key: cord-0321775-dco0o1ey authors: Lyras, Ekaterini Maria; Zimmermann, Karin; Wagner, Lisa Katharina; Dörr, Dorothea; Fischer, Cornelius; Jung, Steffen; Yona, Simon; Hovav, Avi-Hai; Stenzel, Werner; Dommerich, Steffen; Conrad, Thomas; Leutz, Achim; Mildner, Alexander title: Tongue immune compartment analysis reveals spatial macrophage heterogeneity date: 2022-02-10 journal: bioRxiv DOI: 10.1101/2022.02.09.479699 sha: 0200079df7f1f8609d49b398e477deff3ef83fcf doc_id: 321775 cord_uid: dco0o1ey The tongue is a unique muscular organ situated in the oral cavity where it is involved in taste sensation, mastication and articulation. As a barrier organ, which is constantly exposed to environmental pathogens, the tongue is expected to host an immune cell network ensuring local immune defence. However, the composition and the transcriptional landscape of the tongue immune system are currently not completely defined. Here we characterised the tissue-resident immune compartment of the murine tongue during development, health and disease, combining single cell RNA-sequencing with in situ immunophenotyping. We identified distinct local immune cell populations and described two specific subsets of tongue-resident macrophages occupying discrete anatomical niches. Cx3cr1+ macrophages were located specifically in the highly innervated lamina propria beneath the tongue epidermis and at times in close proximity to fungiform papillae. Folr2+ macrophages were detected in deeper muscular tissue. The two macrophage subsets originate from a common proliferative precursor during early postnatal development and responded differently to systemic LPS in vivo. Our description of the under-investigated tongue immune system sets a starting point to facilitate research on tongue immune-physiology and pathology including cancer and taste disorders. The tongue is a highly innervated muscular organ with functions in articulation, mastication and taste perception. Located at the entrance of the gastrointestinal tract, the tongue is constantly exposed to dietary and airborne antigens and therefore acts as a firstline immune organ 1 . Moreover, taste sensation plays a critical role in avoidance of spoiled food and beverages. Accordingly, a tongue-resident immune network would be expected with roles in immune defense, tissue remodeling and tongue homeostasis. However, in the immunological context, the tongue is an understudied organ and the composition of tongue immune cells and their transcriptional status is largely unknown. Here, we define the immune cell landscape of the tongue with a specific focus on mononuclear phagocytes, e.g. tissueresident macrophages (TRM). Until now, the characterization of the mononuclear phagocyte compartment of the tongue mainly focused on Langerhans cells that were first described in the mouse epithelium forty years ago 2 and are identified in humans by their exclusive CD1a immunoreactivity 3, 4 . The characterization of sub-epithelial macrophage subsets is much more enigmatic and so far depended on the histological examination of a few membrane markers 5, 6 . For example, human CD163 + macrophages can be found in subepithelial areas of the tongue 6 , a localization that they share with CD11c + "dendritic" cells 7 . Relying on single markers such as CD11c to identify cells of the "dendritic" cell lineage is problematic since various additional cell types, including monocytes, macrophages and lymphocytes, can express CD11c 8 . Besides histological examinations of mononuclear phagocytes in the healthy tongue, the macrophage involvement in various pathological settings has also been studied. Tongue Langerhans cells were for instance shown to be critically involved in IL17-dependent antifungal immunity in the oral mucosa 9 , play an important role in T cell priming during squamous cell carcinoma development 10 and are depleted in patients with advanced-stage acquired immune deficiency syndrome 11 . Furthermore, an increase of activated ED1 + tongue macrophages was observed in systemic inflammation in rats 12 . The recent COVID-19 pandemic has also suggested a potential link of viral infections with tongue immunity, with loss of taste being one of the hallmark symptoms 13 . However, the lack of knowledge of the tongue immune cell compartment in physiology, hampers our understanding of tongue immune responses following pathogen challenge. Therefore, an unbiased characterization of the tongue immune cells is critical to classify and evaluate tissue-resident cell subsets, e.g. macrophage dynamics during tongue development and pathologies. To this end, we profiled the tongue-resident