Single cell and spatial analysis of immune-hot and immune-cold tumours identifies fibroblast subtypes associated with distinct immunological niches and positive immunotherapy response

Single cell and spatial analysis of immune-hot and immune-cold tumours identifies fibroblast subtypes associated with distinct immunological niches and positive immunotherapy response | 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 Research Article Single cell and spatial analysis of immune-hot and immune-cold tumours identifies fibroblast subtypes associated with distinct immunological niches and positive immunotherapy response Benjamin H Jenkins, Ian Tracy, Maria Fernanda SD Rodrigues, Melanie JL Smith, and 13 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5125055/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Jan, 2025 Read the published version in Molecular Cancer → Version 1 posted 11 You are reading this latest preprint version Abstract Cancer-associated Fibroblasts (CAFs) have emerged as critical regulators of anti-tumour immunity, with both beneficial and detrimental properties that remain poorly characterised. To investigate this, we performed single-cell and spatial transcriptomic analysis, comparing immune-hot and immune-cold HNSCC subgroups (human papillomavirus [HPV] + ve and HPV-ve tumours respectively). This identified six fibroblast subpopulations, including two with immunomodulatory gene expression profiles ( IL-11 + inflammatory [i]CAF and fibroblastic reticular cell [FRC]-like). IL-11 + iCAF were spatially associated with inflammatory monocytes and regulated in vitro through synergistic activation of canonical NF-κB signalling by IL-1β and TNF-α. FRC-like were enriched in HPV + ve tumours, associated with CD4 T-cells and B-cells in tertiary lymphoid structures and regulated through non-canonical NF-κB signalling via lymphotoxin. Pan-cancer analysis revealed several 'iCAF’ subgroups present in both normal and cancer tissues; IL11 + iCAF were found in cancers from the gastrointestinal tract and transcriptomically distinct from iCAFs previously described in pancreatic and breast cancers with greater inflammatory properties; FRC-like fibroblasts, a rare phenotype but present in all tumour types, were associated with significantly better survival in patients receiving checkpoint immunotherapy. This work clarifies and expands current literature on immunomodulatory CAFs, highlighting links with important immunological niches. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Fibroblasts are ubiquitous cells that assume specialised phenotypes and activation states to play a multifaceted role in health and disease [1]. Although historically regarded as structural cells that principally remodel extracellular matrix (ECM) they are now recognised as key immune sentinel cells capable of initiating, maintaining, and suppressing immune responses in response to pathological stimuli [2]. Cancer-associated fibroblasts (CAF) research has mostly focused on cells with a myofibroblast phenotype (myCAF); these are found in most cancer types, have numerous tumour-promoting functions and are akin to myofibroblasts in fibrotic diseases [1,3]. However, recent single cell studies have identified transcriptomically distinct CAF subtypes, including inflammatory CAF (iCAF), antigen presenting CAF (apCAF) and metabolic CAF (meCAF) [4–7]. These phenotypes form part of a plastic population that can change states in response to local stimuli; Biffi and colleagues elegantly demonstrated this, switching pancreatic stellate cells between myCAF and iCAF states in vitro by manipulating TGFβ and IL-1 signalling respectively [4]. This plasticity emphasises the role of fibroblasts as key early response cells in tissues. Although iCAF have been identified in several cancer types, including pancreatic ductal adenocarcinoma (PDAC), breast and lung cancers [5–7], it is not yet clear whether this phenotype is common to all tumours or whether all ‘iCAF’ are the same. Although there are common mechanisms that fibroblasts use to regulate tissue inflammation, inflammatory stimuli have been shown to produce organ-specific immune signatures in fibroblasts from different organs [8] suggesting that inflammatory CAF phenotypes could vary. The clinical success of immune checkpoint inhibitors in treating multiple cancer types is well established. However, only a subset of patients respond favourably [9], and this has generated significant interest in understanding how the tumour microenvironment suppresses anti-tumour immunity. myCAF have several immunosuppressive functions and myCAF-rich tumours are resistant to immunotherapy [6,10,11]. iCAF also express several cytokines associated with immune evasion, including IL6, LIF and CXCL12 [12]. Conversely, in autoimmunity, fibroblasts have been shown to amplify chronic inflammation [13], and a novel population of ‘interferon licenced fibroblasts’ that enhance immunotherapy response has been identified in murine tumour models [14]. Thus, a fibroblast may support or suppress immunity depending on context. Given their plasticity, the concept of generating an immune-supportive phenotype to improve immunotherapy response in cancer is intriguing. In most cancer types, tumours with high levels of tumour-infiltrating lymphocytes (TIL) have better prognosis and show improved response to checkpoint immunotherapy [15]. Head and neck cancer (HNSCC) is subdivided into human papillomavirus (HPV)-related (HPV + ve) tumours and those typically associated with smoking/alcohol (~ 30% and 70% of cases respectively). Around 85% of HPV + ve HNSCC are heavily infiltrated by T- and B-cells and, despite presenting mostly at late stage, are associated with significantly better survival compared with TIL-low HPV-ve tumours [16]. In this study we hypothesised that comparative analysis of ‘immune hot’ (HPV + ve) and ‘immune cold’ (HPV-ve) HNSCC subtypes would identify different inflammatory fibroblast populations that reflect the multifaceted role of fibroblasts in immunity; specifically, that HPV + ve tumours contain a fibroblast phenotype that likely supports anti-tumour immunity. Using single cell and spatial transcriptomics we analysed HPV-positive and HPV-negative HNSCC, identifying six fibroblast subsets, including two characterised by expression of immunomodulatory genes ( IL-11 + inflammatory iCAF and fibroblastic reticular cell [FRC]-like). These fibroblast subsets occupied distinct immunological niches; IL-11 + iCAF were spatially associated with and activated by inflammatory monocytes through canonical NF-κB signalling regulated by IL-1β and TNF-α. FRC-like were enriched in HPV + ve tumours, associated with CD4 T-cells and B-cells in tertiary lymphoid structures, and were regulated through non-canonical NF-κB signalling via lymphotoxin. Pan-cancer analysis showed that ‘iCAF’ is not a single phenotype; IL11 + iCAF, present in HNSCC and other gastrointestinal (GI) tract cancers, are transcriptomically distinct from iCAFs previously described in pancreatic and breast cancers and have heightened inflammatory features; FRC-like represented a rare phenotype, but present in all tumour types and associated with positive response to checkpoint immunotherapy. Methods Human Subjects Ethical approval for the study was obtained through the UK National Research Ethics Service (REC No. 09/H0501/90) and written informed consent was obtained from all subjects. Tumour and matched-normal tissue were obtained from patients undergoing surgical tumour resection at Poole Hospital (Poole, Dorset, UK) for HNSCC. Tissue samples were transported (within 1 hour) to the laboratory on ice in serum-free Dulbecco’s Modified Eagle Medium (DMEM; Sigma-Aldrich). Sample information (including clinical and sample digestions) is shown in Supplementary Table 1. Patient HPV status was confirmed using p16 immunostaining in combination with assessing HPV-encoded gene expression using a human-HPV hybrid reference genome to align and map reads (see details below). HPV-encoded exons were detected in 6 patients using the human-HPV-16 reference genome and 1 patient using the human-HPV-33 reference genome (Supplementary Figure 1A). Primary fibroblast culture and in vitro experiments Please see Supplementary Materials. Sample processing Please see Supplementary Materials. scRNA-Seq For each sample, 5000 single cells were captured on an Illumina 10X Chromium Controller TM system using the Illumina single cell 3’ gene expression and library preparation kits (V3.1 #1000269). Sample capture, sample indexing, and library preparation were carried out according to manufacturer’s instructions. Size distribution, quality control, and quantification of the libraries was assessed using High Sensitivity DNA chips (Agilent Technologies #5067-4626) and KAPA library quantification qPCR kit (Roche #07960140001). Prepared libraries were pooled and sent to Oxford Genomics (UK) for 150-base pair, paired-end sequencing on a Novaseq6000 TM . Sequence alignment and annotation FASTQ files were aligned to the Human reference genome (GRCh38–2020-A) which had the HPV genome concatenated to both the FASTA and .GTF reference files (using cellranger count v6.1.1, 10x Genomics). Human-HPV references were made using the cellranger mkref command (cellranger V6.1.1). scRNA-Seq data was processed with cellranger count (cellranger v6.1.1) generating feature-barcode matrices in which subsequent data analysis was carried out in R (v4.1.1) using Seurat package (v4.1.0). Quality control, normalisation and integration Each patient expression matrix was initially created into a Seurat object with cells requiring expression of at least 200 genes and genes expressed in at least 3 cells. Poor quality cells were removed using a mitochondrial RNA percentage threshold calculated by the median + 3* median absolute deviation [cells above this threshold (~20%) were removed]. Cells expressing >6000 features were removed to reduce potential doublets. Seurat was then used for normalisation and reciprocal PCA (RPCA) integration of scRNA-Seq data (Further details in Supplementary Materials). Principal component analysis (PCA) was then performed on the integrated object followed by Uniform Manifold Approximation and Projection (UMAP) visualisation. Clustering was performed using shared nearest-neighbour (SNN) graph construction (FindNeighbors) followed by FindClusters. Identifying marker genes Differentially expressed genes (DEGs) were identified for each cluster using FindAllMarkers (Wilcoxon rank sum test) with genes selected expressed in ≥25% of cells and log2FC ≥0.5 (adjusted p value <0.05). DEGs were compared to known cell type markers described widely in the literature to annotate broad and finer cell types. HNSCC inter-dataset integration Seurat’s RPCA integration was also used for HNSCC inter-dataset integration with GSE164690 (Further details in Supplementary Materials). Gene module scores and pathway analysis Module scores were calculated using Seurat’s AddModuleScore function calculated by taking the average expression levels of each cluster at the single cell level subtracted by aggregated expression of control feature sets. All gene signatures used in the analysis are shown in Supplementary Table 8. Pathway analysis was performed using over-representation analysis (enrichr v3.2; [48]) and gene set enrichment analysis (GSEA) (clusterprofiler v4.6.2; [49]) using KEGG and MSigDB Hallmark databases. Enriched pathways with P value and adjust p value <0.05 were examined. PROGENy: Pathway RespOnsive GENes for activity inference was used to infer activities of 14 pathways [50]. PROGENy pathway activity scores were calculated for the Seurat object running ‘progeny’ command (organism="Human", top=500, perm=1, return_assay = TRUE). Pathway activity scores were then scaled. Summarised scores (mean) of each activity for each cell cluster were determined and plotted in a heatmap. Differential abundance We utilised MiloR (v1.6.0) to identify differentially abundant phenotypes using KNN graphs [51]. MiloR was run on the integrated objects separately with the following parameters used for buildGraph and makeNhoods: k= 70, d=20, refined = TRUE. To account for cell type abundance differences resulting from use of different digestions, the digest was specified in the design formula along with source (tumour/normal) or HPV status. SpatialFDR threshold (alpha) was set to 0.05 when highlighting differentially abundant neighbourhoods – which were displayed in bee-swarm plots. When calculating relative cell type proportions per sample in the scRNA-Seq data we accounted for digestion differences in select samples that had undergone scRNA-Seq of liberase and col+ digests separately, calculating pseudo-mixtures by combining the relative abundance of cell types for each sample in a 1:9 (liberase:col+) ratio. Trajectory and pseudotime Monocle 3 (v1.0.0) and slingshot (v2.6.0) were used for trajectory analysis and pseudotime calculations. Both methods yielded the same lineages. Monocle 3 was used to determine genes that change as a function of pseudotime, graph_test, specifying the neighbour_graph as ‘prinicpal_graph’ was run. Transcription Factor Analysis DoRothEA regulons, a collection of transcription factors and their targets, were used to infer transcription factor activities in fibroblast populations [20]. Activities were determined using run_wmean from the decoupleR (v2.5.0) package and subsequently scaled. Pan-Cancer Fibroblast Atlas (PCFA) and label transfer Please see Supplementary Materials. Spatial Transcriptomics All pre-sequencing procedures were carried out following the manufacturer's instructions on 6.5mm capture areas using the Visium V2 CytAssist workflow. All samples were processed through the Spaceranger pipeline (v2.0.0) according to 10x Genomics guidelines. Please see Supplementary Materials for details of spatial transcriptomics processing, spot deconvolution and spatially guided ligand-receptor/NicheNet analysis. Bulk RNA sequencing (scRNA-Seq) Counts for the Head and Neck Squamous Cell Carcinoma, TCGA-HNSC (https://www.cancer.gov/tcga) (566 samples: 520 primary solid tumour; 46 solid tissue normal) cohort (Illumina HiSeq platform) were downloaded using TCGAbiolinks and converted to CPM using edgeR. TPM normalised data for TCGA-HNSC was downloaded from GDC data portal. The cBioPortal for cancer genomics [52] was used to obtain additional metadata from the Head and Neck Squamous Cell Carcinoma (TCGA, Firehose Legacy) study. UCSCXenaTools package (v1.4.8) was used to download Batch effects normalized mRNA data (n=11,060) from the Pan-Cancer Atlas Hub and corresponding clinical metadata. Bulk RNA-Seq Deconvolution To investigate the cell type abundance in bulk RNA-Seq data, the immunedeconv R package (v2.1.0) was used to run MCP-counter [53] on TPM normalised HNSCC (TCGA) Bulk RNA-Seq. The ‘deconvolute’ function was run specifying ‘MCP_counter’. ssGSEA We used single sample GSEA (ssGSEA) using the GSVA package (v1.46.0) to calculate enrichment scores for bulk RNA-Seq samples and gene sets. Analysis of Immunotherapy Data The following bulk RNA-Seq datasets were used for analysis of immunotherapy treated patients. HNSCC (GSE159067): 102 patients with advanced HNSCC treated with immunotherapy targeting PD-1/PD-L1. Lung (GSE161537): 82 patients with advanced non-small cell lung cancer (NSCLC) treated with second-line immunotherapy targeting PD-1/PD-L1. Metastatic melanoma (PRJEB23709): 91 patients treated with anti-PD-1 alone or combined anti-PD-1 and anti-CTLA-4 immunotherapy. Overall survival analysis (Kaplan-Meier and cox regression) was performed out using survival package (v3.5-7); patients were split into high and low (based on ssGSEA scores) using the optimal cut points determined by surv_cutpoint (survminer v0.4.9). Multivariate cox regression was carried out using ‘coxph’ function (survival) specifying the fibroblast abundance, sex, and age. Generation of fibroblast subset specific gene signatures (Supplementary Table 8) used in ssGSEA are detailed in Supplementary Materials. Multiplex immunofluorescence using PhenoCycler-Fusion Please see Supplementary Materials. Statistical analysis Statistical analysis was performed using R environment v4.1.1 [ggpubr package (v0.4.0) for plotting graphs] and Graph Pad Prism 9 (v10, GraphPad, San Diego, CA, USA). Wilcoxon rank-sum test (two-sided) or Students t-test (two-sided) were used to evaluate associations between continuous variables. Normality was assessed by Shapiro–Wilk test. One-way ANOVA or Kruskal-Wallis test was used to compare >2 groups. Multiple comparisons were investigated by adjusting the p-value using the Bonferroni method. Correlation analysis was carried out using spearman’s rho (two-sided). A two-sided Fisher’s exact test was used to compare categorical data between groups. Survival analysis was carried out using Kaplan-Meier curves with log-rank test and multivariate cox regression statistics. P < 0.05 was considered to indicate a statistically significant difference. Results HPV+ve HNSCC has an immune hot tumour-immune microenvironment. We performed scRNA-Seq on treatment-naïve HNSCC samples ( n =10; 7 HPV+ve [‘immune-hot']; 3 HPV-ve [‘immune-cold’) with matched normal oropharyngeal mucosa [ n=7 ]; Figure 1A; Supplementary Figure 1B,C; Supplementary Table 1). To increase patient numbers, this (EPG) dataset (82,844 cells after quality control) was integrated with a publicly available dataset (Figure 1A; HNSCC samples of oral cavity/oropharynx; [17]), generating an atlas of 159,826 cells from 24 patients (11 HPV-ve; 13 HPV+ve; 7 normal; Figure 1B; Supplementary Figure 1D,E). Spatial transcriptomic analysis (10x; Visium) and multiplexed immunochemistry (Akoya; PhenoCycler) were performed on tumour sections from the initial ten patients (Figure 1A). Deconvolution of bulk RNA-Seq data from the HNSCC TCGA cohort using MCP-counter confirmed that that HPV+ve tumours contain significantly more CD8+ T-cells (p<0.0001), CD4+ T-cells (p<0.0001) and B-cells (p<0.0001; Supplementary Figure 1F). Spatial transcriptomic analysis using MCP-counter to deconvolute cell type abundance within individual spots, also showed significantly more T-cells (p<0.001), B-cells (p<0.01) and total lymphocytes (p<0.01) in HPV+ve tumours (Figure 1C; Supplementary Figure 1G); also confirmed by multiplexed IHC (MxIHC; Figure 1D). Notably, there was a prominent fibroblast presence in both HPV-ve and HPV+ve tumours (Figure 1C; Supplementary Figure 1G). A more detailed analysis of immune cell subsets (Supplementary Table 2) in the scRNA-Seq data revealed differences in T-cell and NK-cell phenotypes (56,664 