Single-cell RNA sequencing of OSCC primary tumors and lymph nodes reveals distinct origin and phenotype of fibroblasts

Single-cell RNA sequencing of OSCC primary tumors and lymph nodes reveals distinct origin and phenotype of fibroblasts | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Single-cell RNA sequencing of OSCC primary tumors and lymph nodes reveals distinct origin and phenotype of fibroblasts Yuxin Wang, Qian Zhang, Liang Ding, Jingyi Li, Kunyu Liu, Chengwan Xia, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3862426/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Desmoplasia in fibroblasts within metastatic lymph nodes (MLNs) serves as an indicator of extranodal extension (ENE), which is a critical determinant of distant metastasis and mortality in oral squamous cell carcinoma (OSCC). However, systematic studies on fibroblasts in MLNs are lacking. Therefore, this study investigated and characterized the differences in phenotype, function, and origin of fibroblasts between primary tumors (PTs) and lymph nodes (LNs) in OSCC. We generated single-cell maps of PTs and paired MLNs and draining LNs from three OSCC patients. Among 44,052 fibroblasts, we identified two distinct subpopulations of cancer-associated myofibroblastic cells (mCAFs): RGS4+ mCAF1 and COMP+ mCAF2. Notably, mCAF1 and mCAF2 exhibited distinct distributions, with mCAF1 predominantly localized in the PT and mCAF2 in the MLN. Moreover, pseudotime analysis revealed their distinct origins: mCAF1 originated from inherent normal myofibroblastic cells in the PT, whereas mCAF2 originated from fibroblastic reticular cells in the LNs. Further differential analysis and in-vitro functional experiments using primary fibroblasts revealed that, compared to mCAF1, mCAF2 in MLN exhibited weaker crosstalk with immune cells but displayed enhanced extracellular matrix activity, which is closely linked to ENE formation in OSCC. Additionally, we identified two fibroblast subgroups in a transforming state, characterized by relatively low expression of fibroblastic marker genes but expressed specific epithelial canonical markers, indicating a potential epithelial–mesenchymal transition. Our research offers profound insights into the heterogeneity of fibroblasts between the PT and MLN in OSCC, serving as an essential resource for future drug discovery endeavors. Biological sciences/Cancer/Head and neck cancer/Oral cancer Health sciences/Oncology/Cancer/Head and neck cancer Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Lymph node (LN) metastasis portends poor prognosis in oral squamous cell carcinoma (OSCC) 1, 2 . Extensive functional evidence supports the involvement of cancer-associated fibroblasts (CAFs) in various stages of metastasis, including cancer proliferation, local invasion, intravasation, metastatic niche priming, and colonization 3, 4 . Particularly, the infiltration of metastatic cells beyond the LN capsule, known as extranodal extension (ENE), significantly increases the risk of distant metastasis, rendering it the leading cause of mortality in OSCC 5, 6 . Consequently, targeting metastatic LNs (MLNs) exhibiting ENE rather than focusing solely on the primary tumor (PT) is crucial for preventing distant metastasis in patients with advanced OSCC to improve survival rates 5 . Previous studies showed that desmoplasia by CAFs in MLNs is intricately linked to the transformation of the tumor microenvironment (TME) 7, 8 . However, the mechanisms of CAFs in MLN formation and development remain unclear.  CAFs exhibit significant heterogeneity in their phenotypes, functions, and origins. Single-cell RNA sequencing (scRNA-seq) enables the characterization of gene profiles in individual cells and the tracing of cell lineages. Recent advancements in scRNA-seq have facilitated the identification of distinct phenotypic and functional subgroups of CAFs in PTs, including cancer-associated myofibroblasts (mCAFs), inflammatory CAFs (iCAFs), and antigen-presenting CAFs (apCAFs) 9 . Although previous studies have demonstrated a partial overlap among these CAF subgroups 9 , their origins and associations with functional characteristics remain poorly understood. Our research team has long been committed to the study of CAF heterogeneity in OSCC TME and has revealed their distinct characters and functions in tumor progression, invasion, and immune regulation 10-12 . However, systematic studies on CAFs in OSCC TME are lacking and comparison between MLN and PT is sparse 13, 14 . Understanding the mechanisms underlying the origins of key CAF subgroups and their contributions to tumor development within LNs is critical for formulating strategies to target LN metastasis. In this study, we conducted transcriptome analysis using 87,650 single cells from paired PTs, MLNs, and draining lymph nodes (DLNs) with no evidence of metastasis from OSCC patients. We comprehensively characterized the single-cell landscape of fibroblasts from PTs, MLNs, and DLNs, providing insights into the cellular dynamics and molecular features associated with metastasis. Materials And Methods 1.Patients and samples. All samples were obtained from the Nanjing Stomatological Hospital. All involved patients provided informed consent for this work and all procedures were approved by the Institutional Review Bord of Nanjing stomatological hospital, Medical school of Nanjing University (IRB: NJSH 2021NL011). (1) scRNA sequencing sample Three fresh primary oral squamous carcinoma tissues (1 in tongue, 1 in floor of mouth and 1 in buccal mucosa) along with three matched metastatic lymph nodes and draining lymph nodes were obtained for scRNA sequencing (Supplementary Table1, No 1-3#). Their pathological status was proved by intraoperative frozen diagnosis. (2) Tissue chip for immunochemistry 3 tissue chips consisted of OSCC tissue sample from Nanjing Stomatological Hospital were used in this study. 1 tissue chip contains 95 MLN samples of OSCC patients. The other 2 tissue chips contain a total of 124 OSCC tumor samples and paired 68 adjacent normal samples. Their pathological status was confirmed according to HE and CK5/6 immunohistochemistry test. (3) Primary fibroblast cells 2 pairs of fresh tissue sample from OSCC PT and matched MLN and DLN tissue were collected for primary fibroblast cells extraction (Supplementary Table1, No.4-5#). 2. Single cell RNA sequencing (1) Cell preparation Fresh primary tumor and lymph node tissue were processed immediately after obtaining from OSCC patients. Every sample was cut into small pieces(<1mm in diameter) and washed twice before processing. Sterile 1640 medium containing 0.04% BSA was used in washing and dispensing. Then they were incubated with dissociation reagent containing 5mg/ml collagenase I, 1 mg/ml DNases, 2.5mg/ml hyaluronidase, and 1mg/ml dispase in thermostatic shaker (37°C, 180 rpm). After digestion for 15 mins, supernatant was transferred into a new centrifuge tube and preserved on ice. 3ml dissociation reagent was added into residual tissue sediment and incubated in thermostatic shaker (37°C, 180 rpm) for 15 mins. Above procedures were repeated for three times and all obtained supernatant were centrifugated at 250g for 10 mins to achieve cell pellets. The cell pellets were resuspended and filtered by a 40um cell mesh which was repeated 3 times for filtering effect. Then the cell pellet was resuspended in 1ml ice cold red blood cell lysis buffer and incubated at 4°C for 10 mins. Next, 10ml ice-cold PBS was added and centrifuged at 250g for 10 mins to remove red blood cells and resuspended with 1640 medium with 0.04% BSA. Then, most hematopoietic cells were removed using CD45 microbeads according to manufacturer’s protocol (130-45-801, Miltenyi Biotec). Then, 10µl of suspension was counted under an inverted microscope with a hemocytometer. Trypan blue combined with AO/PI (Sorbio Biotec) fluorescent staining was used to assess cell viability. While cell viability was less than 70%, dead cell removal kit (130-090-101, Miltenyi Biotec) was used to improve it. Finally, eligible single cell suspension samples (cell viability more than 85%, total cell counts more than 1x10 5 , cell diameter less than 40um) were involved in following sequencing. (2) 10x scRNA sequencing scRNA sequencing was performed using 10×Genomics platform by OE Biotech Co., Lts(Shanghai, China). According to manufacture’s standard protocol, cell suspension was loaded to the Chromium Single-cell v3 3’ Chemistry Library. Instrument (10×Genomics) to generate single-cell gel beads in emulsions (GEMs). After generation of GEMs, reverse transcription (RT) reactions were engaged to generate barcoded full-length cDNA, which was followed by GEM-RT clean-up and cDNA amplification (DynaBeads). Subsequently, the amplified cDNA was fragmented, end-repaired, A-tailed, and ligated to an index adaptor, and then the library was amplified. Every library was sequenced on a HiSeq X Ten platform (Illumina), and 150bp paired-end reads were generated. (3) Raw data processing and quality control The library construction, sequencing and data analysis were carried out by OE Biotech Co., Lts(Shanghai, China). The size and purity of all libraries were checked by Agilent 2100 Bioanalyzer. The raw data was evaluated using FastQC software to ensure sequencing quality. Then, the Cell Ranger software pipeline(version 3.1.0) was used to demultiplex cellular barcodes, map reads to the genome and transcriptome using the STAR aligner, and down-sample reads as required to generate normalized aggregate data across samples, producing a matrix of gene counts versus cells. We processed the unique molecular identifier (UMI) count matrix using the R package Seurat(version 3.0) 15 . Cells with UMI numbers <1000 or with over 10% mitochondrial-derived UMI counts were considered low-quality cells and were removed. Library size normalization was then performed on the filtered matrix to obtain the normalized count. (4) Dimensionality reduction and clustering Principal component analysis (PCA) was performed to reduce the dimensionality on the log transformed gene-barcode matrices of top variable genes and top top 50 PCs were used to perform the downstream analysis. Cells were clustered based on a graph-based clustering approach, and were visualized in 2-dimension using t-distributed stochastic neighbor embedding (tSNE) or uniform manifold approximation and projection (UMAP). In this present study, the main 23 cell clusters were identified under 0.4 resolution using the FindClusters function in Seurat. (5) Marker gene and cell type identification The Seurat Findallmarker function was performed to identify preferentially expressed genes in each cluster as marker genes. Here, we use the R package SingleR 16 for unbiased cell type recognition of scRNA-seq and identify cell types reference to conventional markers as described in previous studies. (6) Differentially expressed genes(DEGs) and enrichment analysis DEGs were identified using the Seurat package. P value 1was set as the threshold for significantly differential expression. GO enrichment and KEGG pathway enrichment analysis of DEGs were respectively performed using R based on the hypergeometric distribution. GSVA analysis was performed to grade each pathway using GSVA package 17 . Then the differences between different cell groups were analyzed using limma package 18 . (7) Transcription factor inference To assess TF regulation strength, we applied the single-cell regulatory network inference and clustering (pySCENIC, v0.9.5) workflow, using the 20 thousand motifs database for RcisTarget and GRNboost. (8) Pseudotime analysis To map origin heterogeneity of fibroblast cells, we performed pseudotime analysis with Monocle2 19 . For Monocle analysis, positive marker genes for each cluster were used. Based on the expression profiles of these pseudotime-dependent genes, we perform spatial dimensionality reduction and construct a minium spanning tree (MST). Then, using this MST, we identify the longest path that represents the differentiation trajectory of cells with similar transcriptional features. Then the GO terms analysis was performed using clusterProfilter package. (9) CNVs estimation The InferCNV package was used to detect the copy number variations (CNVs) in EPCAM+ cells with default parameters. All epithelial cells were clustered into 4 clusters and CNVs levels were estimated with normal fibroblasts as control group. Then malignant cells were recognized with extra high CNVs score. (10) Cell-cycle discrimination analyses. We used cell cycle-related genes, including a previously defined core set of 43 G1/S and 54 G2/M genes29. For each cell, a cell cycle phase (G1, S, G2/M) was assigned based on its expression of G1/S or G2/M phase genes using the scoring strategy described in CellCycleScoring function in Seurat. Cells in each cell cycle state were also quantified using the which.cells function (11) Cell-cell communication analysis. To identify different potential interactions between different cells, we used CellPhoneDB 20 to analyze ligand-receptor relationships between different proteins on the cell membrane and in the cytosol. Significant mean and cell communication significance (P < 0.05) was calculated based on the interaction and the normalized cell matrix achieved by scran normalization. 3. Correlation to public datasets Transcriptome and clinical data from The Cancer Genome Atlas (TCGA) HNSC datasets were obtained from https://portal.gdc.cancer.gov. To compare gene expression of extracellular matrix organization and muscle contraction between HNSCC tumor and normal samples, 2 GSEA genesets were applied in this study (Supplementary table 2). We perform ssGSEA analysis on the TCGA-HNSC transcriptome data to calculate the enrichment scores (ES) for each sample in the genesets of extracellular matrix organization and muscle contraction. Then the ES scores between normal or tumor groups were compared and P<0.05 was set as significant difference. 4. Immunochemistry staining Tissues were fixed in 4% PFA, dehydrated, paraffin embedded, and sectioned at 7–12 µm thickness. Frozen tissues were sectioned at 12 µm thickness. Immunostaining was performed both on frozen and paraffin sections. Heat-based antigen retrieval was performed when necessary. Tissue sections were blocked in either 3% Goat serum. The primary antibodies used were rabbit anti-COMP (1:500, Abcam, ab300555#), rabbit anti-RGS4 (1:200, Abcam, ab97307#), rabbit anti-CK5/6 (MXB, MAB-0744#), rabbit anti-CD3 (ZSGB, ZA-0503), rabbit anti-CD4 (ZSGB, ZM-0418), rabbit anti-CD8 (ZSGB, ZA-0508), rabbit anti-CD68 (ZSGB, ZM-0060).Tissue chip sections were scanned by Pannoramic slide Scanner system (3DHISTECH) and visualized by CaseViewer2.4 (3DHISTECH). 5. Tumor associated collagen signatures (TACS) (1) TACS-1: No apparent or only a thin layer of collagen. (2) TACS-2: Elongated collagen fibers parallel to the tumor boundary. (3) TACS-3: Collagen fibers deposition with disrupted edges and presence of fiber bundles perpendicular to the tumor boundary. 6. Primary fibroblast cells extraction and culture 2 OSCC patients who underwent surgical treatment at the Department of Oral and Maxillofacial Surgery, Nanjing Stomatological Hospotal, Medical School of Nanjing University were included in this study. Fresh samples of OSCC PTs, paired MLNs and DLNs were diagnosed by frozen biopsy and collected for primary fibroblast cells extraction. All obtained samples were soaked in DMEM culture medium containing 20% penicillin-streptomycin and kept at 4°C for 2-4 hours. Then, they were moved into 2.5cm dish and trimmed into small pieces with 2mm diameter to remove exterior blood, fat and muscle tissues. After transferring them into 15ml centrifuge tube, centrifugation was then performed at 1500rpm at 4°C for 5 mins to discard the supernatant and repeated three times. Next, we added 5ml pre-prepared collagenase solution to tissue sediments and incubated them at 37°C and 200 rpm for 30-60 mins, until the edges of the tissue blocks appear ‘mist-like’. Then, we added 2ml of DMEM culture medium containing 10% FBS in tissue blocks and centrifuge them at 1500rpm for 5 mins, and discard the supernatant and repeated the process three times. Tissue blocks were then transferred to the bottom of a 25cm 2 culture flask and inverted it in a 37°C incubator for 2-4 hours. Next, we added 4ml of complete DMEM culture medium containing 10% FBS and 1% penicillin-streptomycin to the flask. After 7 days culture, spindle-shaped fibroblast cells could be observed to grow outward. Subsequently, we renewed the medium every 2-3 days until the cells covered the flask and trypsinized them for passaging. All primary fibroblast cells were cultured in a mixture of FM (Sciencell) and DMEM culture medium containing 10% FBS at a ratio of 1:1. They were placed in a water-jacketed CO 2 incubator with a stable temperature of 37°C and a CO 2 concentration of 5%. 7. Immunofluorescence staining. All cells were planked with appropriate cell density on cell slides and fixed by iced methanol for 30 mins at -20°C. After blocking with 1% BSA for 30 mins, cells were incubated with required primary antibody at 4°C overnight. After cleaning by PBS for three times, cells were incubated with corresponding second antibody (1:1000, Anti-mouse IgG Alexa FluorÒ 594 conjugate, Anti-rabbit IgG Alexa FluorÒ 488 conjugate, Cell Signaling) for 1 hour at room temperature. Finally, after cleaning by PBS for three times, the slides were sealed and visualized by fluorescent confocal microscopy (Nikon). All primary antibodies used were mouse anti-αSMA (1:500, Abcam, ab7817#), rabbit anti-COL11A1 (1:200, Abcam, ab166606#), rabbit anti-FN (1:150, Abcam, ab32419#), rabbit anti-TNC (1:200, Abcam, ab108930#), rabbit anti-FAP (1:200, Abcam, ab207178) and rabbit anti-FSP1 (1:200, Abcam, ab197896), mouse anti-POSTN (1:4000, Proteintech, 66491-1). 