Developing and Regenerating a Sense of Taste

1-1: Development of the Tongue

The oropharyngeal cavity where taste buds reside is continuous with the gastrointestinal tract. In vertebrate embryos, the mouth forms as an invagination of the most anterior non-neural ectoderm that apposes the foregut endoderm. As these epithelia meet, they form a bilayer that eventually thins to become the buccopharyngeal membrane. Subsequent rupturing of this membrane results in a continuous oral cavity lined anteriorly and posteriorly by ectodermal and endodermal epithelia, respectively ((Dickinson & Sive, 2006) and see (Chen et al., 2017; Soukup et al., 2013)).

The oropharyngeal cavity forms from the first four pharyngeal arches, with the border between ectoderm and endoderm between the first and second arches (Moore et al., 2011). The ectoderm-derived first arch develops a medial and two lateral lingual swellings, which expand and fuse to become the anterior two/thirds of the tongue. Although the second arch contributes to the anterior tongue core, it does not contribute to surface epithelium, as the third arch overgrows it to form the posterior oral tongue; the fourth arch contributes to the pharyngeal tongue (Figure 2) (Moore et al., 2011).

Thus, the anterior tongue epithelium derives from the ectoderm, and the posterior tongue epithelium arises from the endoderm (Rothova et al., 2012) (Figure 3). Connective tissue and vasculature of the tongue are derived from cranial neural crest cells, while lingual muscles originate from hypoglossal somite myoblasts (Chai et al., 2000; Hammond, 1965; Thirumangalathu et al., 2009). Both neural crest and myoblast cells migrate into the tongue during morphogenesis (Chai et al., 2000; Hazelton, 1970). While tongue formation is complete by embryonic day (E)12.5 in mice and gestational week 8 in humans, the tongue continues to grow throughout embryogenesis and early postnatal life (Moore et al., 2011; Parada et al., 2012).

Taste buds develop within the dorsal surface of the tongue and are derived entirely from the surface epithelium rather than from neurogenic or neural crest origins (Barlow & Northcutt, 1995; Stone et al., 1995; Thirumangalathu et al., 2009; Yu et al., 2021). In rodents, taste buds in the anterior 2/3 of the tongue are housed in fungiform papillae (FFP) distributed on the tongue surface, each containing a single taste bud (Figure 3). Taste buds in the posterior tongue are housed in fewer, large papillae comprising epithelial trenches containing hundreds of taste buds each. These trenches are either part of bilateral foliate papillae, containing several trenches along the lateral sides of the tongue, or in a single CVP, containing two trenches situated at the posterior tongue midline (Barlow, 2015; Purves et al., 2001) (Figure 3). Taste buds also reside in the soft palate, pharynx, epiglottis and esophagus, but those are embedded directly in the epithelium and not in papillae (Henkin & Christiansen, 1967; Travers & Nicklas, 1990; Vercauteren Drubbel & Beck, 2023). The embryonic origins of the different papillae correspond with their location on the tongue; FFP on the anterior tongue are derived from ectoderm while foliate and CVP on the posterior tongue are derived from endoderm (Rothova et al., 2012) (Table 1; Figure 3).

The innervation of taste papillae also corresponds with the arches from which they developed. Branches of the trigeminal and facial nerves (V^th and VII^th cranial nerves, respectively) innervate the first two arches and therefore the anterior tongue (Mangold & Hill, 2007; Moore et al., 2011; Ohla et al., 2019). The mandibular branch of the trigeminal nerve carries somatosensory information, i.e., mechano- and thermo-sensation and pain, while the chorda tympani branch of the facial nerve provides gustatory innervation of fungiform papillae (Mbiene, 2004; Walker, 1990; Zhang et al., 2024) (Figure 3). The glossopharyngeal nerve (IX^th cranial nerve) innervates the circumvallate papilla, and is a mixed nerve conveying both somatosensory and taste information to the CNS (Jacobson et al., 2018; Nagai, 1993; Sonne & Lopez-Ojeda, 2024) (Figure 3). Since the foliate papillae are located near the endoderm/ectoderm border, foliate trenches have mixed innervation (Doty et al., 2009; Mangold & Hill, 2007; Yamamoto & Kawamura, 1975) (Figure 3). Finally, the vagus nerve (X^th cranial nerve) innervates the fourth pharyngeal arch and its resident pharyngeal taste buds (Moore et al., 2011; Ohla et al., 2019).

1-2: Taste Placode Specification and Papilla Morphogenesis

Developmental studies have focused primarily on fungiform (FFP) and circumvallate papillae (CVP). The majority have employed rodent models, as mouse and rat taste systems are similar to that of humans; although rodents have one CVP while humans have 8-12 CVPs (Dorland, 1925). Nevertheless, use of rodent embryos allows exploration of processes in early development, including specification and morphogenesis of the taste periphery. Specification of FF and CV taste placodes begins once the tongue has formed, followed by papilla morphogenesis and taste bud cell differentiation around birth. The developmental origins of the anterior and posterior tongue differ (see above), and perhaps because of this difference, molecular regulation of the earliest aspects of induction and patterning differ between FF and CV taste buds.

1-3: Function and development of the CVP/VEG complex

The oral epithelium and its accompanying taste buds are bathed in saliva supplied by the major and minor salivary glands. Thus, the sense of taste is tightly linked to salivary gland function. Saliva is composed of a complex mix of enzymes, proteins and ions secreted by salivary gland acini into the oral cavity (Galaniha & Nolden, 2022; Matsuo, 2000; Muñoz-González et al., 2018) and see (Carpenter, 2013; Feron, 2019; Khramova & Popov, 2022) for review). The von Ebner’s minor salivary glands (VEGs) reside in the posterior tongue and their ducts connect directly to the ventral epithelia of the CVP trenches (Sbarbati et al., 1999), and together they form the CVP/VEG complex (Sbarbati et al., 1999; Sohn, Gwon, et al., 2011). The serous saliva secreted from the VEG travels through ducts into CVP trenches to: 1) initiate digestion of proteins and lipids via amylase and lingual lipase enzymes, 2) supply aqueous ions allowing for greater taste sensitivity upon tastant binding to TRCs, and 3) continually clear tastants, thereby also maintaining taste sensitivity (Fábián et al., 2015; Harada et al., 2023; Matsuo, 2000; Sbarbati et al., 1999). This coordination at both cellular and molecular levels allows for a consistent sense of taste. In fact, when this coordination is disrupted by treatments such as targeted head and neck cancer radiotherapy, patients experience xerostomia, or hyposalivation, as well as taste dysfunction (Deshpande et al., 2018; Galaniha & Nolden, 2022; Jensen et al., 2019; Mese & Matsuo, 2007).