CD45 + hematopoietic cell compartment by single cell RNA-sequencing (scRNA-seq). Amongst tongue innate lymphoid cells, e.g. ILC2, and the specific presence of mast cells in early postnatal tongues, we further identified two main Irf8-independent macrophage populations, which were characterized by Cx3cr1 and Folr2 expression, respectively. Cx3cr1-expressing macrophages were specifically enriched in the lamina propria of the tongue and were detected in fungiform papillae, which harbor taste buds, but were absent from the epidermis. Folr2-expressing tongue macrophages localized in muscular tissue and in the lamina propria. These anatomical niches were colonized during embryonic and early postnatal development from a Cx3cr1expressing precursor of high proliferation capacity. Both macrophage populations showed a robust inflammatory response after in vivo lipopolysaccharide (LPS) administration, including shared and unique pathways. In summary, our data provide a detailed atlas of the immune cells of the tongue that will facilitate future research of this under-investigated barrier organ. To examine the tissue-resident immune compartment of the tongue in an unbiased manner, we performed scRNA-seq of FACS-purified CD45 + hematopoietic tongue cells isolated from PBS-perfused adult wild-type C57BL/6 mice. Two biologically and technically independent 10X Chromium experiments were performed that yielded highly reproducible results (Suppl. Fig. 1a+b ). We sequenced a total of 6773 cells that clustered into 19 transcriptionally distinct subsets (Fig. 1a+b) and used singleR 14 for cell lineage recognition (Suppl. Fig. 1c) Fig. 1c) , these cells had neither a strong DC nor a strong macrophage signature (Fig. 1b) . As the specific cDC2 gene signature 17 was also absent in this cluster (Suppl. Fig. 1d ), the precise ontogeny of these cells will need further investigation. The GSVA analysis revealed three clusters of cells with macrophage identity, of which clusters 0 and 5 are likely end-stage differentiated macrophage subsets (Fig. 1b) . Cells in the third macrophage cluster (cluster 1) seem to be macrophages of intermediate differentiation, as their transcriptomic signature shares features with both cluster 0 and 5 ( Fig. 1c+d ). Of the two terminally differentiated macrophage clusters, cluster 0 (hereafter referred to as tFOLR2-MF) expressed high levels of Folr2, Lyve1, Pf4 and Timd4, while cluster 5 (hereafter referred to as tCX3CR1-MF) expressed high levels of Cx3cr1, Hexb, Ms4a7, Itgax and Pmepa1 (Fig. 1c+d) . A pairwise comparison of tFOLR2-MF and tCX3CR1-MF transcriptomes revealed 602 differentially expressed genes (DEGs), indicating major differences between the two tongue macrophage populations (abs(FC) > 1.5 and adjusted pvalue < 0.05) (Fig. 1e) . Both tCX3CR1-MF and tFOLR2-MF were transcriptionally distinct from tongue Langerhans cells (tLCs; Fig. 1e ), which rather fell under the cDC signature ( Suppl. Data 4). tFOLR2-MF on the other hand were enriched for gene sets associated with blood vessel biology ("regulation of sprouting angiogenesis") and macrophage function ("cytosolic transport", "macrophage differentiation" and "positive regulation of nitric oxide biosynthesis"); tCX3CR1-MF showed gene enrichment for broad immune biological processes such as "positive regulation of ERK1/2 cascade" or "TLR3 signalling pathway" and exhibited similarities to CNS-resident microglia with gene enrichment in the biological process "microglial cell proliferation" (Fig. 1f) . Recent studies have indicated potential interactions of Cx3cr1-expressing macrophages with neurons [18] [19] [20] . However, unlike these CX 3 CR1 + brown adipose tissue and skin nerve-associated macrophages 18-20 , we did not detect a specific and significant enrichment for genes involved in axon guidance, such as Plexina4 in tCX3CR1-MF (Suppl. Altogether, our scRNA-seq data show that the tongue harbours a wide range of tissue-resident immune cells, of which the majority belong to the mononuclear phagocyte system. We identified two terminally differentiated macrophage subsets: tCX3CR1-MF that have a transcriptomic signature associated with innate immune signaling and tFOLR2-MF that seem to function in blood vessel biology and phagocytosis. We next established a protocol for the identification and isolation of tCX3CR1-MF and tFOLR2-MF by flow cytometry in Cx3cr1 Gfp/+ reporter mice 21 . After DNase/Collagenase IV/Hyaluronidase digestion of the tongue (Suppl. Fig. 2a) , we were able to detect CD64 + cells that could further be separated into cells expressing high levels of Cx3cr1-GFP (tCX3CR1-MF) and cells that