8. Western Blot Total proteins were resolved and transferred to polymembranes (Genscript). After they had been blocked with 5% BSA, the membranes were incubated overnight at 4 °C with anti-COMP (1:1000, Affinity, DF13438#), anti-RGS4 (1:1000, Abcam, ab97307#), anti-aSMA (1:5000, Abcam, ab32575#), anti-MMP1 (1:1000, Abcam, ab134184), anti-MMP2 (1:1000, Cell signaling technology, D4M2N#) and anti-PDPN (1:1000, Abcam, ab236529) antibodies. Then, the membranes were incubated with horseradish peroxidase conjugated goat anti-mouse immunoglobulin G or goat anti-rabbit immunoglobulin G (1:5000, Thermo) and developed with an enhanced chemiluminescence kit (E411-04#, Vazyme). 9. CCK-8 cell viability experiment Fibroblast cells (1x10 3 per well) were pre-cultured in 96-well plates in a 37°C incubator at a CO 2 concentration of 5% for 24 hours. Then, 10ul CCK-8 solution (KeyGEN) was added in each well, and the absorbance was measured at 450 nm with a microplate reader to measure the cell viability. 10. Wound healing assays The wound healing assay was done using standard wound healing inserts (Ibidi). After fibroblast cells were seeded and attached, the insert was removed and a standard 500um width scratch between fibroblast cells in both sides was formed. Then DMEM medium containing 5% FBS was added to overlay cell patches. Then wound healing was evaluated after 24h with a light microscope. 11. Transwell migration experiment The tested fibroblast cells were cultured to the logarithmic growth phase, digested and suspended in serum-free medium, counted, and adjusted to a concentration of 2×10 5 cells/ml. 100ul tested cells were added to the upper chamber of 24 well Transwell chambers and 600ul complete medium containing 10%FBS were added in the lower chambers. The Transwell chambers were incubated in a 37°C incubator at a CO 2 concentration of 5% for 24 hours and upper chambers were transferred into culture wells containing approximately 800μl methanol. After fixation at room temperature for 30 mins, chambers were then transferred to wells containing 800ul crystal violet staining solution and stained for 30 mins at room temperature. After rinsing in clean water multiple times, chambers were removed gently and cells at the bottom side were wiped off using moist cotton swab. Finally, the membranes were peeled off with small forceps and visualized under a microscope to count migrated cell numbers at 9 random fields. 12. Statistical analysis. All statistical analyses and graph generation were performed in R (version 3.6.0) and GraphPad Prism (version 7.0). 13. Data availability statement The data reported in this study are available upon NCBI (PRJNA950395, TaxID:9606). Specific code will be available upon reasonable request. Results 1. Cell type identification from primary OSCC tissue and paired lymph nodes To assess the role of fibroblasts in LN status and prognosis, we enrolled 108 OSCC MLN samples obtained between January 2015 and December 2017 at Nanjing Stomatological Hospital, Medical School of Nanjing University. The degree of desmoplasia and TACS was scored to evaluate the stroma ratio and extracellular matrix (ECM) remodeling in MLN tissue. We also assessed the relationship between these factors and prognosis ( Fig. 1A ). Our findings indicated that higher desmoplasia degree in MLNs correlated with poorer prognosis, including overall survival (OS) and disease-free survival (DFS). In addition, MLN samples with higher TACS scores were more likely to exhibit ENE. To elucidate the OSCC TME and compare the heterogeneity of stromal cells between MLNs and PTs at single-cell resolution, we collected three pairs of fresh OSCC PT, MLN, and DLN tissue samples. These tissues were dissociated into single cells, and magnetic cell sorting (MACS) was performed to remove CD45+ cells and reduce the proportion of hematopoietic cells ( Fig. 1B ). Following quality control and batch removal, 87,650 single cells were obtained using the 10×Genomics platform. Using the Seurat package for unsupervised clustering, we identified 23 clusters, and their visualization was achieved through UMAP( Fig. 1C&D ). Based on their canonical markers, we categorized these cells into nine major cell types: B cells (MS4A1+), endothelial cells (CD31+), epithelial cells (EPCAM+), fibroblasts (DCN+), macrophages (LYZ+), mast cells (KIT+), NK cells (KLRD1+), neutrophils (FCGR3B+), and T cells (CD3D+) ( Fig. 1E&F, Supplementary Fig. 1 ). Subsequently, we reclustered 14,405 endothelial cells into seven clusters ( Fig. 1G&H ). Among these clusters, six were identified as vascular endothelial cells (VECs; CD34+ ICAM1+), whereas the remaining cluster was identified as lymphatic endothelial cells (LECs; PDPN+ POX1+) ( Fig. 1I ). Epithelial cells, totaling 2,525, were classified into normal and malignant types based on their copy number variation (CNV) level ( Fig. 1J&K ). Inter-tissue heterogeneity was observed in a proportion of stromal cells among OSCC PTs, MLNs, and DLNs. Fibroblasts were predominant in all three tissues but particularly abundant in OSCC PTs ( Fig. 1D ). LECs were more concentrated in LN tissues, whereas VECs were evenly distributed across all tissues ( Fig. 1H ). Despite the elimination of most CD45+ cells before scRNA-seq, we successfully captured 26,666 immune cells. Comparing their compositions between PTs, MLNs, and DLNs revealed distinct dominant immune cell populations ( Fig. 1D ). Specifically, B and T cells were more enriched in LNs, whereas macrophages aggregated in MLNs. Neutrophils were restricted to tumor tissues, including PTs and MLNs, playing important roles in shaping the TME. Notably, most captured epithelial cells originated from PT, with only a few were present in LNs ( Fig. 1K ). This observation may be attributed to the high differentiation level of OSCC cells, leading to reduced epithelial cell activity after cell suspension preparation. 2. Dynamics of fibroblasts in primary OSCC tissue and paired LNs In total, 44,052 cells expressing the canonical fibroblast markers DCN, COL1A1, and COL3A1 ( Fig. 1E ) were identified as fibroblasts. Subsequently, these fibroblasts were reclustered into 22 subclusters and categorized into three main types based on their dominant locations ( Fig. 2A&B ). Specifically, Clusters 2, 8, 10, 11, 13, 15, and 22 were predominantly found in tumor sites, including PTs and MLNs, and were designated as CAFs. Clusters 3, 4, 16, 17, and 18 were mostly present in LN tissues, including MLNs or DLNs, and were characterized as LN stromal cells (LN-SCs). The remaining subclusters (1, 5, 6, 7, 9, 12, 14, 19, 20, and 21) were present in all three tissue types and were defined as tissue-inherent fibroblasts (iFCs). Distinct high-variable genes were further explored in CAFs, LN-SCs, and iFCs ( Fig. 2C ) and grouped according to their typical gene expression ( Fig. 2D–F ) and enrichment analysis ( Fig. 2G&H ). CAFs were first categorized based on marker gene expression and enrichment analysis. Clusters 2 and 11 were identified as mCAFs, characterized by high levels of α- smooth muscle actin (ACTA2) ( Fig. 2E ) and related genesets associated with actin cytoskeleton regulation, focal adhesion, and ECM score ( Fig. 2I ). iCAFs (Clusters 8, 10, and 22) exhibited high expression of PDGFRA but lacked Acta2. Furthermore, iCAFs exhibited increased scores for genesets involved in cytokine–cytokine receptor interaction ( Fig. 2I ), indicating their ability to secrete cytokines, consistent with a previous report 21 . CAF-A (Cluster 15) exhibited increased expression of APOE ( Fig. 2E ) and was exclusively limited to PT tissue, suggesting its potential roles in lipoprotein regulation ( Fig. 2H ). Cluster 13 fibroblasts expressed relatively low fibroblastic genes (DCN, PDGFRA, PDPN) and ECM scores ( Fig. 2E&I ), indicating their limited capacity for fibroblastic formation. However, they expressed relatively higher levels of epithelial cell genes (KRT19), indicating a potential developmental relationship. Pseudotime analysis further revealed their position at the terminal route of all fibroblasts, leading to their annotation as epithelial–mesenchymal transition (EMT)-like CAFs (eCAFs) ( Fig. 2J ), highlighting their potential involvement in EMT 22 . A comparison of mCAF distribution between PTs and MLNs revealed inter-tissue heterogeneity. Notably, Cluster 2 primarily occurred in PT, whereas Cluster 11 was predominant in MLN ( Fig. 2K ). Consequently, they were designated as mCAF1 and mCAF2, respectively, based on their distinct tissue preferences. Among the iFCs, myofibroblasts (mFCs; Clusters 5, 6, 7, and 12) ( Fig. 2E ) were identified based on their high ACTA2 expression. KRT19+ EMT-like fibroblasts (eFCs; Clusters 9 and 14) ( Fig. 2E ) exhibited a gene signature similar to that of eCAFs, indicating a potential evolutionary relationship in EMT-transformation. Pseudotime analysis revealed that eFCs and eCAFs were located at the end of the evolutionary route of all fibroblast subtypes ( Fig. 2J ). Cells in Clusters 1, 20, and 21 expressed typical fibroblastic genes DCN but lacked activation-associated genes (ACTA2 and PDGFRA), suggesting their resting status and delineating them as resting fibroblasts (rFCs) (Fig. 2E ) 3 . Furthermore, genes related to endothelial cells, including SELE, VWF, PECAM1, and ACKR1, were expressed in rFCs ( Fig. 2D&E ), suggesting their association with blood vessels and a potential role in angiogenesis 9, 23, 24 . LN-SC subsets were distinguished based on the expression levels of reported stromal cell markers and chemokines ( Supplementary Fig. 2 ) 25 . PDPN+ CD31– cells were identified as fibroblastic reticular cells (FRCs) ( Supplementary Fig. 2A ). FRCs were further grouped into CCL19 hi (Cluster 3) and CCL19 lo CD34 hi (Clusters 4 and 17) FRCs based on differences in immune cell recruitment. Among PDPN- CD31- LN-SCs, Cluster 18 was designated as perivascular cells (PvCs) owing to their expression of ITGA7 and ACTA2, whereas the remaining cells were classified as double-negative cells (DNCs) 26 ( Supplementary Fig. 2A ). Both CCL19 hi FRCs and DNCs exhibited high secretion levels of chemokines, including CCL19, CCL21, CXCL12, and CXCL13, indicating their ability to recruit lymphocytes 27 ( Supplementary Fig. 2B ). Comparisons of biological abilities among CAFs, iFCs, and LN-SCs were made ( Fig. 2G ). Gene set enrichment analysis revealed enrichment of epithelial cell proliferation and metabolism reprogramming in iFCs. Additionally, immune-related processes, including adaptive immune responses, immune effector responses, and chemokine signaling, were enriched in LN-SCs. Ossification and cell adhesion were enriched in CAFs. Furthermore, CAFs expressed functions related to antigen presentation, similar to LN-SCs. 3. Epithelial–fibroblast transformation in OSCC TME Two fibroblastic subtypes positioned lower in the fibroblast trajectory were identified and indicated as eCAFs and eFCs (Fig. 2J) . The expression level of the fibroblastic marker gene DCN varied across all fibroblastic clusters, exhibiting relatively low levels in eCAFs (Cluster 13) and eFCs (Clusters 9 and 14) ( Fig. 2E ). Additionally, eCAFs and eFCs expressed the epithelial marker gene KRT19, suggesting a potential translational status to epithelial cells. Upon comparing fibroblastic (DCN, COL1A1, COL1A3, and COL3A1) and epithelial (EPCAM, KRT6, KRT16, and CDH1) gene expression along the trajectory, we observed that downregulation of the fibroblastic marker genes expression was accompanied by the upregulation of the epithelial marker genes expression, coinciding with the presence of eCAFs and eFCs at the end of trajectory ( Fig. 3A ). These findings suggest a potential progression from eFCs and eCAFs towards epithelial cells. The CellCycleScoring function in Seurat was used to assess the proportion of cell cycle phases (G1, S, and G2/M) among all fibroblastic clusters. Notably, eCAFs exhibited a significant enrichment for cells in G2/M ( Fig. 3B ). Furthermore, the marker genes of eCAFs included several cell cycle-related genes, such as MKI67, CENPF, and TOP2A, indicating the active proliferative ability of eCAFs ( Fig. 3C ). We further investigated potential developmental relationships and identified differentially expressed genes among eCAFs, eFCs, and epithelial cells using Monocle 28 . The eCAFs and eFCs clusters were positioned in the lower region of the major trajectory ( Fig. 3D ). Although the definitive polarity of EMT remained undefined, the inferred pseudotime path revealed that eCAFs were in a transitioning state between eFCs and epithelial cells. Additionally, some eCAFs coexisted with epithelial cells in States 1, 4, and 10, suggesting a potential close developmental relationship ( Fig. 3E ). The pseudotime-dependent genes were hierarchically clustered into four modules ( Fig. 3F ). Enrichment analysis ( Fig. 3G ) demonstrated that the expression of genes related to angiogenesis and vascular development increased along the axis from epithelial cells to eFCs, whereas that of genes associated with the cell cycle were reduced. Notably, the expression of genes implicated in collagen degradation, ECM receptor interaction, collagen formation, and ECM organization were particularly elevated during the transition phase, suggesting the acquisition of ECM remodeling abilities during transformation to eCAFs. In addition, transcription factor analysis using SCENIC 29 revealed a highly activated TP63 pathway 30 in both eCAFs and eFCs, indicating a potential role in epithelial–fibroblast transformation ( Supplementary Fig. 3A ). 4. OSCC stroma comprised two distinct mCAF populations from PT and MLN with distinct origins With the high expression of myofibroblast markers (ACTA2, TAGLN) (Fig. 4A ) and enrichment in vasculature development and actin cytoskeleton organization ( Fig. 2I ), two fibroblast clusters (Clusters 2 and 11) were identified as mCAFs. A comparison of their locations revealed that Cluster 2 was concentrated in the PT, whereas Cluster 11 was mostly located in the MLN ( Fig. 2K ). mCAFs were classified as mCAF1 (Cluster 2, RGS4+) and mCAF2 (Cluster 11, COMP+) based on their positional preference and marker gene expression ( Fig. 4A ). To explore the origin of mCAF2 in MLN, we examined whether they developed from existing fibroblasts in MLN or metastasized from PT with tumor cells. We performed pseudotime trajectory analysis of mCAF1, mCAF2, mFCs, and FRCs, which revealed two distinct cell trajectories. In one trajectory, mCAF1 were proximal to mFCs, whereas in the other trajectory, mCAF2 exhibited proximity to FRCs ( Fig. 4B ). This observation suggests their distinct origins from inherent normal fibroblasts in their tissue sources. Further analysis of the differentiation pseudotime genes of FRCs and mCAF2 ( Fig. 4C&D ) revealed that they originated from Cluster 17 in State 1 and diverged into two main paths, with one ending in State 3 and the other ending in States 5 and 6. Although Clusters 3, 4, and 11 were all in a transitioning state, they exhibited diverse priorities, with Cluster 11 of mCAF2 predominantly present in State 5, whereas the other two clusters were mostly located in State 3. To investigate the gene dynamics along the FRCs–mCAF2 trajectory, we identified the pseudotime-dependent genes and categorized them into four modules based on their diverse expression patterns ( Fig. 4E ). Modules 1, 2, and 3 genes exhibited an increasing expression trend along the trajectory towards mCAF2, whereas Module 4 genes exhibited a decreasing trend. Enrichment analysis ( Fig. 4F ) revealed that Module 1 was enriched in ECM organization, vasculature development, and positive regulation of cell adhesion. Module 3 genes were associated with cytoskeletal organization and actin filament-based processes. These results suggest enhanced ECM remodeling and the presence of contractile features when FRCs transitioned to mCAF2. Module 2 genes were involved in ossification, cartilage development, skeletal system development, and proteolysis regulation. In contrast, Module 4 genes were associated with the positive regulation of cell migration, hematopoietic or lymphoid organ development, immune system development, and cytokine signaling in the immune system, which were downregulated along the trajectory. Overall, these findings suggest that FRCs acquired more myofibroblastic features but lost their immune regulatory function during the transition to mCAF2. The dynamic expression trends of marker genes in the four modules (COMP, SFRP2, ACTA2, and CCN3 for Modules 1, 2, 3, and 4, respectively) were aligned using a gene heatmap ( Fig. 4G ). Notably, the expression of COMP, a marker gene for mCAF2, gradually increased towards State 5 but remained steady towards State 3, mirroring the expression pattern of the contractile marker gene Acta2. This suggests that the expression of the mCAF2 marker gene increased following transition and was associated with ECM remodeling. In addition to mCAF2, pseudotime analysis of mFCs and mCAF1 revealed two diverging cell fates ( Fig. 4H ). Starting at Cluster 7, one fate progressed towards Clusters 5 and 12, whereas the other fate moved towards Cluster 2. Cluster 6 acted as a transitional state along the axis ( Fig. 4I ). Among these five clusters, Cluster 2 was identified as mCAF1, and the remaining clusters were identified as mFC. These findings suggest that the activation of mCAFs or normal myofibroblasts in the OSCC microenvironment could be influenced by diverse mechanisms. We further explored pseudotime-dependent genes along the furcate trajectory, categorizing them into four modules based on their expression patterns ( Fig. 4J ). The expression of Module 4 genes was particularly elevated in States 2–6, which were primarily occupied by mCAF1. Conversely, Module 3 genes were prominently expressed in States 7–9, which were occupied by mFCs ( Fig. 4I) . This indicates that the diverse gene expression signatures of Modules 3 and 4 may play a role in determining the fate of fibroblast activation in OSCC microenvironment. Furthermore, Module 2 genes were highly expressed in the pre-branch state, whereas Module 1 genes showed a relatively stable expression in the transitioning state ( Fig. 4J ). Subsequently, we assessed the biological involvements of these pseudotime-dependent genes ( Fig. 4K ). The gene set from Module 3 was involved in the regulation of wound healing, heart contraction, muscle contraction, and vascular processes in the circulatory system, whereas the gene set from Module 4 was enriched in collagen degradation, collagen formation, and ECM organization, consistent with the expression signatures of the marker genes ( Fig. 4J ). We observed distinct expression patterns where ECM remodeling-related genes (POSTN, COL10A1, COL11A1, MMP1, MMP3, MMP7, MMP9, and MMP13) were predominantly expressed in mCAF1 states; however, the expression of contractile feature-associated genes (DCN, ACTA2, and AVPR1A) and lipid metabolism-related genes were elevated in mFC states ( Fig. 4L ). To validate these findings, we further investigated The Cancer Genome Atlas database of head and neck squamous carcinoma. The analysis revealed an upregulation in the expression of ECM remodeling genes in tumors, whereas the levels of muscle contraction-related genes were downregulated ( Fig. 4M ). These results provide compelling evidence that within the primary microenvironment of OSCC, the acquisition of ECM remodeling ability more accurately represents mCAF formation in the TME than contractile features. 