cells) between HPV+ve and HPV-ve tumours (Supplementary Figure 2A,B). CD4 + ICOS + PDCD1 + (PD-1) T-cells, resembling T follicular helper (Tfh) cells were more common in HPV+ve tumours, as were CD4+ naïve-like T-cell clusters and KIT+ NK-cells (Supplementary Figure 2B). Analysis of B- and plasma cells (26,156 cells) showed that germinal centre (GC) B-cells ( RGS13 +, NEIL1 +), cycling B-cells ( UBE2C +, TYMS +) and naïve B-cells ( TCL1A +, IL4R +) were all enriched in HPV+ve tumours compared to HPV-ve tumours (Supplementary Figure 2C,D), while switched B-cell subsets were found in both. There were no differentially abundant myeloid populations (Supplementary Figure 2E). scRNA-Seq reveals distinct subsets of inflammatory fibroblasts in HNSCC. To investigate fibroblast phenotypes, we first broadly identified fibroblasts based on lumican expression ( LUM+; 4,894 cells); fibroblasts clustered closely with RGS5+ mural cells (2,174 cells), which included pericytes and smooth muscle cells (SMCs; Supplementary Figure 3A,B). We identified six clusters of fibroblasts; three confined to tumours (CAF) and three present in both tumours and normal tissue (Figure 2A; Supplementary Figure 3CD; Supplementary Table 3). Overall, individual patient tumours showed significant fibroblast heterogeneity, generally containing a mixture of the six phenotypes (Supplementary Figure 3E). The largest CAF cluster expressed canonical myofibroblastic CAF (myCAF) markers ( POSTN, MMP11, ACTA2 ) and showed highest enrichment for TGFb signalling (Figure 2B; Supplementary Figure 4A). This cluster was characterised by high expression of ECM genes (including COL1A1, FN1, COL1A2, COL6A3 and COL11A1 ); with numerous differentially expressed genes (DEGs) associated with core matrisome components [collagens (n=14), glycoproteins ( n=22 ) and proteoglycans ( n=5 )] (Supplementary Figure 4B; Supplementary Table 3; [18]. An inflammatory CAF (iCAF) population was characterised by high expression of inflammatory cytokines (e.g., IL11, IL6, CXCL8, CXCL1, CXCL5 ; Figure 2B). Notably, iCAF also expressed upregulated ECM genes (albeit at a comparatively lower level than myCAF), with higher levels of genes associated with ECM remodelling ( MMP3, MMP1, PLAU) , glycolysis/hypoxia ( HIF1A, ENO1, GK, CA12, SLC16A3 /MCT4) and neutrophil-recruiting chemokines ( CXCL1, CXCL5, CXCL6, CXCL8) (Supplementary Table 3). This cluster was highly enriched for hypoxia, NF-κB, and TNFa signalling pathways (Supplementary Figure 4A). A further CAF cluster expressed lower levels of myCAF/iCAF marker genes; this ‘proto-CAF’ cluster displayed few unique DEGs ( n= 19) compared with other fibroblast phenotypes (which ranged from n=83-232 unique DEGs) and on the UMAP adjoined normal fibroblast and CAF clusters, likely representing a transition state. myCAF and iCAF were present in both HPV-ve and HPV+ve HNSCC (Supplementary Figure 3E), but HPV-ve HNSCC samples contained greater proportions of CAF/fibroblasts relative to total cell number per sample (Supplementary Figure 3D). Within normal mucosa we identified three fibroblast subtypes (also present in tumours.) Universal (adventitial) fibroblasts expressing PI16 were present in normal mucosa, HPV+ve and HPV-ve tumour samples (Supplementary Figure 3E). These expressed CD34 and distinctive ECM-associated genes, including COL14A1, OGN and TNXB , likely reflecting their vascular-niche function (Figure 2B, Supplementary Figure 4B). ADH1B+ fibroblasts were the most common subtype in normal tissue (54% of fibroblasts) but were infrequent in tumours (5% of fibroblasts; Figure 2C; Supplementary Figure 3E). The core matrisome profile of ADH1B+ fibroblasts was similar to universal ( PI16+) fibroblasts (Supplementary Figure 4B). The third fibroblast subgroup expressed CCL19, CCL21, VCAM1, IL7 and SPIB (Figure 2B) with a phenotype akin to fibroblastic reticular cells (FRC), specialised fibroblast subsets of lymphoid tissues that organise and traffic lymphoid cells. Notably, FRC-like fibroblasts were significantly enriched in HPV+ve tumours (Figure 2D). Levels of FRC-like fibroblasts varied between individual HPV+ve tumours but were uniformly rare in all HPV-ve cases (Supplementary Figure 3E). Notably, when present in tumours, normal fibroblast subtypes expressed activation- ( FAP, FN1, PDPN, COL1A1 ), inflammation- ( CXCL1, ISG15 ) and insulin-like growth factor (IGF)-related ( IGF1, IGFBP2, IGFBP4 ) genes (Supplementary Figure 4C) suggesting early activation (with these genes expressed at higher levels in CAF clusters). Phenotypic regulators of iCAF and FRC-like inflammatory fibroblasts. We next inferred fibroblast lineages arising from universal ( PI16+ ) fibroblasts [1]. Trajectory analysis identified three lineages leading to the formation of FRC-like fibroblasts (through ‘ ADH1B+ ’), myCAF andiCAF (both through ‘proto-CAF’; Figure 2E; Supplementary Figure 4D). Signalling pathways regulating FRC-like and iCAF inflammatory subsets were examined by assessing pathway enrichment in genes changing as a function of pseudotime in KEGG and Hallmarks gene sets (q_value 0.25). Pseudotime analysis of the FRC-like lineage (1) showed increased expression of genes associated with NF-κB signalling pathway (KEGG; p.adjust <0.01) and Allograft rejection (Hallmarks; p.adjust <0.0001) (Figure 2F; Supplementary Table 4). While many genes (e.g., CCL19, C7, IRF8 ) showed a pseudotime-dependent increase in expression through the ADH1B+ cluster to FRC-like, other genes enriched for TNF-alpha Signalling via NF-κB (Hallmarks; p.adjust <0.0001) increased in the ADH1B+ cluster but decreased in the FRC-like cluster (e.g., S OCS3, JUN, IRF1, FOS, JUNB ). Gene set enrichment analysis (GSEA) of the MSigDB Hallmarks gene sets revealed significant enrichment for allograft rejection in FRC-like fibroblasts (Supplementary Figure 4E). In theiCAF lineage (2), pseudotime analysis revealed increased expression of genes associated with TNF-alpha signalling via NF-κB (Hallmarks; p.adjust <0.0001), Epithelial Mesenchymal transition (Hallmarks; p.adjust <0.0001), inflammatory response (Hallmarks; p.adjust <0.0001) and JAK-STAT signalling pathway (KEGG; p.adjust <0.01) (Figure 2F; Supplementary Table 4). GSEA showed enrichment for Glycolysis, Hypoxia, inflammatory response and TNF-alpha ignalling via NK-kB (Supplementary Figure 4E). Intriguingly, while the chief inflammatory pathway, NF-κB, was associated with both iCAF and FRC-like inflammatory phenotypes, the corresponding genes differed. FRC-like NF-κB genes (e.g., CCL21, CCL19, TNFSF13B ) are specifically associated with the alternative NK-κB pathway, commonly triggered through lymphotoxin, LIGHT, CD40-L and BAFF, and are related to lymphoid organ development and adaptive immunity [19]. Conversely , iCAF NF-κB genes (e.g., CXCL1, CXCL8 ) are generally associated with the classical NK-κB pathway, typically activated via IL-1, TNF-α or LPS and associated with inflammation and innate immunity [19]. Accordingly, we examined transcription factor (TF) activity in FRC-like and iCAF subsets by assessing the transcriptomic ‘footprint’ of active transcription factors using the DoRothEA database [20]. FRC-like fibroblasts showed strong activity for RELB and NFKB2 (p100/p52), again providing evidence for alternative NF-κB pathway activation through NF-κB RelB-p52 complexes, wherea s top iCAF active transcription factors included RELA (p65), CEBPB and JUN/FOSL1 (AP1; Figure 2G). FRC-like fibroblasts are found within TLS and colocalise with B-cells and CD4+ T-cells. FRC-like fibroblasts and iCAF possessed distinct inflammatory cytokine profiles (Figure 2H); FRC-like cytokines were associated with lymphocyte recruitment, proliferation, and survival [e.g., CCL19/21, IL7/15, TNFSF14 (LIGHT), TNFSF13B (BAFF), CXCL13 ], with iCAF-specific cytokines related to myeloid/ granulocyte recruitment and differentiation ( CXCL1/5/6/8, CSF2/3 ). This, in addition to their different regulatory pathways, suggested that FRC-like fibroblasts andiCAF were likely associated with distinct immunological niches. To investigate this, we first performed correlative sample-level analysis on scRNA-Seq data using the previously identified immune cell subsets in the integrated HNSCC scRNA-Seq dataset (Supplementary Figure 2). We followed this with spatial transcriptomics (Visium 10x) analysis; using the annotated scRNA-Seq data as a reference to derive cell-type specific gene signatures that were used to deconvolute cell types present within each 55µm spot [21,22]. Applying this integrative approach for myCAF and universal ( PI16+ ) fibroblasts, identified previously described spatial and cellular relationships (Extended Data Figure 1 and 2). In sample-level scRNA-Seq correlations of tumours, FRC-like fibroblasts positively correlated with various B-cell subsets [including cycling B-cells, FCRL4 + B-cells and germinal centre (GC) B-cells], plasma cells, KIT+ NK-cells (Figure 3A), with high correlation with IgM expressing B/plasma cells. FRC-like fibroblasts also correlated with CD4+ T follicular helper (Tfh) cells. Spatial transcriptomic analysis confirmed that each tumour contained multiple fibroblast subsets that were spatially discrete (82.4% of fibroblast-containing spots contained one subset only; Supplementary Figure 5). FRC-like fibroblasts colocalised with B-cells and CD4+ T-cells (non-Treg; p<0.0001), found either in focal areas containing high densities of B/CD4+ cells (non-Treg) (HPV+ve/-ve HNSCC) or occasionally more widespread in two HPV+ve samples (59%/23% total spots) but still colocalising with large numbers of B/CD4+ T-cells (Figure 3BC; Supplementary Figure 6A, B; Supplementary Figure 7A). There was also a spatial correlation between FRC-like fibroblasts and plasma cells, Tregs and CD8+ T-cells (Figure 3B; Supplementary Figure 6A, B). PDPN is commonly utilised as a pan-fibroblast marker and has been specifically employed to identify FRC in lymph nodes [23]. MxIHC on spatial transcriptomic-determined FRC-like regions of interest (ROI) confirmed the presence of PDPN+/CD31- fibroblasts colocalising with CD20+ and CD4+ cells (Figure 3D; Supplementary Figure 7B) within and surrounding CD21+ follicular dendritic cells, along with high densities of B-cells, suggesting formation of tertiary lymphoid structures (TLS). We therefore compared TLS and FRC-like enrichment in spatial transcriptomic data using different TLS gene signatures [25]. Spot deconvolution (RCTD) showed that FRC-like fibroblasts significantly correlated with module enrichment scores for various TLS signatures that have been used in several solid cancer studies (Spearman’s r ≥ 0.5, p<0.0001; Figure 3E, Supplementary Figure 7C, D, E; [24,26–28]). FRC-like fibroblasts are regulated via LTβR signalling. We next investigated potential interactions in the FRC-like niche by examining ligands and receptors that were differentially expressed in Visium spots that contained FRC-like fibroblasts (spots containing at least 5% FRC-like cells imputed by RCTD deconvolution; Supplementary Figure 8A; Supplementary Table 5). LTβR-binding ligands [ TNFSF14 (LIGHT), LTB , LTA, CD40LG ] were amongst cytokine activity-possessing ligands that were spatially associated, expressed by highly correlating cell types (Figure 3F; Supplementary Figure 8B) and known to stimulate alternative NF-κB pathway activation (consistent with previous pathway/TF analysis). LIGHT and CD40LG were top ligands inferred via NicheNet [29] (Supplementary Figure 8C). Lymphotoxin was highly expressed in B-cells, T-cells and DCs (Supplementary Figure 8D), whereas LIGHT was expressed by FRC-like fibroblasts. In common with FRC-like fibroblasts, most other fibroblast subsets also expressed receptors for these (and other) inflammatory ligands (Supplementary Figure 8E), highlighting the potential for immunological plasticity in these cells depending on ligand availability. We then assessed the ability of the LTβR-binding ligand lymphotoxin α1β2 (LT) to regulate the FRC-like phenotype in cultured primary oral fibroblasts (NOFs). Treatment with LT in combination with an ALK5 (TGFBR1) inhibitor induced FRC-like-specific genes CCL19 (p<0.0001) , CCL21 (p<0.0001), SPIB (p<0.0001), RBP5 (p<0.01; Figure 3G). This was confirmed in further primary NOF cultures (n=7; CCL19, p<0.01; CCL21 , p<0.05; SPIB , p<0.01; RBP5, p<0.001; Figure 3H). iCAF colocalise with inflammatory monocytes and neutrophils. iCAF correlated strongly with a subset of CD14 + IL1B high inflammatory monocytes (Spearman’s r=0.72, p<0.0001; Figure 4A; Supplementary Figure 9A). Spatial transcriptomic analysis showed that iCAF were spatially distinct from myCAF (Supplementary Figure 5), located primarily at the tumour periphery, particularly towards the tumour surface. iCAF colocalised with monocytes and neutrophils (p<0.0001), also frequently found at the periphery of tumours (Figure 4B, C; Supplementary Figure 6A, B; Supplementary Figure 9C). MxIHC on iCAF ROI (identified through spatial transcriptomics deconvolution) showed that these areas contained PDPN+/CD31- fibroblasts and were associated with disruptions in surface epithelium (pan-cytokeratin) and CD68+, CD14+ and MPO+ myeloid cells (Figure 4D; Supplementary Figure 9C). iCAF are regulated via IL1 b and TNF a . Next, we examined the expression of spatially located ligands in cell types highly correlating with iCAF (Supplementary Table 5). Top ligands associated with cytokine activity in the iCAF-niche included CXCL8, IL1A, IL6, OSM, IL11 and particularly IL1B which was highly expressed by inflammatory monocytes (Figure 4E; Supplementary Figure 9D). We also inferred ligand regulatory activity using NicheNet, which highlighted IL1B and IL1A as top spatially defined ligands with iCAF (gene set) regulatory potential (Supplementary Figure 9E). Indeed, myeloid cells (monocytes, neutrophils, macrophage) expressed highest levels of these ligands ( IL1B, IL1A, OSM ), supportive of iCAF associating with a myeloid niche (Supplementary Figure 8D). We tested the potential of these ligands for regulating theiCAF phenotype in NOFs. In addition to IL-1b, we included TNF-a due to pathway and TF enrichment for classical NF-κB signalling and high expression in monocytes [although TNF was not spatially differentially expressed in theiCAF niche, likely resulting from expression in multiple immune cell types (Supplementary Figure 9D)]. NOFs were treated with IL-1b (1ng/mL) and TNF-a (1ng/mL), either alone or in combination. We also included TGF-b1 (4ng/mL), a central regulator of the myCAF phenotype for reference. Both IL-1b and TNF-a induced expression of IL6 (p<0.0001), MMP3 (p<0.0001) and MME (p<0.01; Figure 4F), but with limited upregulation of IL11 . However, combining IL-1b with TNF-a increased expression of all inflammatory marker genes compared with individual treatments ( IL6 , p<0.0001; MME , p<0.05), including IL11 , which increased 16-fold (log 2 FC = 4) compared to IL1b alone (p<0.05). This was validated in several primary fibroblast cultures ( n=9 ), where TNF-a and IL-1b robustly induced iCAF gene expression (Figure 4G). IL11 was also induced by TGF-b1 (p<0.0001); conversely, ACTA2 (aSMA; a myCAF marker) was induced by TNF-a/IL-1b (p<0.001); while other iCAF ( IL6, MMP3, MME ) and myCAF ( POSTN, TAGLN, COL1A1 ) genes were more specifically regulated by TNF-a/IL-1b and TGF-b1 respectively (Supplementary Figure 9F). Given the iCAF/monocyte spatial relationship, we investigated whether monocytes regulated the iCAF phenotype. NOF treated with conditioned medium from monocytes activated with LPS induced upregulated expression of iCAF genes ( IL6 , p<0.0001; MMP3 , p<0.0001, IL11 , p<0.001; MME , p<0.001; Figure 4H). Similar to TNF-a/IL-1b treatment, ACTA2 was also increased (p<0.05; Supplementary Figure 9G). Pan-Cancer Fibroblast analysis identifies conserved and semi-conserved inflammatory fibroblast phenotypes. To compare the HNSCC fibroblast phenotypes with other cancers, we generated a scRNA-Seq (10x Chromium) pan-cancer fibroblast atlas (PCFA) from seven cancer types: HNSCC, pancreatic, breast, lung, colon, oesophageal and gastric cancers (Figure 5A). Only datasets containing both tumour and normal samples were included to differentiate between normal (steady-state) and cancer-associated phenotypes (Figure 5B, C, D; Supplementary Figure 10A, B). The PCFA (86,414 fibroblasts; 376 samples), revealed 16 populations, including broadly conserved, as well as tissue-specific subsets. Where possible, fibroblast subgroups were labelled using designations from previous studies (Figure 5C; Supplementary Figure 10A-D; Supplementary Table 6). Conserved phenotypes in normal tissue included universal ( PI16+ ) Fib, stress-response Fib ( DNAJB1+, HSPH1+, HSPA1A+; which incorporated the head & neck ADH1B+ fibroblasts) and CXCL14+ CFD+ Fib (Supplementary Figure 10B). Tissue-specific subsets included CXCL8 + breast fibroblasts [most abundant subset (>85%) in normal breast tissue], NPNT + (alveolar) lung fibroblasts and F3 +/ ADAMDEC1 + colonic/gastric fibroblasts (all found in normal and tumour samples; Figure 5D, E; Supplementary Figure 10B). The FRC-like cluster was mostly composed of cells from the head & neck with a contribution from lung (from normal tissues and cancers; Figure 5D, E). Only subsets found exclusively in cancers were termed CAF (Figure 5D). Conserved common CAF clusters included myCAF (the most abundant CAF subset in all tumour types except gastric cancer; Supplementary Figure 10A, C), IGF1+ CAF and proto-CAF. Other clusters had increased frequency in certain cancer types or were rare across cancers. For example, three clusters had a larger contribution from pancreatic tumour/normal samples ( HAS1+ CAF, metabolic CAF [meCAF] and C7+ Fib [Figure 5E; Supplementary Figure 10A, B]). meCAF expressed markers of glycolysis and hypoxia ( ENO1, ENO2, NDRG1, PGK1, LDHA, SLC2A1 ) consistent with a previously described population [30]. Although this cluster was observed pan-cancer, outside of pancreatic samples it was generally found at very low (Supplementary Figure 10A, C). The HNSCC iCAF resided in the IL11+ CAF cluster (Supplementary Figure 10D); this semi-conserved phenotype was one of the most abundant fibroblast subpopulations in HNSCC, colorectal carcinoma (CRC) and oesophageal squamous cell carcinoma (ESCC) (Supplementary Figure 10C) but was not present in lung or breast cancers. Label transfer using the HNSCC myeloid cells as a reference to identify myeloid phenotypes in CRC/ESCC scRNA-Seq datasets showed that IL11+ CAF similarly correlated with IL1B+ inflammatory monocytes suggesting the same immunological niche present in different cancer types (Supplementary