stained positive for Folr2 (tFOLR2-MF; Fig. 2a ). tCX3CR1-MF also expressed the surface receptors CX 3 CR1, F4/80, MHCII and CD11c, while tFOLR2-MF were additionally characterized by LYVE1 and TIMD4 expression (Fig. 2b) . These surface characteristics could also be used to identify the tCX3CR1-MF and tFOLR2-MF in WT Bl6 animals. To identify tongue-specific signatures of tCX3CR1-MF and tFOLR2-MF and to place them in the context of macrophage biology, we compared the transcriptional profiles of To validate the data quality, we compared expression of common macrophagerelated genes in the different macrophage populations (Fig. 2c) . All cells except Langerhans cells expressed the macrophage genes Cd68, Mertk and Fcgr1. Furthermore, Csf1r expression was detected in almost all macrophage subsets with particularly high levels in microglia, but not in alveolar macrophages 16 . Cx3cr1, Lyve1 or MHCII-related genes such as Cd74 were also expressed according to the expected expression pattern (Fig. 2c) , which confirmed the accuracy of our gating and sorting strategy. Correlation analysis of all macrophage populations revealed that, in line with published data 14, 22 , classical TRM populations such as microglia, splenic macrophages, Langerhans cells and alveolar macrophages each had a very distinct expression profile, indicating the robust tissue imprinting of these cells (Fig. 2d) . Tongue macrophages on the other hand were more related to heart and intestinal macrophages, regardless of their tissue of residence. Morover, even within this group of TRM, MHCII expressing cells like heart MHCII + macrophages, tCX3CR1-MF and colon macrophages showed a higher correlation to each other and were distinct from MHCIIcell populations, including tFOLR2-MF and heart MHCII + macrophages (Fig. 2d) . Thus, tCX3CR1-MF and tFOLR2-MF share similarities with the two main interstitial macrophage populations that have been previously identified across various tissues 19, 22 We focused our analysis on up-regulated genes (FC > 2; adjusted p-value < 0,001) to identify a tissue-specific signature of each macrophage subset and were able to annotate previously described marker genes to macrophage populations isolated from the lung, heart, skin, brain and spleen 16, 23 . We next performed a GO enrichment analysis on these DEGs and found in agreement with previously published work 24 an enrichment of GO terms that facilitate the tissue-specific function of each TRM subset (Fig. 2f) . In comparison, tCX3CR1-MF showed a strong enrichment for genes involved in inflammatory pathways, such 'positive regulation of TNF production' or 'cytokine-mediated signaling pathway', which is in line with our scRNAseq data presented in Fig. 1 . tFOLR2-MF were characterized by weak but significant enrichment for 'cellular hyperosmotic salinity response' and 'response to toxic substances'. Taken together, these data demonstrate that tCX3CR1-MF and tFOLR2-MF tongue macrophages belong to the family of interstitial macrophages. They fall into the two broad categories of Cx3cr1-and Lyve1/Folr2/Timd4-expressing cells 19, 22 , but also show unique transcriptomic signatures that probably reflect the requirements of their local tissue niche. The unique transcriptomic signatures of tCX3CR1-MF and tFOLR2-MF could indicate that these populations inhabit different microanatomical niches within the tongue. To test this hypothesis, we performed immunohistochemistry on adult mouse tongues. Cx3cr1 Gfp/+ animals were perfused and fixed tongue sections were stained with antibodies against GFP and LYVE1. Of note, we have used LYVE1 and FOLR2 markers interchangeably for the identification of tFOLR2-MF. Indeed, GFP + cells were concentrated in the lamina propria, while LYVE1 staining was evident on cells throughout the tongue tissue with the exception of the epidermis (Fig. 3a+b) . Since lymphatic vessels also stain positive for LYVE1 25 , the tissue was counterstained with antibody against CD68, a common pan-macrophage marker. Thus, we could identify tCX3CR1-MF as double-positive CD68 + Cx3cr1-GFP + cells in the lamina propria, but not in the tongue epidermis or muscle (Fig. 3c ) and tFOLR2-MF as double-positive CD68 + LYVE1 + cells in the lamina propria and the underlying muscle ( Fig. 3d ). Of note, tFOLR2-MF were morphologically distinct from LYVE1 + Podoplanin + lymphatics ( Fig. 3e) . EPCAM + Langerhans cells with a ramified morphology localized exclusively in the epidermis of the tongue (Fig. 3f) . To better characterize the tissue localisation of tCX3CR1-MF and tFOLR2-MF, we quantified their distribution in different layers of the tongue. CD68 + LYVE1 + double-positive tFOLR2-MF localized in the muscular layer as well as in the lamina propria ( Fig. 3g+h and Suppl. Fig. 4) , while CD68 + Cx3cr1-GFP + double-positive tCX3CR1-MF were only detected in the lamina propria and were virtually absent in muscular tissue (Fig. 3g) . However, clusters of Cx3cr1-GFP + cells could be detected in the posterior part of the tongue, along Tuj1 + nerves that possibly cater to circumvallate and foliate papillae (Fig. 3i) and in innervated areas of the deep muscle, along the chorda tympani branch of the facial nerve. tCX3CR1-MF were present at the base of both filiform and fungiform papillae (which harbour taste buds) and within the lamina propria, which is densely innervated by sensory fibres (Fig. 3j+k ). Thus, we show here that tongue tCX3CR1-MF and tFOLR2-MF inhabit distinct anatomical regions of the tongue. tCX3CR1-MF localized in the highly innervated lamina propria at the base of filiform papillae and within fungiform papillae, while tFOLR2-MF can additionally be found in deeper layers, often in proximity to blood vessels. Regardless of their tissue-specific roles in homeostasis, macrophages usually also function as first-line responders to pathogens. We therefore tested the inflammatory response of tongue immune cells to the bacteria endotoxin lipopolysaccharide (LPS) . Mice were challenged intraperitoneally (i.p.) with LPS and transcriptomic changes of the tongue hematopoietic system (CD45 + cells) were determined by scRNA-seq 6h later. In total, we sequenced 8165 cells. Integration of the LPS data with the data from steady state mice indicated that all cell populations were present 6 hours after LPS injection (Fig. 4a ). LPS injection led to a general increase of inflammatory gene expression such as Oasl1 and Ifi204 across all cells of the tongue immune system (Fig. 4b) . We focused on tongue-resident macrophages of cluster 0 (tFOLR2-MF) and cluster 5 (tCX3CR1-MF) and examined DEGs between the steady state and LPS conditions in these two populations. 211 DEGs were identified in tFOLR2-MF (the full list of DEGs can be found in Suppl. Data 7), of which 125 genes were upregulated (e.g. Il1b, Relb and Slfn4) and 86 genes downregulated after LPS injection (e.g. Folr2, Lyve1 and Klf4; Fig. 4c+d ). Interestingly, many of the tFOLR2-MF signature genes (i.e. Folr2 and Lyve1) were downregulated after LPS exposure (Fig. 4c) . In tCX3CR1-MF we detected 127 DEGs between the physiological and the pathological state of which 84 were upregulated (e.g. Ifit2, Lgals3 and Usp18) and 43 were downregulated (e.g. Lyz2, Ccr2 and Cd9). LPS induced upregulation of 45 common genes (e.g. Cxcl10, Ccl5, Il1rn and Gbp2) and downregulation of 17 common genes in both tongue macrophage subsets (e.g. Fcrls, S100a10, Lyz1 and Retnla; Fig. 4c+d ). GO enrichment analysis was used to explore potential signaling differences of tFOLR2-MF and tCX3CR1-MF in response to systemic LPS. Shared upregulated genes in the two subsets were involved in 'defense response', 'response to cytokine' and 'response to bacterium'. Genes that were only upregulated in tCX3CR1-MF macrophages were particularly enriched for GO terms associated with type I interferon signaling (Fig. 4e) . On the other hand, tCX3CR1-MF showed downregulation of genes involved in 'mononuclear cell migration' and 'cell population proliferation' after LPS exposure (Fig. 4e) . tFOLR2-MF showed the specific upregulation of genes involved in ‚immune system process' and ‚Nfkb signaling', while they downregulated in response to LPS the 'response to stress', ‚localisation of cell' apoptotic processes. Thus, both tFOLR2-MF and tCX3CR1-MF were activated by systemic administration of bacterial components (LPS) . However, they responded differently, with strong type I interferon signaling response characterizing tCX 3 CR1-MF and a more 'classical' Nfkbmediated macrophage response seen in tFOLR2-MF. We next investigated the spatiotemporal distribution of tongue macrophages over development in an effort to shed light on the origin and timeline of establishment of tFOLR2-MF and tCX3CR1-MF populations. First, we isolated leukocytes from Cx3cr1 Gfp/+ reporter mice at different ages and performed flow cytometry to detect CD11b + CD64 + macrophages that we could further separate according to Cx3cr1-GFP expression and FOLR2 immunoreactivity. We used FOLR2 as a marker since macrophages at this developmental stage do not show LYVE1 surface expression. At embryonic day 17.5 (E17.5), all tongue macrophages were characterized by high Cx3cr1-GFP expression (Fig. 5a) . Of these Cx3cr1-GFP + cells, two-thirds additionally