5. Strong crosstalk occurred between mCAF1 and immune cells in OSCC PT Based on expression levels and discrimination of differential gene analysis between mCAF1 and mCAF2 (Fig. 5A ), we selected RGS4 as a marker gene for mCAF1 and COMP for mCAF2. Employing immunohistochemical (IHC) chip analysis ( Supplementary Fig.4 ), we compared the distribution of mCAF1 and mCAF2 between paired PT and MLN samples. Our results revealed that RGS4+ mCAF1 were more abundant in OSCC PT tissues, whereas COMP+ mCAF2 exhibited greater prevalence in OSCC MLNs ( Fig. 5B ). Subsequently, we conducted a biological analysis of the differentially expressed genes between mCAF1 and mCAF2 ( Fig. 5C) . Genes related to muscle contraction and cell–cell adhesion were more enriched in mCAF2, whereas pathways involved in the regulation of T cell differentiation and macrophage activation were more enriched in mCAF1. To further compare the immunomodulatory abilities of mCAF1 and mCAF2, we performed receptor–ligand analysis to explore the interaction between immune cells and mCAF1 or mCAF2. Despite the removal of most CD45+ cells through MACS prior to scRNA-seq, we obtained a considerable number of immune cells, up to 26,666, including B cells, macrophages, mast cells, NK cells, neutrophils, and T cells ( Fig. 1D ). Owing to their well-documented heterogeneity, T cells were further classified into naïve T, Treg, CD8+ T, and NK-T cells based on the expression levels of typical marker genes ( Fig. 5D & Supplementary Fig. 5 ). The receptor–ligand relationship between immune cells and mCAF1 or mCAF2 was explored using CellPhoneDB analysis 20 . The histogram ( Fig. 5E ) shows that more receptor–ligand pairs were identified in PT than in MLN, particularly between T cells and macrophages, indicating stronger crosstalk between mCAF1 and immune cells. The CXCL12/CXCR4 axis has been widely reported to be involved in the interactions between hematological, solid tumor cells, and stromal cells in several TME 31 . In the present study, we identified high CXCL12 expression in mCAF1 and Cxcr4 expression, its corresponding receptor, in most immune cells ( Fig. 5F ), indicating the potential recruitment of immune cells by mCAF1. In addition, the expression of TGF -β superfamily gene INHBA by mCAF1 induced the expression of ACVRL1 in naïve T cells, indicating potential enhancement of naïve T cell activity 32 . Subsequently, we conducted an IHC analysis to compare the correlation between mCAF1 and mCAF2 and immune cell infiltration in PT or MLN. The results revealed that MLNs with high COMP+ mCAF2 exhibited fewer T cell and macrophage counts than MLNs with relatively low COMP+ mCAF2 ( Fig. 5G ). In contrast, the infiltration of macrophages and T cells in PT was positively correlated with the number of RGS4+ mCAF1 ( Fig. 5H ). Further comparison of T cell subtypes revealed that the presence of RGS4+ mCAF1 was specifically correlated with FOXP3+ cells in PT, suggesting that mCAF1 may be more likely to recruit and activate Treg cells to enhance immunosuppression. 6. mCAF2 from MLNs showed a stronger ECM remodeling ability To further demonstrate the high ECM remodeling ability of mCAF2 in OSCC MLN tissues, we conducted IHC analysis using OSCC tissue chips ( Supplementary Fig. 4 ). A relationship was observed between COMP expression and collagen deposition, where COMP expression corresponded with TACS 33 ( Fig. 6A&B ). To corroborate the findings of scRNA-seq and IHC analysis, we obtained two pairs of OSCC primary fibroblasts from the PTs (CAFs), MLNs fibroblasts (MAFs), and non-cancerous DLN fibroblasts (LFs). Immunofluorescence images revealed that CAFs and MAFs exhibited higher expression of typical fibroblast activation-associated genes, including ACTA2, FAP, and FSP1, than LFs ( Fig. 6C ). Furthermore, functional tests confirmed that CAFs and MAFs demonstrated enhanced proliferation and migration abilities compared to LFs. MAFs exhibited a higher ability for collagen contraction than CAFs and LFs, whereas CAFs exhibited no significant difference in terms of proliferation and migration ability ( Fig. 6D–G ). Immunofluorescence and western blotting experiments substantiated that the expression of ECM remodeling-associated proteins was upregulated in MAFs ( Fig. 6H&I ). This suggests that MAFs exhibited a stronger ECM remodeling ability, independent of activation level. Discussion CAFs are a central component of the TME in OSCC, and their presence is associated with poorer prognosis 9 . Through a retrospective analysis based on IHC, we confirmed the association between an elevated degree of desmoplasia in OSCC MLNs and an increased risk of ENE, as well as an unfavorable prognosis. This finding suggests that fibroblasts play significant roles in MLN formation, which might be potential therapeutic targets for patients with late-stage OSCC. Previous studies showed significant CAF heterogeneity with respect to their phenotypes, origins, and functions 12, 14 . However, the detailed nature of this heterogeneity between MLNs and PTs has remained unknown owing to insufficient fibroblast acquisition in previous sc-RNA seq studies 14, 34 . Therefore, we systematically analyzed the characteristics and differences of fibroblasts in the TME among MLNs, DLNs, and paired PTs in OSCC. To improve fibroblast acquisition, we optimized the single-cell suspension preparation and filtered out most CD45+ cells before sequencing. This approach yielded a total of 87,650 cells, allowing the identification of up to 44,052 fibroblasts. To compare the inter-tissue heterogeneity of the obtained fibroblasts, we initially reclassified them into three major subtypes according to their locations. CAFs were identified as fibroblasts present only in cancerous tissues (PTs and MLNs), LN-SCs were identified as fibroblasts located exclusively in LNs but absent in PTs, and the remaining fibroblasts located in all three tissues were identified as iFCs. Among these three major fibroblast subtypes, we further identified 12 subpopulations based on marker gene expression and functional enrichment. Consistent with previous studies 35, 36 , we identified mCAFs and iCAFs as major CAF subpopulations in the TME. The mCAFs subgroup was characterized by the expression of ACTA2 and enrichment in actin cytoskeleton and focal adhesion, whereas iCAFs subgroup expressed PDGFRA but lacked aSMA and were involved in cytokine–cytokine receptor interactions. In addition, we discovered an emerging CAF subgroup, which was classified as eCAFs. Previous studies have shown that epithelial cells in the heart can generate fibroblasts via EMT 37 . Overexpression of TGF-β has been demonstrated to cause EMT and CAF formation, contributing to the development of a fibrotic TME 38, 39 . In this study, eCAFs expressed relatively low levels of fibroblastic marker genes but expressed certain epithelial marker genes. It also demonstrated active proliferative ability and a potential transforming relationship with the eFC subgroup of iFCs. This finding supports the presence of an EMT state in the TME. Upon comparing the heterogeneity of CAFs between MLNs and PTs, we identified two mCAF subgroups: mCAF1 (RGS4+) and mCAF2 (COMP+). Through sc-RNA seq analysis, substantiated by IHC analysis, we observed a concentration of COMP+ mCAF2 in MLNs, whereas RGS4+ fibroblasts were more predominant in PTs. In PTs, the recruitment and activation of normal fibroblasts towards mCAFs occur through diverse mechanisms, including TGF-b, Wnt, and mechanotransduction signaling, which are well-documented 40 . However, in MLNs, mCAF origin (whether they originate from PTs mCAFs) and activation mechanism (whether they are activated from inherent stromal cells in the LNs) remain unclear. To delineate the evolutionary trajectories of mCAF1 and mCAF2, we performed pseudotime analysis. The results suggested that mCAF1 may originate from inherent mFCs in PTs, whereas mCAF2 may be derived from FRCs in LNs. Previous studies have demonstrated that proper maturation of FRCs determines LN development and immunity organization, whereas YAP/TAZ hyperactivation of FRCs can lead to enhanced myofibroblast characteristics and fibrotic LN structure 41 . However, the mechanism underlying the dysfunction of FRCs in MLN formation remains unclear. We conducted differential and enrichment analyses, which allowed for a comprehensive comparison of the heterogeneity of phenotypes and functions between mCAF1 and mCAF2. Notably, even with similar activation levels, mCAF1 from PTs exhibited stronger crosstalk with immune cells than mCAF2. IHC analysis further corroborated these findings, revealing a higher concentration of T cells and macrophages in PT areas rich in RGS4+ mCAF1; however, fewer T cells and macrophages were present in MLN areas abundant in COMP+ mCAF2. This finding indicates that the abundance of COMP+ mCAF2 in MLNs may restrict immune cell activity. In addition, our study demonstrated that mCAF2 in MLNs was related to cancer-associated collagen remodeling and the occurrence of ENE. A fine network of FRCs is essential for regulating immune cell entry and maintaining LN structure and function 42 . However, when FRCs undergo fibroblastic damage due to chronic inflammation and cancer, they compromise immune response in LNs 7 . A previous study suggested that hyperactivation of YAP/TAZ in FRCs severely impairs their maturation and leads to non-functional and fibrotic LNs 41 . Our study demonstrated that mCAFs in OSCC MLNs may originate from inherent FRCs. During this transformative phase, there was a diminished ability for immune cell recruitment; however, ECM remodeling function was enhanced. Previous studies have shown that CAFs play a central role in the deposition, modification, and degradation of ECM components and that dysregulated ECM remodeling by CAFs can lead to desmoplastic reactions associated with poor outcomes in breast, pancreatic, and lung cancers 43 . Our in vitro study and IHC analysis provided further evidence for the stronger ECM remodeling ability of primary CAFs from OSSC MLNs. Moreover, the presence of desmoplastic fibroblasts and ECM remodeling in MLN significantly influenced ENE formation, which may cause OSCC mortality. This finding suggests potential roles of improperly activated FRCs in MLN formation. In conclusion, our study demonstrated the presence of significant heterogeneity in OSCC stromal cells between the PTs and MLNs. Specifically, as a major subtype of fibroblasts, mCAFs in PT or MLN differed in terms of phenotype, origin, and function. Originating from inherent FRCs, COMP+ mCAFs in OSCC MLNs exhibited a more potent collagen remodeling ability but weakened immune cell recruitment activity, which may contribute to ENE formation and portend poor outcomes. This provides new insight into potential treatment strategy that targets specific CAF subgroups in MLNs of advanced OSCC. However, further research is required to elucidate the mechanisms underlying mCAF activation in OSCC MLNs and its association with ENE formation. Declarations Acknowledgement  This work was supported by the National Natural Science Foundation of China (No. 82103516) and “3456” Cultivation Program For Junior Talents of Nanjing Stomatological School, Medical School of Nanjing Univeristy(NO. 0222R207, 0222C108), the Jiangsu Province Science & Technology Department (BE2021609), Key Project supported by Medical Science and technology development Foundation，Nanjing Department of Health(ZKX23056). We also thank Xuan Zhou from OE Biotech Co., Ltd (Shanghai, China) to help us with the 10x sing-cell RNA sequencing. Conflict of Interest Supplementary Information is available for this paper. The authors declare that there was no conflict of interest. References Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021. Zanoni DK, Montero PH, Migliacci JC, Shah JP, Wong RJ, Ganly I, et al. Survival outcomes after treatment of cancer of the oral cavity (1985-2015). Oral Oncol. 2019;90:115-21. Kalluri R. The biology and function of fibroblasts in cancer. Nat Rev Cancer. 2016;16(9):582-98. Louault K, Li RR, DeClerck YA. Cancer-Associated Fibroblasts: Understanding Their Heterogeneity. Cancers (Basel). 2020;12(11). Cho JK, Hyun SH, Choi N, Kim MJ, Padera TP, Choi JY, et al. Significance of lymph node metastasis in cancer dissemination of head and neck cancer. Transl Oncol. 2015;8(2):119-25. Matsumoto F, Mori T, Matsumura S, Matsumoto Y, Fukasawa M, Teshima M, et al. Prognostic significance of surgical extranodal extension in head and neck squamous cell carcinoma patients. Jpn J Clin Oncol. 2017;47(8):699-704. Riedel A, Shorthouse D, Haas L, Hall BA, Shields J. Tumor-induced stromal reprogramming drives lymph node transformation. Nat Immunol. 2016;17(9):1118-27. Nakayama H, Ohuchida K, Yoshida M, Miyazaki T, Takesue S, Abe T, et al. Degree of desmoplasia in metastatic lymph node lesions is associated with lesion size and poor prognosis in pancreatic cancer patients. Oncol Lett. 2017;14(3):3141-7. Lavie D, Ben-Shmuel A, Erez N, Scherz-Shouval R. Cancer-associated fibroblasts in the single-cell era. Nat Cancer. 2022;3(7):793-807. Zhao X, Ding L, Lu Z, Huang X, Jing Y, Yang Y, et al. Diminished CD68(+) Cancer-Associated Fibroblast Subset Induces Regulatory T-Cell (Treg) Infiltration and Predicts Poor Prognosis of Oral Squamous Cell Carcinoma Patients. Am J Pathol. 2020;190(4):886-99. Ding L, Fu Y, Zhu N, Zhao M, Ding Z, Zhang X, et al. OXTR(High) stroma fibroblasts control the invasion pattern of oral squamous cell carcinoma via ERK5 signaling. Nat Commun. 2022;13(1):5124. Ding L, Ren J, Zhang D, Li Y, Huang X, Hu Q, et al. A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous cell carcinoma via LncRNA-CAF/interleukin-33. Carcinogenesis. 2018;39(3):397-406. Cillo AR, Kürten CHL, Tabib T, Qi Z, Onkar S, Wang T, et al. Immune Landscape of Viral- and Carcinogen-Driven Head and Neck Cancer. Immunity. 2020;52(1):183-99.e9. Puram SV, Tirosh I, Parikh AS, Patel AP, Yizhak K, Gillespie S, et al. Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer. Cell. 2017;171(7):1611-24.e24. Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol. 2018;36(5):411-20. Aran D, Looney AP, Liu L, Wu E, Fong V, Hsu A, et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat Immunol. 2019;20(2):163-72. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. Qiu X, Mao Q, Tang Y, Wang L, Chawla R, Pliner HA, et al. Reversed graph embedding resolves complex single-cell trajectories. Nat Methods. 2017;14(10):979-82. Vento-Tormo R, Efremova M, Botting RA, Turco MY, Vento-Tormo M, Meyer KB, et al. Single-cell reconstruction of the early maternal-fetal interface in humans. Nature. 2018;563(7731):347-53. Chen Z, Zhou L, Liu L, Hou Y, Xiong M, Yang Y, et al. Single-cell RNA sequencing highlights the role of inflammatory cancer-associated fibroblasts in bladder urothelial carcinoma. Nat Commun. 