The VEG, like the CVP, are derived from endoderm, and as the CVP undergoes morphogenesis, the ventral most tips of the CVP trenches give rise to the VEG excretory ductal cells (Jitpukdeebodintra et al., 2002; Kist et al., 2014; Sohn, Gwon, et al., 2011) (Figure 4B). Like major salivary glands, the VEG is composed of secretory acinar cell clusters wrapped by myoepithelial cells connected via small, intercalated ducts (Adrian Florin et al., 2024; Chibly et al., 2022). These smaller ducts connect back to a main excretory duct. Studies in mouse have investigated embryonic VEG formation in the context of CVP development. CVP/VEG complex morphogenesis follows CVP placode specification around E12.5 and taste papilla morphogenesis commencing at E15.5 (Figure 4B) (Jitpukdeebodintra et al., 2002). At E16.5, clusters of VEG epithelial cells are observed beneath the ventral-most CVP epithelial trenches (Jitpukdeebodintra et al., 2002) (Figure 4B). The presence of PI3K and Ki67-positive cells in the CVP trench termini suggests that proliferation is critical in the formation of VEG duct primordia (Sohn, Gwon, et al., 2011). Already at E16, nerve fibers are evident in the subjacent mesenchyme and epithelium of the VEG anlagen suggesting innervation also plays a role in VEG epithelial development (Sohn, Gwon, et al., 2011). CVP trench termini then undergo branching morphogenesis beginning at E17.5-18 that continues through postnatal day (PN) 21, where a complex network of VEG acini with small salivary ducts all feed into the main excretory duct continuous with the CVP trench epithelia (Jitpukdeebodintra et al., 2002; Sbarbati et al., 1999; Sohn, Gwon, et al., 2011) (Figure 4B). It should also be noted that factors like FGF10 play a critical role in both CVP trench invagination (see above) and branching morphogenesis in the major salivary glands, suggesting the testable hypothesis that VEG morphogenesis may also require FGF10 (Adhikari et al., 2017; Oginuma et al., 2017; Ohuchi et al., 2000; Petersen et al., 2011; Ray & Soriano, 2023; Zhang et al., 2018). Saliva production begins at birth, with accumulation of salivary enzymes beginning between PN1-3 and continuing through PN17 in rats, suggesting the secretory function of VEG cells is limited until near weaning (Jitpukdeebodintra et al., 2002). Though similar work has yet to be performed in mice, data from rats suggest a role for serous saliva in taste sensitivity prior to weaning.

While both CVP and VEG epithelia are derived from endoderm, it is unclear if VEG acini arise from developing CVP epithelial termini or if CVP and VEG have separate early origins and then connect in late gestation. In fact, the foliate papillae (FolP) may also be an embryonic source of VEGs. FolP comprise 4-7 bilateral taste bud-bearing trenches, each of which are also connected to the VEG via secretory ducts (Kist et al., 2014; Matsuo, 2000). Whether VEG acini emerge from FolP epithelial termini has not been explored. Further, while developmental and functional relationships of the CVP and VEG have been described, the molecular mechanisms regulating VEG development and adult renewal are mostly unexplored – topics that have been well investigated for the major salivary glands (see (Rocchi et al., 2021; Suzuki et al., 2021) for review). The extent to which VEG development is regulated similarly to the major salivary glands is an area for further investigation.

2-1: Multiple progenitor populations are present in adult taste epithelium.

Mitotically active progenitor cells reside outside taste buds at the basement membrane. These progenitors are bi-potent, giving rise to both taste cells and surrounding non-taste epithelium (see (Barlow, 2015) for review). Several specific markers of progenitors have been identified in FFP and CVP using inducible genetic lineage tracing in mice. Both cytokeratin (KRT) 14 and KRT5 are expressed by basal progenitors, and contribute new cells to taste buds and non-taste epithelium in both FFP and CVP (Asano-Miyoshi et al., 2008; Gaillard et al., 2015; Okubo et al., 2009). The hedgehog target gene GLI1 also marks mitotic basal keratinocytes adjacent to taste buds in FFP and CVP and inducible lineage tracing reveals GLI1-descdendant taste and non-taste progeny in fungiform taste papillae (Kumari et al., 2024; Liu et al., 2013).

Lgr5 and Lgr6 (leucine-rich repeat-containing G-protein coupled receptor 5 and 6) are WNT/ß-catenin target genes, as well as transmembrane receptors that enhance WNT signaling when bound by R-spondin ligands [REFs] (and see below). Both Lgr5 and Lgr6 are well known progenitor markers in other epithelia, e.g. intestine and hairy skin (Barker et al., 2013), and are also expressed by adult taste progenitors (Ren et al., 2014; Ren et al., 2020; Takeda et al., 2013; Yee et al., 2013). However, Lgr5 expression in progenitors depends upon age. In neonatal mice, Lgr5 marks both FFP and CVP progenitors but becomes restricted to CVP progenitors in adults (Ren et al., 2020; Takeda et al., 2013). In both CVP and FFP, basal epithelial cells adjacent to buds express Lgr6, and lineage tracing results in labeled taste and non-taste progeny (Ren et al., 2014). A recent study, however, has suggested both Lgr5 expression and lineage contribution to taste and non-taste cells persist in adult FFP, while Lgr6+ progenitor contribution to adult CVP taste epithelium is limited (Kim et al., 2024). These studies collectively indicate Lgr6+ and Lgr5+ progenitors are indeed taste competent, but their relative contributions in adult FFP and CVP remains to be resolved.