expressed FOLR2 (G2 in Fig. 5a ). Similar proportions of macrophage subsets were observed in the mouse tongue at postnatal day 3 (p3). At this time-point, an intermediate Cx3cr1-GFP int FOLR2 + subset was also present (G3 in Fig. 5a+b) . In subsequent developmental stages (p11 and p28), the proportion of doublepositive Cx3cr1-GFP + FOLR2 + cells progressively decreased and two main Cx3cr1-GFP + FOLR2and Cx3cr1-GFP int FOLR2 + macrophage populations (corresponding to tCX3CR1-MF and tFOLR2-MF respectively) were established by 8 weeks of age (Fig. 5b) . To correlate the flow cytometry data to spatial localization, we performed immunohistochemical analysis of tongue sections from Cx3cr1 Gfp/+ mice during early developmental stages. As mentioned above, tongue macrophages did not express Lyve1 during development and the FOLR2 antibody used for FACS did not work for immunohistochemical staining of the tissue. We thus relied on Cx3cr1-GFP as a marker of macrophages. Tissue sections were counterstained with antibodies against CD31 to identify blood vessels, and, since various nerves (e.g. VII th , IX th and X th cranial nerve ganglia) innervate the tongue tissue during embryogenesis, we also stained tissue sections for antibeta Tubulin III (Tuj1) to visualize neurons. At E14.5 Cx3cr1-GFP + macrophages could be detected throughout the whole tongue tissue with no specific localization pattern (Fig. 5c) . This continued through postnatal day p0 and until p10, whereby Cx3cr1-GFP + macrophages were still dispersed throughout the tongue but started to align along the lamina propria. At these stages, taste bud maturation is observed 26, 27 and we detected the first Cx3cr1-GFP + macrophages in proximity to fungiform papillae (Fig. 5c) . These data demonstrate the progression of a dynamic tongue macrophage compartment during mouse tongue development and that the specific localization of tongue macrophages in distinct anatomical niches is only established during postnatal stages. The histological analysis reveals the unordered distribution of Cx3cr1-GFP + macrophages during early developmental stages towards distinct localization at adulthood. However, it is not clear from our histological data whether this reflects a sequential developmental maturation of macrophages or rather a progressive exchange of embryonic populations with adult, bone marrow-derived monocytes. To gain further insight into this question, we investigated the tongue immune compartment from mice at post-natal day 3 (p3) with scRNA-seq. We chose p3 since it is the stage at which we could first identify Cx3cr1-GFP int FOLR2 + macrophages by flow cytometry (Fig. 5b) . We sequenced 13898 cells (from a pool of n=7 mice) and overlaid the data on the scRNA-seq data derived from adult mice (Fig. 6a) . All adult tongue immune cell populations were also present at p3, however, we noticed major differences in the frequency of several populations. Notably, p3 tongues harbored more mast cells (clusters 11 & 13) and fewer Fn1+ myeloid cells (cluster 6) than adult tongues (Fig. 6b) . We also detected a newly appearing cluster of cells in p3 tongues (cluster 2), which was almost absent in adult mice (Fig. 6b) . These cluster 2 cells expressed high levels of proliferation-related genes, including Top2a, Ccnb2, Mki67, and further showed expression of typical macrophage lineage genes such as Cx3cr1, Tlr4, Pf4 and Lyve1, which suggests they represent proliferating macrophage precursor cells (Fig. 6c) . Indeed, when the UMAP is projected in 3D, cluster 2 cells seemed to incorporate into tFOLR2-MF ( Fig. 6d; Suppl. Data 8) . To investigate the connection of cluster 2 precursors with tongue macrophages we used Slingshot, which models developmental trajectories in scRNA-seq data 28 . Slingshot analysis revealed a possible bifurcation of the precursor cells at the cluster 1 level at which trajectories either split into cluster 0 or cluster 5 cells (Fig. 6e) , which might indicate that both tFOLR2-MF and tCX3CR1-MF populations derive from a common precursor. We confirmed the proliferation activity of macrophage precursor cells at p3 by EdU in vivo labelling. One day after injection, EdU incorporation could be readily observed in about 30-40% of CD68 + FOLR2 + cells, while both adult tongue macrophage subsets showed no signs of homeostatic proliferation (Fig. 6f) . We The proximity of tCX3CR1-MF to nerves raises the question of whether they perform specific nerve-associated functions. In the skin, large, peripheral nerve-associated Cx3cr1 + macrophages were recently described 18 that were involved in the maintenance of myelin sheath integrity