2020;11(1):5077. Zhang M, Yang H, Wan L, Wang Z, Wang H, Ge C, et al. Single-cell transcriptomic architecture and intercellular crosstalk of human intrahepatic cholangiocarcinoma. J Hepatol. 2020;73(5):1118-30. Cai W, Sun X, Jin F, Xiao D, Li H, Sun H, et al. PERK-eIF2α-ERK1/2 axis drives mesenchymal-endothelial transition of cancer-associated fibroblasts in pancreatic cancer. Cancer Lett. 2021;515:86-95. Bartoschek M, Oskolkov N, Bocci M, Lövrot J, Larsson C, Sommarin M, et al. Spatially and functionally distinct subclasses of breast cancer-associated fibroblasts revealed by single cell RNA sequencing. Nat Commun. 2018;9(1):5150. Rodda LB, Lu E, Bennett ML, Sokol CL, Wang X, Luther SA, et al. Single-Cell RNA Sequencing of Lymph Node Stromal Cells Reveals Niche-Associated Heterogeneity. Immunity. 2018;48(5):1014-28.e6. Malhotra D, Fletcher AL, Astarita J, Lukacs-Kornek V, Tayalia P, Gonzalez SF, et al. Transcriptional profiling of stroma from inflamed and resting lymph nodes defines immunological hallmarks. Nat Immunol. 2012;13(5):499-510. Cyster JG, Ansel KM, Reif K, Ekland EH, Hyman PL, Tang HL, et al. Follicular stromal cells and lymphocyte homing to follicles. Immunol Rev. 2000;176:181-93. Trapnell C, Cacchiarelli D, Grimsby J, Pokharel P, Li S, Morse M, et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat Biotechnol. 2014;32(4):381-6. Aibar S, González-Blas CB, Moerman T, Huynh-Thu VA, Imrichova H, Hulselmans G, et al. SCENIC: single-cell regulatory network inference and clustering. Nat Methods. 2017;14(11):1083-6. Dotto GP, Rustgi AK. Squamous Cell Cancers: A Unified Perspective on Biology and Genetics. Cancer Cell. 2016;29(5):622-37. Domanska UM, Kruizinga RC, Nagengast WB, Timmer-Bosscha H, Huls G, de Vries EG, et al. A review on CXCR4/CXCL12 axis in oncology: no place to hide. Eur J Cancer. 2013;49(1):219-30. Qi J, Sun H, Zhang Y, Wang Z, Xun Z, Li Z, et al. Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat Commun. 2022;13(1):1742. Brett EA, Sauter MA, Machens HG, Duscher D. Tumor-associated collagen signatures: pushing tumor boundaries. Cancer Metab. 2020;8:14. Lambrechts D, Wauters E, Boeckx B, Aibar S, Nittner D, Burton O, et al. Phenotype molding of stromal cells in the lung tumor microenvironment. Nat Med. 2018;24(8):1277-89. ChallaSivaKanaka S, Vickman RE, Kakarla M, Hayward SW, Franco OE. Fibroblast heterogeneity in prostate carcinogenesis. Cancer Lett. 2022;525:76-83. Öhlund D, Handly-Santana A, Biffi G, Elyada E, Almeida AS, Ponz-Sarvise M, et al. Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer. J Exp Med. 2017;214(3):579-96. Gittenberger-de Groot AC, Vrancken Peeters MP, Mentink MM, Gourdie RG, Poelmann RE. Epicardium-derived cells contribute a novel population to the myocardial wall and the atrioventricular cushions. Circ Res. 1998;82(10):1043-52. Peng D, Fu M, Wang M, Wei Y, Wei X. Targeting TGF-β signal transduction for fibrosis and cancer therapy. Mol Cancer. 2022;21(1):104. Su J, Morgani SM, David CJ, Wang Q, Er EE, Huang YH, et al. TGF-β orchestrates fibrogenic and developmental EMTs via the RAS effector RREB1. Nature. 2020;577(7791):566-71. Kuzet SE, Gaggioli C. Fibroblast activation in cancer: when seed fertilizes soil. Cell Tissue Res. 2016;365(3):607-19. Choi SY, Bae H, Jeong SH, Park I, Cho H, Hong SP, et al. YAP/TAZ direct commitment and maturation of lymph node fibroblastic reticular cells. Nat Commun. 2020;11(1):519. Fletcher AL, Acton SE, Knoblich K. Lymph node fibroblastic reticular cells in health and disease. Nat Rev Immunol. 2015;15(6):350-61. Sahai E, Astsaturov I, Cukierman E, DeNardo DG, Egeblad M, Evans RM, et al. A framework for advancing our understanding of cancer-associated fibroblasts. Nat Rev Cancer. 2020;20(3):174-86. Additional Declarations (Not answered) Supplementary Files SupplementaryFigure1.jpg SupplementaryFigure2.jpg SupplementaryFigure3.jpg SupplementaryFigure4.jpg SupplementaryFigure5.jpg SupplementaryTable1.docx SupplementaryTable2.xlsx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3862426","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":269819058,"identity":"f282b915-7da6-4961-9914-93cc1d52cf44","order_by":0,"name":"Yuxin Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8UlEQVRIiWNgGAWjYBADHgZm5gMHPlRIyPETr4W9LfHgjDMWxpINxNtzxvgwb1tF4gZCWgyOnz38mrftsIy5RILBAd55EowbGJgfPrqBT8uZvDRr3rY0HssZCQkHJLdJMJszsBkb5+DRYnYgx8yYt82Gx+BGwoEDhtsk2CwbeNik8Wo5/wakRQKoJbHhQOIcIOMAIS03cowfg205c5jhwMEGCQmCWuxvvDFjnHMujcfgeBvDwYZjEgaSzQT8ItmfY/zhTdlhe4PD/J8//6mpq+9nb374GJ8WIGCT4kHhM+NXDlby8QdhRaNgFIyCUTCSAQCMtE+bTQ64fQAAAABJRU5ErkJggg==","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":true,"prefix":"","firstName":"Yuxin","middleName":"","lastName":"Wang","suffix":""},{"id":269819059,"identity":"f8236f2c-887c-4149-8590-e60f1e25c5e4","order_by":1,"name":"Qian Zhang","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Zhang","suffix":""},{"id":269819060,"identity":"94948b02-47e5-493f-99f6-f4d3ef6a2c65","order_by":2,"name":"Liang Ding","email":"","orcid":"https://orcid.org/0000-0002-1043-0923","institution":"Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Liang","middleName":"","lastName":"Ding","suffix":""},{"id":269819061,"identity":"367faa39-01df-4a8b-93a6-30a69d99413e","order_by":3,"name":"Jingyi Li","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Jingyi","middleName":"","lastName":"Li","suffix":""},{"id":269819062,"identity":"8eb405fc-d826-45b7-b308-d391672a0492","order_by":4,"name":"Kunyu Liu","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Kunyu","middleName":"","lastName":"Liu","suffix":""},{"id":269819063,"identity":"3f114751-63d6-46b0-9a08-0ab37b379ff2","order_by":5,"name":"Chengwan Xia","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Chengwan","middleName":"","lastName":"Xia","suffix":""},{"id":269819064,"identity":"a0958c8d-9a1c-479c-8063-73a0f27d0ea7","order_by":6,"name":"Sheng Chen","email":"","orcid":"","institution":"Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Sheng","middleName":"","lastName":"Chen","suffix":""},{"id":269819065,"identity":"31f3ed7b-bd8b-464a-86cc-09560e4dcf41","order_by":7,"name":"Xiaofeng Huang","email":"","orcid":"","institution":"The Institute and Hospital of Stomatology, Nanjing University Medical School","correspondingAuthor":false,"prefix":"","firstName":"Xiaofeng","middleName":"","lastName":"Huang","suffix":""},{"id":269819066,"identity":"1bb256cf-e5af-470b-a22e-e99fb32fd095","order_by":8,"name":"Yumei Pu","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Yumei","middleName":"","lastName":"Pu","suffix":""},{"id":269819067,"identity":"608b2d98-78bc-4257-a14f-1bd78c8547ba","order_by":9,"name":"Qingang Hu","email":"","orcid":"","institution":"Nanjing Stomatological Hospital, Medical School of Nanjing University","correspondingAuthor":false,"prefix":"","firstName":"Qingang","middleName":"","lastName":"Hu","suffix":""}],"badges":[],"createdAt":"2024-01-14 06:50:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3862426/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3862426/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":50385699,"identity":"fdc221b1-9f44-4fb5-966f-1af382914405","added_by":"auto","created_at":"2024-01-30 17:35:12","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1683687,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCell type identification of OSCC PTs, MLNs, and DLNs using \u0026nbsp;scRNA-seq.\u003c/strong\u003e A. Pathological analysis of OSCC patients with high Des and TACS scores revealed poor OS and DFS and a high ENE incidence. B. Workflow of sample collection and data analysis processes in this study. C. UMAP of 87,650 cells profiled, with each cell color-coded for cluster number (left) and origin location (right). D. UMAP of nine identified cell types and their distribution in distinct locations. E. Dot plots of conserved and specific cell markers in identified cells. F. Expression of marker genes for the cell types identified above each panel. G. UMAP of 14,405 identified endothelial cells profiled, with each cell color-coded for cluster number, origin location, and identified endothelial cell type (VEC and LEC). H. Histogram of endothelial cell type distribution in distinct locations. I. Expression of marker genes for VEC and LEC identified above each panel. J. UMAP of 2,525 identified epithelial cells profiled, with each cell color-coded for cluster number, origin location, CNV level, and identified normal or malignant type. K. Histogram of normal or malignant cell distribution in distinct locations.\u003c/p\u003e\n\u003cp\u003eVEC, vascular endothelial cells; LEC, lymphatic endothelial cells; OSCC, oral squamous cell carcinoma; PT, primary tumor; MLN, metastatic lymph nodes; DLN, draining lymph nodes; scRNA-seq, single-cell RNA sequencing; Des, desmoplasia; TACS, tumor-associated collagen signature; OS, overall survival; DFS, disease-free survival; ENE, extranodal extension; UMAP, uniform manifold approximation and projection; CNV, copy number variation\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/0cb0581d69e4930c94951168.jpg"},{"id":50383919,"identity":"7e0e1424-3da6-44ab-bdd8-6041cac7d002","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2024862,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacter identification of fibroblasts\u003c/strong\u003e. A. tSNE map of all fibroblasts profiled, with each cell color-coded for cluster number (left), origin location (middle), and three identified main types (right). B. Histogram showing the distribution of three main cell types in distinct locations. C. Heatmap depicting the expression of top high-variable genes from CAFs, iFCs, and LN-SCs. D. Dot plot of typical gene expression of all identified fibroblast subtypes. E. tSNE representation of typical marker gene expression in each panel. F. tSNE map of identified fibroblast subtypes (left) and their distribution in distinct locations (right). G. Heat map of GSVA score comparison of GO biological processes and KEGG pathway among PTs, MLNs, and DLNs. H. Dot plot of enrichment analysis for distinct CAF subtypes. I. tSNE representation of the distribution of distinct geneset scores for all fibroblasts. J. Pseudotime analysis of all identified fibroblast subtypes. K. tSNE (left) and histogram (right) of two mCAF subtypes (mCAF1 and mCAF2), showing different dominant distributions in PTs or MLNs.\u003c/p\u003e\n\u003cp\u003etSNE, t-distributed stochastic neighbor embedding; CAFs, cancer-associated fibroblasts; iFCs, inherent fibroblasts; LN-SCs, lymph node stromal cells; GSVA, gene set variation analysis; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PTs, primary tumors; MLNs, metastatic lymph nodes; DLNs, distant lymph nodes; mCAFs, cancer-associated myofibroblastic cells\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/8f598841767214887451ae94.jpg"},{"id":50383922,"identity":"9b066cd2-6a52-468c-b231-d44e0307ac4b","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1834949,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePseudotime trajectory analysis among eFCs, eCAFs, and epithelial cells.\u003c/strong\u003e A. \u0026nbsp;Expression patterns of fibroblastic marker genes (DCN, COL1A1, COL1A2, and COL3A1) and epithelial marker genes (EPCAM, CDH1, KRT8, and KRT18) along the inferred pseudotime, color-coded by fibroblast subtypes. B. tSNE plot of the expression of cell cycle-related genes (MKI67, CENPF, and TOP2A) visualized among fibroblast clusters. C. Histogram displaying distinct cell-cycle phage distribution among fibroblast subtypes. D. Pseudotime trajectory of eCAFs, eFCs, and epithelial cells with distinct color coding for cell type (left), state (middle), and pseudotime (right). E. Histogram showing eCAF, eFC, and epithelial cell distribution among pseudotime states. F. Heatmap depicting distinct gene expression signatures of four identified gene modules (Modules 1–4) along epithelial cell–eCAF–eFC trajectory. G. Dot plot of enrichment analysis for distinct gene modules performed on pseudotime from epithelial cell-eCAF-eFC trajectory.\u003c/p\u003e\n\u003cp\u003eeCAFs, epithelial–mesenchymal transition (EMT)-like CAFs; eFCs, EMT-like fibroblasts; tSNE, t-distributed stochastic neighbor embedding\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/304f5da09ae45b8c540b3c0d.jpg"},{"id":50385700,"identity":"c8773bc8-93cf-49d0-9f3a-60ad58f9dd18","added_by":"auto","created_at":"2024-01-30 17:35:12","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1922618,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTwo distinct mCAF populations revealed distinct origins in OSCC PTs and MLNs.\u003c/strong\u003e A. tSNE plot showing two distinct mCAF subtypes (mCAF1 and mCAF2) expressing similar high myofibroblastic genes (ACTA2, TAGLN) but exhibiting distinct preference for RGS4 and COMP.\u003c/p\u003e\n\u003cp\u003eB. Pseudotime reveals two distinct cell trajectories for mCAF1 and mCAF2. C. Pseudotime ordering on mCAF2 and FRCs, arranging them into one major trajectory with three minor bifurcations. D. Histogram of mCAF2 and FRC cluster distribution in each pseudotime states. E. Heatmap of four gene modules (Modules 1 to 4) of pseudotime-dependent genes in mCAF2 and FRC. F. Dot plots illustrating enrichment analysis of four pseudotime-dependent gene modules in mCAF2 and FRC. G. Dynamic gene expression of COMP, SFRP2, ACTA2, and CCN3 along the trajectory of mCAF2 and FRCs. H. Pseudotime ordering on mCAF1 and mFCs, arranging them into two diverse trajectories. I. Histogram depicting the distribution of mCAF1 and mFC clusters in each pseudotime state. J. Heatmap showing four gene modules (Modules 1 to 4) of pseudotime-dependent genes in mCAF1 and mFCs. K. Dot plots illustrating enrichment analysis of four pseudotime-dependent gene modules of mCAF1 and mFCs. L. Heatmap displaying the expression of representative genes in pseudotime-dependent gene modules of mCAF–mFC. M. Analysis based on TCGA HNSCC bulk RNAseq cohort revealing significantly higher enrichment score of ECM organization in tumor tissue compared to normal tissue.\u003c/p\u003e\n\u003cp\u003ePTs, primary tumors; MLNs, metastatic lymph nodes; tSNE, t-distributed stochastic neighbor embedding; FRCs, fibroblastic reticular cells; TCGA, The Cancer Genome Atlas; HNSCC, head and neck squamous cell carcinoma; ECM, extracellular matrix; mCAFs, cancer-associated myofibroblasts; mFCs, myofibroblasts\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/20551491ec2cc7f8fb896513.jpg"},{"id":50385701,"identity":"6ca79631-b68b-43b6-b8e7-25038e1e07f9","added_by":"auto","created_at":"2024-01-30 17:35:13","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":2786097,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInverse crosstalk between mCAF1 or mCAF2 and immune cells. \u003c/strong\u003eA. Heatmap showing high-variation genes of mCAF1 and mCAF2. B. IHC analysis revealed distinct concentrations of RGS4+ mCAF1 in PTs with COMP+ mCAF2 in corresponding paired MLNs. C. Histogram depicting distinct GO enrichment analysis between mCAF1 and mCAF2. D. tSNE of NK/T cell classification (left) and distribution (right) among PT, MLN, and DLN. E. Histogram showing significant ligand-receptor cell pairs between immune cells and mCAF1 or mCAF2. F. Dot plots illustrating the top significant receptor–ligand pairs between mCAF1 and immune cells. G. IHC analysis revealed significantly lower aggregation of macrophages (CD68+) and T cells (CD3+) in MLN tissues with rich COMP+ mCAF2 population. H. IHC analysis indicated significantly higher infiltration of macrophages (CD68+), T cells (CD3+), and Treg cells (FOXP3+) in PT tissues with an abundance of RGS4+ mCAF1. * P \u0026lt; 0.05\u003c/p\u003e\n\u003cp\u003ePTs, primary tumors; MLNs, metastatic lymph nodes; GO, Gene Ontology; tSNE, t-distributed stochastic neighbor embedding; NKs, natural killer cells; DLNs, draining lymph nodes; mCAFs, cancer-associated myofibroblastic cells; IHC, immunohistochemical\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/6202284a31455634b2715c8e.jpg"},{"id":50383929,"identity":"1200dbb9-26ae-412c-a387-69ebd33101fd","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":966415,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003emCAF2 in MLNs showed a stronger collagen remodeling ability and influenced ENE formation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA. IHC analysis showed distinct COMP expression levels in OSCC MLNs with different collagen content. B. Histogram depicting higher COMP IHC scores in MLNs with higher TACS score or ENE+ MLNs. C. Immunofluorescence images showing higher expression of ACTA2, FAP, and FSP1 in CAFs compared to that in MAFs and LFs. D. CCK-8 test demonstrated higher cell viability of CAFs and MAFs compared to LFs. Transwell migration examination (E) and wound healing assay (F) revealed higher migration ability of CAFs and MAFs compared to LFs. G. MAFs exhibited higher collagen contraction ability compared to CAFs and LFs. H. Immunofluorescence images of ECM remodeling genes (COL11A1, FN, TNC, and POSTN) in CAFs, MAFs, and LFs. I. Western blotting showing the expression of COMP, MMP1, MMP2, ACTA2, PDPN, and RGS4 in CAFs, MAFs, and LFs.