SOX2 and SOX10 are SRY-related HMG box family transcription factors and are expressed by adult taste progenitors. Proliferative cells outside taste buds are SOX2+, and long-term lineage trace results in tomato expression in both FFP and CVP taste and non-taste epithelial cells up to 21 months after tamoxifen induction (Ohmoto et al., 2017). Further characterization using a Confetti reporter mouse found uniquely labeled cells of multiple colors within each taste bud, indicating several different SOX2+ progenitors contribute progeny to each taste bud (Ohmoto et al., 2022). SOX10 tracing also reveals lineage labeled cells in taste buds in SOX10Cre;Rosa26tdTomao mice with constitutive Cre expression in SOX10+ cells. Sox10 is expressed by neural crest-derived mesenchyme adjacent to CVP epithelium, but these Sox10+ cells are unlikely to contribute to buds given the recent demonstration that neural crest does not give rise to taste cells (see above and (Yu et al., 2021)). Nor is Sox10 expressed by basal progenitors within the taste epithelium proper. However, Sox10 expression is evident in epithelial cells of the Von Ebner’s gland ducts that connect to the CVP epithelium (Yu et al., 2020). These data raise the intriguing possibility that SOX10+ progenitors in VEG ducts give rise to taste progeny that move into the CVP and contribute to taste buds.

The development of lingual organoid technology has provided a new way to assess lineage potential of taste progenitor populations in vitro. Adult stem cells are, by definition, able to replace themselves as well as give rise to functional mature cell types. To determine the lineage competency of Lgr5+ and Lgr6+ progenitors, Ren et al. used Lgr5CreER-GFP and Lgr6CreER-GFP mice. Culture of isolated GFP+ cells from the CVP of either strain revealed that single Lgr5+ and Lgr6+ progenitors proliferate and differentiate into organoids that contain all taste cell types as well as non-taste epithelium (Ren et al., 2014). These experiments support previous work showing Lgr5+ and Lgr6+ cells are taste progenitors in vivo (see above). Interestingly, Lgr5 lineage competency may be dependent on age; single Lgr5+ cells from FFP in neonatal mice generate taste organoids, but to date taste organoids generated by Lgr5+ cells from adult FFP have not been reported (Ren et al., 2020). Lingual organoids have also been generated from single SOX2+ cells derived from the CVP of Sox2GFP mice. These organoids contain all three taste cell types and non-taste epithelium, complementing in vivo evidence that SOX2 is a progenitor marker in the CVP (Shechtman et al., 2023).

Although many markers of taste progenitors have been identified, it is unclear if these mark the same, unique, or overlapping progenitor populations. At a populational level, KRT5, KRT14 and SOX2 have been reported to be co-expressed by basal keratinocytes in taste epithelium (Okubo et al., 2009), as are Lgr5 and SOX2 (Ohmoto et al., 2017; Shechtman et al., 2023; Takeda et al., 2013), Lgr5 and Lgr6 (Ren et al., 2014), and Lgr5 and KRT14 (Yee et al., 2013). However, quantifying the extent of progenitor marker co-expression at a cellular rather than populational level is necessary to assess the extent of progenitor heterogeneity. It is possible one progenitor population gives rise to more lineage-restricted progenitors, and/or that different progenitor populations have limited potential, producing a subset of taste epithelial cell types. For example, lineage tracing of SOX10+ cells located in the ductal cells in the mouse CVP/VEG complex (see above) results in preferential labeling of type III taste cells, suggesting SOX10+ progenitors are biased towards type III cell production (Yu et al., 2020). Lineage potential may also vary with marker gene expression levels, as is the case for SOX2+ progenitors. Cells expressing high SOX2 generate taste organoids that, like organoids derived from Lgr5+ progenitors, contain differentiated taste cells of all 3 types as well as non-taste epithelial cells. SOX2^low+ cells by contrast generate exclusively non-taste organoids (Shechtman et al., 2023). This suggests that SOX2 expression levels marks sub-populations of progenitors, with SOX2^high cells capable of producing the entire taste epithelial lineage, and low SOX2 expression marking progenitors with limited potential.

2-2: Fate specification and differentiation of taste cells.

Progenitor progeny destined for a taste fate become post-mitotic taste precursor cells (classically referred to as type IV cells), that express Shh within 12 hours of their final division (Miura et al., 2006; Miura et al., 2001). Subsequently, Shh+ precursors can differentiate into each of the taste cell types (Miura et al., 2014). While embryonically Shh signaling represses taste fate (see above), Shh promotes taste cell differentiation in adult lingual epithelium (Figure 5; Box1). Ectopic expression of Shh in KRT14+ progenitor cells induces formation of ectopic taste buds that contain all three taste cell types (Castillo et al., 2014). Ablation of Shh signaling through expression of a dominant-negative GLI repressor in KRT5+ progenitors, or through deletion of Shh from KRT5+ progenitors and nerve fibers, leads to taste bud loss but does not affect epithelial proliferation (Castillo-Azofeifa et al., 2017; Liu et al., 2013). Pharmacological inhibition of Shh signaling leads to similar results discussed below (“Cancer treatments”). Thus, Shh signaling drives taste cell differentiation and is necessary to maintain differentiated taste buds (and see (Kumari & Mistretta, 2023; Mistretta & Kumari, 2017; Mistretta & Kumari, 2019)).

Wnt signaling regulates both proliferation and differentiation of progenitor cells in adult mice. Deleting β-catenin or Wnt10a in KRT5+ progenitors inhibits their proliferation and depletes the progenitor pool (Gaillard et al., 2017; Xu et al., 2017). Shh+ type IV cells and all three taste cell types are steadily lost in β-catenin cKO mice, suggesting addition of new taste cells into buds is hindered (Gaillard et al., 2017). Stabilizing β-catenin, thus activating Wnt signaling, in KRT5+ progenitors causes a burst of proliferation, followed by rapid and excess differentiation of taste cells at the expense of non-taste epithelium. The excess TRCs are predominantly type I, and to a lesser extent type II cells, while type III TRCs are unchanged. This suggests a model where Wnt levels regulate TRC fate choice and is sufficient to drive differentiation into some but not all taste cell types (Gaillard et al., 2015) (Figure 5).