Figure 10E). ‘iCAF’ gene signature highlights different normal fibroblast and CAF populations Given the restriction of IL11+ CAF to certain cancers, we performed enrichment analysis using a previously described PDAC iCAF gene signature [5] to investigate whether other PCFA subgroups expressed iCAF markers. This highlighted several fibroblast clusters from both normal and tumour tissues; in normal tissues these included Stress-response Fib, CXCL8+ Breast Fib and universal ( PI16+ ) fibroblasts; in tumours, IL11+ CAF, IGF1+ CAF and proto-CAF (Figure 6A). We examined the iCAF signature-enriched clusters in tumours in more detail (Supplementary Figure 10A). IGF1+ CAF expressed all iCAF markers (Supplementary Figure 10F; [5], and had the greatest transcriptional resemblance to PDAC iCAF (p<0.0001, Fisher’s exact test; Supplementary Figure 10G). IGF1+ CAF were present in most cancer types (e.g., breast, oesophageal, lung), including high levels in PDAC (comparable to myCAF in abundance; Figure 6B; Supplementary Figure 10C). Transcriptionally, IGF1+ CAF differed considerably from IL11+ CAF (Figure 6C); clustering close to universal ( PI16+ ) Fib and maintaining expression of universal ( PI16+ ) Fib genes ( PI16, CFD, COL14A1 ; Figure 6C). From the initial HNSCC analysis, ~50% of (universal) PI16+ labelled fibroblasts from tumour samples were labelled as IGF1+ CAF in the larger PCFA dataset, showing that universal ( PI16+ ) fibroblasts from tumours show evidence of activation and inflammatory changes (including expression of FAP, COL1A1, IGF1 ; Supplementary Figure 4C; Supplementary Figure 10D). Given the transcriptional similarity between universal ( PI16+ ) fibroblasts and IGF1+ CAF, we examined DEGs between these phenotypes (Supplementary Table 7). IGF1+ CAF showed increased expression of both inflammatory ( CXCL8, CXCL2, CCL5 ) and myofibroblastic genes ( POSTN, MMP11, COL1A1 ), but these were expressed at lower levels than in other CAF subsets (Figure 6D). Inflammatory genes ( CXCL8, CSF3, CCL5; as well as MMP1, INHBA and MMP3), were most highly expressed in IL11+ CAF (which had highest expression of all iCAF marker genes; IL6 (p<0.0001), CXCL8 (p<0.0001), IL11 (p<0.0001); Figure 6E). ECM genes ( COL1A1, COL10A1, POSTN) were most highly expressed in myCAF. Thus, although, IGF1+ CAF upregulate both inflammatory and myofibroblastic genes, these are expressed at markedly lower levels than IL11+ CAF and myCAF respectively. FRC-like fibroblasts are present across cancers at low frequency and are associated with positive response to immunotherapy. Across the seven cancers included in the PCFA, FRC-like cells were a relatively rare fibroblast population, with 77.8% of cells contributed from head and neck (62.1% normal; 15.6% tumour). Within tumour samples, FRC-like fibroblasts were found with highest average relative abundance in HNSCC (11.2%) followed by lung cancer (1.7%) but were present in all cancers at low frequencies (Figure 7A). FRC-like cells were detected in several normal tissue sites, likely representing mucosa-associated lymphoid tissues (MALT; Figure 7A). In breast, oesophageal, lung and pancreatic cancers, FRC-like fibroblasts were enriched in tumour samples, while were less common in HNSCC, gastric and colon cancer relative to normal samples. Given the potential role of apCAF in modulating anti-tumour immunity [5,6,31], we analysed the PCFA to determine which fibroblast subtypes expressed MHC-II genes. Although FRC-like fibroblasts showed the highest expression of MHC-II genes (Supplementary Figure 11A), unlike murine fibroblasts, expression was not restricted to a specific phenotype and was found in several clusters (Supplementary figure 11A) similar to other human studies [31,32]. Compared to professional antigen-presenting cells, FRC-like fibroblasts show reduced expression and an incomplete repertoire of MHC-II genes (Supplementary Figure 11B). The positive correlation between the FRC-like fibroblast and a TLS gene signature [24] across bulk RNA-Seq datasets (Figure 7B), suggested that an FRC-like-containing immune hub exists in different cancers. Given that TLS have been linked with positive response to checkpoint immunotherapy in several cancer types [24,25,33] we investigated whether FRC-like fibroblasts were similarly associated. First, we utilised a dataset (GSE159067) consisting of pre-treatment samples from 102 patients with advanced HNSCC treated with anti-PD-1/PD-L1 immunotherapy. Samples were scored using ssGSEA for fibroblast subset-specific genes (see methods; Supplementary Table 8). We found patients with higher FRC-like scores had significantly improved survival (p<0.01) (Figure 7C). In contrast, iCAF were associated with significantly poorer survival (Supplementary Figure 11C). We performed the same analysis on datasets from lung cancer and melanoma patients. Similarly, we observed higher FRC-like scores were associated with significantly improved survival in ICI treated patients with lung cancer (p<0.01; anti-PD1/PD-L1) and melanoma (p<0.001; anti-CTLA4/PD-1); unlike HNSCC, iCAF were not prognostic (Figure 7D; Supplementary Figure 11D, E, F). Discussion Single cell analysis of 'immune-hot’ and ‘immune-cold’ HNSCC tumours, including normal mucosa (not included in recent HNSCC scRNA-Seq datasets; [17,34] identified six major fibroblast subgroups; universal (PI16+ ) fibroblasts, ADH1B + fibroblasts and FRC-like fibroblasts were present in normal tissue and tumours. myCAF, IL11 + iCAF and proto-CAF were limited to tumours. Of these, proto-CAF fibroblasts were likely a transition state as cells differentiated towards myCAF/iCAF phenotypes. HPV + ve and HPV-ve cancers contained mixtures of all fibroblast subgroups, although proportions and relative abundance varied in individual tumours. In support of our hypothesis, we found significantly higher numbers of FRC-like fibroblasts expressing CCL19 and CCL21 in HPV + ve cancers. These tumours were situated in the oropharynx, an anatomical site that contains secondary lymphoid organs (SLO; tonsils) and considerable numbers of FRC-like cells were also present in matched normal oropharyngeal tissue. However, scRNASeq did not identify FRC-like fibroblasts in HPV-ve HNSCC tumours at the same site, suggesting either that FRC-like fibroblasts are retained within HPV + ve tumours or arise de novo . In lymph nodes, FRC play a central role in structural organisation; attracting and maintaining T-cells, supporting B-cell survival, promoting dendritic cell migration, and controlling permeability of high endothelial venules [23]. Similar cells arise in autoimmune disease, where they transdifferentiate from local fibroblasts and play a central role in supporting TLS formation and maintenance [35]. Consistent with this, within tumours we found FRC-like fibroblasts located with B-cells and (Tfh) CD4 T-cells in TLS structures, correlating with several well-described TLS gene signatures. Development of mature FRCs from precursor cells in SLO is driven by LTβR signalling [23], and TLS-forming FRC-like ‘immunofibroblasts’ have been shown to be regulated by LTα1b2 and IL22 in Sjogren’s syndrome [35]. We found FRC-like fibroblasts to be similarly regulated through non-canonical NF-κB signalling, showing strong activity for transcription factors RELB and NKFB2 (p100/p52). LTβR signals via alternative NF-κB, binding to ligands such LTα1β2 and LIGHT [36]. LTA , LTB , LIGHT/ TNFSF14 (all signalling via LTβR) and CD40L were all spatially associated with FRC-like cells, expressed by B- and T-cells (with LIGHT expressed by FRC-like), potentially driving the FRC-like phenotypic transition. These ligands have all been strongly implicated in TLS neogenesis [37]. Notably, LTBR was expressed by all fibroblast subsets (including CAF), suggesting a common capability to respond to LTBR ligands. In vitro , treatment of primary fibroblasts with lymphotoxin induced FRC-like genes ( CCL19, CCL21, SPIB ) which was enhanced by inhibiting TGFβ signalling. TLS have been reported in a variety of cancers including HNSCC and NSCLC [38,39], but their occurrence likely differs between cancer types. This perhaps is reflected in our pan-cancer fibroblast atlas; FRC-like cells were found with highest average relative abundance in HPV + ve HNSCC (11.2%) followed by lung cancer (1.7%). Rarer phenotypes are under-represented in scRNASeq and the distinct FRC-like clusters in HNSCC and pan-cancer analysis was likely aided by inclusion of an FRC-like-rich cancer type (HPV + ve HNSCC). The pan-cancer analysis demonstrated that FRC-like fibroblasts are present in all cancer types but with low abundance, and thus probably do not cluster discretely when datasets are analysed separately. It is also noteworthy that we detected FRC-like fibroblasts in HPV-ve Visium sections, but not using scRNA-Seq. This highlights the power of deriving cell type specific gene signatures from scRNA-Seq data and using this to deconvolute spatial transcriptomic analysis of tissue sections: enabling far greater numbers of cells to be profiled and avoiding the challenges associated with isolating stromal cells from tissue through disaggregation. The presence of TLS is associated with favourable prognosis in many cancer types, including HNSCC [25,38], in part reflecting the presence of an ongoing, antigen-dependent immune response [39]. Moreover, the presence of TLS has been shown to predict for response to immunotherapy response in several cancer types [24,33]. Our analysis of HNSCC, lung cancer and melanoma patients treated with immune checkpoint blockade shows that high levels of FRC-like fibroblasts in tumours are associated with significantly improved survival suggesting that higher levels of FRC-like fibroblasts may identify likely responders. Furthermore, given their central role in TLS organisation and maintenance, generating FRC-like fibroblasts could be an attractive therapeutic strategy to potentiate immunotherapy response. Recent studies have identified iCAF in several tumour types, including pancreatic cancer and breast cancer [5,6,30,40], using a variety of markers that encompass inflammatory cytokines and other genes ( CXCL1, CXCL8, CXCL12, IL6, CFD, DPT ) [5]. Using a frequently used iCAF gene signature [5], we identified several fibroblast clusters enriched in the pan-cancer analysis that shared expression of genes such as CXCL8, CXCL1, CXCL2, IL6 , with some phenotypes present in normal tissue (e.g., CXCL8 + breast fibroblasts; a similar fibroblast population has been described previously in breast tissue as ‘Fibro-major’ [41]). Of the iCAF signature expressing subsets specific to cancers, IGF1 + CAF and IL11 + CAF were abundant in tumours. IGF1 + CAF were present in all tumour types and expressed all iCAF markers originally identified in PDAC. IGF1 + CAF were transcriptionally similar to universal ( PI16+ ) fibroblasts, maintaining expression of universal fibroblast genes ( PI16, PLA2G2A, CFD ) but showed evidence of activation (expression of FAP, COL1A1, IGF1 ) and expression of iCAF markers ( IL6, CXCL8, CXCL2 ). Within the HNSCC dataset, ~ 50% of PI16 labelled fibroblasts from tumour samples were labelled as IGF1 + CAF in the pan-cancer analysis, suggesting that this phenotype is an early/low activation phenotype consistent with previous studies [1,32]. IGF1 has been reported to mark iCAF in several cancer types [40,42], and this low activation subset probably represents the most commonly referenced ‘iCAF’ phenotype in the literature currently. IL11 + CAF expressed significantly higher levels of inflammatory genes compared to IGF1 + CAF. These were prevalent in GI tumours (HNSCC, CRC, ESCC), but not detected in breast or lung cancers. Although IL11 + CAF could be found with epithelial cells (unlike previous work highlighting iCAF to be distant to epithelial cells; [43]), they especially correlated with inflammatory monocytes and neutrophils. A recent large scRNA-Seq analysis of colorectal tumours revealed a ‘myeloid-cell-attracting’ hub consisting of inflammatory monocytes, neutrophils and MMP3 + CAF [44] hypothesised to be associated with tissue damage and microbial products. An association between inflammatory fibroblasts and myeloid cells has also been described in autoimmune inflammatory bowel disease [45] and periodontitis [46], suggesting this inflammatory niche exists beyond cancer. Gene enrichment and pseudotime analyses identified canonical NF-κB and JAK/STAT signalling as regulating the IL11 + CAF phenotype, with IL-1β and TNF-α as likely ligands. Treatment of normal oropharyngeal fibroblasts with IL1β and TNF-α combination upregulated genes expressed by this phenotype in vivo . Consistent with the spatial analysis, treatment of fibroblasts with conditioned media from inflammatory monocytes treatment produced similar results. The immunological role of IL11 + CAF in cancer is not clear, but highly expressed genes, including IL-6 cytokine family ( IL6, IL11, OSM ) are associated with immunotherapy resistance [12]. The IL11 + CAF subset was associated with significantly poorer overall survival in immunotherapy-treated HNSCC patients. We did not detect bona fide apCAF, either in HNSCC or pan-cancer. apCAF expressing MHC-II/ CD74 were originally identified In KPC PDAC murine models [5] and have subsequently been reported in a several human cancers e.g., breast cancer [40]. However, other studies have failed to identify apCAF. In PDAC, these have been shown to arise from mesothelial cells acquiring fibroblastic features through IL1 and TGFβ signalling [47]; this may explain the absence of apCAF in this study where the curation of the PCFA strictly excluded cells expressing markers denoting alternative cell types, including mural cells and mesothelial cells. MHC-II-related genes were expressed in several fibroblast clusters, with FRC-like cells having highest expression (albeit far lower than professional antigen presenting cells). Dominguez and colleagues similarly found all human CAF to express CD74/HLA-DRA [31]. In breast cancer, a CD74 -expressing CAF cluster (IFNγ-iCAF) has been depicted as apCAF; this population also expressed CCL19 [6]. Conclusion In conclusion, single cell analysis of HNSCC identifies inflammatory fibroblast subsets that are associated with distinct immune cell niches: FRC-like with CD4 + T-cells and B-cells; IL11 + CAF with inflammatory monocytes and neutrophils. HPV + ve HNSCC contain significantly higher levels of FRC-like fibroblasts; their spatial location within TLS, and their positive association with immunotherapy response suggests that these cells support anti-tumour immunity. We also identify transcriptionally discrete iCAF phenotypes including a low activation/transition phenotype ( IGF1+ ), likely the predominant iCAF in the current literature, as well as a more highly inflammatory IL11 + CAF subtype found within cancers of the GI tract. Distinguishing between these phenotypes and dissecting functional differences will be important considerations going forwards. It is intriguing that immunological differences within tumours may be tied to fibroblast phenotypes, and the association of fibroblast subtypes with both negative and positive effects on anti-tumour immunity raises intriguing therapeutic possibilities. Abbreviations CAF : Cancer-associated fibroblast HNSCC : Head and neck squamous cell carcinoma HPV : Human papillomavirus iCAF : Inflammatory CAF FRC : Fibroblastic reticular cell ECM : Extracellular matrix myCAF : Myofibroblastic CAF apCAF : Antigen-presenting CAF meCAF : Metabolic CAF PDAC : Pancreatic ductal adenocarcinoma TIL : Tumour infiltrating lymphocytes GI : Gastrointestinal UMAP : Uniform Manifold Approximation and Projection DEG : Differentially expressed gene NSCLC : Non-small cell lung cancer GSEA : Gene set enrichment analysis ssGSEA : single sample gene set enrichment analysis scRNA-Seq : single cell RNA-Sequencing MxIHC : Multiplexed immunohistochemistry Tfh : T follicular helper GC : Geminal centre TF : Transcription factor ROI : Region of interest RCTD : Robust cell type deconvolution TLS : Tertiary lymphoid structure LT bR : Lymphotoxin beta receptor LT : lymphotoxin α1β2 NOF : Normal primary oral fibroblast PCFA : Pan-cancer fibroblast atlas Fib : Fibroblast CRC : Colorectal carcinoma ESCC : Oesophageal squamous cell carcinoma MALT : Mucosa-associated lymphoid tissues MHC-II : Major histocompatibility complex class II SLO : Secondary lymphoid organ Declarations Ethical Approval Ethical approval for the study was obtained through the UK National Research Ethics Service (South Central - Hampshire B Research Ethics Committee) and written informed consent was obtained from all subjects (REC No. 09/H0501/90). Competing interests The authors declare that they have no competing interests. Authors' contributions BHJ, GJT, CJH, JGB, wrote the manuscript. GJT, CJH, JGB, project led the study. BHJ, IT, MFSR, MJS, BRM, SM, AR, HZ, KL, AA, LL, LD, experimental design. ME, EVK, patient selection and sample acquisition. BHJ, IT, MFSR, MJS, BRM, SM, AR, HZ, KL, AA, LL, LD, ME, EVK, performed the research. IT, BHJ, tissue processing and isolation of primary fibroblasts. IT, MFSR, MJS, BRM, performed in vitro experiments. SM, AR, HZ, KL, AA, LL, LD, generated spatial transcriptomics and MxIHC (Phenocycler) data. IT carried out single cell RNA Sequencing. BHJ, KL, CJH, performed bioinformatics analysis. All authors reviewed the manuscript. Funding This work was funded through a Cancer Research UK Programme grant (DRVRPG-Jun\100004; GJT/CJH), a Cancer Research UK Centres Network Accelerator Award Grant (A20256; GJT), the CRUK and NIHR Experimental Cancer Medicine Center (ECMC) Southampton (A15581 & ECMCQQR-2022/100018), a Pathological Society of Great Britain & Ireland Visiting Fellowship (grant no. VF 1002 03; GJT/MFSR), Sao Paulo Research Foundation (FAPESP) grant 2022/05364-0 (MFSR) and a Project grant from Gilead Sciences Inc. BHJ was funded through an AstraZeneca PhD studentship Availability of data and materials All research data will be submitted on review. Sequencing data generated in this study (scRNA-Seq and Spatial Transcriptomics) will be available in the Gene Expression Omnibus (GEO). Code will be available on Github. MxIHC (Pheno-cycler) data will be made available. The data analysed in this study obtained from GEO: GSE164690, GSE161529, GSE150290, GSE160269, GSE178341, GSE129455, GSE159067 and GSE161537. PRJCA001063 was retrieved from the Genome Sequence Archive. PRJEB23709 was accessed through http://tide.dfci.harvard.edu/download/. Acknowledgements The authors gratefully acknowledge the support of the Faculty of Medicine Tissue Bank, University of Southampton. References Buechler MB, Pradhan RN, Krishnamurty AT, Cox C, Calviello AK, Wang AW, et al. Cross-tissue organization of the fibroblast lineage. Nature. 