\u003c/p\u003e\n\u003cp\u003eMLNs, metastatic lymph nodes; ENE, extracapsular extension; TACS, tumor-associated collagen signatures; CAFs, cancer-associated fibroblasts; LFs, lymph node fibroblasts; CCK-8, Cell Counting Kit-8; ECM, extracellular matrix; MAFs, metastatic lymph node fibroblasts; IHC, immunohistochemistry; OSCC, oral squamous cell carcinoma\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/49c932136d9fd1962c3fd5d3.jpg"},{"id":50386350,"identity":"9c362311-c211-438a-89dc-740a502cbff7","added_by":"auto","created_at":"2024-01-30 17:43:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1853023,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/c1ee61d7-28dc-41dd-8955-a3e533488f21.pdf"},{"id":50383918,"identity":"e832edc4-06a4-4b80-a11a-3f047cc88afb","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1689954,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementaryFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/b84d3fd69e9b4498f0006398.jpg"},{"id":50383920,"identity":"4414f7ff-0059-4caf-9a9a-65ebc9819020","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":429873,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/fa99a7dceb3b74f6744ca9e7.jpg"},{"id":50383921,"identity":"600fc340-f724-4e7d-81ba-6a0be26c3dc8","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1934935,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/0afc1d5f06b6e26ebdae51ff.jpg"},{"id":50383924,"identity":"5b241ef9-005f-45d7-8ad8-ab14b6d30585","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":1418761,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/569f29d0ab2d507e36249750.jpg"},{"id":50383928,"identity":"15707942-3200-42bd-ac2b-635bbef611c1","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"jpg","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":1281819,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/eec796a37a13c8acfe54c520.jpg"},{"id":50383930,"identity":"f5349f92-fba8-4aa3-9d13-a500cd56a48f","added_by":"auto","created_at":"2024-01-30 17:27:13","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":15853,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/110f7b92492cacc1482bb078.docx"},{"id":50383926,"identity":"b17e05c7-e0d2-4b0b-a19b-38eabc415867","added_by":"auto","created_at":"2024-01-30 17:27:12","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":11289,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-3862426/v1/29dbb6246a45bb883376ecf9.xlsx"}],"financialInterests":"(Not answered)","formattedTitle":"Single-cell RNA sequencing of OSCC primary tumors and lymph nodes reveals distinct origin and phenotype of fibroblasts","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLymph node (LN) metastasis portends poor prognosis in oral squamous cell carcinoma (OSCC)\u003csup\u003e1, 2\u003c/sup\u003e. Extensive functional evidence supports the involvement of cancer-associated fibroblasts (CAFs) in various stages of metastasis, including cancer proliferation, local invasion, intravasation, metastatic niche priming, and colonization\u003csup\u003e3, 4\u003c/sup\u003e. Particularly, the infiltration of metastatic cells beyond the LN capsule, known as extranodal extension (ENE), significantly increases the risk of distant metastasis, rendering it the leading cause of mortality in OSCC\u003csup\u003e5, 6\u003c/sup\u003e. Consequently, targeting metastatic LNs (MLNs)\u0026nbsp;exhibiting ENE rather than focusing solely on the primary tumor (PT) is crucial for preventing distant metastasis in patients with advanced OSCC to improve survival rates\u003csup\u003e5\u003c/sup\u003e. Previous studies showed that desmoplasia by CAFs in MLNs is intricately linked to the transformation of the tumor microenvironment (TME)\u003csup\u003e7, 8\u003c/sup\u003e.\u0026nbsp;However, the mechanisms of CAFs in MLN formation and development remain unclear.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCAFs exhibit significant heterogeneity in their phenotypes, functions, and origins. Single-cell RNA sequencing (scRNA-seq) enables the characterization of gene profiles in individual cells and the tracing of cell lineages. Recent advancements in scRNA-seq have facilitated the identification of distinct phenotypic and functional subgroups of CAFs in PTs, including cancer-associated myofibroblasts (mCAFs), inflammatory CAFs (iCAFs), and antigen-presenting CAFs (apCAFs)\u003csup\u003e9\u003c/sup\u003e. Although previous studies have demonstrated a partial overlap among these CAF subgroups\u003csup\u003e9\u003c/sup\u003e, their origins and associations with functional characteristics remain poorly understood.\u003c/p\u003e\n\u003cp\u003eOur research team has long been committed to the study of CAF heterogeneity in OSCC TME and has revealed their distinct characters and functions in tumor progression, invasion, and immune regulation\u003csup\u003e10-12\u003c/sup\u003e. However, systematic studies on CAFs in OSCC TME are lacking and comparison between MLN and PT is sparse\u003csup\u003e13, 14\u003c/sup\u003e.\u0026nbsp;Understanding the mechanisms underlying the origins of key CAF subgroups and their contributions to tumor development within LNs is critical for formulating strategies to target LN metastasis.\u003c/p\u003e\n\u003cp\u003eIn this study, we conducted transcriptome analysis using 87,650 single cells from paired PTs, MLNs, and draining lymph nodes (DLNs) with no evidence of metastasis from OSCC patients. We comprehensively characterized the single-cell landscape of fibroblasts from PTs, MLNs, and DLNs, providing insights into the cellular dynamics and molecular features associated with metastasis.\u003c/p\u003e"},{"header":"Materials And Methods","content":"\u003cp\u003e1.Patients and samples.\u003c/p\u003e\n\u003cp\u003eAll samples were obtained from the Nanjing Stomatological Hospital. All involved patients provided informed consent for this work and all procedures were approved by the Institutional Review Bord of Nanjing stomatological hospital, Medical school of Nanjing University (IRB: NJSH 2021NL011).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(1) scRNA sequencing sample\u003c/p\u003e\n\u003cp\u003eThree fresh primary oral squamous carcinoma tissues (1 in tongue, 1 in floor of mouth and 1 in buccal mucosa) along with three matched metastatic lymph nodes and draining lymph nodes were obtained for scRNA sequencing (Supplementary Table1, No 1-3#). Their pathological status was proved by intraoperative frozen diagnosis.\u003c/p\u003e\n\u003cp\u003e(2) Tissue chip for immunochemistry\u003c/p\u003e\n\u003cp\u003e3 tissue chips consisted of OSCC tissue sample from Nanjing Stomatological Hospital were used in this study. 1 tissue chip contains 95 MLN samples of OSCC patients. The other 2 tissue chips contain a total of 124 OSCC tumor samples and paired 68 adjacent normal samples. Their pathological status was confirmed according to HE and CK5/6 immunohistochemistry test. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(3) Primary fibroblast cells\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2 pairs of fresh tissue sample from OSCC PT and matched MLN and DLN tissue were collected for primary fibroblast cells extraction (Supplementary Table1, No.4-5#).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2. Single cell RNA sequencing\u003c/p\u003e\n\u003cp\u003e(1) Cell preparation\u003c/p\u003e\n\u003cp\u003eFresh primary tumor and lymph node tissue were processed immediately after obtaining from OSCC patients. Every sample was cut into small pieces(\u0026lt;1mm in diameter) and washed twice before processing. Sterile 1640 medium containing 0.04% BSA was used in washing and dispensing. Then they were incubated with dissociation reagent containing 5mg/ml\u0026nbsp;collagenase I, 1 mg/ml DNases, 2.5mg/ml hyaluronidase, and 1mg/ml dispase in thermostatic shaker (37\u0026deg;C, 180 rpm). After digestion for 15 mins, supernatant was transferred into a new centrifuge tube and preserved on ice. 3ml dissociation reagent was added into residual tissue sediment and incubated in thermostatic shaker (37\u0026deg;C, 180 rpm) for 15 mins. Above procedures were repeated for three times and all obtained supernatant were centrifugated at 250g for 10 mins to achieve cell pellets. The cell pellets were resuspended and filtered by a 40um cell mesh which was repeated 3 times for filtering effect. Then the cell pellet was resuspended in 1ml ice cold red blood cell lysis buffer and incubated at 4\u0026deg;C for 10 mins. Next, 10ml ice-cold PBS was added and centrifuged at 250g for 10 mins to remove red blood cells and resuspended with 1640 medium with 0.04% BSA.\u003c/p\u003e\n\u003cp\u003eThen, most hematopoietic cells were removed using CD45 microbeads according to manufacturer\u0026rsquo;s protocol (130-45-801, Miltenyi Biotec). Then, 10\u0026micro;l of suspension was counted under an inverted microscope with a hemocytometer. Trypan blue combined with AO/PI (Sorbio Biotec) fluorescent staining was used to assess cell viability. While cell viability was less than 70%, dead cell removal kit (130-090-101, Miltenyi Biotec) was used to improve it. Finally, eligible single cell suspension samples (cell viability more than 85%, total cell counts more than 1x10\u003csup\u003e5\u003c/sup\u003e, cell diameter less than 40um) were involved in following sequencing.\u003c/p\u003e\n\u003cp\u003e(2) 10x scRNA sequencing\u0026nbsp;\u003c/p\u003e\n\u003cp\u003escRNA sequencing was performed using 10\u0026times;Genomics platform by OE Biotech Co., Lts(Shanghai, China). According to manufacture\u0026rsquo;s standard protocol, cell suspension was loaded to the Chromium Single-cell v3 3\u0026rsquo; Chemistry Library.\u0026nbsp;Instrument (10\u0026times;Genomics) to generate single-cell gel beads in emulsions (GEMs).\u0026nbsp;After generation of GEMs, reverse transcription (RT) reactions were engaged to generate barcoded full-length cDNA, which was followed by GEM-RT clean-up and cDNA amplification (DynaBeads). Subsequently, the amplified cDNA was fragmented, end-repaired, A-tailed, and ligated to an index adaptor, and then the library was amplified. Every library was sequenced on a HiSeq X Ten platform (Illumina), and 150bp paired-end reads were generated.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;(3) Raw data processing and quality control\u003c/p\u003e\n\u003cp\u003eThe library construction, sequencing and data analysis were carried out by OE Biotech Co., Lts(Shanghai, China). The size and purity of all libraries were checked by Agilent 2100 Bioanalyzer. The raw data was evaluated using FastQC software to ensure sequencing quality. Then, the Cell Ranger software pipeline(version 3.1.0) was used to demultiplex cellular barcodes, map reads to the genome and transcriptome using the STAR aligner, and down-sample reads as required to generate normalized aggregate data across samples, producing a matrix of gene counts versus cells.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe processed the unique molecular identifier (UMI) count matrix using the R package Seurat(version 3.0)\u003csup\u003e15\u003c/sup\u003e.\u0026nbsp;Cells with UMI numbers \u0026lt;1000 or with over 10% mitochondrial-derived UMI counts were considered low-quality cells and were removed.\u0026nbsp;Library size normalization was then performed on the filtered matrix to obtain the normalized count.\u003c/p\u003e\n\u003cp\u003e(4) Dimensionality reduction and clustering\u003c/p\u003e\n\u003cp\u003ePrincipal component analysis (PCA) was performed to reduce the dimensionality on the log transformed gene-barcode matrices of top variable genes and top\u0026nbsp;top 50 PCs were used to perform the downstream analysis. Cells were clustered based on a graph-based clustering approach, and were visualized in 2-dimension using t-distributed stochastic neighbor embedding\u0026nbsp;(tSNE) or uniform manifold approximation and projection (UMAP). In this present study, the main 23 cell clusters were identified under 0.4 resolution using the FindClusters function in Seurat.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(5) Marker gene and cell type identification\u003c/p\u003e\n\u003cp\u003eThe Seurat Findallmarker function was performed to identify preferentially expressed genes in each cluster as marker genes. Here, we use the R package SingleR\u003csup\u003e16\u003c/sup\u003e for unbiased cell type recognition of scRNA-seq and identify cell types reference to\u0026nbsp;conventional markers as described in previous studies.\u003c/p\u003e\n\u003cp\u003e(6) Differentially expressed genes(DEGs) and enrichment analysis\u003c/p\u003e\n\u003cp\u003eDEGs were identified using the Seurat package. P value \u0026lt; 0.05 and |log\u003csub\u003e2\u003c/sub\u003efoldchange| \u0026gt; 1was set as the threshold for significantly differential expression. GO enrichment and KEGG pathway enrichment analysis of DEGs were respectively performed using R based on the hypergeometric distribution. GSVA analysis was performed to grade each pathway using GSVA package\u003csup\u003e17\u003c/sup\u003e. Then the differences between different cell groups were analyzed using limma package\u003csup\u003e18\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(7) Transcription factor inference\u003c/p\u003e\n\u003cp\u003eTo assess TF regulation strength, we applied the single-cell regulatory network inference and clustering (pySCENIC, v0.9.5) workflow, using the 20 thousand motifs database for RcisTarget and GRNboost.\u003c/p\u003e\n\u003cp\u003e(8) Pseudotime analysis\u003c/p\u003e\n\u003cp\u003eTo map origin heterogeneity of fibroblast cells, we performed pseudotime analysis with Monocle2\u003csup\u003e19\u003c/sup\u003e. For Monocle analysis, positive marker genes for each cluster were used. Based on the expression profiles of these pseudotime-dependent genes, we perform spatial dimensionality reduction and construct a minium spanning tree (MST). Then, using this MST, we identify the longest path that represents the differentiation trajectory of cells with similar transcriptional features. Then the GO terms analysis was performed using clusterProfilter package.\u003c/p\u003e\n\u003cp\u003e(9) CNVs estimation\u003c/p\u003e\n\u003cp\u003eThe InferCNV package was used to detect the copy number variations (CNVs) in EPCAM+ cells with default parameters. All epithelial cells were clustered into 4 clusters and CNVs levels were estimated with normal fibroblasts as control group. Then malignant cells were recognized with extra high CNVs score.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(10)\u0026nbsp;Cell-cycle discrimination analyses.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe used cell cycle-related genes, including a previously defined core set of 43 G1/S and 54 G2/M genes29. For each cell, a cell cycle phase (G1, S, G2/M) was assigned based on its expression of G1/S or G2/M phase genes using the scoring strategy described in CellCycleScoring function in Seurat. Cells in each cell cycle state were also quantified using the which.cells function\u003c/p\u003e\n\u003cp\u003e(11) Cell-cell communication analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo identify different potential interactions between different cells, we used CellPhoneDB\u003csup\u003e20\u003c/sup\u003e to analyze ligand-receptor relationships between different proteins on the cell membrane and in the cytosol. Significant mean and cell communication significance (P \u0026lt; 0.05) was calculated based on the interaction and the normalized cell matrix achieved by scran normalization.\u003c/p\u003e\n\u003cp\u003e3. Correlation to public datasets\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTranscriptome and clinical data from The Cancer Genome Atlas (TCGA) HNSC datasets were obtained from https://portal.gdc.cancer.gov. To compare gene expression of extracellular matrix organization and muscle contraction between HNSCC tumor and normal samples, 2 GSEA genesets were applied in this study (Supplementary table 2). We perform ssGSEA analysis on the TCGA-HNSC transcriptome data to calculate the enrichment scores (ES) for each sample in the genesets of extracellular matrix organization and muscle contraction. Then the ES scores between normal or tumor groups were compared and P\u0026lt;0.05 was set as significant difference.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e4. Immunochemistry staining\u003c/p\u003e\n\u003cp\u003eTissues were fixed in 4% PFA, dehydrated, paraffin embedded, and sectioned at 7\u0026ndash;12 \u0026micro;m thickness. Frozen tissues were sectioned at 12 \u0026micro;m thickness. Immunostaining was performed both on frozen and paraffin sections. Heat-based antigen retrieval was performed when necessary. Tissue sections were blocked in either 3% Goat serum. The primary antibodies used were rabbit anti-COMP (1:500, Abcam, ab300555#), rabbit anti-RGS4 (1:200, Abcam, ab97307#), rabbit anti-CK5/6 (MXB, MAB-0744#), rabbit anti-CD3 (ZSGB, ZA-0503), rabbit anti-CD4 (ZSGB, ZM-0418), rabbit anti-CD8 (ZSGB, ZA-0508), rabbit anti-CD68 (ZSGB, ZM-0060).Tissue chip sections were scanned by Pannoramic slide Scanner system (3DHISTECH) and visualized by CaseViewer2.4 (3DHISTECH).\u003c/p\u003e\n\u003cp\u003e5. Tumor associated collagen signatures (TACS)\u003c/p\u003e\n\u003cp\u003e(1) TACS-1: No apparent or only a thin layer of collagen.\u003c/p\u003e\n\u003cp\u003e(2) TACS-2: Elongated collagen fibers parallel to the tumor boundary.\u003c/p\u003e\n\u003cp\u003e(3) TACS-3: Collagen fibers deposition with disrupted edges and presence of fiber bundles perpendicular to the tumor boundary.\u003c/p\u003e\n\u003cp\u003e6. Primary fibroblast cells extraction and culture\u003c/p\u003e\n\u003cp\u003e2 OSCC patients who underwent surgical treatment at the Department of Oral and Maxillofacial Surgery, Nanjing Stomatological Hospotal, Medical School of Nanjing University were included in this study. Fresh samples of OSCC PTs, paired MLNs and DLNs were diagnosed by frozen biopsy and collected for primary fibroblast cells extraction.