METTL3-mediated m6A RNA methylation regulates dorsal lingual epithelium homeostasis

The dorsal lingual epithelium, which is composed of taste buds and keratinocytes differentiated from K14+ basal cells, discriminates taste compounds and maintains the epithelial barrier. N6-methyladenosine (m6A) is the most abundant mRNA modification in eukaryotic cells. How METTL3-mediated m6A modification regulates K14+ basal cell fate during dorsal lingual epithelium formation and regeneration remains unclear. Here we show knockout of Mettl3 in K14+ cells reduced the taste buds and enhanced keratinocytes. Deletion of Mettl3 led to increased basal cell proliferation and decreased cell division in taste buds. Conditional Mettl3 knock-in mice showed little impact on taste buds or keratinization, but displayed increased proliferation of cells around taste buds in a protective manner during post-irradiation recovery. Mechanically, we revealed that the most frequent m6A modifications were enriched in Hippo and Wnt signaling, and specific peaks were observed near the stop codons of Lats1 and FZD7. Our study elucidates that METTL3 is essential for taste bud formation and could promote the quantity recovery of taste bud after radiation.


INTRODUCTION
The lingual epithelium consists of non-taste epithelium and taste epithelium. The non-taste epithelium covers a large proportion of the tongue's surface. In mice, fungiform papillae (FFP), surrounded by mechanosensory filiform papillae (FLP), are distributed in the front part of the dorsal lingual epithelium 1 . The single circumvallate papilla (CVP) in mice, containing numerous taste buds, is located in the midline of the posterior lingual epithelium, whereas singular FFP houses only one taste bud 2 .
The taste system is mediated by the taste buds and innervated sensory neurons. Murine taste buds contain 50-100 elongated epithelial cells, which can be categorized into several types (types I, type II, and type III) 3 . Through different specific receptors, taste buds can detect five taste qualities: bitter, salt, sweet, sour, and umami (savory) [4][5][6] . Dysgeusia is common in patients undergoing head and neck radiotherapy [7][8][9] . Some patients may recover from taste dysfunction after some months or years, but a small minority of patients may suffer from permanent taste loss 10,11 . Loss of taste buds after radiation is caused by natural taste cell death and the interruption of taste cell replenishment 8 .
Previous studies using lineage tracing indicate that basal cells expressing cytokeratin 5 (K5 or Krt5) and K14 are progenitors of both non-taste epithelium and taste epithelium in mice 12 . Taste buds undergo continuous turnover, with an average life span of 10-14 days, while the non-taste epithelium takes 5-7 days to be renewed 13,14 . In non-taste epithelium, K5/K14 + basal progenitors differentiate into K13 + (KRT13) keratinocytes, which make up the suprabasal epithelial layers of the FLP and FFP 15,16 . K5/K14 + basal progenitors can also generate new cells into taste buds, subsequently developing into mature taste cell types 1 . Type I cells resemble glia and are the most abundant cells present in taste buds. However, the specific function of type I cells remains elusive 17,18 . Some researchers regard type I cells as salt detector 4 . Type II cells can transduce different signals by detecting bitter, sweet, and umami tasting stimuli, whereas type III cells can transduce sour flavors [19][20][21] . These three different cell types can be identified via distinctive markers: type I cells express ecto-ATPase, NTPdase2; type II cells express α-gustducin, phospholipase Cβ2; and type III cells express NCAM and SNAP-25 [22][23][24] .
Here, our group generated an epidermis-specific Mettl3 knockout mouse model and found that METTL3 was an essential RNA methyltransferase that regulated lingual epithelium progenitor differentiation and was crucial for taste bud development. Moreover, overexpression of Mettl3 promoted taste bud recovery from radiation injury. We also discovered a mechanistic pathway by identifying downstream target genes and signals.