2021;593:575–9. Davidson S, Coles M, Thomas T, Kollias G, Ludewig B, Turley S, et al. Fibroblasts as immune regulators in infection, inflammation and cancer. Nat Rev Immunol. 2021;21:704–17. Hanley CJ, Waise S, Ellis MJ, Lopez MA, Pun WY, Taylor J, et al. Single-cell analysis reveals prognostic fibroblast subpopulations linked to molecular and immunological subtypes of lung cancer. Nat Commun. 2023;14:387. Biffi G, Oni TE, Spielman B, Hao Y, Elyada E, Park Y, et al. IL1-Induced JAK/STAT Signaling Is Antagonized by TGFβ to Shape CAF Heterogeneity in Pancreatic Ductal Adenocarcinoma. Cancer Discov. 2019;9:282–301. Elyada E, Bolisetty M, Laise P, Flynn WF, Courtois ET, Burkhart RA, et al. Cross-Species Single-Cell Analysis of Pancreatic Ductal Adenocarcinoma Reveals Antigen-Presenting Cancer-Associated Fibroblasts. Cancer Discov. 2019;9:1102–23. Kieffer Y, Hocine HR, Gentric G, Pelon F, Bernard C, Bourachot B, et al. Single-Cell Analysis Reveals Fibroblast Clusters Linked to Immunotherapy Resistance in Cancer. Cancer Discov. 2020;10:1330–51. 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;24:1277–89. Krausgruber T, Fortelny N, Fife-Gernedl V, Senekowitsch M, Schuster LC, Lercher A, et al. Structural cells are key regulators of organ-specific immune response. Nature. 2020;583:296–302. Dammeijer F, Gulijk M van, Mulder EE, Lukkes M, Klaase L, Bosch T van den, et al. The PD-1/PD-L1-Checkpoint Restrains T cell Immunity in Tumor-Draining Lymph Nodes. Cancer Cell. 2020;38:685-700.e8. Ford K, Hanley CJ, Mellone M, Szyndralewiez C, Heitz F, Wiesel P, et al. NOX4 Inhibition Potentiates Immunotherapy by Overcoming Cancer-Associated Fibroblast-Mediated CD8 T-cell Exclusion from Tumors. Cancer Res. 2020;canres;0008-5472.CAN-19-3158v2. Mariathasan S, Turley SJ, Nickles D, Castiglioni A, Yuen K, Wang Y, et al. TGFβ attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature. 2018;554:544–8. Tsukamoto H, Fujieda K, Miyashita A, Fukushima S, Ikeda T, Kubo Y, et al. Combined Blockade of IL6 and PD-1/PD-L1 Signaling Abrogates Mutual Regulation of Their Immunosuppressive Effects in the Tumor Microenvironment. Cancer Research. 2018;78:5011–22. Croft AP, Campos J, Jansen K, Turner JD, Marshall J, Attar M, et al. Distinct fibroblast subsets drive inflammation and damage in arthritis. Nature. 2019;570:246–51. Grauel AL, Nguyen B, Ruddy D, Laszewski T, Schwartz S, Chang J, et al. TGFβ-blockade uncovers stromal plasticity in tumors by revealing the existence of a subset of interferon-licensed fibroblasts. Nat Commun. 2020;11:6315. Galon J, Bruni D. Approaches to treat immune hot, altered and cold tumours with combination immunotherapies. Nat Rev Drug Discov. 2019;18:197–218. Ward MJ, Thirdborough SM, Mellows T, Riley C, Harris S, Suchak K, et al. Tumour-infiltrating lymphocytes predict for outcome in HPV-positive oropharyngeal cancer. Br J Cancer. 2014;110:489–500. Kürten CHL, Kulkarni A, Cillo AR, Santos PM, Roble AK, Onkar S, et al. Investigating immune and non-immune cell interactions in head and neck tumors by single-cell RNA sequencing. Nat Commun. 2021;12:7338. Hynes RO, Naba A. Overview of the matrisome--an inventory of extracellular matrix constituents and functions. Cold Spring Harb Perspect Biol. 2012;4:a004903. Bonizzi G, Karin M. The two NF-κB activation pathways and their role in innate and adaptive immunity. Trends in Immunology. 2004;25:280–8. Garcia-Alonso L, Holland CH, Ibrahim MM, Turei D, Saez-Rodriguez J. Benchmark and integration of resources for the estimation of human transcription factor activities. Genome Res. 2019;29:1363–75. Cable DM, Murray E, Zou LS, Goeva A, Macosko EZ, Chen F, et al. Robust decomposition of cell type mixtures in spatial transcriptomics. Nat Biotechnol. 2022;40:517–26. Dong R, Yuan G-C. SpatialDWLS: accurate deconvolution of spatial transcriptomic data. Genome Biol. 2021;22:145. Fletcher AL, Acton SE, Knoblich K. Lymph node fibroblastic reticular cells in health and disease. Nat Rev Immunol. 2015;15:350–61. Cabrita R, Lauss M, Sanna A, Donia M, Skaarup Larsen M, Mitra S, et al. Tertiary lymphoid structures improve immunotherapy and survival in melanoma. Nature. 2020;577:561–5. Sautès-Fridman C, Petitprez F, Calderaro J, Fridman WH. Tertiary lymphoid structures in the era of cancer immunotherapy. Nat Rev Cancer. 2019;19:307–25. Coppola D, Nebozhyn M, Khalil F, Dai H, Yeatman T, Loboda A, et al. Unique ectopic lymph node-like structures present in human primary colorectal carcinoma are identified by immune gene array profiling. Am J Pathol. 2011;179:37–45. Gu-Trantien C, Loi S, Garaud S, Equeter C, Libin M, de Wind A, et al. CD4+ follicular helper T cell infiltration predicts breast cancer survival. J Clin Invest. 2013;123:2873–92. Hennequin A, Derangère V, Boidot R, Apetoh L, Vincent J, Orry D, et al. Tumor infiltration by Tbet+ effector T cells and CD20+ B cells is associated with survival in gastric cancer patients. Oncoimmunology. 2016;5:e1054598. Browaeys R, Saelens W, Saeys Y. NicheNet: modeling intercellular communication by linking ligands to target genes. Nat Methods. 2020;17:159–62. Ma C, Yang C, Peng A, Sun T, Ji X, Mi J, et al. Pan-cancer spatially resolved single-cell analysis reveals the crosstalk between cancer-associated fibroblasts and tumor microenvironment. Mol Cancer. 2023;22:170. Dominguez CX, Müller S, Keerthivasan S, Koeppen H, Hung J, Gierke S, et al. Single-Cell RNA Sequencing Reveals Stromal Evolution into LRRC15+ Myofibroblasts as a Determinant of Patient Response to Cancer Immunotherapy. Cancer Discov. 2020;10:232–53. Grout JA, Sirven P, Leader AM, Maskey S, Hector E, Puisieux I, et al. Spatial positioning and matrix programs of cancer-associated fibroblasts promote T cell exclusion in human lung tumors. Cancer Discov. 2022;12:2606–25. Helmink BA, Reddy SM, Gao J, Zhang S, Basar R, Thakur R, et al. B cells and tertiary lymphoid structures promote immunotherapy response. Nature. 2020;577:549–55. 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;171:1611-1624.e24. Nayar S, Campos J, Smith CG, Iannizzotto V, Gardner DH, Mourcin F, et al. Immunofibroblasts are pivotal drivers of tertiary lymphoid structure formation and local pathology. Proc Natl Acad Sci USA. 2019;116:13490–7. Remouchamps C, Boutaffala L, Ganeff C, Dejardin E. Biology and signal transduction pathways of the Lymphotoxin-αβ/LTβR system. Cytokine & Growth Factor Reviews. 2011;22:301–10. Kang W, Feng Z, Luo J, He Z, Liu J, Wu J, et al. Tertiary Lymphoid Structures in Cancer: The Double-Edged Sword Role in Antitumor Immunity and Potential Therapeutic Induction Strategies. Front Immunol. 2021;12:689270. Ruffin AT, Cillo AR, Tabib T, Liu A, Onkar S, Kunning SR, et al. B cell signatures and tertiary lymphoid structures contribute to outcome in head and neck squamous cell carcinoma. Nat Commun. 2021;12:3349. Schumacher TN, Thommen DS. Tertiary lymphoid structures in cancer. Science. 2022;375:eabf9419. Cords L, Tietscher S, Anzeneder T, Langwieder C, Rees M, de Souza N, et al. Cancer-associated fibroblast classification in single-cell and spatial proteomics data. Nat Commun. 2023;14:4294. Kumar T, Nee K, Wei R, He S, Nguyen QH, Bai S, et al. A spatially resolved single-cell genomic atlas of the adult human breast. Nature. 2023;620:181–91. 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;11:5077. 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;214:579–96. Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, et al. Spatially organized multicellular immune hubs in human colorectal cancer. Cell. 2021;184:4734-4752.e20. Smillie CS, Biton M, Ordovas-Montañes J, Sullivan KM, Burgin G, Graham DB, et al. Cellular and inter-cellular rewiring of the human colon during ulcerative colitis. Cell. 2019;178:714-730.e22. Williams DW, Greenwell-Wild T, Brenchley L, Dutzan N, Overmiller A, Sawaya AP, et al. Human oral mucosa cell atlas reveals a stromal-neutrophil axis regulating tissue immunity. Cell. 2021;184:4090-4104.e15. Huang H, Wang Z, Zhang Y, Pradhan RN, Ganguly D, Chandra R, et al. Mesothelial cell-derived antigen-presenting cancer-associated fibroblasts induce expansion of regulatory T cells in pancreatic cancer. Cancer Cell. 2022;40:656-673.e7. Kuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016;44:W90–7. Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb). 2021;2:100141. Schubert M, Klinger B, Klünemann M, Sieber A, Uhlitz F, Sauer S, et al. Perturbation-response genes reveal signaling footprints in cancer gene expression. Nat Commun. 2018;9:20. Dann E, Henderson NC, Teichmann SA, Morgan MD, Marioni JC. Differential abundance testing on single-cell data using k-nearest neighbor graphs. Nat Biotechnol. 2022;40:245–53. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401–4. Becht E, Giraldo NA, Lacroix L, Buttard B, Elarouci N, Petitprez F, et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 2016;17:218. Additional Declarations No competing interests reported. Supplementary Files S1tableClinical.xlsx S2tablescRNASeqimmunemarkers.xlsx S3tableFibromarkers.xlsx S4tablePTPathwayenrichment.xlsx S5tableSpatialDELR.xlsx S6tablePCFAmarkersmax1000cells.xlsx S7tablepseudobulkIGF1CAFPI16Fib.xlsx S8tableallgenesignatures.xlsx SupplementaryMaterialsMolCancer.docx Cite Share Download PDF Status: Published Journal Publication published 06 Jan, 2025 Read the published version in Molecular Cancer → Version 1 posted Editorial decision: Revision requested 14 Nov, 2024 Reviews received at journal 14 Nov, 2024 Reviews received at journal 08 Nov, 2024 Reviewers agreed at journal 06 Nov, 2024 Reviewers agreed at journal 06 Nov, 2024 Reviewers agreed at journal 03 Nov, 2024 Reviewers agreed at journal 03 Nov, 2024 Reviewers invited by journal 03 Nov, 2024 Editor assigned by journal 23 Sep, 2024 Submission checks completed at journal 23 Sep, 2024 First submitted to journal 20 Sep, 2024 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-5125055","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":378327260,"identity":"fdd5f573-95e6-4b64-9803-00d01184a9bc","order_by":0,"name":"Benjamin H Jenkins","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Benjamin","middleName":"H","lastName":"Jenkins","suffix":""},{"id":378327261,"identity":"550fd240-713b-47ab-a552-7fbff86fe419","order_by":1,"name":"Ian Tracy","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Ian","middleName":"","lastName":"Tracy","suffix":""},{"id":378327262,"identity":"50dff411-ccab-45c1-a57f-b73110abc976","order_by":2,"name":"Maria Fernanda SD Rodrigues","email":"","orcid":"","institution":"Universidade Nove de Julho","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"Fernanda SD","lastName":"Rodrigues","suffix":""},{"id":378327264,"identity":"cc396441-b243-4f7d-a5ce-25363eb2c0a1","order_by":3,"name":"Melanie JL Smith","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Melanie","middleName":"JL","lastName":"Smith","suffix":""},{"id":378327266,"identity":"a13dd4f6-409f-4e93-8f0c-7cc1a1d3dc79","order_by":4,"name":"Begoña R Martinez","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Begoña","middleName":"R","lastName":"Martinez","suffix":""},{"id":378327268,"identity":"3dbb7968-cd84-4ca8-b8a5-ea01491fdf65","order_by":5,"name":"Mark Edmond","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Mark","middleName":"","lastName":"Edmond","suffix":""},{"id":378327270,"identity":"5ca41369-f560-43f2-95fa-5d1519584ed7","order_by":6,"name":"Sangeetha Mahadevan","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Sangeetha","middleName":"","lastName":"Mahadevan","suffix":""},{"id":378327274,"identity":"6cbf396c-947b-4b7f-b8cd-c848e758ad07","order_by":7,"name":"Anjali Rao","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Anjali","middleName":"","lastName":"Rao","suffix":""},{"id":378327277,"identity":"cec7da28-15be-4f1c-b737-0108cae5d6ab","order_by":8,"name":"Hailing Zong","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Hailing","middleName":"","lastName":"Zong","suffix":""},{"id":378327279,"identity":"e76187d3-c335-4ac3-8918-024a77ab7527","order_by":9,"name":"Kai Liu","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Kai","middleName":"","lastName":"Liu","suffix":""},{"id":378327281,"identity":"73718148-184a-4071-a675-36c693976527","order_by":10,"name":"Abhishek Aggarwal","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Abhishek","middleName":"","lastName":"Aggarwal","suffix":""},{"id":378327284,"identity":"00925242-b8c5-4ff1-9a82-57e6d5f277f2","order_by":11,"name":"Li Li","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Li","suffix":""},{"id":378327287,"identity":"ecf24ddb-b1ca-4f83-89da-ebe58b98ef98","order_by":12,"name":"Lauri Diehl","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Lauri","middleName":"","lastName":"Diehl","suffix":""},{"id":378327289,"identity":"e10864a4-a189-400d-8e9d-e09a53aa4f7a","order_by":13,"name":"Emma V King","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Emma","middleName":"V","lastName":"King","suffix":""},{"id":378327292,"identity":"a07feb08-95f1-4c34-8c55-901f900ca51e","order_by":14,"name":"Jamie G Bates","email":"","orcid":"","institution":"Gilead Sciences Inc.","correspondingAuthor":false,"prefix":"","firstName":"Jamie","middleName":"G","lastName":"Bates","suffix":""},{"id":378327293,"identity":"c5c84577-cf4a-4722-b938-7c7719c70242","order_by":15,"name":"Christopher J Hanley","email":"","orcid":"","institution":"University of Southampton","correspondingAuthor":false,"prefix":"","firstName":"Christopher","middleName":"J","lastName":"Hanley","suffix":""},{"id":378327294,"identity":"28afb2da-c63f-4559-894a-6a2fec1e95b4","order_by":16,"name":"Gareth J Thomas","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1klEQVRIiWNgGAWjYHACNjDJjyTCTEAHM0SLZAMDYwNpWgwOEKtF3oH/2IOPOXb2xucPH3/wcQeDPH8Dj7EBPi2GB5jZDWduS07cdiMtsXHmGQbDGQd4jBPwamlgZpPm3cacYHaDx7CZt42BcQMDj/EBIrTU2xv3nwFrsSeoRZ4BrOUw0PAcsJZEkBa8DjNgZjaTnLnteOIMoF9mzmyTSJ5xmK0Yr/fl2xufSXzcVm3P33/4wIePbTa2/e3NmyXw2nIYlS9BOCLlGwgoGAWjYBSMglHAAABCYkCC38qxqAAAAABJRU5ErkJggg==","orcid":"","institution":"University of Southampton","correspondingAuthor":true,"prefix":"","firstName":"Gareth","middleName":"J","lastName":"Thomas","suffix":""}],"badges":[],"createdAt":"2024-09-20 16:52:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5125055/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5125055/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12943-024-02191-9","type":"published","date":"2025-01-06T15:57:08+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":70947383,"identity":"6d4a97ec-771e-41f0-9836-b4b76653161c","added_by":"auto","created_at":"2024-12-09 13:09:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1742289,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHPV+ve HNSCC tumours display an immune-hot tumour immune microenvironment.\u003c/strong\u003e A) Schematic of workflow for integrative single cell and spatial analysis. B) Plot showing UMAP embeddings for integrated (Seurat RPCA) HNSCC scRNA-Seq dataset comprising HPV-ve HNSCC (n=11; 59,907 cells), HPV+ve HNSCC (n=13; 69,967 cells) and normal oropharyngeal tissue (n=7; 29,952 cells). UMAP plots displaying 12 clusters are accompanied with bar plots showing relative proportions of broad cell types per patient sample. Clusters are annotated based on expression of marker genes as shown in Supplementary Figure 1E. C) H\u0026amp;E images with spatial feature plots showing spatial transcriptomics MCP-counter deconvoluted abundance for B-cells, T-cells and CAF in representative examples of HPV+ve and HPV-ve patients. Cell type abundance within each Visium (10x) spot estimated by MCP-counter is displayed. D) MxIHC (Phenocycler-Fusion) examples of DAPI, CD3, CD20 and Pan-Cytokeratin staining in representative HPV+ve and HPV-ve patients.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/bfcaa49a07f89d70469223a6.png"},{"id":70948235,"identity":"3aef1f94-65de-4e29-811d-38d2007b8a30","added_by":"auto","created_at":"2024-12-09 13:17:57","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1389693,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003escRNA-Seq reveals distinct subsets of inflammatory fibroblasts in HNSCC. \u003c/strong\u003eA) UMAP of fibroblasts from integrated HNSCC dataset showing six clusters (4,894 cells; n=24 HNSCC; n=7 normal). Clusters identified through graph-based clustering are coloured and annotated. B) Heatmap showing the average expression of selected differentially expressed genes for each fibroblast cluster. C) Differential abundance testing of between HNSCC and normal samples. Highlighting differentially abundant neighborhoods. LogFC is displayed via colour of neighborhoods on UMAP plot and bee-swarm plots. D) Differential abundance testing of between HPV+ve and HPV-ve HNSCC. LogFC is displayed via colour of neighborhoods on UMAP plot and bee-swarm plots. E) Trajectory analysis showing curves (overlayed onto UMAP plot) reflecting fibroblast lineages arising from universal (PI16+) fibroblasts. Lineage reconstruction and pseudotime inference using Slingshot package. F) Examination of potential signalling pathways regulating iCAF and FRC-like inflammatory subsets by assessing pathway enrichment in genes that change as a function of pseudotime in the KEGG and Hallmarks gene sets (determined using Monocle 3 trajectory; q_value \u0026lt; 0.05 \u0026amp; morans_I \u0026gt; 0.25). Over-representation analysis showing selected enriched pathways of pseudotime-dependent genes. G) Heatmap of activity of the top 25 transcription factors using DoRothEA regulons (wmean). Clustered scaled activity scores are shown. Below the heatmap shows scaled activity of RELA and RELB for each cell on the UMAP plot. H) Average expression of genes with cytokine activity across fibroblast clusters. Differentially expressed genes filtered for GOMF_CYTOKINE ACTIVITY MSigDB gene set.