Aside from extrinsic signals, several transcription factors regulate taste cell renewal. SOX2 is expressed in type I cells and in taste progenitor cells (see above) (Castillo-Azofeifa et al., 2018; Okubo et al., 2009; Okubo et al., 2006; Shechtman et al., 2023; Suzuki, 2008; Takeda et al., 2013). SOX2 function downstream of Shh signaling is required for maintaining addition of new cells into taste buds (Figure 5), as loss of SOX2 causes loss of taste buds even with excess Shh ligand (Castillo-Azofeifa et al., 2018; Ohmoto et al., 2020). GLI3, a Shh-related transcription factor, also regulates TRC differentiation. Knockout of GLI3 in Lgr5+ progenitors increases a subset of type II cells and decreases type III cells, showing GLI3 may promote type III over type II cell fate (Qin et al., 2018). Conversely, the transcription factor Pou2F3 (Skn1a) is expressed in type II cells. Global knockout of Pou2F3 leads to loss of type II cells and an overabundance of type III cells, suggesting Pou2F3 promotes type II cell fate at the expense of type III cells (Matsumoto et al., 2011) (Figure 5). It is not known whether GLI3 and Pou2F3 mutually repress each other, nor how expression of each transcription factor is precisely controlled. Identifying signaling cascades or other intrinsic factors that control expression of these transcription factors will provide insight into type II and type III cell fate choice.

Type II cells are thought to comprise separate bitter, umami and sweet functional subtypes and the molecular regulators controlling selection of these distinct fates is beginning to be explored. From published RNA-sequencing data from mouse CVP (Lee et al., 2017), Ohmoto et al. found Eya1, a transcription factor, is enriched in bitter vs sweet TRCs. In situ hybridization confirmed this expression pattern, establishing Eya1 as a candidate regulator of bitter subtype differentiation (Ohmoto et al., 2021). Eya2 is also expressed in the mature taste epithelium, although both Eya1 and Eya2 are only expressed in endoderm-derived CVP and foliate papillae. Knockout of Eya1 or Eya2 each lead to fewer type II cells in CVP taste buds, supporting a role of these transcription factors in regulating type II cell homeostasis (Zhang & Xu, 2023). Another transcription factor, Etv1, is expressed in sweet and umami type II cells. Global knockout of Etv1 decreases both subtypes without affecting the overall number of type II cells, suggesting Etv1 is dispensable for initial type II cell fate but promotes differentiation of sweet and umami subtypes (Ohmoto et al., 2023). Gli3, discussed above as a regulator of type II vs type III cell fate, also may regulate fate choice between type II cell subtypes. Knockout of Gli3 from Lgr5+ progenitors increases the number of sweet and umami cells, while having no effect on bitter type II cells (Qin et al., 2018). This suggests that Gli3 represses sweet and umami fate; opposite to the function of ETV1 (Figure 5). However, upstream genetic regulators of these transcription factors that govern type II cell fate have yet to be identified.

Ascl1, a Notch pathway transcription factor, is expressed in a subset of Shh+ taste precursor cells and by all differentiated type III cells (Seta et al., 2006). Lineage tracing of Ascl1+ cells in newborn mice revealed reporter expression mostly in type III cells reflecting their expression of Ascl1, but some type II cells were also labeled, indicating that Ascl1+ precursor cells can differentiate into either TRC type (Matsuyama et al., 2023). Global knockout of Ascl1 assessed in newborn mice causes loss of type III cells, suggesting Ascl1 function is necessary for differentiation of type III cells (Seta et al., 2011). How Ascl1 deletion affects adult TRC differentiation has yet to be tested, however, Ascl1+ cell lineage tracing labels mostly type III and occasional type II cells, similar to newborn mice (C.-C. Hsu et al., 2021; Matsuyama et al., 2023). Taken together, these data indicate that Ascl1 is expressed in a sub-population of taste precursor cells capable of generating either type II or type III cells in both neonates and adults. Since type III cells are the only differentiated cells that express Ascl1 in adults, precursors that maintain Ascl1 expression may be biased towards a type III cell fate (Figure 5). Ascl1 is a downstream transcription factor associated with Notch signaling, suggesting a role for Notch in adult taste cell differentiation during renewal, a role yet to be studied adult mice (but see below).

Lingual organoids have become a powerful in vitro tool to test the role of signaling pathways in taste cell differentiation. Organoids cultured in media with Wnt3a ligand grow larger and contain significantly more type I, II and III cells than organoids grown without it (Ren et al., 2017). Pharmacological enhancement of β-catenin activity in organoids also increases expression of taste cell markers (Shechtman et al., 2023), supporting in vivo results that Wnt positively regulates proliferation and TRC differentiation. Although activating Shh signaling in organoids does not affect taste cell differentiation (Ren et al., 2014; Shechtman et al., 2023), Hh pathway inhibition reduces differentiation of taste cells (Ren et al., 2017), consistent with Shh’s role in vivo. The role of Notch signaling has not been tested in mice but has been explored in organoids. Inhibition of Notch signaling with DBZ (a γ-secretase inhibitor) in organoids decreases differentiation of type I cells and increases differentiation of type II and III cells, suggesting Notch signaling promotes type I cell fate (Ren et al., 2017) (Figure 5). Notch inhibits Ascl1 expression in other systems (Kageyama et al., 2005), supporting a model in which Notch signaling drives type I cell fate by repressing Asc1. Genetic experiments in adult mice will be needed to fully investigate Notch function in adult taste renewal.

Organoids have also been used to identify additional signaling pathways that regulate taste homeostasis, at least in vitro. Adding A8301 (TGFβ pathway inhibitor) and SB202190 (MAPK inhibitor) to organoid culture promotes organoid growth and inhibits taste cell differentiation, while removing these drugs from the media allows robust differentiation of all taste cell types (Shechtman et al., 2021). These results suggest that TGFβ and MAPK signaling repress proliferation and promote differentiation of taste progenitors (Figure 5). Modulating BMP signaling with the antagonist Noggin also affects differentiation. Organoids cultured in Noggin conditioned media contain more differentiated taste cells than organoids grown in media lacking Noggin, suggesting BMP signaling inhibits taste cell differentiation (Ren et al., 2017) (Figure 5). Organoid experiments identifying TGFβ, MAPK and BMP pathways as regulators of taste cell differentiation in vitro can inform future work to elucidate how these pathways regulate taste renewal in vivo.