RESULTS
Deletion of Mettl3 in epidermal progenitors leads to taste bud defects As previously described, basal progenitors develop into taste cells (K8) and differentiated keratinocytes (K13) (Fig. 1a) 15,16,40 . To explore whether deletion of Mettl3 affected taste bud development, we crossed Mettl3 fl/fl mice 28 with K14-Cre transgenic mice to conditionally delete Mettl3 from epithelium basal progenitors. K14-Cre;Mettl3 fl/fl mice were viable and born at Mendelian's ratio. Most of them survived by postnatal day 4 (P4), and few could survive by P7. We also generated the K14-Cre;tdTomato;Mettl3 fl/fl mice, in which K14 + cells and their daughter cells were labeled with tdTomato fluorescence, and confirmed that METTL3 was largely abolished within Tomato + cells in the CVP (Fig. 1b).
Because of their short lifespan, we sacrificed K14-Cre; Mettl3 fl/fl mice at P4 to explore whether taste bud differentiation was affected. Compared to control CVPs, the taste buds in mutant CVPs could not be recognized by hematoxylin and eosin (H&E) staining (Fig. 1c). K8 is a marker of differentiated taste bud cells 40 . Immunofluorescence staining revealed that the number of K8 + cells in the CVP was significantly decreased (Fig. 1d). We also observed a reduced number of type II cells (marked by gustducin) and type III cells (marked by SNAP25) in mutant CVP (Fig. 1e, f). Consistent with the taste bud loss phenotype, the innervated areas (marked by PGP9.5) of mutant CVP were remarkably reduced (Fig. 1g).
Deletion of Mettl3 leads to abnormal keratinization of lingual epithelium We then examined whether METTL3 regulated keratinization of the lingual epithelium. Scanning electron microscopy (SEM) showed that epithelial-specific Mettl3 deletion caused morphological abnormalities in FLP at P4 (Fig. 2a). Excessive keratinized fragments were observed on the surface of the tongue epithelium in mutant mice (Fig. 2a). H&E staining showed that the thickness of the epithelium was increased and the cell alignment was irregular compared to that of control mice (Fig. 2a). Quantitative analysis of epithelium thickness showed that the entire epithelium thickness of the mutants was almost double that of the controls (Fig. 2e). This observation was confirmed by immunofluorescence staining of PAN-CK and K13 cells (Fig. 2b, c, f, g). Although there was no obvious difference in the appearance of FFP by SEM, the number of K8 + taste cells in FFP decreased in the mutant mice (Fig. 2d).
Deletion of Mettl3 increases basal cell proliferation but decreases cell proliferation around taste buds Transcription factor p63 is a crucial regulator of epidermal development and marks basal stem cells 41 . We noted that the number of P63 + cells in the CVP increased in the mutants (Fig. 3a). According to immunofluorescence results for P63 and 5-ethynyl-2′-deoxyuridine (EdU), the proliferation of P63 + cells in mutant CVP was more active than in control mice (Fig. 3a). Consistent results were observed in the FLP analysis (Fig. 3b). Deletion of Mettl3 led to an increase in the number of P63 + cells and promoted the proliferation of basal cells (Fig. 3b). In contrast, cell proliferation around taste buds reduced in Mettl3 knockout mice (Fig. 3c). Deletion of Mettl3 did not affect apoptosis of taste or non-taste cells (Fig. 3d, e).
Overexpression of Mettl3 promotes taste bud recovery after irradiation Next, we investigated whether Mettl3 overexpression could prevent lingual epithelial disorders. To this end, we generated K14-Cre driven Mettl3-tdTomato knock-in mice (K14-Cre;Mettl3 KI/KI ) to conditionally overexpress Mettl3 in epidermal progenitor cells 28 . As K14-Cre;Mettl3 KI/KI mice did not exhibit a significant change in taste bud and epithelium development, we challenged them with 15 Gy irradiation (Fig. 4a). Both knock-in and control mice exhibited a severe loss of taste buds at 7 days post-irradiation (dpi) (Fig. 4b, c), indicating that overexpression of Mettl3 could not protect mice from epithelial injury due to irradiation. Notably, at 14 dpi, there were more recovered taste buds in the knock-in mice (Fig. 4b, c). Furthermore, overexpression of Mettl3 increased the proliferation of cells around taste buds at 7 dpi (Fig. 4d).

METTL3-mediated m 6 A RNA methylation regulates Hippo and Wnt pathways
To explore the underlying mechanisms, we performed m 6 A RNA immunoprecipitation sequencing (m 6 ARIP-seq) of the lingual epithelium in K14-Cre; Mettl3 fl/fl mice and their Mettl3 fl/fl littermates at P4. Consistent with previous reports 28,42 , m 6 A peaks shared a common GGACU motif (Fig. 5a) and were enriched around the stop codon (Fig. 5b). Gene pathway analysis revealed that the most frequent changes involving m 6 A modifications after Mettl3 deletion were enriched in the Hippo and Wnt signaling pathways (Fig. 5c).
Specific peaks were observed near the stop codons of the large tumor suppressor kinase 1 (Lats1) and Frizzled class receptor 7 (Fzd7) (Fig. 5d). The deletion of Mettl3 significantly decreased the abundance of m 6 A modifications (Fig. 5d). Immunofluorescence staining and western blot analysis confirmed reduced LATS1 and FZD7 protein levels, respectively (Fig. 5e, f). Loss of METTL3 led to increased nuclear localization of YAP and TAZ in the basal epithelium ( Fig. 5g, h). FZD7, a Wnt receptor, transduces signals and activates the Wnt signaling pathway 43,44 . Immunofluorescence staining for β-catenin indicated that nuclear β-catenin expression by K14-Cre; Mettl3 fl/fl mice at P4 was significantly decreased in terms of both taste and non-taste epithelia (Fig. 5i). In addition, LEF1 expression was decreased in K14-Cre; Mettl3 fl/fl mice (Fig. 5j).