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/bb805b7250611f753245a73c.png"},{"id":70947398,"identity":"edc3f745-15dd-4c47-b103-0dbe71e51e91","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1896293,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFRC-like fibroblasts colocalise with B-cells and CD4+ T-cells, found within TLS and are regulated via LTβR signalling. \u003c/strong\u003eA) FRC-like fibroblast and immune cell sample-level scRNA-Seq correlations (spearman; p\u0026lt;0.05). For HNSCC samples only, fibroblast proportions (relative to total fibroblasts) per sample were correlated against immune cell cluster proportions (relative to total immune cells). Only significant positive associations are shown. B) Spatial transcriptomics cell type correlations (spearman) using RCTD imputed abundance (normalised weights ≥ 0.05). Visium (10x) spots were deconvoluted using RCTD. Annotated scRNA-Seq data was used as a reference for derivation of cell-type specific gene signatures that were used to deconvolute the cell types present within each 55uM spot. Spearman correlation of normalised weights was carried out on each patient separately. Correlation coefficients are plotted for each of 10 patients, median displayed as vertical line in boxplot and mean as star symbol. Weighted Fisher’s method was used to combine p values. C) Spatial feature plot of deconvoluted values of FRC-like fibroblasts, B-cells and CD4+ T-cells in a HPV+ve HNSCC sample. D) MxIHC (Phenocycler-Fusion) showing staining (DAPI, Pan-cytokeratin, PDPN, CD31, aSMA, CD21, CD20, CD4) in FRC-like containing region of interest identified by RCTD deconvolution. MxIHC markers are shown separately and accompanied by composite image of all markers. PDPN+CD31- cells marking fibroblasts. CD21 (CR2) marking follicular dendritic cells found with dense aggregates of B-cells (CD20) and CD4+ T-cells (CD4). E) FRC-like abundance (RCTD) and TLS signature [24]enrichment (AddModuleScore) spatial feature plot with spearman correlation of RCTD normalised weights. Correlations for each Visium (10x) spot across all 10 patients. Top 5 correlations shown, including FRC-like fibroblasts with highest correlation coefficient. F) Volcano and ligand-receptor interaction plots showing spatially differentially expressed ligands (Log2FC ≥ 1; padj \u0026lt; 0.0001). Differentially expressed ligands identified using FindMarkers on FRC-like containing spots (normalised weight ≥ 0.05) filtered for ligands. Ligands displayed are those found within the GOMF_CYTOKINE_ACTIVITY MSigDB gene set; expressed in FRC-like fibroblasts, B-cells or CD4+ T-cells; and have expression of corresponding receptor in FRC-like fibroblasts. G) qPCR analysis of FRC-like fibroblast markers (\u003cem\u003eCCL19, CCL21, SPIB and RBP5\u003c/em\u003e) in primary NOF treated with a TGFBR1 inhibitor (ALKi; 1µM) and 50ng/ml LTa1β2 for 7 days. Results show mean ± SD of 3 independent experiments in n=1 primary NOF cell line. One-way ANOVA with Bonferroni correction. H) qPCR analysis of FRC-like fibroblast markers (\u003cem\u003eCCL19, CCL21, SPIB and RBP5\u003c/em\u003e) in n=7 primary NOF lines treated with 100ng/ml LTa1β2 + ALKi (1µM) for 48 hours. Results show mean ± SD of 9 independent experiments, colours of points correspond to primary NOF line. Paired Student t test (two-tailed). † = Ct undetermined, assumed Ct=40. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001. ****p \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/75b330effa523ee70d04643d.png"},{"id":70948232,"identity":"e4898993-9d67-4141-a92f-607bd880768d","added_by":"auto","created_at":"2024-12-09 13:17:56","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1682501,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eiCAF colocalise inflammatory monocytes and neutrophils and are regulated via IL1β and TNF𝛼signalling.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA) iCAF and immune cell sample-level scRNA-Seq correlations (spearman; p\u0026lt;0.05). For HNSCC samples only, fibroblast proportions (relative to total fibroblasts) per sample were correlated against immune cell cluster proportions (relative to total immune cells). Only significant positive associations are shown. B) Spatial transcriptomics cell type correlations (spearman) using RCTD imputed abundance (normalised weights ≥ 0.05). Visium (10x) spots were deconvoluted using RCTD. Annotated scRNA-Seq data was used as a reference for derivation of cell-type specific gene signatures that were used to deconvolute the cell types present within each 55uM spot. Spearman correlation of normalised weights was carried out on each patient separately. Correlation coefficients are plotted for each of 10 patients, median displayed as vertical line in boxplot and mean as star symbol. Weighted Fisher’s method was used to combine p values. C) Spatial feature plot of deconvoluted values of iCAF, monocytes and neutrophils in an HPV-ve HNSCC sample. D) MxIHC (Phenocycler-Fusion) showing staining (DAPI, Pan-cytokeratin, PDPN, CD31, aSMA, MPO, CD68, CD14) in iCAF containing region of interest identified by RCTD deconvolution. MxIHC markers are shown separately and accompanied by composite image of all markers. E) Volcano and ligand-receptor interaction plots showing spatially differentially expressed ligands (Log2FC ≥ 1; padj \u0026lt; 0.0001). Differentially expressed ligands identified using FindMarkers on iCAF containing spots (normalised weight ≥ 0.05) filtered for ligands. Ligands displayed are those found within the GOMF_CYTOKINE_ACTIVITY MSigDB gene set; expressed in iCAF, monocytes or neutrophils; and have expression of corresponding receptor in iCAF. F) qPCR analysis of iCAF markers (\u003cem\u003eIL6, MMP3, IL11\u003c/em\u003e and \u003cem\u003eMME\u003c/em\u003e) in primary NOF treated with IL1β (1ng/mL), TNF𝛼(1ng/mL), IL1β (1ng/mL)+TNF𝛼(1ng/mL) and TGFβ (4ng/mL) for 48 hours. Results show mean ± SD of 3 biological replicates in n=1 primary NOF cell line. One-way ANOVA with Bonferroni correction. P values marked by asterisk under bars reflect comparisons with CTL. G) qPCR analysis of iCAF markers (\u003cem\u003eIL6, MMP3, IL11\u003c/em\u003e and \u003cem\u003eMME\u003c/em\u003e) in n=5 primary NOF lines treated with TGFβ (4ng/mL) or IL1β (1ng/mL)+TNF𝛼(1ng/mL) for 72 hours. Results show mean ± SD of n=9 independent experiments, colours of points correspond to primary NOF line. One-way ANOVA with Bonferroni correction. H) qPCR analysis of iCAF markers (\u003cem\u003eIL6, MMP3, IL11\u003c/em\u003e and \u003cem\u003eMME\u003c/em\u003e) in n=3/4 primary NOF lines treated with TGFβ (4ng/mL), IL1β (1ng/mL)+TNF𝛼(1ng/mL), monocyte conditioned media (CM) or LPS-activated monocyte conditioned media for 72 hours. Results show mean ± SD of n≥3 independent experiments, colours of points correspond to primary NOF line. One-way ANOVA with Bonferroni correction. *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001. ****p \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/5bfcc87268d2214b4f89a434.png"},{"id":70947385,"identity":"3c745f45-6c75-4b6d-8b03-a229d5907a42","added_by":"auto","created_at":"2024-12-09 13:09:56","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1773660,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePan-Cancer Fibroblast analysis identifies conserved and semiconserved inflammatory fibroblast phenotypes. \u003c/strong\u003eA) Schematic of Pan-Cancer Fibroblast Atlas (PCFA) including anatomical sites, sample/fibroblast numbers and original publications. This integrated PCFA contained 86,414 fibroblasts from 376 samples. B) UMAP plot of PCFA displaying 16 clusters, and to the right, UMAPs coloured by anatomical site and source of sample (tumour or normal). Samples were integrated using harmony via Seurat v5 sketch-based integration. C) PCFA UMAP split by anatomical site and tumour/normal samples with density of fibroblasts highlighted on UMAP. D) Relative proportion of each cluster in down-sampled (to same number of cells from each anatomical site and same number of cells from tumour/normal samples) normal and tumour samples. E) Relative proportion of each cluster in down-sampled (to same number) anatomical sites (including normal and tumour samples).\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/fb55b86806b6ae8d277ae0d6.png"},{"id":70947387,"identity":"fb7fc41c-bd20-4549-b34d-f9f057062579","added_by":"auto","created_at":"2024-12-09 13:09:56","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1127233,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e‘iCAF’ gene signature highlights different normal fibroblast and CAF populations. \u003c/strong\u003eA) Feature plot (UMAP) showing expression of iCAF signature enrichment in PCFA, split by tumour or normal samples. AddModuleScore using the 12-gene iCAF signature from Elyada \u003cem\u003eet al.,\u003c/em\u003e (2019) [5]. Multiple PCFA clusters show enrichment for iCAF signature. B) Proportion of \u003cem\u003eIL11+\u003c/em\u003e CAF, proto-CAF and \u003cem\u003eIGF1+ \u003c/em\u003eCAF in tumour samples across cancer types. C) Volcano plot showing differentially expressed genes (DEGs; log2FC ≥ 1 or log2FC ≤ -1; adjusted p value \u0026lt; 0.05; identified using pseudobulk sample-level analysis) between \u003cem\u003eIGF1+\u003c/em\u003e CAF and \u003cem\u003eIL11+ \u003c/em\u003eCAF. Selected genes are labelled. D) Heatmap showing average expression of DEGs upregulated in \u003cem\u003eIGF1+ \u003c/em\u003eiCAF compared to universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts. Clustering of rows form gene modules. E) Selected iCAF gene (\u003cem\u003eIL6, CXCL8, IL11, LIF\u003c/em\u003e) expression across clusters (sample-level). Wilcoxon rank-sum test (two-sided) compared to IGF1+ CAF. ns p ≥ 0.05, *p \u0026lt; 0.05; **p \u0026lt; 0.01; ***p \u0026lt; 0.001. ****p \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/240dee1e5bf75d30f40bfdb8.png"},{"id":70949326,"identity":"6acc75ff-7ffa-4b40-8adb-ab8cf1dd1c9e","added_by":"auto","created_at":"2024-12-09 13:25:57","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":650349,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFRC-like fibroblasts are present across cancers at low frequency and are associated with positive response to immunotherapy. \u003c/strong\u003eA) Abundance of FRC-like fibroblasts across anatomical sites (normal only) and cancer types (tumour only). Log10 scale used due to extremely low abundance of FRC-like fibroblasts in non-head \u0026amp; neck tissue/tumours. FRC-like fibroblasts are rare however present across cancer types. B) Correlation of FRC-like fibroblast signature and TLS signature [24]enrichment across selected cancer types in TCGA Bulk RNA-Seq data. ssGSEA run using batch effects normalized mRNA data from the Pan-Cancer Atlas Hub (UCSCXena). Spearman correlation coefficients and p-values displayed. C) Kaplan-Meier (overall) survival plot showing anti-PD-1/PD-L1 treated HNSCC cohort (GSE159067; n=102), stratified by FRC-like fibroblast (ssGSEA) scores. Below, forest plot for multivariate cox regression model using FRC-like level (high or low), patient sex and patient age. Hazard ratio estimates along with confidence intervals (95%) and p-values are plotted for each variable. Statistical significance shown on Kaplan-Meier plot assessed using a log-rank test. D) Kaplan-Meier (overall) survival plot showing anti-PD-1/PD-L1 treated NSCLC cohort (GSE161537; n=82) and anti-CTLA-4 + anti-PD-1 or anti-PD-1 treated melanoma cohort (PRJEB23709; n=91) stratified by FRC-like fibroblast (ssGSEA) scores. Below, forest plot for multivariate cox regression model using FRC-like level (high or low), patient sex and patient age. Hazard ratio estimates along with confidence intervals (95%) and p-values are plotted for each variable. Statistical significance shown on Kaplan-Meier plot assessed using a log-rank test. p\u0026lt;0.05 considered statistically significant.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/9dbaa1b28478571430fd6d91.png"},{"id":73694160,"identity":"a6bd3e0e-d017-4627-8ea2-6f059712ab20","added_by":"auto","created_at":"2025-01-13 16:11:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":13138202,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/ab1e0ecc-f94d-4b8f-b2b0-7061f05e94aa.pdf"},{"id":70947384,"identity":"fbc3c413-8cca-4aa9-aa71-1f2dae5b493a","added_by":"auto","created_at":"2024-12-09 13:09:56","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":17965,"visible":true,"origin":"","legend":"","description":"","filename":"S1tableClinical.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/ccbacf7a9124e74b346f1490.xlsx"},{"id":70948233,"identity":"16575045-bbff-4cfc-a846-96f8d8d3e9ac","added_by":"auto","created_at":"2024-12-09 13:17:56","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":2175638,"visible":true,"origin":"","legend":"","description":"","filename":"S2tablescRNASeqimmunemarkers.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/2b34e89bad31d113e8be2585.xlsx"},{"id":70948234,"identity":"aa1b1e69-df20-45ea-8069-1793b5484491","added_by":"auto","created_at":"2024-12-09 13:17:56","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":364176,"visible":true,"origin":"","legend":"","description":"","filename":"S3tableFibromarkers.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/34813cca872ff761b47bd25f.xlsx"},{"id":70948237,"identity":"148f86d5-21cd-4323-b0c5-c6e4a3a39226","added_by":"auto","created_at":"2024-12-09 13:17:57","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":18871,"visible":true,"origin":"","legend":"","description":"","filename":"S4tablePTPathwayenrichment.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/05437f63b9688caa93f0dce8.xlsx"},{"id":70947389,"identity":"ea08545e-d831-43bf-b0ca-29e5803045df","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":57444,"visible":true,"origin":"","legend":"","description":"","filename":"S5tableSpatialDELR.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/4e9b9a522dbc80d1feac3c7a.xlsx"},{"id":70947393,"identity":"995f6646-c9fe-4631-969d-b9d5fbed720c","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":750859,"visible":true,"origin":"","legend":"","description":"","filename":"S6tablePCFAmarkersmax1000cells.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/a659688569b3132dfb783a15.xlsx"},{"id":70947396,"identity":"0aa20463-d191-4d11-974b-8de12b40aba8","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":681083,"visible":true,"origin":"","legend":"","description":"","filename":"S7tablepseudobulkIGF1CAFPI16Fib.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/99a92cde5de5570447df7afc.xlsx"},{"id":70947390,"identity":"b867ba3b-e300-4f4b-864d-4bad1529a6fc","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":18698,"visible":true,"origin":"","legend":"","description":"","filename":"S8tableallgenesignatures.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/5e4bc3e364b90f16c7d891e8.xlsx"},{"id":70947397,"identity":"d042c080-fd0a-4723-bc71-68675e7193ca","added_by":"auto","created_at":"2024-12-09 13:09:57","extension":"docx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":8780133,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterialsMolCancer.docx","url":"https://assets-eu.researchsquare.com/files/rs-5125055/v1/8ecc7660893a0186446408be.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Single cell and spatial analysis of immune-hot and immune-cold tumours identifies fibroblast subtypes associated with distinct immunological niches and positive immunotherapy response","fulltext":[{"header":"Introduction","content":"\u003cp\u003eFibroblasts are ubiquitous cells that assume specialised phenotypes and activation states to play a multifaceted role in health and disease [1]. Although historically regarded as structural cells that principally remodel extracellular matrix (ECM) they are now recognised as key immune sentinel cells capable of initiating, maintaining, and suppressing immune responses in response to pathological stimuli [2].\u003c/p\u003e \u003cp\u003eCancer-associated fibroblasts (CAF) research has mostly focused on cells with a myofibroblast phenotype (myCAF); these are found in most cancer types, have numerous tumour-promoting functions and are akin to myofibroblasts in fibrotic diseases [1,3]. However, recent single cell studies have identified transcriptomically distinct CAF subtypes, including inflammatory CAF (iCAF), antigen presenting CAF (apCAF) and metabolic CAF (meCAF) [4\u0026ndash;7]. These phenotypes form part of a plastic population that can change states in response to local stimuli; Biffi and colleagues elegantly demonstrated this, switching pancreatic stellate cells between myCAF and iCAF states \u003cem\u003ein vitro\u003c/em\u003e by manipulating TGFβ and IL-1 signalling respectively [4]. This plasticity emphasises the role of fibroblasts as key early response cells in tissues. Although iCAF have been identified in several cancer types, including pancreatic ductal adenocarcinoma (PDAC), breast and lung cancers [5\u0026ndash;7], it is not yet clear whether this phenotype is common to all tumours or whether all \u0026lsquo;iCAF\u0026rsquo; are the same. Although there are common mechanisms that fibroblasts use to regulate tissue inflammation, inflammatory stimuli have been shown to produce organ-specific immune signatures in fibroblasts from different organs [8] suggesting that inflammatory CAF phenotypes could vary.\u003c/p\u003e \u003cp\u003eThe clinical success of immune checkpoint inhibitors in treating multiple cancer types is well established. However, only a subset of patients respond favourably [9], and this has generated significant interest in understanding how the tumour microenvironment suppresses anti-tumour immunity. myCAF have several immunosuppressive functions and myCAF-rich tumours are resistant to immunotherapy [6,10,11]. iCAF also express several cytokines associated with immune evasion, including IL6, LIF and CXCL12 [12]. Conversely, in autoimmunity, fibroblasts have been shown to amplify chronic inflammation [13], and a novel population of \u0026lsquo;interferon licenced fibroblasts\u0026rsquo; that enhance immunotherapy response has been identified in murine tumour models [14]. Thus, a fibroblast may support or suppress immunity depending on context. Given their plasticity, the concept of generating an immune-supportive phenotype to improve immunotherapy response in cancer is intriguing.\u003c/p\u003e \u003cp\u003eIn most cancer types, tumours with high levels of tumour-infiltrating lymphocytes (TIL) have better prognosis and show improved response to checkpoint immunotherapy [15]. Head and neck cancer (HNSCC) is subdivided into human papillomavirus (HPV)-related (HPV\u0026thinsp;+\u0026thinsp;ve) tumours and those typically associated with smoking/alcohol (~\u0026thinsp;30% and 70% of cases respectively). Around 85% of HPV\u0026thinsp;+\u0026thinsp;ve HNSCC are heavily infiltrated by T- and B-cells and, despite presenting mostly at late stage, are associated with significantly better survival compared with TIL-low HPV-ve tumours [16].