2-3: The embryo to adult transition coincides with birth

Following embryogenesis, Shh⁺ placode cells differentiate into the first taste cells at birth; these are steadily lost, presumably replaced by adult progenitors. Inducible labeling of embryonic Shh⁺ taste placode cells at mid-gestation revealed robust TRC labeling at P0, but these genetically marked taste bud cells are lost within a few months. Shh⁺ placodal cells thus do not give rise to adult taste progenitors (Thirumangalathu et al., 2009); rather new postnatal TRCs must arise from another lingual epithelial population activated after birth. Until recently, when this transition from embryonic development to adult homeostasis occurs has been unknown.

In adult mice, KRT14⁺ progenitors contribute to adult taste buds (Okubo et al., 2009 and see above). Embryonically KRT14 is expressed throughout the lingual epithelium prior to taste placode formation (Golden et al., 2021). However, as Shh⁺ placodes form (E12.5-13.5), they downregulate KRT14, and lineage tracing of KRT14⁺ basal cells after this time does not label taste bud primordia (Golden et al., 2021). Instead, KRT14 lineage tracing shortly after birth results in labeled cells in postnatal buds, indicating that the onset of new taste cell contribution of KRT14⁺ progenitors begin at birth (Golden et al., 2021; Okubo et al., 2009). In mid-gestation mouse embryos, Shh represses taste fates, while during late gestation Shh has no impact on taste bud development. In adult homeostasis, by contrast, Shh promotes TRC differentiation. This shift in Shh function coincides with the onset of taste bud cell renewal at birth, as genetic Shh overexpression in embryos repressed taste placode development but induced ectopic taste buds when initiated at P2 (Golden et al., 2021). How the Hh pathway switches roles is likely due to changes in downstream pathway effector gene expression. For example, expression of Hh target transcription factors, FoxA1 and FoxA2 is dynamic, and their expression shifts from the embryonic to the adult pattern around birth (Besnard et al., 2004; Golden et al., 2021; Luo et al., 2009).

To date the Shh pathway is the only one demonstrated to switch function at the embryo-to-adult transition, yet Wnt, Notch and Bmp pathway components are also expressed in the taste epithelium, and their expression patterns frequently differ between embryos and adults (Gaillard et al., 2017; Gaillard et al., 2015; Nguyen & Barlow, 2010; Seta et al., 2003; Seta et al., 2006; Zhou et al., 2006). In contrast to the changing role of Shh, Wnt/ß-catenin signaling has a pro-taste role in embryos that is conserved during homeostasis (Gaillard & Barlow, 2011; Gaillard et al., 2017; Gaillard et al., 2015; Iwatsuki et al., 2007; F. Liu et al., 2007). The BMP pathway promotes or represses taste fate at different stages of embryogenesis while its role during homeostasis has yet to be explored (Beites et al., 2009; Ishan et al., 2023; Zhou et al., 2006). Expression of Notch signaling components in both embryonic and adult taste epithelia is consistent with Notch functioning in both contexts. As discussed above, Notch has been implicated in type III TRC differentiation in mice and the development of 5HT⁺ type III-like TRCs in zebrafish (C. C. Hsu et al., 2021; Kapsimali et al., 2011; Seta et al., 2011; Seta et al., 2003; Seta et al., 2006). Several studies suggest a role for Notch in type II TRCs during development, where the Notch target Hes1 represses type II TRC fate (Ota et al., 2009). Thus, in addition to the Shh signaling pathway, the roles of others may also change as the epithelium transitions from embryonic development to adult homeostasis at birth.

3-1: Cancer treatments impact taste

Loss of taste (ageusia) or taste dysfunction (dysgeusia) are common side effects for patients undergoing cancer treatments. Depending on the treatment, between 20-80% of cancer patients report troubling dysgeusia (Berteretche et al., 2004; Cheng, 2007; Kaizu et al., 2021; Obayashi, Sugita, et al., 2023; Sevryugin et al., 2021; Vigarios et al., 2017; Zabernigg et al., 2010). This range is likely an underestimate, since ageusia/dysgeusia are generally underreported and underdiagnosed (Bernhardson et al., 2009). Patients with taste alterations report severe impacts on quality of life and can have decreased nutritional intake and worse long-term outcomes (Epstein & Barasch, 2010; Nolden et al., 2019). To mitigate these symptoms, increase drug tolerance and improve patient quality of life, we must understand the cellular and molecular mechanisms driving taste alterations. Here, we review cancer treatments known to cause taste loss/dysfunction and how these treatments affect the taste system (and see (Gaillard & Barlow, 2021) for detailed review).

Head and neck cancer patients receive tumor-targeted radiotherapy that nonetheless frequently damages taste papillae. These patients have taste loss that generally recovers, but a distorted sense of taste can last months or years (Bolze et al., 1982; Chen et al., 2019; Chencharick & Mossman, 1983; da Silva et al., 2019; Deshpande et al., 2018; Irune et al., 2014; Maes et al., 2002; Nelson, 1998; Ray-Chaudhuri et al., 2013). Acutely, radiation leads to loss of taste pores in head and neck cancer patients, which was inferred as loss of taste buds (Just et al., 2005). Radiation induces DNA damage in dividing cells and thus reduces tumor proliferation (Surova & Zhivotovsky, 2013; Wang, 2019). In mice, a single non-lethal dose of radiation to the head and neck likewise increased taste progenitor cell death, and transiently reduced proliferation (Figure 6), which in turn caused a transient interruption of new taste cell supply and reduced sensitivity to sweet tastant (Jewkes et al., 2018; Nguyen et al., 2012). However, rather than a single dose, patients are treated with fractionated radiotherapy, i.e. small daily doses of irradiation over several weeks. When mice are treated with a 5-day fractionated radiation paradigm, the taste epithelium response is more complex: proliferation is diminished and progenitor number drops (Figure 6), but apoptotic progenitors were not observed. Instead, TUNEL+ cells were evident within buds. We speculated irradiation forces any damaged progenitors to exit the cell cycle and then undergo partial taste differentiation, ultimately undergoing cell death after entering taste buds (Gaillard et al., 2019); this hypothesis remains to be tested. Importantly, in mice the taste epithelium recovers following either single or fractionated dosing, although recovery is more rapid after one dose. What underlies persistent taste dysfunction in head and neck cancer patients weeks, months or years after fractionated radiotherapy is not understood. Already we observed noticeable changes in cellular responses in fractionated versus single dosing of mice; it is plausible that multiweek fractionated radiotherapy that patients receive drives additional and more complex tissue level responses, as well as distinct mechanisms of recovery.