DISCUSSION
In recent years, many studies have reported a role for METTL3 in stem cell differentiation. METTL3-mediated m 6 A modification regulates the development of the hematopoietic system, spermatogenesis, and other organs 45,46 . Our research group also found that m 6 A modification mediated by METTL3 plays a key role in regulating the fate of bone marrow mesenchymal stem cells 28 . In this study, we uncovered an essential role for METTL3 in lingual epithelial homeostasis. METTL3 was widely expressed in the lingual epithelium. Deletion of Mettl3 led to severe defects in taste bud development and abnormal epithelial thickening. In addition, overexpression of Mettl3 promoted taste bud recovery from radiation damage. Furthermore, the Wnt and Hippo signaling pathways may be responsible for the striking phenotypic defects in taste buds and keratinizing epithelium.
The Wnt/β-catenin pathway is required for taste bud development and taste cell renewal 15,47,48 . Basal cells with high expression of β-catenin can give rise to taste cells, whereas lower levels of β-catenin expression promote keratinocyte fate 1,15 . Conditional β-catenin knockout in progenitor cells causes a decrease in taste cells 48 . Here, we showed that the deletion of Mettl3 resulted in the downregulation of β-catenin and LEF1, which led to defects in taste cell development.
However, a study demonstrated that conditional knockout of Ctnnb1 in the epidermis led to FLP and FFP developmental defects and thinner epithelium, which did not match our phenotype 49 . In our study, β-catenin expression was extremely low after Mettl3 knockout, whereas the lingual epithelium was thickened. This finding reminded us that different mechanisms regulate the nontaste epithelium. METTL3-mediated m 6 A modification has been identified as the most abundant mRNA modification that regulates biological processes in mRNA 27,50 . Deletion of Mettl3 reduces the m 6 A peaks of many mRNAs, which changes their fate choices. The m 6 ARIP-seq analysis of the lingual epithelium showed that m 6 A modifications were mainly enriched in the Hippo and Wnt signaling pathways (Fig. 5c). LATS1 is a crucial kinase that    phosphorylates and inactivates the transcriptional coactivators YAP and TAZ [51][52][53] . In our study, we found that loss of METTL3 reduced the m 6 A modification of LATS1 and further inhibited the Hippo pathway, resulting in abnormal proliferation of keratinizing epithelium.
Another area for investigation is how METTL3 regulates cellular physiological functions and pathological progression through m 6 A modification. A number of studies have elaborated on the role of RNA m 6 A modifications in alternative splicing 42,45 . A recent study also pointed out that m 6 A methylation regulates an array of   chromosome-associated regulatory RNAs (carRNAs) to globally tune chromatin state and gene transcription 54 . In some studies, METTL3 and m 6 A modifications have been shown to enhance mRNA stability to promote cell proliferation 37,55 . m 6 A has been shown to promote mRNA translation in certain cell types 56,57 . In our previous study, the deletion of Mettl3 slowed down the translation of target proteins and further inhibited downstream signaling 28 . m 6 A modifications are important and function depending on the cellular context. Thus, we did not further investigate mRNA metabolism in this study. Dysgeusia is common and significant in patients after receiving head and neck radiotherapy. A previous study found that radiation interrupts the renewal of taste bud cells by inhibiting the proliferation and differentiation of basal progenitor cells, resulting in taste bud injury in mice 8 . In subsequent studies, they found that activation of the Wnt/β-catenin signaling pathway promotes the recovery of taste cells from radiation 58 . To determine the function of METTL3 in radiation-induced gustation dysfunction, conditional Mettl3 knock-in mice were exposed to radiation and analyzed for taste bud maintenance and recovery. We found that overexpression of Mettl3 promoted taste bud recovery from radiation damage by increasing the proliferation of taste bud progenitor cells. Interestingly, overexpression of Mettl3 did not protect taste buds from radiation injury.
In conclusion, we elucidated that METTL3 was an essential regulator of lingual epithelial homeostasis by regulating m 6 A modification. Deletion of Mettl3 in the epidermis reduced the expression of LATS1 and FZD7 and further blocked downstream pathways, which led to taste bud defects and epithelial thickening. In addition, overexpression of Mettl3 promoted taste bud recovery from radiation damage by increasing the proliferation of taste bud progenitor cells.
All mice were housed in specific pathogen-free (SPF) facilities with a 12-hour light-dark illumination cycle. All studies performed on mice were approved by the Subcommittee on Research and Animal Care (SRAC) at Sichuan University.
Tissue preparation After anesthesia with xylazine (10 mg·kg −1 ) and ketamine (80 mg·kg −1 ), mice were perfused transcardially with normal saline and 4% paraformaldehyde (PFA) in 0.1 mol·L −1 phosphate buffer. Tongues were dissected from the mandible and fixed in 4% PFA overnight at 4°C. For frozen sections, tissues were transferred to 20% sucrose in 0.1 mol·L −1 phosphate buffer overnight at 4°C. The samples were embedded in OCT compound (Sakura Finetek, Torrance, USA) and cryosectioned to 12 µm. For paraffin sections, the samples were dehydrated in graded ethanol and xylene and then embedded in 5-µm-thick paraffin sections using a microtome (Leica, RM2255, Wetzlar, German).
Histology and immunofluorescence For hematoxylin and eosin (H&E) staining, the paraffin sections were de-waxed using graded xylene solutions. Staining was performed according to the manufacturer's instructions (Solarbio Science and Technology, Beijing, China).
For immunohistochemistry, paraffin sections were prepared by the above procedures, microwaved in sodium citrate buffer, and incubated with primary and secondary antibodies (Boster Biological Technology, Wuhan, China). Finally, the sections were developed using the AEC (3-amino-9-ethylcarbazole) Staining Kit (Boster Biological Technology, Wuhan, China).
For immunofluorescence, cryosections were thawed at room temperature (26°C), rehydrated in 0.1 mol·L −1 phosphate-buffered saline and microwaved in sodium citrate buffer. After incubation with the primary antibody, the sections were incubated with Alexa Fluor-labeled secondary antibodies (Jackson Laboratory, Pennsylvania, USA).

