mediated Tastes. Neuron. 2018;98(3):547-561.e 10. doi: 10.1016/j.neuron.2018.03.043 29681531 PMC 5934295 · doi ↗ · pubmed ↗
- 3Romanov RA, Lasher RS, High B, Savidge LE, Lawson A, Rogachevskaja OA, et al. Chemical synapses without synaptic vesicles: Purinergic neurotransmission through a CALHM 1 channel-mitochondrial signaling complex. Sci Signal. 2018;11(529):eaao 1815. doi: 10.1126/scisignal.aao 1815 29739879 PMC 5966022 · doi ↗ · pubmed ↗
- 4Rothova M, Thompson H, Lickert H, Tucker AS. Lineage tracing of the endoderm during oral development. Dev Dyn. 2012;241(7):1183–91. doi: 10.1002/dvdy.23804 22581563 · doi ↗ · pubmed ↗
- 5Jitpukdeebodintra S, Chai Y, Snead ML. Developmental patterning of the circumvallate papilla. Int J Dev Biol. 2002;46(5):755–63. doi: 10.1387/ijdb.12216988 12216988 · doi ↗ · pubmed ↗
- 6Lee M-J, Kim J-Y, Lee S-I, Sasaki H, Lunny DP, Lane EB, et al. Association of Shh and Ptc with keratin localization in the initiation of the formation of circumvallate papilla and von Ebner’s gland. Cell Tissue Res. 2006;325(2):253–61. doi: 10.1007/s 00441-006-0160-1 16552524 · doi ↗ · pubmed ↗
- 7Sbarbati A, Crescimanno C, Bernardi P, Benati D, Merigo F, Osculati F. Postnatal development of the intrinsic nervous system in the circumvallate papilla-von Ebner gland complex. Histochem J. 2000;32(8):483–8. doi: 10.1023/a:1004144317336 11095073 · doi ↗ · pubmed ↗
- 8Sbarbati A, Crescimanno C, Merigo F, Benati D, Bernardi P, Bertini M, et al. A brief survey of the modifications in sensory-secretory organs of the neonatal rat tongue. Biol Neonate. 2001;80(1):1–6. doi: 10.1159/000047111 11474141 · doi ↗ · pubmed ↗