and axon sprouting after injury and had a transcriptional signature related to nervous system functions. This particular signature was absent from tCX3CR1-MF, although tCX3CR1-MF did express some genes that are enriched in microglia (e.g. Hexb, Apoe) that also reside in close proximity to neurons. Various cranial nerve ganglia innervate the tongue tissue during embryogenesis and axons are guided by multiple chemoattractive factors to the tongue epithelium such as brain derived neurotrophic factor (BDNF) 35 . It is known that microglia-derived BDNF plays an important role in synapse formation and plasticity in the adult brain 36 . It is therefore possible that the subsequent postnatal maturation of taste receptor cells and the synaptic interconnectivity with neurons might be influenced by tCX3CR1-MF. The interesting topic of whether and how macrophages interact with nerves in the tongue needs further investigation. It is also possible that tongue macrophages contribute directly or indirectly to instances of taste dysfunction. Various infections including Covid-19 or middle ear infections [37] [38] [39] , different medical treatment regimens 40 and aging 41 have all been associated with taste disruption. Taste bud maintenance relies on a continuous renewal of differentiated taste receptor cells 42 and it has been shown that systemic inflammation like peripheral LPS injection increases TNFα and IL10 production by taste cells 43, 44 , inhibits taste progenitor cell proliferation and interferes with taste cell renewal 45 . Such a mechanism might explain taste disorders associated with infections in general. We showed that systemic inflammation also causes a direct response from tongue-resident macrophages. We did not investigate the precise contribution of inflammation to taste sensation, yet consider the possibility that tongue macrophages may release cytokines and potentially neurotoxic products that could cause nerve damage. The reverse has also been shown, that tissue macrophages can also mediate neuronal protection and therefore limit neuronal damage upon infection 46 . Whether and how the various tongue mononuclear phagocytes are involved in taste perception remains to be explored and could be of clinical relevance in conditions such as anorexia or cancer, where appetite-and weight-loss is often aggravated by taste-dysfunctions. The question remains of the role of tongue immunity in cases where tongue homeostasis is disrupted. It was recently shown that tongue CD163 + macrophages infiltrate tumor tissue in squamous cell carcinoma at a high frequency 5 . CD163 + macrophages were interpreted to be "M2 macrophages", and increased infiltration correlated with worse outcomes compared to patients with a high infiltration of CD11c + "M1" macrophages 5 . Regardless of whether the "M1/M2" classification really applies to in vivo situations 47 , our data corroborate the notion of the existence of distinct subsets of tongue macrophages, which might respond differently to tumor-specific environmental cues. We present a comprehensive catalog of immune cells in the murine tongue in physiological conditions and upon LPS-induced systemic inflammation. We identified two novel macrophage subsets in the tongue, namely tCX3CR1-MF and tFOLR2-MF, and place these findings in the context of mammalian macrophage biology. We hope that these data will encourage and support further investigations of the tongue as barrier and as an underrated immunological organ. This study was performed in strict accordance with national and international guidelines for Adult mice were anesthetized by intraperitoneal injection of 150mg/kg body weight pentobarbital sodium (WDT) and intracardially perfused with PBS. Brain: Brains were dissected, and the olfactory bulb and cerebellum were removed. The remaining brain was minced and filtered through a 70µm cell strainer in ice-cold high- Colon: Cells from Cx3cr1 Gfp/+ mice were isolated as previously described 48 . In brief, the intestine was removed and feces were flushed with cold PBS without calcium and magnesium. The intestine was longitudinally opened and cut into 0,5 cm pieces. Intestinal The MARS-seq protocol was used for bulk RNA sequencing 49 . After reverse transcription, samples were analyzed by qPCR and samples with similar Ct values were pooled. Samples were treated with Exonuclease I (New England BioLabs (NEB)) for 30 min at 37°C and for 10 min at 80°C followed by a 1.2X AMPure XP beads (Beckman Coulter) cleanup. The second strand synthesis kit (NEB) at 16°C for 2 hours was used for cDNA synthesis followed by a Mice were injected intraperitoneally