Many chemotherapies also target proliferating cells (Figure 6). Taxol-based chemotherapies, which block proliferation by stabilizing microtubules, e.g. docetaxel, paclitaxel, nab-paclitaxel (Kaizu et al., 2021) are highly associated with taste dysfunction (Amézaga et al., 2018; Buttiron Webber et al., 2023). Up to 60% of patients taking taxol-based drugs experience dysgeusia with decreased sensitivity to sweet, bitter and sour (Kaizu et al., 2021). Cisplatin and cyclophosphamide, alkylating agents that damage DNA in dividing cells, also cause dysgeusia (Campagna et al., 2018; Gamper et al., 2012; I et al., 2016; IJpma et al., 2017; Minakata et al., 2002; Sicchieri et al., 2019). In mice, both drugs decrease taste progenitor proliferation and induce apoptosis in taste epithelium (Mukherjee et al., 2017; Ren et al., 2023). Moreover, following the initial wave of cell death, the impact of cyclophosphamide on taste progenitor proliferation persists for up to 2 weeks. This two-stage process causes biphasic (early and late) loss of taste sensitivity to sweet, umami and salty tastants (B. C. Jewkes et al., 2018; Mukherjee et al., 2013a, 2013b; Mukherjee & Delay, 2011). Initially, decreased taste sensitivity is caused by direct damage to taste buds. Taste sensitivity recovers transiently but then again is diminished, consistent with decreased taste bud cell number assumed to be due to decreased progenitor proliferation (Mukherjee et al., 2013a, 2013b; Mukherjee & Delay, 2011; Mukherjee et al., 2017). In addition to directly impacting the taste epithelium, cyclophosphamide treatment results in reduced innervation of taste buds likely contributing to decreased taste sensitivity (Wood et al., 2024). Compared to a single dose, fractionated cyclophosphamide dosing in mice prolongs disruption of cell proliferation, delays recovery of taste bud cell number, and increases salt detection thresholds in ways that are longer lasting and more cyclic (Delay et al., 2019; B. C. Jewkes et al., 2018). These results suggest that fractionated drug dosing of mice, which is more like what patients receive, causes a more complex and sustained response to injury than a single drug treatment. The effects of cisplatin and cyclophosphamide (both single-injection and fractionated dosing) on taste epithelium are ameliorated somewhat by co-treatment with amifostine, which reduces free-radicals and increases DNA repair (Delay et al., 2019; Mukherjee et al., 2013a, 2013b; Ren et al., 2023; Sarkar et al., 2021). Thus, protecting taste epithelium from DNA damage by co-treatment with amifostine may be a plausible strategy to mitigate taste dysfunction induced by cancer therapies.

Activating mutations in Hedgehog (Hh) pathway genes underlies most basal cell carcinomas, and patients are often treated with Hh pathway inhibitors (Lin & Matsui, 2012; Teglund & Toftgård, 2010). Most of these target the Hh effector Smoothened (Smo) (Box 1B). As Hh signaling is a key regulator of taste homeostasis (Figure 5 and above), it is not surprising that patients frequently report taste dysfunction (Poggi & Kolesar, 2013; Sanmartín et al., 2021; Sekulic et al., 2012; Tang et al., 2012). Consistent with taste loss in patients, mice treated with Smo antagonists gradually lose taste buds over a span of several weeks (Castillo-Azofeifa et al., 2017; Kumari et al., 2015; Kumari et al., 2018; Lu et al., 2018; Yang et al., 2015). Gustatory nerve responses to tastants are also lost in drug treated mice (Kumari et al., 2015). However, innervation remains intact and nerve responses to tactile and cold stimuli are unaffected (Donnelly et al., 2022; Kumari et al., 2015; Kumari et al., 2018). After treatment, taste buds regenerate and gustatory nerve responses are restored, consistent with patients report of recovery of taste after cessation of drug (Fife et al., 2017; Kumari et al., 2017; Lu et al., 2018). Molecular genetic studies in mice have revealed that Hh signaling is required for taste bud homeostasis, specifically that the ligand, Sonic hedgehog (Shh), promotes taste cell differentiation (Figure 5 and see above). Indeed, lineage tracing of KRT5+ progenitors in mice treated with Smo antagonist revealed fewer new labeled taste cells compared to controls (Castillo-Azofeifa et al., 2017; Lu et al., 2018). Several studies suggest these antagonists also impact progenitor cell proliferation (Kumari et al., 2018; Lu et al., 2018). Thus, Smo inhibitors impact both progenitor proliferation and differentiation of new taste cells (Figure 6), and this combination leads to taste bud loss and presumably underlies taste dysfunction in patients.

Most chemotherapies disrupt taste, but how they do so is unexplored for most drugs. Patients taking regimens that include irinotecan report significantly more dysgeusia than regimens without it (Zabernigg et al., 2010). Irinotecan is a DNA topoisomerase inhibitor that targets proliferating cells during G2 and S phases (de Man et al., 2018). Surprisingly, taste progenitor proliferation and taste bud cell number were unaffected in irinotecan-treated mice. Instead, expression of the sodium channel ENaC was specifically reduced by drug treatment (Obayashi, Sakayori, et al., 2023) (Figure 6). ENaC is expressed in salt-responsive taste cells and required for salt taste detection; irinotecan-treated mice with decreased ENaC expression showed decreased behavioral responses to salt (Chandrashekar et al., 2010; Obayashi, Sakayori, et al., 2023). Why irinotecan did not inhibit taste progenitor proliferation and how it specifically affects ENaC+ salt-sensitive taste cells are important questions for future investigation.

Tyrosine kinase inhibitors (TKIs) comprise a large class of anti-cancer drugs, with high incidence of dysgeusia (Vigarios et al., 2017). TKIs inhibit receptor tyrosine kinases (RTKs), although the role of RTKs in taste homeostasis is under-explored. Expression of RTKs, including Ret, c-Kit, EGFRs, FGFRs and IGFRs have been reported within the taste system (Biggs et al., 2016; Choo & Dando, 2021; Donnelly et al., 2018; McLaughlin, 2000; Sukumaran et al., 2017). While IGFRs do not appear important in taste cell homeostasis, the role of the other RTKs needs to be defined (Biggs et al., 2016). Understanding how RTKs influence taste homeostasis will be important for identifying mechanisms of and potentially mitigating TKI-induced dysgeusia.