\u003c/p\u003e \u003cp\u003eIn this study we hypothesised that comparative analysis of \u0026lsquo;immune hot\u0026rsquo; (HPV\u0026thinsp;+\u0026thinsp;ve) and \u0026lsquo;immune cold\u0026rsquo; (HPV-ve) HNSCC subtypes would identify different inflammatory fibroblast populations that reflect the multifaceted role of fibroblasts in immunity; specifically, that HPV\u0026thinsp;+\u0026thinsp;ve tumours contain a fibroblast phenotype that likely supports anti-tumour immunity. Using single cell and spatial transcriptomics we analysed HPV-positive and HPV-negative HNSCC, identifying six fibroblast subsets, including two characterised by expression of immunomodulatory genes (\u003cem\u003eIL-11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;inflammatory iCAF and fibroblastic reticular cell [FRC]-like). These fibroblast subsets occupied distinct immunological niches; \u003cem\u003eIL-11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;iCAF were spatially associated with and activated by inflammatory monocytes through canonical NF-κB signalling regulated by IL-1β and TNF-α. FRC-like were enriched in HPV\u0026thinsp;+\u0026thinsp;ve tumours, associated with CD4 T-cells and B-cells in tertiary lymphoid structures, and were regulated through non-canonical NF-κB signalling via lymphotoxin. Pan-cancer analysis showed that \u0026lsquo;iCAF\u0026rsquo; is not a single phenotype; \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;iCAF, present in HNSCC and other gastrointestinal (GI) tract cancers, are transcriptomically distinct from iCAFs previously described in pancreatic and breast cancers and have heightened inflammatory features; FRC-like represented a rare phenotype, but present in all tumour types and associated with positive response to checkpoint immunotherapy.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch3\u003eHuman Subjects\u003c/h3\u003e\n\u003cp\u003eEthical approval for the study was obtained through the UK National Research Ethics Service (REC No. 09/H0501/90) and written informed consent was obtained from all subjects. Tumour and matched-normal tissue were obtained from patients undergoing surgical tumour resection at Poole Hospital (Poole, Dorset, UK) for HNSCC. Tissue samples were transported (within 1\u0026thinsp;hour) to the laboratory on ice in serum-free Dulbecco\u0026rsquo;s Modified Eagle Medium (DMEM; Sigma-Aldrich). Sample information (including clinical and sample digestions) is shown in Supplementary Table 1. Patient HPV status was confirmed using p16 immunostaining in combination with assessing HPV-encoded gene expression using a human-HPV hybrid reference genome to align and map reads (see details below). HPV-encoded exons were detected in 6 patients using the human-HPV-16 reference genome and 1 patient using the human-HPV-33 reference genome (Supplementary Figure 1A).\u003c/p\u003e\n\u003ch3\u003ePrimary fibroblast culture and \u003cem\u003ein vitro\u003c/em\u003e experiments\u003c/h3\u003e\n\u003cp\u003ePlease see Supplementary Materials.\u003c/p\u003e\n\u003ch3\u003eSample processing\u003c/h3\u003e\n\u003cp\u003ePlease see Supplementary Materials.\u003c/p\u003e\n\u003ch3\u003escRNA-Seq\u003c/h3\u003e\n\u003cp\u003eFor each sample, 5000 single cells were captured on an Illumina 10X Chromium Controller\u003csup\u003eTM\u003c/sup\u003e system using the Illumina single cell 3\u0026rsquo; gene expression and library preparation kits (V3.1 #1000269). Sample capture, sample indexing, and library preparation were carried out according to manufacturer\u0026rsquo;s instructions. Size distribution, quality control, and quantification of the libraries was assessed using High Sensitivity DNA chips (Agilent Technologies #5067-4626) and KAPA library quantification qPCR kit (Roche #07960140001). Prepared libraries were pooled and sent to Oxford Genomics (UK) for 150-base pair, paired-end sequencing on a Novaseq6000\u003csup\u003eTM\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eSequence alignment and annotation\u003c/h3\u003e\n\u003cp\u003eFASTQ files were aligned to the Human reference genome (GRCh38\u0026ndash;2020-A) which had the HPV genome concatenated to both the FASTA and .GTF reference files (using cellranger count v6.1.1, 10x Genomics). Human-HPV references were made using the cellranger mkref command (cellranger V6.1.1). scRNA-Seq data was processed with cellranger count (cellranger v6.1.1) generating feature-barcode matrices in which subsequent data analysis was carried out in R (v4.1.1) using Seurat package (v4.1.0).\u003c/p\u003e\n\u003ch3\u003eQuality control, normalisation and integration\u003c/h3\u003e\n\u003cp\u003eEach patient expression matrix was initially created into a Seurat object with cells requiring expression of at least 200 genes and genes expressed in at least 3 cells. Poor quality cells were removed using a mitochondrial RNA percentage threshold calculated by the median + 3* median absolute deviation [cells above this threshold (~20%) were removed]. Cells expressing \u0026gt;6000 features were removed to reduce potential doublets. Seurat was then used for normalisation and\u0026nbsp;reciprocal PCA (RPCA)\u0026nbsp;integration of scRNA-Seq data (Further details in Supplementary Materials). Principal component analysis (PCA) was then performed on the integrated object followed by Uniform Manifold Approximation and Projection (UMAP) visualisation. Clustering was performed using shared nearest-neighbour (SNN) graph construction (FindNeighbors) followed by FindClusters.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eIdentifying marker genes\u003c/h3\u003e\n\u003cp\u003eDifferentially expressed genes (DEGs) were identified for each cluster using FindAllMarkers (Wilcoxon rank sum test) with genes selected expressed in \u0026ge;25% of cells and log2FC \u0026ge;0.5 (adjusted p value \u0026lt;0.05). DEGs were compared to known cell type markers described widely in the literature to annotate broad and finer cell types. \u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eHNSCC inter-dataset integration\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eSeurat\u0026rsquo;s RPCA integration was also used for HNSCC inter-dataset integration with GSE164690 (Further details in Supplementary Materials).\u003c/p\u003e\n\u003ch3\u003eGene module scores and pathway analysis\u003c/h3\u003e\n\u003cp\u003eModule scores were calculated using Seurat\u0026rsquo;s AddModuleScore function calculated by taking the average expression levels of each cluster at the single cell level subtracted by aggregated expression of control feature sets. All gene signatures used in the analysis are shown in Supplementary Table 8. Pathway analysis was performed using over-representation analysis (enrichr v3.2; [48]) and gene set enrichment analysis (GSEA) (clusterprofiler v4.6.2; [49]) using KEGG and MSigDB Hallmark databases. Enriched pathways with P value and adjust\u0026nbsp;p value \u0026lt;0.05 were examined. PROGENy: Pathway RespOnsive GENes for activity inference was used to infer activities of 14 pathways\u0026nbsp;[50]. PROGENy pathway activity scores were calculated for the Seurat object running \u0026lsquo;progeny\u0026rsquo; command (organism=\u0026quot;Human\u0026quot;, top=500, perm=1, return_assay = TRUE). Pathway activity scores were then scaled. Summarised scores (mean) of each activity for each cell cluster were determined and plotted in a heatmap.\u003c/p\u003e\n\u003ch3\u003eDifferential abundance\u003c/h3\u003e\n\u003cp\u003eWe utilised MiloR (v1.6.0) to identify differentially abundant phenotypes using KNN graphs [51]. MiloR was run on the integrated objects separately with the following parameters used for buildGraph and makeNhoods: k= 70, d=20, refined = TRUE. To account for cell type abundance differences resulting from use of different digestions, the digest was specified in the design formula along with source (tumour/normal) or HPV status. SpatialFDR threshold (alpha) was set to 0.05 when highlighting differentially abundant neighbourhoods \u0026ndash; which were displayed in bee-swarm plots.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhen calculating relative cell type proportions per sample in the scRNA-Seq data we accounted for digestion differences in select samples that had undergone scRNA-Seq of liberase and col+ digests separately, calculating pseudo-mixtures by combining the relative abundance of cell types for each sample in a 1:9 (liberase:col+) ratio.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eTrajectory and pseudotime\u003c/h3\u003e\n\u003cp\u003eMonocle 3 (v1.0.0) and slingshot (v2.6.0) were used for trajectory analysis and pseudotime calculations. Both methods yielded the same lineages. Monocle 3 was used to determine genes that change as a function of pseudotime, graph_test, specifying the neighbour_graph as \u0026lsquo;prinicpal_graph\u0026rsquo; was run.\u003c/p\u003e\n\u003ch3\u003eTranscription Factor Analysis\u003c/h3\u003e\n\u003cp\u003eDoRothEA regulons, a collection of transcription factors and their targets, were used to infer transcription factor activities in fibroblast populations [20]. Activities were determined using run_wmean from the decoupleR (v2.5.0) package and subsequently scaled.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003ePan-Cancer Fibroblast Atlas (PCFA) and label transfer\u003c/h3\u003e\n\u003cp\u003ePlease see Supplementary Materials.\u003c/p\u003e\n\u003ch3\u003eSpatial Transcriptomics\u003c/h3\u003e\n\u003cp\u003eAll pre-sequencing procedures were carried out following the manufacturer\u0026apos;s instructions on 6.5mm capture areas using the Visium V2 CytAssist workflow. All samples were processed through the Spaceranger pipeline (v2.0.0) according to 10x Genomics guidelines. Please see Supplementary Materials for details of spatial transcriptomics processing, spot deconvolution and spatially guided ligand-receptor/NicheNet analysis.\u003c/p\u003e\n\u003ch3\u003eBulk RNA sequencing (scRNA-Seq)\u003c/h3\u003e\n\u003cp\u003eCounts for the Head and Neck Squamous Cell Carcinoma, TCGA-HNSC (https://www.cancer.gov/tcga) (566 samples: 520 primary solid tumour; 46 solid tissue normal) cohort (Illumina HiSeq platform) were downloaded using TCGAbiolinks and converted to CPM using edgeR. TPM normalised data for TCGA-HNSC was downloaded from GDC data portal. The cBioPortal for cancer genomics [52] was used to obtain additional metadata from the Head and Neck Squamous Cell Carcinoma (TCGA, Firehose Legacy) study.\u0026nbsp;UCSCXenaTools package (v1.4.8) was used to download Batch effects normalized mRNA data (n=11,060) from the Pan-Cancer Atlas Hub and corresponding clinical metadata.\u003c/p\u003e\n\u003ch3\u003eBulk RNA-Seq Deconvolution\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eTo investigate the cell type abundance in bulk RNA-Seq data, the immunedeconv R package (v2.1.0) was used to run MCP-counter [53] on TPM normalised HNSCC (TCGA) Bulk RNA-Seq. The \u0026lsquo;deconvolute\u0026rsquo; function was run specifying \u0026lsquo;MCP_counter\u0026rsquo;.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003essGSEA\u003c/h3\u003e\n\u003cp\u003eWe used single sample GSEA (ssGSEA) using the GSVA package (v1.46.0) to calculate enrichment scores for bulk RNA-Seq samples and gene sets.\u003c/p\u003e\n\u003ch3\u003eAnalysis of Immunotherapy Data\u003c/h3\u003e\n\u003cp\u003eThe following bulk RNA-Seq datasets were used for analysis of immunotherapy treated patients. HNSCC (GSE159067): 102 patients with advanced HNSCC treated with immunotherapy targeting PD-1/PD-L1. Lung (GSE161537): 82 patients with advanced non-small cell lung cancer (NSCLC) treated with second-line immunotherapy targeting PD-1/PD-L1. Metastatic melanoma (PRJEB23709): 91 patients treated with anti-PD-1 alone or combined anti-PD-1 and anti-CTLA-4 immunotherapy. Overall survival analysis (Kaplan-Meier and cox regression) was performed out using survival package (v3.5-7); patients were split into high and low (based on ssGSEA scores) using the optimal cut points determined by surv_cutpoint (survminer v0.4.9). Multivariate cox regression was carried out using \u0026lsquo;coxph\u0026rsquo; function (survival) specifying the fibroblast abundance, sex, and age. Generation of fibroblast subset specific gene signatures (Supplementary Table 8) used in ssGSEA are detailed in Supplementary Materials.\u003c/p\u003e\n\u003ch3\u003eMultiplex immunofluorescence using PhenoCycler-Fusion\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003ePlease see Supplementary Materials.\u003c/p\u003e\n\u003ch3\u003eStatistical analysis\u003c/h3\u003e\n\u003cp\u003eStatistical analysis was performed using R environment v4.1.1 [ggpubr package (v0.4.0) for plotting graphs] and Graph Pad Prism 9 (v10, GraphPad, San Diego, CA, USA). Wilcoxon rank-sum test (two-sided) or Students t-test (two-sided) were used to evaluate associations between continuous variables. Normality was assessed by Shapiro\u0026ndash;Wilk test. One-way ANOVA or Kruskal-Wallis test was used to compare \u0026gt;2 groups. Multiple comparisons were investigated by adjusting the p-value using the Bonferroni method. Correlation analysis was carried out using spearman\u0026rsquo;s rho (two-sided). A two-sided Fisher\u0026rsquo;s exact test was used to compare categorical data between groups. Survival analysis was carried out using Kaplan-Meier curves with log-rank test and multivariate cox regression statistics. P \u0026lt; 0.05 was considered to indicate a statistically significant difference.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eHPV+ve HNSCC has an immune hot tumour-immune microenvironment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed scRNA-Seq on treatment-na\u0026iuml;ve HNSCC samples (\u003cem\u003en\u003c/em\u003e=10; 7 HPV+ve [\u0026lsquo;immune-hot\u0026apos;]; 3 HPV-ve [\u0026lsquo;immune-cold\u0026rsquo;) with matched normal oropharyngeal mucosa [\u003cem\u003en=7\u003c/em\u003e]; Figure 1A; Supplementary Figure 1B,C; Supplementary Table 1). To increase patient numbers, this (EPG) dataset (82,844 cells after quality control) was integrated with a publicly available dataset (Figure 1A; HNSCC samples of oral cavity/oropharynx; [17]), generating an atlas of 159,826\u0026nbsp;cells from 24 patients (11 HPV-ve; 13 HPV+ve; 7 normal; Figure 1B; Supplementary Figure 1D,E). Spatial transcriptomic analysis (10x; Visium) and multiplexed immunochemistry (Akoya; PhenoCycler) were performed on tumour sections from the initial ten patients (Figure 1A).\u003c/p\u003e\n\u003cp\u003eDeconvolution of bulk RNA-Seq data from the HNSCC TCGA cohort using MCP-counter confirmed that that HPV+ve tumours contain significantly more CD8+ T-cells (p\u0026lt;0.0001), CD4+ T-cells (p\u0026lt;0.0001) and B-cells (p\u0026lt;0.0001; Supplementary Figure 1F). Spatial transcriptomic analysis using MCP-counter to deconvolute cell type abundance within individual spots, also showed significantly more T-cells (p\u0026lt;0.001), B-cells (p\u0026lt;0.01) and total lymphocytes (p\u0026lt;0.01) in HPV+ve tumours (Figure 1C; Supplementary Figure 1G); also confirmed by multiplexed IHC (MxIHC; Figure 1D). Notably, there was a prominent fibroblast presence in both HPV-ve and HPV+ve tumours (Figure 1C; Supplementary Figure 1G).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA more detailed analysis of immune cell subsets (Supplementary Table 2) in the scRNA-Seq data revealed differences in T-cell and NK-cell phenotypes (56,664 cells) between HPV+ve and HPV-ve tumours (Supplementary Figure 2A,B). \u003cem\u003eCD4\u003c/em\u003e+ \u003cem\u003eICOS\u003c/em\u003e+ \u003cem\u003ePDCD1\u003c/em\u003e+ (PD-1) T-cells, resembling T follicular helper (Tfh) cells were more common in HPV+ve tumours, as were CD4+ na\u0026iuml;ve-like T-cell clusters and \u003cem\u003eKIT+\u003c/em\u003e NK-cells (Supplementary Figure 2B). Analysis of B- and plasma cells (26,156 cells) showed that germinal centre (GC) B-cells (\u003cem\u003eRGS13\u003c/em\u003e+, \u003cem\u003eNEIL1\u003c/em\u003e+), cycling B-cells (\u003cem\u003eUBE2C\u003c/em\u003e+, \u003cem\u003eTYMS\u003c/em\u003e+) and na\u0026iuml;ve B-cells (\u003cem\u003eTCL1A\u003c/em\u003e+,\u0026nbsp;\u003cem\u003eIL4R\u003c/em\u003e+)\u0026nbsp;were all enriched in HPV+ve tumours compared to HPV-ve tumours\u0026nbsp;(Supplementary Figure 2C,D), while switched B-cell subsets were found in both. There were no differentially abundant myeloid populations\u0026nbsp;(Supplementary Figure 2E).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003escRNA-Seq reveals distinct subsets of inflammatory fibroblasts in HNSCC.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate fibroblast phenotypes, we first broadly identified fibroblasts based on lumican expression (\u003cem\u003eLUM+;\u0026nbsp;\u003c/em\u003e4,894 cells); fibroblasts clustered closely with \u003cem\u003eRGS5+\u003c/em\u003e mural cells (2,174 cells), which included\u0026nbsp;pericytes and smooth muscle cells (SMCs; Supplementary Figure 3A,B). We identified six clusters of fibroblasts; three confined to tumours (CAF) and three present in both tumours and normal tissue (Figure 2A; Supplementary Figure 3CD; Supplementary Table 3). Overall, individual patient tumours showed significant fibroblast heterogeneity, generally containing a mixture of the six phenotypes (Supplementary Figure 3E).\u003c/p\u003e\n\u003cp\u003eThe largest CAF cluster expressed canonical myofibroblastic CAF (myCAF) markers (\u003cem\u003ePOSTN, MMP11, ACTA2\u003c/em\u003e) and showed highest enrichment for TGFb\u0026nbsp;signalling (Figure 2B; Supplementary Figure 4A). This cluster was characterised by high expression of ECM genes (including \u003cem\u003eCOL1A1, FN1, COL1A2, COL6A3\u0026nbsp;\u003c/em\u003eand \u003cem\u003eCOL11A1\u003c/em\u003e); with numerous differentially expressed genes (DEGs) associated with core matrisome components [collagens (n=14), glycoproteins (\u003cem\u003en=22\u003c/em\u003e) and proteoglycans (\u003cem\u003en=5\u003c/em\u003e)] (Supplementary Figure 4B; Supplementary Table 3; [18].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAn inflammatory CAF (iCAF) population was characterised by high expression of inflammatory cytokines (e.g.,\u003cem\u003e\u0026nbsp;IL11, IL6, CXCL8, CXCL1,\u003c/em\u003e \u003cem\u003eCXCL5\u003c/em\u003e; Figure 2B). Notably, iCAF also expressed upregulated ECM genes (albeit at a comparatively lower level than myCAF), with higher levels of genes associated with ECM remodelling (\u003cem\u003eMMP3, MMP1, PLAU)\u003c/em\u003e, glycolysis/hypoxia (\u003cem\u003eHIF1A, ENO1, GK, CA12, SLC16A3\u003c/em\u003e/MCT4) and neutrophil-recruiting chemokines (\u003cem\u003eCXCL1, CXCL5, CXCL6, CXCL8)\u003c/em\u003e (Supplementary Table 3). This cluster was highly enriched for hypoxia, NF-\u0026kappa;B, and TNFa signalling pathways (Supplementary Figure 4A).