Antibodies
The following antibodies were used: rabbit anti-METTL3 (

Irradiation
After anesthesia, the mice were covered with a custom-made lead shield, leaving the head and neck exposed. The mice were irradiated with 15 Gy in an X-ray irradiator (~1.25 Gy per min, Accela, X-RAD 160). Irradiated mice were returned to their cages for recovery.
TUNEL assay To assess cell death, the TUNEL assay was performed using the In Situ Cell Death Detection Kit TMR Red (Boster Biological Technology, MK 1012-100, Wuhan, China). The sections were digested with proteinase K for 5 min and then washed in tris-HCl buffered saline three times. Sections were incubated with labeling buffer for 2 h at 37°C prior to TUNEL reactions. Labeling buffer was prepared according to the manufacturer's instructions. After two washes, the sections were incubated in blocking solution at room temperature for 30 min. The blocking solution was then removed. The fluorescence probes were used to detect cell death. Sections were counterstained with DAPI and imaged using laser scanning confocal microscopy (LSCM; Olympus FV3000, Tokyo, Japan).
EdU labeling For EdU labeling, mice were injected with 25 µg of EdU per gram of body weight and euthanized after 1 h. Tongues were fixed overnight in 4% PFA and embedded in paraffin. After de-waxing, paraffin sections were incubated with the Click-iT EdU Imaging Kit (Invitrogen, CA, USA).

SEM
Tongue samples were fixed in 4% PFA overnight and dehydrated in a graded series of ethanol concentrations. Dehydrated samples were then incubated in 50% hexamethyldisilazane (Sigma-Aldrich, St. Louis, USA) for 20 min, followed by three solvent changes to 100% hexamethyldisilazane. After air-drying overnight, the samples were sputter-coated with gold-palladium. Specimens were examined and photographed using a SEM 49 .
Statistical analysis All data are presented as means ± standard error. For comparison between two independent groups, statistical differences were analyzed using unpaired two-tailed Student's t-test. Statistical significance was set at p < 0.05.

DATA AVAILABILITY
All data are available in the main text.