Developing cancer treatments with high efficacy in eliminating cancer cells is vital, but it is also important to consider how adverse side effects like ageusia/dysgeusia could be mitigated. Characterizing mechanisms by which specific cancer treatments cause dysgeusia will provide insight into development of strategies to mitigate this side effect, as well as cellular and molecular mechanisms that regulate taste homeostasis in health.

3-2: Response of the taste system to nerve Injury

For over 100 years we have known that maintenance of adult taste buds is nerve-dependent. Transection of the chorda tympani or glossopharyngeal nerve causes degeneration of FFP and CVP taste buds, respectively. When taste papillae are re-innervated, taste buds fully regenerate (Guagliardo & Hill, 2007; Guth, 1957, 1958; Miura et al., 2004; Oakley et al., 1993; Oakley & Witt, 2004; Oliver & Whitehead, 1992; Seta et al., 1999; Sollars et al., 2002; Torrey, 1931; Uchida et al., 2003; Vintschgau & Hönigschmied, 1877; Whitehead et al., 1987; Yee et al., 2005). These studies suggest that gustatory nerves supply trophic factors that support adult taste renewal. In fact, recent studies have now identified several specific proteins that function in neurotrophic support of taste buds.

Shh is necessary for adult taste bud homeostasis (see Figure 5 and above), and Shh expressed by taste precursor cells was presumed the required source of this ligand (Miura et al., 2014). More recently, several reports revealed that Shh is also produced by gustatory neurons that innervate fungiform taste buds (Castillo-Azofeifa et al., 2017; Kumari et al., 2017; Lu et al., 2018). However, neuron-specific knockout of Shh ligand does not affect taste homeostasis (Castillo-Azofeifa et al., 2017; Golden et al., 2021). While genetic deletion of the Hh effector Smo in lingual epithelium leads to taste bud loss (Kumari et al., 2017), epithelial knockout of Shh does not affect taste bud homeostasis. Only coincident knockout of Shh from both gustatory neurons and lingual epithelium causes taste bud loss in FFP (Castillo-Azofeifa et al., 2017). While glossopharyngeal gustatory neurons also express Shh, the role of neural Shh has not been assessed for CVP taste buds (Lu et al., 2018). In sum, these data indicate that together neural and epithelial Shh support taste bud homeostasis, and that each tissue-level Shh source can compensate for the other if one is lost. How this compensation occurs is unknown; each source alone may provide sufficient ligand, or the remaining source may upregulate Shh expression in compensation. Finally, and importantly, because loss of neural Shh did not result in taste bud degeneration, additional nerve-supplied factors are likely required to maintain taste buds.

R-Spondin’s comprise a family of 4 secreted factors (RSPO-1, −2, −3 and −4) that bind Lgr4/5/6 transmembrane receptors, which augments WNT/ß-catenin signaling [refs]. Recently RSPO-2 has been identified as a nerve-supplied trophic factor that supports taste bud renewal. Specifically, RSPO-2 is expressed by gustatory neurons in both the VII^th and IX^th cranial nerve ganglia, and while CVP taste buds degenerate following glossopharyngeal nerve transection (see above), systemic supply of adenovirally encoded RSPO-1 or RSPO-2 prevents degeneration of denervated buds (Lin et al., 2021). Additionally, RSPO ligands are important components of culture media for taste organoids (Ren et al., 2014; Shechtman et al., 2021). Removal of RSPO from the media prevents differentiation of taste cells in organoids (Lin et al., 2021). In addition to Lgr4/5/6 receptors, RSPOs bind the transmembrane E3 ligases RNF43/ZNRF3, all of which are expressed by taste progenitors (Lu et al., 2022). Binding of RSPO to Lgr homologs positively regulates WNT/β-catenin signaling, while RNF43/ZNRF3 negatively regulate WNT by targeting Frizzled receptors for endocytosis (de Lau et al., 2014; Hao et al., 2012; Planas-Paz et al., 2016). In taste tissue, RNF43/ZNRF3 double knockout results in taste epithelium hyperplasia and increased taste cell differentiation, consistent with the impact of increased WNT/ß-catenin signaling [Gaillard et al, PLOS Gen 2015]. Further, the double knockout phenotype is rescued by Wnt inhibition (Lu et al., 2022), again suggesting RNF43/ZNRF3 serve as a break on taste homeostasis. Interestingly, CVP taste buds do not degenerate after glossopharyngeal nerve transection in the RNF43/ZNRF3 double-knockout mice. This supports a model where RSPO from nerves binds RNF43/ZNRF3 to release inhibition of WNT signaling that in turn promotes taste cell renewal (Lu et al., 2022).

Shh and RSPOs are trophic factors supplied by gustatory nerves to maintain taste epithelium, but taste buds in turn also produce trophic factors that support gustatory neurons. BDNF expressed by developing taste primordia is required for their initial innervation (Lopez & Krimm, 2006; Ma et al., 2009; Nosrat et al., 1997; Ringstedt et al., 1999). BDNF is also required for survival of gustatory neurons and maintenance of taste buds during development (Mistretta et al., 1999; Nosrat et al., 1997; Patel et al., 2010; Zhang et al., 1997). In adults, BDNF is expressed by taste bud cells (Meng et al., 2015; Yee et al., 2003). Conditional, ubiquitous knockout of BDNF in adult mice decreases innervation of FFP taste buds without causing neuron death and results in smaller taste buds. Epithelium-specific knockout of BDNF in adults also decreases innervation of taste buds, suggesting BDNF from adult taste buds is required to maintain gustatory innervation but not neuron survival (Meng et al., 2015). BDNF binds the receptor TrkB, which is expressed by a subset of gustatory neurons. BDNF is required to maintain innervation by TrkB+ neurons, while TrkB- neurons are unaffected by BDNF knockout (Rios-Pilier & Krimm, 2019; Tang et al., 2017). BDNF expression is highest in type III TRCs, suggesting TrkB+ gustatory neurons preferentially innervate type III cells while TrkB- neurons may innervate type II cells (Huang et al., 2015; Yee et al., 2003). Trophic factors within adult taste buds that maintain innervation by TrkB- gustatory neurons have yet to be identified. Whether BDNF supports innervation by the glossopharyngeal nerve in adult CVP is also not known. Since BDNF is needed for initial innervation of both FFP and CVP taste buds during development, BDNF function may also be similar in FFP and CVP in adults.