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA further CAF cluster expressed lower levels of myCAF/iCAF marker genes; this \u0026lsquo;proto-CAF\u0026rsquo; cluster displayed few unique DEGs (\u003cem\u003en=\u003c/em\u003e19) compared with other fibroblast phenotypes (which ranged from \u003cem\u003en=83-232\u003c/em\u003e unique DEGs) and on the UMAP adjoined normal fibroblast and CAF clusters, likely representing a transition state. myCAF and iCAF were present in both HPV-ve and HPV+ve HNSCC (Supplementary Figure 3E), but HPV-ve HNSCC samples contained greater proportions of CAF/fibroblasts relative to total cell number per sample (Supplementary Figure 3D).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWithin normal mucosa we identified three fibroblast subtypes (also present in tumours.) Universal (adventitial) fibroblasts expressing \u003cem\u003ePI16\u0026nbsp;\u003c/em\u003ewere present in normal mucosa, HPV+ve and HPV-ve tumour samples (Supplementary Figure 3E). These expressed \u003cem\u003eCD34\u0026nbsp;\u003c/em\u003eand distinctive ECM-associated genes, including \u003cem\u003eCOL14A1, OGN\u003c/em\u003e and \u003cem\u003eTNXB\u003c/em\u003e, likely reflecting their vascular-niche function (Figure 2B, Supplementary Figure 4B). \u003cem\u003eADH1B+\u003c/em\u003e fibroblasts were the most common subtype in normal tissue (54% of fibroblasts) but were infrequent in tumours (5% of fibroblasts; Figure 2C; Supplementary Figure 3E). The core matrisome profile of \u003cem\u003eADH1B+\u003c/em\u003e fibroblasts was similar to universal (\u003cem\u003ePI16+)\u003c/em\u003e fibroblasts (Supplementary Figure 4B). The third fibroblast subgroup expressed \u003cem\u003eCCL19, CCL21, VCAM1, IL7\u0026nbsp;\u003c/em\u003eand \u003cem\u003eSPIB\u003c/em\u003e (Figure 2B) with a phenotype akin to fibroblastic reticular cells (FRC), specialised fibroblast subsets of lymphoid tissues that organise and traffic lymphoid cells. Notably, FRC-like fibroblasts were significantly enriched in HPV+ve tumours (Figure 2D). Levels of FRC-like fibroblasts varied between individual HPV+ve tumours but were uniformly rare in all HPV-ve cases (Supplementary Figure 3E).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNotably, when present in tumours, normal fibroblast subtypes expressed activation- (\u003cem\u003eFAP, FN1, PDPN, COL1A1\u003c/em\u003e), inflammation- (\u003cem\u003eCXCL1, ISG15\u003c/em\u003e) and insulin-like growth factor (IGF)-related (\u003cem\u003eIGF1, IGFBP2, IGFBP4\u003c/em\u003e) genes (Supplementary Figure 4C) suggesting early activation (with these genes expressed at higher levels in CAF clusters).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhenotypic regulators of iCAF and FRC-like inflammatory fibroblasts.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next inferred fibroblast lineages arising from universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts [1]. Trajectory analysis identified three lineages leading to the formation of FRC-like fibroblasts (through \u0026lsquo;\u003cem\u003eADH1B+\u003c/em\u003e\u0026rsquo;), myCAF andiCAF (both through \u0026lsquo;proto-CAF\u0026rsquo;; Figure 2E; Supplementary Figure 4D).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSignalling pathways regulating FRC-like and iCAF inflammatory subsets were examined by assessing pathway enrichment in genes changing as a function of pseudotime in KEGG and Hallmarks gene sets (q_value \u0026lt; 0.05 \u0026amp; morans_I \u0026gt; 0.25). Pseudotime analysis of the FRC-like lineage (1) showed increased expression of genes associated with NF-\u0026kappa;B signalling pathway (KEGG; p.adjust \u0026lt;0.01) and Allograft rejection (Hallmarks; p.adjust \u0026lt;0.0001) (Figure 2F; Supplementary Table 4). While many genes (e.g., \u003cem\u003eCCL19, C7, IRF8\u003c/em\u003e) showed a pseudotime-dependent increase in expression through the \u003cem\u003eADH1B+\u003c/em\u003e cluster to FRC-like, other genes enriched for TNF-alpha Signalling via NF-\u0026kappa;B (Hallmarks; p.adjust \u0026lt;0.0001) increased in the \u003cem\u003eADH1B+\u003c/em\u003e cluster but decreased in the FRC-like cluster (e.g., S\u003cem\u003eOCS3, JUN, IRF1, FOS, JUNB\u003c/em\u003e). Gene set enrichment analysis (GSEA) of the MSigDB Hallmarks gene sets revealed significant enrichment for allograft rejection in FRC-like fibroblasts (Supplementary Figure 4E).\u003c/p\u003e\n\u003cp\u003eIn theiCAF lineage (2), pseudotime analysis revealed increased expression of genes associated with TNF-alpha signalling via NF-\u0026kappa;B (Hallmarks; p.adjust \u0026lt;0.0001), Epithelial Mesenchymal transition (Hallmarks; p.adjust \u0026lt;0.0001), inflammatory response (Hallmarks; p.adjust \u0026lt;0.0001) and JAK-STAT signalling pathway (KEGG; p.adjust \u0026lt;0.01) (Figure 2F; Supplementary Table 4). GSEA showed enrichment for Glycolysis, Hypoxia, inflammatory response and TNF-alpha ignalling via NK-kB (Supplementary Figure 4E).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIntriguingly, while the chief inflammatory pathway, NF-\u0026kappa;B, was associated with both iCAF and FRC-like inflammatory phenotypes, the corresponding genes differed. FRC-like NF-\u0026kappa;B genes (e.g., \u003cem\u003eCCL21, CCL19, TNFSF13B\u003c/em\u003e) are specifically associated with the alternative NK-\u0026kappa;B pathway, commonly triggered through lymphotoxin, LIGHT, CD40-L and BAFF, and are related to lymphoid organ development and adaptive immunity [19]. Conversely\u003cem\u003e,\u0026nbsp;\u003c/em\u003eiCAF NF-\u0026kappa;B genes (e.g., \u003cem\u003eCXCL1, CXCL8\u003c/em\u003e) are generally associated with the classical NK-\u0026kappa;B pathway, typically activated via IL-1, TNF-\u0026alpha; or LPS and associated with inflammation and innate immunity [19]. Accordingly, we examined transcription factor (TF) activity in FRC-like and iCAF subsets by assessing the transcriptomic \u0026lsquo;footprint\u0026rsquo; of active transcription factors using the DoRothEA database [20]. FRC-like fibroblasts showed strong activity for RELB and NFKB2 (p100/p52), again providing evidence for alternative NF-\u0026kappa;B pathway activation through NF-\u0026kappa;B RelB-p52 complexes, wherea\u003cem\u003es\u0026nbsp;\u003c/em\u003etop iCAF active transcription factors included RELA (p65), CEBPB and JUN/FOSL1 (AP1; Figure 2G).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFRC-like fibroblasts are found within TLS and colocalise with B-cells and CD4+ T-cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFRC-like fibroblasts and iCAF possessed distinct inflammatory cytokine profiles (Figure 2H); FRC-like cytokines were associated with lymphocyte recruitment, proliferation, and survival [e.g., \u003cem\u003eCCL19/21, IL7/15, TNFSF14\u003c/em\u003e (LIGHT), \u003cem\u003eTNFSF13B\u0026nbsp;\u003c/em\u003e(BAFF), \u003cem\u003eCXCL13\u003c/em\u003e], with iCAF-specific cytokines related to myeloid/ granulocyte recruitment and differentiation (\u003cem\u003eCXCL1/5/6/8, CSF2/3\u003c/em\u003e). This, in addition to their different regulatory pathways, suggested that FRC-like fibroblasts andiCAF were likely associated with distinct immunological niches. To investigate this, we first performed correlative sample-level analysis on scRNA-Seq data using the previously identified immune cell subsets in the integrated HNSCC scRNA-Seq dataset (Supplementary Figure 2). We followed this with spatial transcriptomics (Visium 10x) analysis; using the annotated scRNA-Seq data as a reference to derive cell-type specific gene signatures that were used to deconvolute cell types present within each 55\u0026micro;m spot [21,22]. Applying this integrative approach for myCAF and universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts, identified previously described spatial and cellular relationships (Extended Data Figure 1 and 2).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn sample-level scRNA-Seq correlations of tumours, FRC-like fibroblasts positively correlated with various B-cell subsets [including cycling B-cells, \u003cem\u003eFCRL4\u003c/em\u003e+ B-cells and germinal centre (GC) B-cells], plasma cells, \u003cem\u003eKIT+\u003c/em\u003e NK-cells (Figure 3A), with high correlation with IgM expressing B/plasma cells. FRC-like fibroblasts also correlated with CD4+ T follicular helper (Tfh) cells. Spatial transcriptomic analysis confirmed that each tumour contained multiple fibroblast subsets that were spatially discrete (82.4% of fibroblast-containing spots contained one subset only; Supplementary Figure 5). FRC-like fibroblasts colocalised with B-cells and CD4+ T-cells (non-Treg; p\u0026lt;0.0001), found either in focal areas containing high densities of B/CD4+ cells (non-Treg) (HPV+ve/-ve HNSCC) or occasionally more widespread in two HPV+ve samples (59%/23% total spots) but still colocalising with large numbers of B/CD4+ T-cells (Figure 3BC; Supplementary Figure 6A, B; Supplementary Figure 7A). There was also a spatial correlation between FRC-like fibroblasts and plasma cells, Tregs and CD8+ T-cells (Figure 3B; Supplementary Figure 6A, B). PDPN is commonly utilised as a pan-fibroblast marker and has been specifically employed to identify FRC in lymph nodes [23]. MxIHC on spatial transcriptomic-determined FRC-like regions of interest (ROI) confirmed the presence of PDPN+/CD31- fibroblasts colocalising with CD20+ and CD4+ cells (Figure 3D;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSupplementary Figure 7B) within and surrounding CD21+ follicular dendritic cells, along with high densities of B-cells, suggesting formation of tertiary lymphoid structures (TLS). We therefore compared TLS and FRC-like enrichment in spatial transcriptomic data using different TLS gene signatures [25]. Spot deconvolution (RCTD) showed that FRC-like fibroblasts significantly correlated with module enrichment scores for various TLS signatures that have been used in several solid cancer studies (Spearman\u0026rsquo;s r \u0026ge; 0.5, p\u0026lt;0.0001; Figure 3E, Supplementary Figure 7C, D, E; [24,26\u0026ndash;28]).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFRC-like fibroblasts are regulated via LT\u0026beta;R signalling.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next investigated potential interactions in the FRC-like niche by examining ligands and receptors that were differentially expressed in Visium spots that contained FRC-like fibroblasts (spots containing at least 5% FRC-like cells imputed by RCTD deconvolution; Supplementary Figure 8A; Supplementary Table 5). LT\u0026beta;R-binding ligands [\u003cem\u003eTNFSF14\u0026nbsp;\u003c/em\u003e(LIGHT), \u003cem\u003eLTB\u003c/em\u003e, \u003cem\u003eLTA, CD40LG\u003c/em\u003e] were amongst cytokine activity-possessing ligands that were spatially associated, expressed by highly correlating cell types (Figure 3F; Supplementary Figure 8B) and known to stimulate alternative NF-\u0026kappa;B pathway activation (consistent with previous pathway/TF analysis). LIGHT and CD40LG were top ligands inferred via NicheNet [29] (Supplementary Figure 8C). Lymphotoxin was highly expressed in B-cells, T-cells and DCs (Supplementary Figure 8D), whereas LIGHT was expressed by FRC-like fibroblasts. In common with FRC-like fibroblasts, most other fibroblast subsets also expressed receptors for these (and other) inflammatory ligands (Supplementary Figure 8E), highlighting the potential for immunological plasticity in these cells depending on ligand availability.\u003c/p\u003e\n\u003cp\u003eWe then assessed the ability of the LT\u0026beta;R-binding ligand lymphotoxin \u0026alpha;1\u0026beta;2 (LT) to regulate the FRC-like phenotype in cultured primary oral fibroblasts (NOFs). Treatment with LT in combination with an ALK5 (TGFBR1) inhibitor induced FRC-like-specific genes \u003cem\u003eCCL19\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001)\u003cem\u003e, CCL21\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001), \u003cem\u003eSPIB\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001), \u003cem\u003eRBP5\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.01; Figure 3G). This was confirmed in further primary NOF cultures (n=7; \u003cem\u003eCCL19,\u0026nbsp;\u003c/em\u003ep\u0026lt;0.01; \u003cem\u003eCCL21\u003c/em\u003e, p\u0026lt;0.05; \u003cem\u003eSPIB\u003c/em\u003e, p\u0026lt;0.01; \u003cem\u003eRBP5,\u0026nbsp;\u003c/em\u003ep\u0026lt;0.001; Figure 3H).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eiCAF colocalise with inflammatory monocytes and neutrophils.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eiCAF correlated strongly with a subset of \u003cem\u003eCD14\u003c/em\u003e+ \u003cem\u003eIL1B\u003c/em\u003e\u003csup\u003ehigh\u003c/sup\u003e inflammatory monocytes (Spearman\u0026rsquo;s r=0.72, p\u0026lt;0.0001; Figure 4A; Supplementary Figure 9A). Spatial transcriptomic analysis showed that iCAF were spatially distinct from myCAF (Supplementary Figure 5), located primarily at the tumour periphery, particularly towards the tumour surface. iCAF colocalised with monocytes and neutrophils (p\u0026lt;0.0001), also frequently found at the periphery of tumours (Figure 4B, C; Supplementary Figure 6A, B; Supplementary Figure 9C). MxIHC on iCAF ROI (identified through spatial transcriptomics deconvolution) showed that these areas contained PDPN+/CD31- fibroblasts and were associated with disruptions in surface epithelium (pan-cytokeratin) and CD68+, CD14+ and MPO+ myeloid cells (Figure 4D; Supplementary Figure 9C). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eiCAF are regulated via IL1\u003c/strong\u003e\u003cstrong\u003eb\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and TNF\u003c/strong\u003e\u003cstrong\u003ea\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, we examined the expression of spatially located ligands in cell types highly correlating with iCAF (Supplementary Table 5). Top ligands associated with cytokine activity in the iCAF-niche included \u003cem\u003eCXCL8, IL1A, IL6, OSM, IL11\u003c/em\u003e and particularly \u003cem\u003eIL1B\u0026nbsp;\u003c/em\u003ewhich was highly expressed by inflammatory monocytes (Figure 4E; Supplementary Figure 9D). We also inferred ligand regulatory activity using NicheNet, which highlighted IL1B and IL1A as top spatially defined ligands with iCAF (gene set) regulatory potential (Supplementary Figure 9E). Indeed, myeloid cells (monocytes, neutrophils, macrophage) expressed highest levels of these ligands (\u003cem\u003eIL1B, IL1A, OSM\u003c/em\u003e), supportive of iCAF associating with a myeloid niche (Supplementary Figure 8D).\u003c/p\u003e\n\u003cp\u003eWe tested the potential of these ligands for regulating theiCAF phenotype in NOFs. In addition to IL-1b, we included TNF-a due to pathway and TF enrichment for classical NF-\u0026kappa;B signalling and high expression in monocytes [although \u003cem\u003eTNF\u003c/em\u003e was not spatially differentially expressed in theiCAF niche, likely resulting from expression in multiple immune cell types (Supplementary Figure 9D)]. NOFs were treated with IL-1b (1ng/mL) and TNF-a (1ng/mL), either alone or in combination. We also included TGF-b1 (4ng/mL), a central regulator of the myCAF phenotype for reference. Both IL-1b and TNF-a induced expression of \u003cem\u003eIL6\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001), \u003cem\u003eMMP3\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001) and \u003cem\u003eMME\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.01; Figure 4F), but with limited upregulation of \u003cem\u003eIL11\u003c/em\u003e. However, combining IL-1b\u0026nbsp;with TNF-a\u0026nbsp;increased expression of all inflammatory marker genes compared with individual treatments (\u003cem\u003eIL6\u003c/em\u003e, p\u0026lt;0.0001; \u003cem\u003eMME\u003c/em\u003e, p\u0026lt;0.05), including \u003cem\u003eIL11\u003c/em\u003e, which increased 16-fold (log\u003csub\u003e2\u003c/sub\u003eFC = 4) compared to IL1b\u0026nbsp;alone (p\u0026lt;0.05). This was validated in several primary fibroblast cultures (\u003cem\u003en=9\u003c/em\u003e), where TNF-a and IL-1b robustly induced iCAF gene expression (Figure 4G). \u003cem\u003eIL11\u0026nbsp;\u003c/em\u003ewas also induced by TGF-b1 (p\u0026lt;0.0001); conversely, \u003cem\u003eACTA2\u0026nbsp;\u003c/em\u003e(aSMA; a myCAF marker) was induced by TNF-a/IL-1b\u0026nbsp;(p\u0026lt;0.001); while other iCAF (\u003cem\u003eIL6, MMP3, MME\u003c/em\u003e) and myCAF (\u003cem\u003ePOSTN, TAGLN, COL1A1\u003c/em\u003e) genes were more specifically regulated by TNF-a/IL-1b\u0026nbsp;and TGF-b1\u0026nbsp;respectively (Supplementary Figure 9F).\u003c/p\u003e\n\u003cp\u003eGiven the iCAF/monocyte spatial relationship, we investigated whether monocytes regulated the iCAF phenotype. NOF treated with conditioned medium from monocytes activated with LPS induced upregulated expression of iCAF genes (\u003cem\u003eIL6\u003c/em\u003e, p\u0026lt;0.0001; \u003cem\u003eMMP3\u003c/em\u003e, p\u0026lt;0.0001, \u003cem\u003eIL11\u003c/em\u003e, p\u0026lt;0.001; \u003cem\u003eMME\u003c/em\u003e, p\u0026lt;0.001; Figure 4H). Similar to TNF-a/IL-1b\u0026nbsp;treatment, \u003cem\u003eACTA2\u0026nbsp;\u003c/em\u003ewas also increased (p\u0026lt;0.05; Supplementary Figure 9G).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePan-Cancer Fibroblast analysis identifies conserved and semi-conserved inflammatory fibroblast phenotypes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo compare the HNSCC fibroblast phenotypes with other cancers, we generated a scRNA-Seq (10x Chromium) pan-cancer fibroblast atlas (PCFA) from seven cancer types: HNSCC, pancreatic, breast, lung, colon, oesophageal and gastric cancers (Figure 5A). Only datasets containing both tumour and normal samples were included to differentiate between normal (steady-state) and cancer-associated phenotypes (Figure 5B, C, D; Supplementary Figure 10A, B).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe PCFA (86,414 fibroblasts; 376 samples), revealed 16 populations, including broadly conserved, as well as tissue-specific subsets. Where possible, fibroblast subgroups were labelled using designations from previous studies (Figure 5C; Supplementary Figure 10A-D; Supplementary Table 6). Conserved phenotypes in normal tissue included universal (\u003cem\u003ePI16+\u003c/em\u003e) Fib, stress-response Fib (\u003cem\u003eDNAJB1+,\u0026nbsp;\u003c/em\u003e\u003cem\u003eHSPH1+, HSPA1A+;\u003c/em\u003e which incorporated the head \u0026amp; neck \u003cem\u003eADH1B+\u003c/em\u003e fibroblasts) and \u003cem\u003eCXCL14+ CFD+\u003c/em\u003e Fib (Supplementary Figure 10B). Tissue-specific subsets included \u003cem\u003eCXCL8\u003c/em\u003e+ breast fibroblasts [most abundant subset (\u0026gt;85%) in normal breast tissue], \u003cem\u003eNPNT\u003c/em\u003e+ (alveolar) lung fibroblasts and \u003cem\u003eF3\u003c/em\u003e+/\u003cem\u003eADAMDEC1\u003c/em\u003e+ colonic/gastric fibroblasts (all found in normal and tumour samples; Figure 5D, E; Supplementary Figure 10B). The FRC-like cluster was mostly composed of cells from the head \u0026amp; neck with a contribution from lung (from normal tissues and cancers; Figure 5D, E).