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll obtained samples were soaked in DMEM culture medium containing 20% penicillin-streptomycin and kept at 4\u0026deg;C for 2-4 hours.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThen, they were moved into 2.5cm dish and trimmed into small pieces with 2mm diameter to remove exterior blood, fat and muscle tissues. After transferring them into 15ml centrifuge tube, centrifugation was then performed at 1500rpm at 4\u0026deg;C for 5 mins to discard the supernatant and repeated three times. Next, we added 5ml pre-prepared collagenase solution to tissue sediments and incubated them at 37\u0026deg;C and 200 rpm for 30-60 mins, until the edges of the tissue blocks appear \u0026lsquo;mist-like\u0026rsquo;. Then, we added 2ml of DMEM culture medium containing 10% FBS in tissue blocks and centrifuge them at 1500rpm for 5 mins, and discard the supernatant and repeated the process three times. Tissue blocks were then transferred to the bottom of a 25cm\u003csup\u003e2\u003c/sup\u003e culture flask and inverted it in a 37\u0026deg;C incubator for 2-4 hours.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNext, we added 4ml of complete DMEM culture medium containing 10% FBS and 1% penicillin-streptomycin to the flask. After 7 days culture, spindle-shaped fibroblast cells could be observed to grow outward. Subsequently, we renewed the medium every 2-3 days until the cells covered the flask and trypsinized them for passaging.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll primary fibroblast cells were cultured in a mixture of FM (Sciencell) and DMEM culture medium containing 10% FBS at a ratio of 1:1. They were placed in a water-jacketed CO\u003csub\u003e2\u003c/sub\u003e incubator with a stable temperature of 37\u0026deg;C and a CO\u003csub\u003e2\u003c/sub\u003e concentration of 5%.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e7. Immunofluorescence staining.\u003c/p\u003e\n\u003cp\u003eAll cells were planked with appropriate cell density on cell slides and fixed by iced methanol for 30 mins at -20\u0026deg;C. After blocking with 1% BSA for 30 mins, cells were incubated with required primary antibody at 4\u0026deg;C overnight.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAfter cleaning by PBS for three times, cells were incubated with corresponding second antibody (1:1000, Anti-mouse IgG Alexa Fluor\u0026Ograve;\u0026nbsp;594 conjugate, Anti-rabbit IgG Alexa Fluor\u0026Ograve;\u0026nbsp;488 conjugate, Cell Signaling) for 1 hour at room temperature. Finally, after cleaning by PBS for three times, the slides were sealed and visualized by fluorescent confocal microscopy (Nikon).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll primary antibodies used were mouse anti-\u0026alpha;SMA (1:500, Abcam, ab7817#), rabbit anti-COL11A1 (1:200, Abcam, ab166606#), rabbit anti-FN (1:150, Abcam, ab32419#), rabbit anti-TNC (1:200, Abcam, ab108930#), rabbit anti-FAP (1:200, Abcam, ab207178) and rabbit anti-FSP1 (1:200, Abcam, ab197896), mouse anti-POSTN (1:4000, Proteintech, 66491-1).\u003c/p\u003e\n\u003cp\u003e8. Western Blot\u003c/p\u003e\n\u003cp\u003eTotal proteins were resolved and transferred to polymembranes (Genscript). After they had been blocked with 5% BSA, the membranes were incubated overnight at 4 \u0026deg;C with anti-COMP (1:1000, Affinity, DF13438#), anti-RGS4 (1:1000, Abcam, ab97307#), anti-aSMA (1:5000, Abcam, ab32575#), anti-MMP1 (1:1000, Abcam, ab134184), anti-MMP2 (1:1000, Cell signaling technology, D4M2N#) and anti-PDPN (1:1000, Abcam, ab236529) antibodies. Then, the membranes were incubated with horseradish peroxidase conjugated goat anti-mouse immunoglobulin G or goat anti-rabbit immunoglobulin G (1:5000, Thermo) and developed with an enhanced chemiluminescence kit (E411-04#, Vazyme).\u003c/p\u003e\n\u003cp\u003e9. CCK-8 cell viability experiment\u003c/p\u003e\n\u003cp\u003eFibroblast cells (1x10\u003csup\u003e3\u003c/sup\u003e per well) were pre-cultured in 96-well plates in a 37\u0026deg;C incubator at a CO\u003csub\u003e2\u003c/sub\u003e concentration of 5% for 24 hours. Then, 10ul CCK-8 solution (KeyGEN) was added in each well, and the absorbance was measured at 450 nm with a microplate reader to measure the cell viability.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e10. Wound healing assays\u003c/p\u003e\n\u003cp\u003eThe wound healing assay was done using standard wound healing inserts (Ibidi). After fibroblast cells were seeded and attached, the insert was removed and a standard 500um width scratch between fibroblast cells in both sides was formed. Then DMEM medium containing 5% FBS was added to overlay cell patches. Then wound healing was evaluated after 24h with a light microscope.\u003c/p\u003e\n\u003cp\u003e11. Transwell migration experiment\u003c/p\u003e\n\u003cp\u003eThe tested fibroblast cells were cultured to the logarithmic growth phase, digested and suspended in serum-free medium, counted, and adjusted to a concentration of 2\u0026times;10\u003csup\u003e5\u003c/sup\u003e cells/ml. 100ul tested cells were added to the upper chamber of 24 well Transwell chambers and 600ul complete medium containing 10%FBS were added in the lower chambers. The Transwell chambers were incubated in a 37\u0026deg;C incubator at a CO\u003csub\u003e2\u003c/sub\u003e concentration of 5% for 24 hours and upper chambers were transferred into culture wells containing approximately 800\u0026mu;l methanol. After fixation at room temperature for 30 mins, chambers were then transferred to wells containing 800ul crystal violet staining solution and stained for 30 mins at room temperature. After rinsing in clean water multiple times, chambers were removed gently and cells at the bottom side were wiped off using moist cotton swab. Finally, the membranes were peeled off with small forceps and visualized under a microscope to count migrated cell numbers at 9 random fields. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e12. Statistical analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll statistical analyses and graph generation were performed in R (version 3.6.0) and GraphPad Prism (version 7.0).\u003c/p\u003e\n\u003cp\u003e13. Data availability statement\u003c/p\u003e\n\u003cp\u003eThe data reported in this study are available upon NCBI (PRJNA950395, TaxID:9606). Specific code will be available upon reasonable request.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e1. Cell type identification from primary OSCC tissue and paired lymph nodes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo assess the role of fibroblasts in LN status and prognosis, we enrolled 108 OSCC MLN samples obtained between January 2015 and December 2017 at Nanjing Stomatological Hospital, Medical School of Nanjing University. The degree of desmoplasia and TACS was scored to evaluate the stroma ratio and extracellular matrix (ECM) remodeling in MLN tissue. We also assessed the relationship between these factors and prognosis (\u003cstrong\u003eFig. 1A\u003c/strong\u003e). Our findings indicated that higher desmoplasia degree in MLNs correlated with poorer prognosis, including overall survival (OS) and disease-free survival (DFS). In addition, MLN samples with higher TACS scores were more likely to exhibit ENE.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo elucidate the OSCC TME and compare the heterogeneity of stromal cells between MLNs and PTs at single-cell resolution, we collected three pairs of fresh OSCC PT, MLN, and DLN tissue samples.\u0026nbsp;These tissues were dissociated into single cells, and magnetic cell sorting (MACS) was performed to remove CD45+ cells and reduce the proportion of hematopoietic cells\u0026nbsp;(\u003cstrong\u003eFig. 1B\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFollowing quality control and batch removal, 87,650 single cells were obtained using the 10\u0026times;Genomics platform. Using the Seurat package for unsupervised clustering, we identified 23 clusters, and their visualization was achieved through\u0026nbsp;UMAP(\u003cstrong\u003eFig. 1C\u0026amp;D\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBased on their canonical markers, we categorized these cells into nine major cell types: B cells (MS4A1+), endothelial cells (CD31+), epithelial cells (EPCAM+), fibroblasts (DCN+), macrophages (LYZ+), mast cells (KIT+), NK cells (KLRD1+), neutrophils (FCGR3B+), and T cells (CD3D+) (\u003cstrong\u003eFig. 1E\u0026amp;F, Supplementary Fig. 1\u003c/strong\u003e). Subsequently, we reclustered 14,405 endothelial cells into seven clusters (\u003cstrong\u003eFig. 1G\u0026amp;H\u003c/strong\u003e). Among these clusters, six were identified as vascular endothelial cells (VECs; CD34+ ICAM1+), whereas the remaining cluster was identified as lymphatic endothelial cells (LECs; PDPN+ POX1+) (\u003cstrong\u003eFig. 1I\u003c/strong\u003e). Epithelial cells, totaling 2,525, were classified into normal and malignant types based on their copy number variation (CNV) level (\u003cstrong\u003eFig. 1J\u0026amp;K\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eInter-tissue heterogeneity was observed in a proportion of stromal cells among OSCC PTs, MLNs, and DLNs. Fibroblasts were predominant in all three tissues but particularly abundant in OSCC PTs (\u003cstrong\u003eFig. 1D\u003c/strong\u003e). LECs were more concentrated in LN tissues, whereas VECs were evenly distributed across all tissues (\u003cstrong\u003eFig. 1H\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDespite the elimination of most CD45+ cells before scRNA-seq, we successfully captured 26,666 immune cells. Comparing their compositions between PTs, MLNs, and DLNs revealed distinct dominant immune cell populations (\u003cstrong\u003eFig. 1D\u003c/strong\u003e). Specifically, B and T cells were more enriched in LNs, whereas macrophages aggregated in MLNs. Neutrophils were restricted to tumor tissues, including PTs and MLNs, playing important roles in shaping the TME.\u003c/p\u003e\n\u003cp\u003eNotably, most captured epithelial cells originated from PT, with only a few were present in LNs (\u003cstrong\u003eFig. 1K\u003c/strong\u003e). This observation may be attributed to the high differentiation level of OSCC cells, leading to reduced epithelial cell activity after cell suspension preparation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2. Dynamics of fibroblasts in primary OSCC tissue and paired LNs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn total, 44,052 cells expressing the canonical fibroblast markers DCN, COL1A1, and COL3A1 (\u003cstrong\u003eFig. 1E\u003c/strong\u003e) were identified as fibroblasts. Subsequently, these fibroblasts were reclustered into 22 subclusters and categorized into three main types based on their dominant locations (\u003cstrong\u003eFig. 2A\u0026amp;B\u003c/strong\u003e). Specifically, Clusters 2, 8, 10, 11, 13, 15, and 22 were predominantly found in tumor sites, including PTs and MLNs, and were designated as CAFs. Clusters 3, 4, 16, 17, and 18 were mostly present in LN tissues, including MLNs or DLNs, and were characterized as LN stromal cells (LN-SCs). The remaining subclusters (1, 5, 6, 7, 9, 12, 14, 19, 20, and 21) were present in all three tissue types and were defined as tissue-inherent fibroblasts (iFCs).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDistinct high-variable genes were further explored in CAFs, LN-SCs, and iFCs (\u003cstrong\u003eFig. 2C\u003c/strong\u003e) and grouped according to their typical gene expression (\u003cstrong\u003eFig. 2D\u0026ndash;F\u003c/strong\u003e) and enrichment analysis (\u003cstrong\u003eFig. 2G\u0026amp;H\u003c/strong\u003e). CAFs were first categorized based on marker gene expression and enrichment analysis. Clusters 2 and 11 were identified as mCAFs, characterized by high levels of\u0026nbsp;\u0026alpha;- smooth muscle actin (ACTA2) (\u003cstrong\u003eFig. 2E\u003c/strong\u003e) and related genesets associated with actin cytoskeleton regulation, focal adhesion, and ECM score (\u003cstrong\u003eFig. 2I\u003c/strong\u003e). iCAFs (Clusters 8, 10, and 22) exhibited high expression of PDGFRA but lacked Acta2. Furthermore, iCAFs exhibited increased scores for genesets involved in cytokine\u0026ndash;cytokine receptor interaction (\u003cstrong\u003eFig. 2I\u003c/strong\u003e), indicating their ability to secrete cytokines, consistent with a previous report\u003csup\u003e21\u003c/sup\u003e. CAF-A (Cluster 15) exhibited increased expression of APOE (\u003cstrong\u003eFig. 2E\u003c/strong\u003e) and was exclusively limited to PT tissue, suggesting its potential roles in lipoprotein regulation (\u003cstrong\u003eFig. 2H\u003c/strong\u003e). Cluster 13 fibroblasts expressed relatively low fibroblastic genes (DCN, PDGFRA, PDPN) and ECM scores (\u003cstrong\u003eFig. 2E\u0026amp;I\u003c/strong\u003e), indicating their limited capacity for fibroblastic formation. However, they expressed relatively higher\u0026nbsp;levels of epithelial cell genes (KRT19), indicating a potential developmental relationship. Pseudotime analysis further revealed their position at the terminal route of all fibroblasts, leading to their annotation as epithelial\u0026ndash;mesenchymal transition (EMT)-like\u0026nbsp;CAFs (eCAFs) (\u003cstrong\u003eFig. 2J\u003c/strong\u003e), highlighting their potential involvement in EMT\u003csup\u003e22\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eA comparison of mCAF distribution between PTs and MLNs revealed inter-tissue heterogeneity. Notably, Cluster 2 primarily occurred in PT, whereas Cluster 11 was predominant in MLN (\u003cstrong\u003eFig. 2K\u003c/strong\u003e). Consequently, they were designated as mCAF1 and mCAF2, respectively, based on their distinct tissue preferences.\u003c/p\u003e\n\u003cp\u003eAmong the iFCs, myofibroblasts (mFCs; Clusters 5, 6, 7, and 12) (\u003cstrong\u003eFig. 2E\u003c/strong\u003e) were identified based on their high ACTA2 expression.\u0026nbsp;KRT19+ EMT-like fibroblasts (eFCs; Clusters 9 and 14) (\u003cstrong\u003eFig. 2E\u003c/strong\u003e) exhibited a gene signature similar to that of eCAFs, indicating a potential evolutionary relationship in EMT-transformation. Pseudotime analysis revealed that eFCs and eCAFs were located at the end of the evolutionary route of all fibroblast subtypes (\u003cstrong\u003eFig. 2J\u003c/strong\u003e). Cells in Clusters 1, 20, and 21 expressed typical fibroblastic genes DCN but lacked activation-associated genes (ACTA2 and PDGFRA), suggesting their resting status and delineating them as resting fibroblasts (rFCs)\u003cstrong\u003e\u0026nbsp;(Fig. 2E\u003c/strong\u003e)\u003csup\u003e3\u003c/sup\u003e. Furthermore, genes related to\u0026nbsp;endothelial cells, including SELE, VWF, PECAM1, and ACKR1, were expressed in rFCs (\u003cstrong\u003eFig. 2D\u0026amp;E\u003c/strong\u003e), suggesting their association with blood vessels and a potential role in angiogenesis\u003csup\u003e9, 23, 24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eLN-SC subsets were distinguished based on the expression levels of reported stromal cell markers and chemokines (\u003cstrong\u003eSupplementary Fig. 2\u003c/strong\u003e)\u003csup\u003e25\u003c/sup\u003e. PDPN+ CD31\u0026ndash; cells were identified as fibroblastic reticular cells (FRCs) (\u003cstrong\u003eSupplementary Fig. 2A\u003c/strong\u003e). FRCs were further grouped into CCL19\u003csup\u003ehi\u003c/sup\u003e (Cluster 3) and CCL19\u003csup\u003elo\u003c/sup\u003eCD34\u003csup\u003ehi\u003c/sup\u003e (Clusters 4 and 17) FRCs based on differences in immune cell recruitment. Among PDPN- CD31- LN-SCs, Cluster 18 was designated as perivascular cells (PvCs) owing to their expression of ITGA7 and ACTA2, whereas the remaining cells were classified as double-negative cells (DNCs)\u003csup\u003e26\u003c/sup\u003e(\u003cstrong\u003eSupplementary Fig. 2A\u003c/strong\u003e). Both CCL19\u003csup\u003ehi\u003c/sup\u003e FRCs and DNCs exhibited high secretion levels of chemokines, including CCL19, CCL21, CXCL12, and CXCL13, indicating their ability to recruit lymphocytes\u003csup\u003e27\u003c/sup\u003e (\u003cstrong\u003eSupplementary Fig. 2B\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eComparisons of biological abilities among CAFs, iFCs, and LN-SCs were made (\u003cstrong\u003eFig. 2G\u003c/strong\u003e). Gene set enrichment analysis revealed enrichment of epithelial cell proliferation and metabolism reprogramming in iFCs. Additionally, immune-related processes, including adaptive immune responses, immune effector responses, and chemokine signaling, were enriched in LN-SCs. Ossification and cell adhesion were enriched in CAFs. Furthermore, CAFs expressed functions related to antigen presentation, similar to LN-SCs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3. Epithelial\u0026ndash;fibroblast transformation in OSCC TME\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo fibroblastic subtypes positioned lower in the fibroblast trajectory were identified and indicated as eCAFs and eFCs\u003cstrong\u003e\u0026nbsp;(Fig. 2J)\u003c/strong\u003e. The expression level of the fibroblastic\u0026nbsp;marker gene DCN varied across all fibroblastic clusters, exhibiting relatively low levels in eCAFs (Cluster 13) and eFCs (Clusters 9 and 14) (\u003cstrong\u003eFig. 2E\u003c/strong\u003e). Additionally, eCAFs and eFCs expressed the epithelial marker gene KRT19, suggesting a potential translational status to epithelial cells. Upon comparing fibroblastic (DCN, COL1A1, COL1A3, and COL3A1) and epithelial (EPCAM, KRT6, KRT16, and CDH1) gene expression along the trajectory, we observed that downregulation of the fibroblastic marker genes expression was accompanied by the upregulation of the epithelial marker genes expression, coinciding with the presence of eCAFs and eFCs at the end of trajectory (\u003cstrong\u003eFig. 3A\u003c/strong\u003e). These findings suggest a potential progression from eFCs and eCAFs towards epithelial cells.