To summarize, taste bud homeostasis requires innervation by gustatory neurons, which supply Shh and RSPOs needed for proper Shh and Wnt signaling, respectively. When innervation is interrupted by infection or injury, taste buds likely degenerate due in part to dysregulation of these pathways. BDNF from taste buds acts as a chemoattractant to target gustatory neurites to taste papillae during development, and in adulthood maintains the gustatory innervation.

3-3: Inflammation and immune surveillance in the taste system

Inflammation has wide-sweeping effects on numerous systems and can impact a multitude of processes from basic cellular functions to tissue-level responses to a range of inflammation-induced signals (Chen et al., 2018; Medzhitov, 2021; Soares et al., 2023). Recent work has identified several key players in inflammatory response pathways in taste tissue, including interferon receptors, Toll-like receptors and corresponding downstream signaling components, suggesting inflammatory signaling plays a role in taste homeostasis (Wang et al., 2009). Several different models of inducing an inflammatory response have now been used to assess how inflammation affects taste cell biology and taste function.

Early work revealed that in mice, the response to taste stimuli is altered when an immune response is mounted following nerve transection (Cavallin & McCluskey, 2005; Hill & Phillips, 1994; Phillips & Hill, 1996). Another common inflammation model is treatment of animals or cell cultures with lipopolysaccharide (LPS), a component of gram-negative bacterial membranes, which induces the release of proinflammatory cytokines (Gauthier et al., 2022; Page et al., 2022; Skrzypczak-Wiercioch & Sałat, 2022). Immune cells are present in the taste system following an inflammatory response (Cavallin & McCluskey, 2005; Feng et al., 2020; Phillips & Hill, 1996), suggesting inflammatory pathways may contribute to altered taste perception (Figure 7). More recent studies using both in vivo and in vitro models reveal that type II and III TRCs as well as their associated nerve fibers respond to LPS treatment by producing nitric oxide (Wu et al., 2022). Changes in taste epithelium at the molecular genetic level have recently been assessed via comparison of scATAC-seq data from CVP epithelium from LPS-treated vs control mice, to identify changes in chromatin landscape in response to inflammation (Lin et al., 2023). Unsurprisingly, chromatin accessibility varied across taste epithelial cell types; however, a large change was observed in type II TRCs with more than 20 bitter receptor genes (the type 2 taste receptors (Tas2rs) displaying increased accessibility, correlating with a previously demonstrated change in bitter perception with LPS treatment (Wu et al., 2022) (Figure 7). LPS-treated mice had decreased taste nerve responses to sour, though changes in sour sensing type III TRC chromatin states were not discussed. Regardless, these data suggest that inflammation induces changes in TRC chromatin accessibility, which is correlated with changes in taste function. Interestingly, taste progenitors had increased accessibility of key immune regulator genes, i.e., Nfkb1, and markers of immune response pathways, i.e., MAPK, TNF, IF-y, and others. These results suggest progenitors are also an immune-responsive population which in turn could alter taste epithelial cell composition (Figure 7). It remains to be determined how changes in genome accessibility impact progenitor and TRC transcriptomes, and how these changes impact TRC function and/or if cell renewal is affected.

The role of inflammation in the taste system has also been investigated in the context of obesity, a known driver inflammation in many tissues (Soták et al., 2024). In humans, obesity and increased body mass index are correlated with blunted sweet and salty taste (Overberg et al., 2012; Park et al., 2015; Vignini et al., 2019), though a clear link to between diminished taste and inflammation in humans has not been reported. However, mice fed a high fat diet that become obese have increased expression of the inflammatory marker TNFa in taste epithelium (Kaufman et al., 2018). Obese mice had fewer and smaller CVP taste buds as well as increased cell death and reduced progenitor proliferation compared to controls. Intriguingly, taste bud number and size were rescued in TNFa knock out mice fed a high fat diet that nonetheless became obese. The authors proposed a model where obesity causes TNFa-induced inflammation, resulting in fewer and smaller taste buds, which they suggested could also occur in humans, and underly reduced taste perception in high BMI individuals.

Tuft cells of the intestine serve numerous roles including secretory functions, epithelial regeneration in response to injury, and pathogen-sensing to trigger an immune response (Gerbe et al., 2016; Howitt et al., 2016; Huang et al., 2024; von Moltke et al., 2016). Tuft cells in many epithelia express markers of a mounted immune response both in homeostasis and in response to parasitic insult (expression of IL-25, IL-13, IL-33). These cells also have interactions with group 2 innate lymphoid cells (Hendel et al., 2022; Schneider et al., 2019). Type II TRCs share many similarities with tuft cells in intestine and airway, most notably, all require Pou2F3 for differentiation and share markers expressed by type II TRCs. Most recently, type II TRCs have been suggested to function as pathogen-sensors with immune responsiveness (Qin et al., 2023). This study revealed similarities between type II TRCs and intestinal microfold cells (M cells); the latter interface directly with underlying immune cells. These authors found that specifically sweet/umami type II TRCs express makers of M Cells (GP2, Spib and CCL9), and that their expression increases upon administration of RANKL, a factor required for M-cell differentiation in the intestine (de Lau et al., 2012; Knoop et al., 2009). This group also found that Pou2f3 knock out mice have decreased expression of M-cell genes in taste epithelia consistent with type II cell loss in the genetic model [Matsumoto et al], further supporting a role for type II cells in immune surveillance. These similarities between intestinal M-cells and sweet/bitter type II cells suggest that sweet/bitter type II TRCs may similarly communicate with underlying resident immune cells during oral immune surveillance (Figure 7) (and see (Xi et al., 2022). Ultimately, these similarities between type II TRCs and other tissue resident tuft-like cells present a new line of inquiry for researchers moving forward. Future work could address how potential immune dynamics mediated by sweet/bitter type II TRCs may affect changes in taste homeostasis.