with 1 mg/kg LPS (E. coli 0111:B4) in 200 µl PBS 6 hours prior to sacrifice. The EdU Click 488 Kit from BaseClick were used. In brief, 0.5 mg/g of EDU in PBS was injected intraperitoneally to adult mice or subcutaneously to pups 15 hours before to sacrifice. Single cell suspensions were obtained from tongues (see relevant section) and cells were processed according to the manufacturer's instructions. Bulk sequencing data analysis. Data that were generated within this study have been deposited in Gene Expression Omnibus (GEO) with the accession code GSEXXXX. We would like to thank Jermaine Voß for excellent technical support, as well as the MDC Tongue as a first-line immune organ Interepithelial cells of the oral mucosa in mice. An ultrastructural classification with reflections on the origin of the Langerhans cell Geographic tongue: assessment of peripheral nerve status, Langerhans cell, and HLA-DR expression. Oral Surg Oral Med Oral Pathol Oral Radiol 124 Regional variation in Langerhans cell distribution and density in normal human oral mucosa determined using monoclonal antibodies against CD1, HLADR, HLADQ and HLADP Prognostic Relevance of Macrophage Phenotypes in High-grade Oral Tongue Squamous Cell Carcinomas Alteration of Macrophage Infiltrating Compartment: A Novel View on Oral Carcinogenesis Immune cells of the human peripheral taste system: dominant dendritic cells and CD4 T cells The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting Langerin+ DCs regulate innate IL-17 production in the oral mucosa during Candida albicans-mediated infection Early antitumor activity of oral Langerhans cells is compromised by a carcinogen Depletion of Langerhans cells in the tongue from patients with advanced-stage acquired immune deficiency syndrome: relation to opportunistic infections Lipopolysaccharide-induced up-regulation of activated macrophages in the degenerating taste system Smell and Taste Dysfunction in Patients With COVID-19: A Systematic Review and Metaanalysis Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage GSVA: gene set variation analysis for microarray and RNA-seq data Gene-expression profiles and transcriptional regulatory pathways that underlie the identity and diversity of mouse tissue macrophages Identification of cDC1-and cDC2-committed DC progenitors reveals early lineage priming at the common DC progenitor stage in the bone marrow A Subset of Skin Macrophages Contributes to the Surveillance and Regeneration of Local Nerves Two distinct interstitial macrophage populations coexist across tissues in specific subtissular niches Brown-adipose-tissue macrophages control tissue innervation and homeostatic energy expenditure Analysis of fractalkine receptor CX(3)CR1 function by targeted deletion and green fluorescent protein reporter gene insertion Three tissue resident macrophage subsets coexist across organs with conserved origins and life cycles Network analysis of transcriptomic diversity amongst resident tissue macrophages and dendritic cells in the mouse mononuclear phagocyte system Tissue-resident macrophage enhancer landscapes are shaped by the local microenvironment Mouse LYVE-1 is an endocytic receptor for hyaluronan in lymphatic endothelium Development and maturation of taste buds of the palatal epithelium of the rat: histological and immunohistochemical study Maturation of taste buds on the soft palate of the postnatal rat Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics Essential role of the IRF8-KLF4 transcription factor cascade in murine monocyte differentiation Neuronal, stromal, and T-regulatory cell crosstalk in murine skeletal muscle Innate lymphoid cells promote lung-tissue homeostasis after infection with influenza virus Phenotype and Tissue Residency of Lymphocytes in the Murine Oral Mucosa Oral CD103-CD11b+ classical dendritic cells present sublingual antigen and induce Foxp3+ regulatory T cells in draining lymph nodes Mapping of the lingual immune system reveals the presence of both regulatory and effector CD4+ T cells Epithelial-derived brain-derived neurotrophic factor is required for gustatory neuron targeting during a critical developmental period Microglia Promote Learning-Dependent Synapse Formation through Brain-Derived Neurotrophic Factor Pathogenesis of dysgeusia in COVID-19 patients: a scoping review A New Symptom of COVID-19: Loss of Taste and Smell Chorda tympani injury: operative findings and postoperative symptoms Taste changes experienced by patients receiving chemotherapy Taste-related sensations in old age Taste buds: cells, signals and synapses Expression and secretion of TNF-α in mouse taste buds: a novel function of a specific subset of type II taste cells Interleukin-10 is produced by a specific subset of taste receptor cells and critical for maintaining structural integrity of mouse taste buds Lipopolysaccharide-induced inflammation attenuates taste progenitor cell proliferation and shortens the life span of taste bud cells Adrenergic Signaling in Muscularis Macrophages Limits Infection-Induced Neuronal Loss Abandoning M1/M2 for a Network Model of Macrophage Function IL-23-mediated mononuclear phagocyte crosstalk protects mice from Citrobacter rodentium-induced colon immunopathology Massively Parallel Single-Cell RNA-Seq for Marker-Free Decomposition of Tissues into Cell Types Fiji: an open-source platform for biological-image analysis Distinct localization of mouse tongue macrophages (a) Panoramic image of a tongue sagittal section from a perfused Cx3cr1 Gfp/+ mouse. The tongue was stained for anti-GFP (CX3CR1; green) and anti-LYVE1 (violet). (b) Magnification of the area depicted in (a) that includes a fungiform papilla with annotation of different tissue layers. EPI = epidermis, LP = lamina propria, MUS = muscle. See also Suppl. Fig. 4 for definition of layers. (c) Co-staining of anti-GFP (green) and anti-LYVE1 (violet) with anti-CD68 (white). The dashed white line indicates the border between the epidermis and the lamina propria + (violet) podoplanin -cells. (f) Anti-Epcam (violet) staining identifies Langerhans cells in the epidermis of the tongue. Sections were stained with anti-Lyve1 (white) and DAPI (blue) Quantification of tFOLR2-MF and tCX3CR1-MF in lamina propria (LP) and muscle layer (MUS) in adult female mouse tongues. Each dot represents one animal Rare Cx3cr1-GFP + (green) cell clusters could also be detected in innervated Tuj1 + (violet) areas in posterior regions of the tongue. (j) 3D reconstruction of Cx3cr1 Gfp/+ tongue tissue sections stained for anti-CD31 (red), anti-Tuj1 (white), anti-GFP (Cx3cr1; green) and DAPI (blue). (k) Localization of tCX3CR1-MF in fungiform (right) and at the base of filiform (left) papillae. Sections were stained for anti-CD31 (violet) Supplementary Figure 2: Different isolation methods for the identification of tongue leukocytes (a) Digestion of tongues from PBS perfused mice with collagenase IV, hyaluronidase and DNase. Shown are examples for Bl6 mice (left) and Cx3cr1 Gfp/+ mice (right) Supplementary Figure 3: Gating strategy for the isolation of tissue resident macrophages Shown are the gating strategies that allow the identification of (a) Langerhans cells in the skin, (b) splenic macrophages, (c) Cx3cr1 + colonic macrophages The insets represent different areas of the tongue that were quantified. (b) Imaris-based mask for the identification of epithelium, lamina propria and muscle. Analysis of cell density was performed within these areas. (c) Macrophage subset cell counts in the respective regions of the tongue as represented in (a) Each dot represents one animal. The experiment was repeated twice. (c) Histological analysis of an adult Cx3cr1-GFP Irf8 -/-tongue reveals the presence of GFP + tCX3CR1-MF in the lamina propria. Sections were stained with anti-GFP (green) and anti-LYVE1 (violet) antibodies. (d) 9047 CD45 + cells were profiled from adult Irf8-deficient mice (pool of n=5 mice) by scRNA-seq. Shown are UMAP dimension reductions of WT (left) and Irf8-deficient CD45 + tongue cells (right). (e) Gene expression examples in wt and Irf8-deficient cells Each cluster can be found in separated tabs. As test method we used MAST, the log2fc threshold was set to 0.25 and a Bonferroni mutiple testing correction was applied. We only considered genes, that were expressed in at least 20% of the cells in at least one of the groups Supplementary data 2: The file contains the average expression value for each cluster based on the normalized counts of the cells in the respective sample. Additionally, a column for each cluster indicates, whether a gene is a conserved marker for this cluster (see Suppl. 3: Core gene signatures of cDC (tab 1), macrophages (tab 2), cDC1 (tab 3) and cDC2 (tab 4). The gene lists were derived from 16 This table includes the full bulk RNA-seq read counts for all isolated tissue resident macrophage populations that are shown in Fig. 2. Each subset is represented in a separate tab Listed are the upregulated genes that could be extracted from the bulk RNAseq data. These genes were used for GO annotations represented in Fig Listed are all significant differential expressed genes between untreated and LPS-treated tFOLR2-MF (tab1) and tCX3CR1-MF (tab2) The authors declare no competing interests.