\u003c/p\u003e\n\u003cp\u003eThe\u0026nbsp;CellCycleScoring function in Seurat was used to assess the proportion of cell cycle phases (G1, S, and G2/M) among all fibroblastic clusters. Notably, eCAFs exhibited a significant enrichment for cells in G2/M (\u003cstrong\u003eFig. 3B\u003c/strong\u003e). Furthermore, the marker genes of eCAFs included several cell cycle-related genes, such as MKI67, CENPF, and TOP2A, indicating the active proliferative ability of eCAFs (\u003cstrong\u003eFig. 3C\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe further investigated potential developmental relationships and identified differentially expressed genes among eCAFs, eFCs, and epithelial cells using\u0026nbsp;Monocle\u003csup\u003e28\u003c/sup\u003e. The eCAFs and eFCs clusters were positioned in the lower region of the major trajectory (\u003cstrong\u003eFig. 3D\u003c/strong\u003e). Although the definitive polarity of EMT remained undefined, the inferred pseudotime path revealed that eCAFs were in a transitioning state between eFCs and epithelial cells. Additionally, some eCAFs coexisted with epithelial cells in States 1, 4, and 10, suggesting a potential close developmental relationship (\u003cstrong\u003eFig. 3E\u003c/strong\u003e). The pseudotime-dependent genes were hierarchically clustered into four modules (\u003cstrong\u003eFig. 3F\u003c/strong\u003e). Enrichment analysis (\u003cstrong\u003eFig. 3G\u003c/strong\u003e) demonstrated that the expression of genes related to angiogenesis and vascular development increased along the axis from epithelial cells to eFCs, whereas that of genes associated with the cell cycle were reduced. Notably, the expression of genes implicated in collagen degradation, ECM receptor interaction, collagen formation, and ECM organization were particularly elevated during the transition phase, suggesting the acquisition of ECM remodeling abilities during transformation to eCAFs. In addition,\u0026nbsp;transcription factor analysis using SCENIC\u003csup\u003e29\u003c/sup\u003e revealed a highly activated TP63 pathway\u003csup\u003e30\u003c/sup\u003e in both eCAFs and eFCs, indicating a potential role in epithelial\u0026ndash;fibroblast transformation (\u003cstrong\u003eSupplementary\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eFig. 3A\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4. OSCC stroma comprised two distinct mCAF populations from PT and MLN with distinct origins\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWith the high expression of myofibroblast markers (ACTA2, TAGLN)\u003cstrong\u003e\u0026nbsp;(Fig. 4A\u003c/strong\u003e) and enrichment in vasculature development and actin cytoskeleton organization (\u003cstrong\u003eFig. 2I\u003c/strong\u003e), two fibroblast clusters (Clusters 2 and 11) were identified as mCAFs. A comparison of their locations revealed that Cluster 2 was concentrated in the PT, whereas Cluster 11 was mostly located in the MLN (\u003cstrong\u003eFig. 2K\u003c/strong\u003e). mCAFs were classified as mCAF1 (Cluster 2, RGS4+) and mCAF2 (Cluster 11, COMP+) based on their positional preference and marker gene expression (\u003cstrong\u003eFig. 4A\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo explore the origin of mCAF2 in MLN, we examined whether they developed from existing fibroblasts in MLN or metastasized from PT with tumor cells. We performed pseudotime trajectory analysis of mCAF1, mCAF2, mFCs, and FRCs, which revealed two distinct cell trajectories. In one trajectory, mCAF1 were proximal to mFCs, whereas in the other trajectory, mCAF2 exhibited proximity to FRCs (\u003cstrong\u003eFig. 4B\u003c/strong\u003e). This observation suggests their distinct origins from inherent normal fibroblasts in their tissue sources.\u003c/p\u003e\n\u003cp\u003eFurther analysis of the differentiation pseudotime genes of FRCs and mCAF2 (\u003cstrong\u003eFig. 4C\u0026amp;D\u003c/strong\u003e) revealed that they originated from Cluster 17 in State 1 and diverged into two main paths, with one ending in State 3 and the other ending in States 5 and 6. Although Clusters 3, 4, and 11 were all in a transitioning state, they exhibited diverse priorities, with Cluster 11 of mCAF2 predominantly present in State 5, whereas the other two clusters were mostly located in State 3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo investigate the gene dynamics along the FRCs\u0026ndash;mCAF2 trajectory, we identified the pseudotime-dependent genes and categorized them into four modules based on their diverse expression patterns (\u003cstrong\u003eFig. 4E\u003c/strong\u003e). Modules 1, 2, and 3 genes exhibited an increasing expression trend along the trajectory towards mCAF2, whereas Module 4 genes exhibited a decreasing trend. Enrichment analysis (\u003cstrong\u003eFig. 4F\u003c/strong\u003e) revealed that Module 1 was enriched in ECM organization, vasculature development, and positive regulation of cell adhesion. Module 3 genes were associated with cytoskeletal organization and actin filament-based processes. These results suggest enhanced ECM remodeling and the presence of contractile features when FRCs transitioned to mCAF2. Module 2 genes were involved in ossification, cartilage development, skeletal system development, and proteolysis regulation. In contrast, Module 4 genes were associated with the positive regulation of cell migration, hematopoietic or lymphoid organ development, immune system development, and cytokine signaling in the immune system, which were downregulated along the trajectory. Overall, these findings suggest that FRCs acquired more myofibroblastic features but lost their immune regulatory function during the transition to mCAF2.\u003c/p\u003e\n\u003cp\u003eThe dynamic expression trends of marker genes in the four modules (COMP, SFRP2, ACTA2, and CCN3 for Modules 1, 2, 3, and 4, respectively) were aligned using a gene heatmap (\u003cstrong\u003eFig. 4G\u003c/strong\u003e). Notably, the expression of COMP, a marker gene for mCAF2, gradually increased towards State 5 but remained steady towards State 3, mirroring the expression pattern of the contractile marker gene Acta2. This suggests that the expression of the mCAF2 marker gene increased following transition and was associated with ECM remodeling.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn addition to mCAF2, pseudotime analysis of mFCs and mCAF1 revealed two diverging cell fates (\u003cstrong\u003eFig. 4H\u003c/strong\u003e). Starting at Cluster 7, one fate progressed towards Clusters 5 and 12, whereas the other fate moved towards Cluster 2. Cluster 6 acted as a transitional state along the axis (\u003cstrong\u003eFig. 4I\u003c/strong\u003e). Among these five clusters, Cluster 2 was identified as mCAF1, and the remaining clusters were identified as mFC. These findings suggest that the activation of mCAFs or normal myofibroblasts in the OSCC microenvironment could be influenced by diverse mechanisms.\u003c/p\u003e\n\u003cp\u003eWe further explored pseudotime-dependent genes along the furcate trajectory, categorizing them into four modules based on their expression patterns (\u003cstrong\u003eFig. 4J\u003c/strong\u003e). The expression of Module 4 genes was particularly elevated in States 2\u0026ndash;6, which were primarily occupied by mCAF1. Conversely, Module 3 genes were prominently expressed in States 7\u0026ndash;9, which were occupied by mFCs (\u003cstrong\u003eFig. 4I)\u003c/strong\u003e. This indicates that the diverse gene expression signatures of Modules 3 and 4 may play a role in determining the fate of fibroblast activation in OSCC microenvironment. Furthermore, Module 2 genes were highly expressed in the pre-branch state, whereas Module 1 genes showed a relatively stable expression in the transitioning state (\u003cstrong\u003eFig. 4J\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubsequently, we assessed the biological involvements of these pseudotime-dependent genes (\u003cstrong\u003eFig. 4K\u003c/strong\u003e). The gene set from Module 3 was involved in the regulation of wound healing, heart contraction, muscle contraction, and vascular processes in the circulatory system, whereas the gene set from Module 4 was enriched in collagen degradation, collagen formation, and ECM organization, consistent with the expression signatures of the marker genes (\u003cstrong\u003eFig. 4J\u003c/strong\u003e). We observed distinct expression patterns where ECM remodeling-related genes (POSTN, COL10A1, COL11A1, MMP1, MMP3, MMP7, MMP9, and MMP13) were predominantly expressed in mCAF1 states; however, the expression of contractile feature-associated genes (DCN, ACTA2, and AVPR1A) and lipid metabolism-related genes were elevated in mFC states (\u003cstrong\u003eFig. 4L\u003c/strong\u003e). To validate these findings, we further investigated The Cancer Genome Atlas database of head and neck squamous carcinoma. The analysis revealed an upregulation in the expression of ECM remodeling genes in tumors, whereas the levels of muscle contraction-related genes were downregulated (\u003cstrong\u003eFig. 4M\u003c/strong\u003e). These results provide compelling evidence that within the primary microenvironment of OSCC, the acquisition of ECM remodeling ability more accurately represents mCAF formation in the TME than contractile features.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e5. Strong crosstalk occurred between mCAF1 and immune cells in OSCC PT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBased on expression levels and discrimination of differential gene analysis between mCAF1 and mCAF2 \u003cstrong\u003e(Fig. 5A\u003c/strong\u003e), we selected RGS4 as a marker gene for mCAF1 and COMP for mCAF2. Employing immunohistochemical (IHC) chip analysis (\u003cstrong\u003eSupplementary Fig.4\u003c/strong\u003e), we compared the distribution of mCAF1 and mCAF2 between paired PT and MLN samples. Our results revealed that RGS4+ mCAF1 were more abundant in OSCC PT tissues, whereas COMP+ mCAF2 exhibited greater prevalence in OSCC MLNs (\u003cstrong\u003eFig. 5B\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eSubsequently, we conducted a biological analysis of the differentially expressed genes between mCAF1 and mCAF2 (\u003cstrong\u003eFig. 5C)\u003c/strong\u003e. Genes related to muscle contraction and cell\u0026ndash;cell adhesion were more enriched in mCAF2, whereas pathways involved in the regulation of T cell differentiation and macrophage activation were more enriched in mCAF1.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo further compare the immunomodulatory abilities of mCAF1 and mCAF2, we performed receptor\u0026ndash;ligand analysis to explore the interaction between immune cells and mCAF1 or mCAF2. Despite the removal of most CD45+ cells through MACS prior to scRNA-seq, we obtained a considerable number of immune cells, up to 26,666, including B cells, macrophages, mast cells, NK cells, neutrophils, and T cells (\u003cstrong\u003eFig. 1D\u003c/strong\u003e). Owing to their well-documented heterogeneity, T cells were further classified into na\u0026iuml;ve T, Treg, CD8+ T, and NK-T cells based on the expression levels of typical marker genes (\u003cstrong\u003eFig. 5D\u0026nbsp;\u003c/strong\u003e\u0026amp; \u003cstrong\u003eSupplementary Fig. 5\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe receptor\u0026ndash;ligand relationship between immune cells and mCAF1 or mCAF2 was explored using CellPhoneDB analysis\u003csup\u003e20\u003c/sup\u003e. The histogram (\u003cstrong\u003eFig. 5E\u003c/strong\u003e) shows that more receptor\u0026ndash;ligand pairs were identified in PT than in MLN, particularly between T cells and macrophages, indicating stronger crosstalk between mCAF1 and immune cells.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe CXCL12/CXCR4 axis has been widely reported to be involved in the interactions between\u0026nbsp;hematological, solid tumor cells, and stromal cells in several TME\u003csup\u003e31\u003c/sup\u003e. In the present study, we identified high CXCL12 expression in mCAF1 and Cxcr4 expression, its corresponding receptor, in most immune cells (\u003cstrong\u003eFig. 5F\u003c/strong\u003e), indicating the potential recruitment of immune cells by mCAF1. In addition, the expression of TGF\u0026nbsp;-\u0026beta;\u0026nbsp;superfamily gene INHBA by mCAF1 induced the expression of ACVRL1 in na\u0026iuml;ve T cells, indicating potential enhancement of na\u0026iuml;ve T cell activity\u003csup\u003e32\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubsequently, we conducted an IHC analysis to compare the correlation between mCAF1 and mCAF2 and immune cell infiltration in PT or MLN. The results revealed that MLNs with high COMP+ mCAF2 exhibited fewer T cell and macrophage counts than MLNs with relatively low COMP+ mCAF2 (\u003cstrong\u003eFig. 5G\u003c/strong\u003e). In contrast, the infiltration of macrophages and T cells in PT was positively correlated with the number of RGS4+ mCAF1 (\u003cstrong\u003eFig. 5H\u003c/strong\u003e). Further comparison of T cell subtypes revealed that the presence of RGS4+ mCAF1 was specifically correlated with FOXP3+ cells in PT, suggesting that mCAF1 may be more likely to recruit and activate Treg cells to enhance immunosuppression.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6. mCAF2 from MLNs showed a stronger ECM remodeling ability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further demonstrate the high ECM remodeling ability of mCAF2 in OSCC MLN tissues, we conducted IHC analysis using OSCC tissue chips (\u003cstrong\u003eSupplementary\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eFig. 4\u003c/strong\u003e). A relationship was observed between COMP expression and collagen deposition, where COMP expression corresponded with TACS\u003csup\u003e33\u0026nbsp;\u003c/sup\u003e(\u003cstrong\u003eFig. 6A\u0026amp;B\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo corroborate the findings of scRNA-seq and IHC analysis, we obtained two pairs of OSCC primary fibroblasts from the PTs (CAFs), MLNs fibroblasts (MAFs), and non-cancerous DLN fibroblasts (LFs). Immunofluorescence images revealed that CAFs and MAFs exhibited higher expression of typical fibroblast activation-associated genes, including ACTA2, FAP, and FSP1, than LFs (\u003cstrong\u003eFig. 6C\u003c/strong\u003e). Furthermore, functional tests confirmed that CAFs and MAFs demonstrated enhanced proliferation and migration abilities compared to LFs. MAFs exhibited a higher ability for collagen contraction than CAFs and LFs, whereas CAFs exhibited no significant difference in terms of proliferation and migration ability (\u003cstrong\u003eFig. 6D\u0026ndash;G\u003c/strong\u003e). Immunofluorescence and western blotting experiments substantiated that the expression of ECM remodeling-associated proteins was upregulated in MAFs (\u003cstrong\u003eFig. 6H\u0026amp;I\u003c/strong\u003e). This suggests that MAFs exhibited a stronger ECM remodeling ability, independent of activation level.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eCAFs are a central component of the TME in OSCC, and their presence is associated with poorer prognosis\u003csup\u003e9\u003c/sup\u003e.\u0026nbsp;Through a retrospective analysis based on IHC, we confirmed the association between an elevated\u0026nbsp;degree of desmoplasia\u0026nbsp;in OSCC MLNs and an increased risk of ENE, as well as an unfavorable prognosis. This finding suggests that fibroblasts play significant roles\u0026nbsp;in MLN formation, which might be potential therapeutic targets for patients with late-stage OSCC.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrevious studies showed significant CAF heterogeneity with respect to their phenotypes, origins, and functions\u003csup\u003e12, 14\u003c/sup\u003e. However, the detailed nature of this heterogeneity between MLNs and PTs has remained unknown owing to insufficient fibroblast acquisition in previous sc-RNA seq studies\u003csup\u003e14, 34\u003c/sup\u003e. Therefore, we systematically analyzed the characteristics and differences of fibroblasts in the TME among MLNs, DLNs, and paired PTs in OSCC. To improve fibroblast acquisition, we optimized the single-cell suspension preparation and filtered out most CD45+ cells before sequencing. This approach yielded a total of 87,650 cells, allowing the identification of up to 44,052 fibroblasts.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo compare the inter-tissue heterogeneity of the obtained fibroblasts, we initially reclassified them into three major subtypes according to their locations. CAFs were identified as fibroblasts present only in cancerous tissues (PTs and MLNs), LN-SCs were identified as fibroblasts located exclusively in LNs but absent in PTs, and the remaining fibroblasts located in all three tissues were identified as iFCs. Among these three major fibroblast subtypes, we further identified 12\u0026nbsp;subpopulations based on marker gene expression and functional enrichment. Consistent with previous studies\u003csup\u003e35, 36\u003c/sup\u003e, we identified mCAFs and iCAFs as major CAF subpopulations in the TME. The mCAFs subgroup was characterized by the expression of ACTA2 and enrichment in\u0026nbsp;actin cytoskeleton and focal adhesion, whereas iCAFs subgroup expressed PDGFRA but lacked\u0026nbsp;aSMA\u0026nbsp;and were involved in cytokine\u0026ndash;cytokine receptor interactions. In addition, we discovered an emerging CAF subgroup, which was classified as eCAFs. Previous studies have shown that epithelial cells in the heart can generate fibroblasts via EMT\u003csup\u003e37\u003c/sup\u003e. Overexpression of TGF-\u0026beta; has been demonstrated to cause EMT and CAF formation, contributing to the development of a fibrotic TME\u003csup\u003e38, 39\u003c/sup\u003e. In this study, eCAFs expressed relatively low levels of fibroblastic marker genes but expressed certain epithelial marker genes. It also demonstrated active proliferative ability and a potential transforming relationship with the eFC subgroup of iFCs. This finding supports the presence of an EMT state in the TME.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUpon comparing the heterogeneity of CAFs between MLNs and PTs, we identified two mCAF subgroups: mCAF1 (RGS4+) and mCAF2 (COMP+). Through sc-RNA seq analysis, substantiated by IHC analysis, we observed a concentration of COMP+ mCAF2 in MLNs, whereas RGS4+ fibroblasts were more predominant in PTs. In PTs, the recruitment and activation of normal fibroblasts towards mCAFs occur through diverse mechanisms, including TGF-b, Wnt, and mechanotransduction signaling, which are well-documented\u003csup\u003e40\u003c/sup\u003e. However, in MLNs, mCAF origin (whether they originate from PTs mCAFs) and activation mechanism (whether they are activated from inherent stromal cells in the LNs) remain unclear. To delineate the evolutionary trajectories of mCAF1 and mCAF2, we performed pseudotime analysis. The results suggested that mCAF1 may originate\u0026nbsp;from inherent mFCs in PTs, whereas mCAF2 may be derived from FRCs in LNs. Previous studies have demonstrated that proper maturation of FRCs determines LN development and immunity organization, whereas YAP/TAZ hyperactivation of FRCs can lead to enhanced myofibroblast characteristics and fibrotic LN structure\u003csup\u003e41\u003c/sup\u003e. However, the mechanism underlying the dysfunction of FRCs in MLN formation remains unclear.\u003c/p\u003e\n\u003cp\u003eWe conducted differential and enrichment analyses, which allowed for a comprehensive comparison of the heterogeneity of phenotypes and functions between mCAF1 and mCAF2. Notably, even with similar activation levels, mCAF1 from PTs exhibited stronger crosstalk with immune cells than mCAF2. IHC analysis further corroborated these findings, revealing a higher concentration of T cells and macrophages in PT areas rich in RGS4+ mCAF1; however, fewer T cells and macrophages were present in MLN areas abundant in COMP+ mCAF2. This finding indicates that the abundance of COMP+ mCAF2 in MLNs may restrict immune cell activity. In addition, our study demonstrated that mCAF2 in MLNs was related to cancer-associated collagen remodeling and the occurrence of ENE. A fine network of FRCs is essential for regulating immune cell entry and maintaining LN structure and function\u003csup\u003e42\u003c/sup\u003e. However, when FRCs undergo fibroblastic damage due to chronic inflammation and cancer, they compromise immune response in LNs\u003csup\u003e7\u003c/sup\u003e. A previous study suggested that\u0026nbsp;hyperactivation of YAP/TAZ in FRCs severely impairs their maturation and leads to non-functional and fibrotic LNs\u003csup\u003e41\u003c/sup\u003e. Our study demonstrated that mCAFs in OSCC MLNs may originate from inherent FRCs. During this transformative phase, there was a diminished ability for immune cell recruitment; however, ECM remodeling function was enhanced. Previous studies have shown that\u0026nbsp;CAFs play a central role in the deposition, modification, and degradation of ECM components and that dysregulated ECM remodeling by CAFs can lead to desmoplastic reactions associated with poor outcomes in breast, pancreatic, and lung cancers\u003csup\u003e43\u003c/sup\u003e. Our in vitro study and IHC analysis provided further evidence for the stronger ECM remodeling ability of primary CAFs from OSSC MLNs. Moreover, the presence of desmoplastic fibroblasts and ECM remodeling in MLN significantly influenced ENE formation, which may cause OSCC mortality. This finding suggests potential roles of improperly activated FRCs in MLN formation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn conclusion, our study demonstrated the presence of significant heterogeneity in OSCC stromal cells between the PTs and MLNs. Specifically, as a major subtype of fibroblasts, mCAFs in PT or MLN differed in terms of phenotype, origin, and function. Originating from inherent FRCs, COMP+ mCAFs in OSCC MLNs exhibited a more potent collagen remodeling ability but weakened immune cell recruitment activity, which may contribute to ENE formation and portend poor outcomes. This provides new insight into potential treatment strategy that targets specific CAF subgroups in MLNs of advanced OSCC. However, further research is required to elucidate the mechanisms underlying mCAF activation in OSCC MLNs and its association with ENE formation.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003c/strong\u003eThis work was supported by the National Natural Science Foundation of China (No. 82103516) and \u0026ldquo;3456\u0026rdquo; Cultivation Program For Junior Talents of Nanjing Stomatological School, Medical School of Nanjing Univeristy(NO. 0222R207, 0222C108), the Jiangsu Province Science \u0026amp; Technology Department (BE2021609), Key Project supported by Medical Science and technology development Foundation，Nanjing Department of Health(ZKX23056). We also thank Xuan Zhou from OE Biotech Co., Ltd (Shanghai, China) to help us with the 10x sing-cell RNA sequencing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSupplementary Information is available for this paper. The authors declare that there was no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021.\u003c/li\u003e\n\u003cli\u003eZanoni DK, Montero PH, Migliacci JC, Shah JP, Wong RJ, Ganly I, et al. Survival outcomes after treatment of cancer of the oral cavity (1985-2015). Oral Oncol. 2019;90:115-21.\u003c/li\u003e\n\u003cli\u003eKalluri R. The biology and function of fibroblasts in cancer. Nat Rev Cancer. 2016;16(9):582-98.\u003c/li\u003e\n\u003cli\u003eLouault K, Li RR, DeClerck YA. Cancer-Associated Fibroblasts: Understanding Their Heterogeneity. Cancers (Basel). 2020;12(11).\u003c/li\u003e\n\u003cli\u003eCho JK, Hyun SH, Choi N, Kim MJ, Padera TP, Choi JY, et al. Significance of lymph node metastasis in cancer dissemination of head and neck cancer. Transl Oncol. 2015;8(2):119-25.\u003c/li\u003e\n\u003cli\u003eMatsumoto F, Mori T, Matsumura S, Matsumoto Y, Fukasawa M, Teshima M, et al. Prognostic significance of surgical extranodal extension in head and neck squamous cell carcinoma patients. Jpn J Clin Oncol. 2017;47(8):699-704.\u003c/li\u003e\n\u003cli\u003eRiedel A, Shorthouse D, Haas L, Hall BA, Shields J. Tumor-induced stromal reprogramming drives lymph node transformation. Nat Immunol. 2016;17(9):1118-27.\u003c/li\u003e\n\u003cli\u003eNakayama H, Ohuchida K, Yoshida M, Miyazaki T, Takesue S, Abe T, et al. Degree of desmoplasia in metastatic lymph node lesions is associated with lesion size and poor prognosis in pancreatic cancer patients. Oncol Lett. 2017;14(3):3141-7.\u003c/li\u003e\n\u003cli\u003eLavie D, Ben-Shmuel A, Erez N, Scherz-Shouval R. Cancer-associated fibroblasts in the single-cell era. Nat Cancer. 2022;3(7):793-807.\u003c/li\u003e\n\u003cli\u003eZhao X, Ding L, Lu Z, Huang X, Jing Y, Yang Y, et al. Diminished CD68(+) Cancer-Associated Fibroblast Subset Induces Regulatory T-Cell (Treg) Infiltration and Predicts Poor Prognosis of Oral Squamous Cell Carcinoma Patients. Am J Pathol. 2020;190(4):886-99.\u003c/li\u003e\n\u003cli\u003eDing L, Fu Y, Zhu N, Zhao M, Ding Z, Zhang X, et al. OXTR(High) stroma fibroblasts control the invasion pattern of oral squamous cell carcinoma via ERK5 signaling. Nat Commun. 2022;13(1):5124.\u003c/li\u003e\n\u003cli\u003eDing L, Ren J, Zhang D, Li Y, Huang X, Hu Q, et al. A novel stromal lncRNA signature reprograms fibroblasts to promote the growth of oral squamous cell carcinoma via LncRNA-CAF/interleukin-33. Carcinogenesis. 2018;39(3):397-406.\u003c/li\u003e\n\u003cli\u003eCillo AR, K\u0026uuml;rten CHL, Tabib T, Qi Z, Onkar S, Wang T, et al. Immune Landscape of Viral- and Carcinogen-Driven Head and Neck Cancer. Immunity. 2020;52(1):183-99.e9.\u003c/li\u003e\n\u003cli\u003ePuram SV, Tirosh I, Parikh AS, Patel AP, Yizhak K, Gillespie S, et al. Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer. Cell. 2017;171(7):1611-24.e24.\u003c/li\u003e\n\u003cli\u003eButler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol. 2018;36(5):411-20.\u003c/li\u003e\n\u003cli\u003eAran D, Looney AP, Liu L, Wu E, Fong V, Hsu A, et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat Immunol. 2019;20(2):163-72.\u003c/li\u003e\n\u003cli\u003eH\u0026auml;nzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7.\u003c/li\u003e\n\u003cli\u003eRitchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47.\u003c/li\u003e\n\u003cli\u003eQiu X, Mao Q, Tang Y, Wang L, Chawla R, Pliner HA, et al. Reversed graph embedding resolves complex single-cell trajectories. Nat Methods. 2017;14(10):979-82.\u003c/li\u003e\n\u003cli\u003eVento-Tormo R, Efremova M, Botting RA, Turco MY, Vento-Tormo M, Meyer KB, et al. Single-cell reconstruction of the early maternal-fetal interface in humans. Nature. 2018;563(7731):347-53.\u003c/li\u003e\n\u003cli\u003eChen Z, Zhou L, Liu L, Hou Y, Xiong M, Yang Y, et al. Single-cell RNA sequencing highlights the role of inflammatory cancer-associated fibroblasts in bladder urothelial carcinoma. Nat Commun. 2020;11(1):5077.\u003c/li\u003e\n\u003cli\u003eZhang M, Yang H, Wan L, Wang Z, Wang H, Ge C, et al. Single-cell transcriptomic architecture and intercellular crosstalk of human intrahepatic cholangiocarcinoma. J Hepatol. 2020;73(5):1118-30.\u003c/li\u003e\n\u003cli\u003eCai W, Sun X, Jin F, Xiao D, Li H, Sun H, et al. PERK-eIF2\u0026alpha;-ERK1/2 axis drives mesenchymal-endothelial transition of cancer-associated fibroblasts in pancreatic cancer. Cancer Lett. 2021;515:86-95.\u003c/li\u003e\n\u003cli\u003eBartoschek M, Oskolkov N, Bocci M, L\u0026ouml;vrot J, Larsson C, Sommarin M, et al. Spatially and functionally distinct subclasses of breast cancer-associated fibroblasts revealed by single cell RNA sequencing. Nat Commun. 2018;9(1):5150.\u003c/li\u003e\n\u003cli\u003eRodda LB, Lu E, Bennett ML, Sokol CL, Wang X, Luther SA, et al. Single-Cell RNA Sequencing of Lymph Node Stromal Cells Reveals Niche-Associated Heterogeneity. Immunity. 2018;48(5):1014-28.e6.\u003c/li\u003e\n\u003cli\u003eMalhotra D, Fletcher AL, Astarita J, Lukacs-Kornek V, Tayalia P, Gonzalez SF, et al. Transcriptional profiling of stroma from inflamed and resting lymph nodes defines immunological hallmarks. Nat Immunol. 2012;13(5):499-510.\u003c/li\u003e\n\u003cli\u003eCyster JG, Ansel KM, Reif K, Ekland EH, Hyman PL, Tang HL, et al. Follicular stromal cells and lymphocyte homing to follicles. Immunol Rev. 2000;176:181-93.\u003c/li\u003e\n\u003cli\u003eTrapnell C, Cacchiarelli D, Grimsby J, Pokharel P, Li S, Morse M, et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat Biotechnol. 2014;32(4):381-6.\u003c/li\u003e\n\u003cli\u003eAibar S, Gonz\u0026aacute;lez-Blas CB, Moerman T, Huynh-Thu VA, Imrichova H, Hulselmans G, et al. SCENIC: single-cell regulatory network inference and clustering. Nat Methods. 2017;14(11):1083-6.\u003c/li\u003e\n\u003cli\u003eDotto GP, Rustgi AK. Squamous Cell Cancers: A Unified Perspective on Biology and Genetics. Cancer Cell. 2016;29(5):622-37.\u003c/li\u003e\n\u003cli\u003eDomanska UM, Kruizinga RC, Nagengast WB, Timmer-Bosscha H, Huls G, de Vries EG, et al. A review on CXCR4/CXCL12 axis in oncology: no place to hide. Eur J Cancer. 2013;49(1):219-30.\u003c/li\u003e\n\u003cli\u003eQi J, Sun H, Zhang Y, Wang Z, Xun Z, Li Z, et al. Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat Commun. 2022;13(1):1742.\u003c/li\u003e\n\u003cli\u003eBrett EA, Sauter MA, Machens HG, Duscher D. Tumor-associated collagen signatures: pushing tumor boundaries. Cancer Metab. 2020;8:14.\u003c/li\u003e\n\u003cli\u003eLambrechts D, Wauters E, Boeckx B, Aibar S, Nittner D, Burton O, et al. Phenotype molding of stromal cells in the lung tumor microenvironment. Nat Med. 2018;24(8):1277-89.\u003c/li\u003e\n\u003cli\u003eChallaSivaKanaka S, Vickman RE, Kakarla M, Hayward SW, Franco OE. Fibroblast heterogeneity in prostate carcinogenesis. Cancer Lett. 2022;525:76-83.\u003c/li\u003e\n\u003cli\u003e\u0026Ouml;hlund D, Handly-Santana A, Biffi G, Elyada E, Almeida AS, Ponz-Sarvise M, et al. Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer. J Exp Med. 2017;214(3):579-96.\u003c/li\u003e\n\u003cli\u003eGittenberger-de Groot AC, Vrancken Peeters MP, Mentink MM, Gourdie RG, Poelmann RE. Epicardium-derived cells contribute a novel population to the myocardial wall and the atrioventricular cushions. Circ Res. 1998;82(10):1043-52.\u003c/li\u003e\n\u003cli\u003ePeng D, Fu M, Wang M, Wei Y, Wei X. Targeting TGF-\u0026beta; signal transduction for fibrosis and cancer therapy. Mol Cancer. 2022;21(1):104.\u003c/li\u003e\n\u003cli\u003eSu J, Morgani SM, David CJ, Wang Q, Er EE, Huang YH, et al. TGF-\u0026beta; orchestrates fibrogenic and developmental EMTs via the RAS effector RREB1. Nature. 2020;577(7791):566-71.\u003c/li\u003e\n\u003cli\u003eKuzet SE, Gaggioli C. Fibroblast activation in cancer: when seed fertilizes soil. Cell Tissue Res. 2016;365(3):607-19.\u003c/li\u003e\n\u003cli\u003eChoi SY, Bae H, Jeong SH, Park I, Cho H, Hong SP, et al. YAP/TAZ direct commitment and maturation of lymph node fibroblastic reticular cells. Nat Commun. 2020;11(1):519.\u003c/li\u003e\n\u003cli\u003eFletcher AL, Acton SE, Knoblich K. Lymph node fibroblastic reticular cells in health and disease. Nat Rev Immunol. 2015;15(6):350-61.\u003c/li\u003e\n\u003cli\u003eSahai E, Astsaturov I, Cukierman E, DeNardo DG, Egeblad M, Evans RM, et al. A framework for advancing our understanding of cancer-associated fibroblasts. Nat Rev Cancer. 2020;20(3):174-86.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-3862426/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3862426/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Desmoplasia in fibroblasts within metastatic lymph nodes (MLNs) serves as an indicator of extranodal extension (ENE), which is a critical determinant of distant metastasis and mortality in oral squamous cell carcinoma (OSCC). However, systematic studies on fibroblasts in MLNs are lacking. Therefore, this study investigated and characterized the differences in phenotype, function, and origin of fibroblasts between primary tumors (PTs) and lymph nodes (LNs) in OSCC. We generated single-cell maps of PTs and paired MLNs and draining LNs from three OSCC patients. Among 44,052 fibroblasts, we identified two distinct subpopulations of cancer-associated myofibroblastic cells (mCAFs): RGS4+ mCAF1 and COMP+ mCAF2. Notably, mCAF1 and mCAF2 exhibited distinct distributions, with mCAF1 predominantly localized in the PT and mCAF2 in the MLN. Moreover, pseudotime analysis revealed their distinct origins: mCAF1 originated from inherent normal myofibroblastic cells in the PT, whereas mCAF2 originated from fibroblastic reticular cells in the LNs. Further differential analysis and in-vitro functional experiments using primary fibroblasts revealed that, compared to mCAF1, mCAF2 in MLN exhibited weaker crosstalk with immune cells but displayed enhanced extracellular matrix activity, which is closely linked to ENE formation in OSCC. Additionally, we identified two fibroblast subgroups in a transforming state, characterized by relatively low expression of fibroblastic marker genes but expressed specific epithelial canonical markers, indicating a potential epithelial–mesenchymal transition. Our research offers profound insights into the heterogeneity of fibroblasts between the PT and MLN in OSCC, serving as an essential resource for future drug discovery endeavors.","manuscriptTitle":"Single-cell RNA sequencing of OSCC primary tumors and lymph nodes reveals distinct origin and phenotype of fibroblasts","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-30 17:27:07","doi":"10.21203/rs.3.rs-3862426/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"042836b4-7624-4370-a1cd-a994dcc27e8e","owner":[],"postedDate":"January 30th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":28419508,"name":"Biological sciences/Cancer/Head and neck cancer/Oral cancer"},{"id":28419509,"name":"Health sciences/Oncology/Cancer/Head and neck cancer"}],"tags":[],"updatedAt":"2024-01-30T17:27:07+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-30 17:27:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3862426","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3862426","identity":"rs-3862426","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}