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOnly subsets found exclusively in cancers were termed CAF (Figure 5D). Conserved common CAF clusters included myCAF (the most abundant CAF subset in all tumour types except gastric cancer; Supplementary Figure 10A, C), \u003cem\u003eIGF1+\u003c/em\u003e CAF and proto-CAF. Other clusters had increased frequency in certain cancer types or were rare across cancers. For example, three clusters had a larger contribution from pancreatic tumour/normal samples (\u003cem\u003eHAS1+\u003c/em\u003e CAF, metabolic CAF [meCAF] and \u003cem\u003eC7+\u0026nbsp;\u003c/em\u003eFib [Figure 5E; Supplementary Figure 10A, B]). meCAF expressed markers of glycolysis and hypoxia (\u003cem\u003eENO1, ENO2, NDRG1, PGK1, LDHA, SLC2A1\u003c/em\u003e) consistent with a previously described population [30]. Although this cluster was observed pan-cancer, outside of pancreatic samples it was generally found at very low (Supplementary Figure 10A, C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe HNSCC iCAF resided in the \u003cem\u003eIL11+\u003c/em\u003e CAF cluster\u0026nbsp;(Supplementary Figure 10D); this semi-conserved phenotype was one of the most abundant fibroblast subpopulations in HNSCC, colorectal carcinoma (CRC) and oesophageal squamous cell carcinoma (ESCC) (Supplementary Figure 10C) but was not present in lung or breast cancers. Label transfer using the HNSCC myeloid cells as a reference to identify myeloid phenotypes in CRC/ESCC scRNA-Seq datasets showed that \u003cem\u003eIL11+\u003c/em\u003e CAF similarly correlated with \u003cem\u003eIL1B+\u003c/em\u003e inflammatory monocytes suggesting the same immunological niche present in different cancer types (Supplementary Figure 10E).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026lsquo;iCAF\u0026rsquo; gene signature highlights different normal fibroblast and CAF populations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGiven the restriction of \u003cem\u003eIL11+\u003c/em\u003e CAF to certain cancers, we performed enrichment analysis using a previously described PDAC iCAF gene signature [5] to investigate whether other PCFA subgroups expressed iCAF markers. This highlighted several fibroblast clusters from both normal and tumour tissues; in normal tissues these included Stress-response Fib, \u003cem\u003eCXCL8+\u003c/em\u003e Breast Fib and universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts; in tumours, \u003cem\u003eIL11+\u003c/em\u003e CAF, \u003cem\u003eIGF1+\u003c/em\u003e CAF and proto-CAF (Figure 6A). We examined the iCAF signature-enriched clusters in tumours in more detail (Supplementary Figure 10A). \u003cem\u003eIGF1+\u003c/em\u003e CAF expressed all iCAF markers (Supplementary Figure 10F; [5], and had the greatest transcriptional resemblance to PDAC iCAF (p\u0026lt;0.0001, Fisher\u0026rsquo;s exact test; Supplementary Figure 10G).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eIGF1+\u003c/em\u003e CAF were present in most cancer types (e.g., breast, oesophageal, lung), including high levels in PDAC (comparable to myCAF in abundance;\u0026nbsp;Figure 6B; Supplementary Figure 10C). Transcriptionally, \u003cem\u003eIGF1+\u003c/em\u003e CAF differed considerably from \u003cem\u003eIL11+\u003c/em\u003e CAF (Figure 6C);\u0026nbsp;clustering close to universal (\u003cem\u003ePI16+\u003c/em\u003e) Fib and maintaining expression of universal (\u003cem\u003ePI16+\u003c/em\u003e) Fib genes (\u003cem\u003ePI16, CFD, COL14A1\u003c/em\u003e; Figure 6C). From the initial HNSCC analysis, ~50% of (universal) \u003cem\u003ePI16+\u003c/em\u003e labelled fibroblasts from tumour samples were labelled as \u003cem\u003eIGF1+\u003c/em\u003e CAF in the larger PCFA dataset, showing that universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts from tumours show evidence of activation and inflammatory changes (including expression of \u003cem\u003eFAP, COL1A1, IGF1\u003c/em\u003e; Supplementary Figure 4C; Supplementary Figure 10D).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGiven the transcriptional similarity between universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts and \u003cem\u003eIGF1+\u003c/em\u003e CAF, we examined DEGs between these phenotypes (Supplementary Table 7). \u003cem\u003eIGF1+\u003c/em\u003e CAF showed increased expression of both inflammatory (\u003cem\u003eCXCL8, CXCL2, CCL5\u003c/em\u003e) and myofibroblastic genes (\u003cem\u003ePOSTN, MMP11, COL1A1\u003c/em\u003e), but these were expressed at lower levels than in other CAF subsets (Figure 6D). Inflammatory genes (\u003cem\u003eCXCL8, CSF3, CCL5;\u003c/em\u003e as well as \u003cem\u003eMMP1, INHBA\u003c/em\u003e and \u003cem\u003eMMP3),\u0026nbsp;\u003c/em\u003ewere most highly expressed in \u003cem\u003eIL11+\u003c/em\u003e CAF (which had highest expression of all iCAF marker genes; \u003cem\u003eIL6\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001), \u003cem\u003eCXCL8\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001), \u003cem\u003eIL11\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.0001); Figure 6E). ECM genes (\u003cem\u003eCOL1A1, COL10A1, POSTN)\u003c/em\u003e were most highly expressed in myCAF. Thus, although, \u003cem\u003eIGF1+\u003c/em\u003e CAF upregulate both inflammatory and myofibroblastic genes, these are expressed at markedly lower levels than \u003cem\u003eIL11+\u003c/em\u003e CAF and myCAF respectively.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFRC-like fibroblasts are present across cancers at low frequency and are associated with positive response to immunotherapy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAcross the seven cancers included in the PCFA, FRC-like cells were a relatively rare fibroblast population, with 77.8% of cells contributed from head and neck (62.1% normal; 15.6% tumour). Within tumour samples, FRC-like fibroblasts were found with highest average relative abundance in HNSCC (11.2%) followed by lung cancer (1.7%) but were present in all cancers at low frequencies (Figure 7A). FRC-like cells were detected in several normal tissue sites, likely representing mucosa-associated lymphoid tissues (MALT; Figure 7A). In breast, oesophageal, lung and pancreatic cancers, FRC-like fibroblasts were enriched in tumour samples, while were less common in HNSCC, gastric and colon cancer relative to normal samples.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGiven the potential role of apCAF in modulating anti-tumour immunity [5,6,31], we analysed the PCFA to determine which fibroblast subtypes expressed MHC-II genes. Although FRC-like fibroblasts showed the highest expression of MHC-II genes (Supplementary Figure 11A), unlike murine fibroblasts, expression was not restricted to a specific phenotype and was found in several clusters (Supplementary figure 11A) similar to other human studies [31,32]. Compared to professional antigen-presenting cells, FRC-like fibroblasts show reduced expression and an incomplete repertoire of MHC-II genes (Supplementary Figure 11B).\u003c/p\u003e\n\u003cp\u003eThe positive correlation between the FRC-like fibroblast and a TLS gene signature [24] across bulk RNA-Seq datasets (Figure 7B), suggested that an FRC-like-containing immune hub exists in different cancers. Given that TLS have been linked with positive response to checkpoint immunotherapy in several cancer types [24,25,33] we investigated whether FRC-like fibroblasts were similarly associated. First, we utilised a dataset (GSE159067) consisting of pre-treatment samples from 102 patients with advanced HNSCC treated with anti-PD-1/PD-L1 immunotherapy. Samples were scored using ssGSEA for fibroblast subset-specific genes (see methods; Supplementary Table 8). We found patients with higher FRC-like scores had significantly improved survival (p\u0026lt;0.01) (Figure 7C). In contrast, iCAF were associated with significantly poorer survival (Supplementary Figure 11C). We performed the same analysis on datasets from lung cancer and melanoma patients. Similarly, we observed higher FRC-like scores were associated with significantly improved survival in ICI treated patients with lung cancer (p\u0026lt;0.01; anti-PD1/PD-L1) and melanoma (p\u0026lt;0.001; anti-CTLA4/PD-1); unlike HNSCC, iCAF were not prognostic (Figure 7D; Supplementary Figure 11D, E, F).\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eSingle cell analysis of 'immune-hot\u0026rsquo; and \u0026lsquo;immune-cold\u0026rsquo; HNSCC tumours, including normal mucosa (not included in recent HNSCC scRNA-Seq datasets; [17,34] identified six major fibroblast subgroups; universal \u003cem\u003e(PI16+\u003c/em\u003e) fibroblasts, \u003cem\u003eADH1B\u003c/em\u003e\u0026thinsp;+\u0026thinsp;fibroblasts and FRC-like fibroblasts were present in normal tissue and tumours. myCAF, \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;iCAF and proto-CAF were limited to tumours. Of these, proto-CAF fibroblasts were likely a transition state as cells differentiated towards myCAF/iCAF phenotypes. HPV\u0026thinsp;+\u0026thinsp;ve and HPV-ve cancers contained mixtures of all fibroblast subgroups, although proportions and relative abundance varied in individual tumours. In support of our hypothesis, we found significantly higher numbers of FRC-like fibroblasts expressing \u003cem\u003eCCL19\u003c/em\u003e and \u003cem\u003eCCL21\u003c/em\u003e in HPV\u0026thinsp;+\u0026thinsp;ve cancers. These tumours were situated in the oropharynx, an anatomical site that contains secondary lymphoid organs (SLO; tonsils) and considerable numbers of FRC-like cells were also present in matched normal oropharyngeal tissue. However, scRNASeq did not identify FRC-like fibroblasts in HPV-ve HNSCC tumours at the same site, suggesting either that FRC-like fibroblasts are retained within HPV\u0026thinsp;+\u0026thinsp;ve tumours or arise \u003cem\u003ede novo\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eIn lymph nodes, FRC play a central role in structural organisation; attracting and maintaining T-cells, supporting B-cell survival, promoting dendritic cell migration, and controlling permeability of high endothelial venules [23]. Similar cells arise in autoimmune disease, where they transdifferentiate from local fibroblasts and play a central role in supporting TLS formation and maintenance [35]. Consistent with this, within tumours we found FRC-like fibroblasts located with B-cells and (Tfh) CD4 T-cells in TLS structures, correlating with several well-described TLS gene signatures. Development of mature FRCs from precursor cells in SLO is driven by LTβR signalling [23], and TLS-forming FRC-like \u0026lsquo;immunofibroblasts\u0026rsquo; have been shown to be regulated by LTα1b2 and IL22 in Sjogren\u0026rsquo;s syndrome [35]. We found FRC-like fibroblasts to be similarly regulated through non-canonical NF-κB signalling, showing strong activity for transcription factors RELB and NKFB2 (p100/p52). LTβR signals via alternative NF-κB, binding to ligands such LTα1β2 and LIGHT [36]. \u003cem\u003eLTA\u003c/em\u003e, \u003cem\u003eLTB\u003c/em\u003e, LIGHT/\u003cem\u003eTNFSF14\u003c/em\u003e (all signalling via LTβR) and \u003cem\u003eCD40L\u003c/em\u003e were all spatially associated with FRC-like cells, expressed by B- and T-cells (with LIGHT expressed by FRC-like), potentially driving the FRC-like phenotypic transition. These ligands have all been strongly implicated in TLS neogenesis [37]. Notably, \u003cem\u003eLTBR\u003c/em\u003e was expressed by all fibroblast subsets (including CAF), suggesting a common capability to respond to LTBR ligands. \u003cem\u003eIn vitro\u003c/em\u003e, treatment of primary fibroblasts with lymphotoxin induced FRC-like genes (\u003cem\u003eCCL19, CCL21, SPIB\u003c/em\u003e) which was enhanced by inhibiting TGFβ signalling.\u003c/p\u003e \u003cp\u003eTLS have been reported in a variety of cancers including HNSCC and NSCLC [38,39], but their occurrence likely differs between cancer types. This perhaps is reflected in our pan-cancer fibroblast atlas; FRC-like cells were found with highest average relative abundance in HPV\u0026thinsp;+\u0026thinsp;ve HNSCC (11.2%) followed by lung cancer (1.7%). Rarer phenotypes are under-represented in scRNASeq and the distinct FRC-like clusters in HNSCC and pan-cancer analysis was likely aided by inclusion of an FRC-like-rich cancer type (HPV\u0026thinsp;+\u0026thinsp;ve HNSCC). The pan-cancer analysis demonstrated that FRC-like fibroblasts are present in all cancer types but with low abundance, and thus probably do not cluster discretely when datasets are analysed separately. It is also noteworthy that we detected FRC-like fibroblasts in HPV-ve Visium sections, but not using scRNA-Seq.\u0026nbsp;This highlights the power of deriving cell type specific gene signatures from scRNA-Seq data and using this to deconvolute spatial transcriptomic analysis of tissue sections: enabling far greater numbers of cells to be profiled and avoiding the challenges associated with isolating stromal cells from tissue through disaggregation.\u003c/p\u003e \u003cp\u003eThe presence of TLS is associated with favourable prognosis in many cancer types, including HNSCC [25,38], in part reflecting the presence of an ongoing, antigen-dependent immune response [39]. Moreover, the presence of TLS has been shown to predict for response to immunotherapy response in several cancer types [24,33]. Our analysis of HNSCC, lung cancer and melanoma patients treated with immune checkpoint blockade shows that high levels of FRC-like fibroblasts in tumours are associated with significantly improved survival suggesting that higher levels of FRC-like fibroblasts may identify likely responders. Furthermore, given their central role in TLS organisation and maintenance, generating FRC-like fibroblasts could be an attractive therapeutic strategy to potentiate immunotherapy response.\u003c/p\u003e \u003cp\u003eRecent studies have identified iCAF in several tumour types, including pancreatic cancer and breast cancer [5,6,30,40], using a variety of markers that encompass inflammatory cytokines and other genes (\u003cem\u003eCXCL1, CXCL8, CXCL12, IL6, CFD, DPT\u003c/em\u003e) [5]. Using a frequently used iCAF gene signature [5], we identified several fibroblast clusters enriched in the pan-cancer analysis that shared expression of genes such as \u003cem\u003eCXCL8, CXCL1, CXCL2, IL6\u003c/em\u003e, with some phenotypes present in normal tissue (e.g., \u003cem\u003eCXCL8\u003c/em\u003e\u0026thinsp;+\u0026thinsp;breast fibroblasts; a similar fibroblast population has been described previously in breast tissue as \u0026lsquo;Fibro-major\u0026rsquo; [41]). Of the iCAF signature expressing subsets specific to cancers, \u003cem\u003eIGF1\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF and \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF were abundant in tumours. \u003cem\u003eIGF1\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF were present in all tumour types and expressed all iCAF markers originally identified in PDAC. \u003cem\u003eIGF1\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF were transcriptionally similar to universal (\u003cem\u003ePI16+\u003c/em\u003e) fibroblasts, maintaining expression of universal fibroblast genes (\u003cem\u003ePI16, PLA2G2A, CFD\u003c/em\u003e) but showed evidence of activation (expression of \u003cem\u003eFAP, COL1A1, IGF1\u003c/em\u003e) and expression of iCAF markers (\u003cem\u003eIL6, CXCL8, CXCL2\u003c/em\u003e). Within the HNSCC dataset, ~\u0026thinsp;50% of PI16 labelled fibroblasts from tumour samples were labelled as \u003cem\u003eIGF1\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF in the pan-cancer analysis, suggesting that this phenotype is an early/low activation phenotype consistent with previous studies [1,32]. \u003cem\u003eIGF1\u003c/em\u003e has been reported to mark iCAF in several cancer types [40,42], and this low activation subset probably represents the most commonly referenced \u0026lsquo;iCAF\u0026rsquo; phenotype in the literature currently.\u003c/p\u003e \u003cp\u003e \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF expressed significantly higher levels of inflammatory genes compared to \u003cem\u003eIGF1\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF. These were prevalent in GI tumours (HNSCC, CRC, ESCC), but not detected in breast or lung cancers. Although \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF could be found with epithelial cells (unlike previous work highlighting iCAF to be distant to epithelial cells; [43]), they especially correlated with inflammatory monocytes and neutrophils. A recent large scRNA-Seq analysis of colorectal tumours revealed a \u0026lsquo;myeloid-cell-attracting\u0026rsquo; hub consisting of inflammatory monocytes, neutrophils and \u003cem\u003eMMP3\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF [44] hypothesised to be associated with tissue damage and microbial products. An association between inflammatory fibroblasts and myeloid cells has also been described in autoimmune inflammatory bowel disease [45] and periodontitis [46], suggesting this inflammatory niche exists beyond cancer.\u003c/p\u003e \u003cp\u003eGene enrichment and pseudotime analyses identified canonical NF-κB and JAK/STAT signalling as regulating the \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF phenotype, with IL-1β and TNF-α as likely ligands. Treatment of normal oropharyngeal fibroblasts with IL1β and TNF-α combination upregulated genes expressed by this phenotype \u003cem\u003ein vivo\u003c/em\u003e. Consistent with the spatial analysis, treatment of fibroblasts with conditioned media from inflammatory monocytes treatment produced similar results. The immunological role of \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF in cancer is not clear, but highly expressed genes, including IL-6 cytokine family (\u003cem\u003eIL6, IL11, OSM\u003c/em\u003e) are associated with immunotherapy resistance [12]. The \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF subset was associated with significantly poorer overall survival in immunotherapy-treated HNSCC patients.\u003c/p\u003e \u003cp\u003eWe did not detect \u003cem\u003ebona fide\u003c/em\u003e apCAF, either in HNSCC or pan-cancer. apCAF expressing MHC-II/\u003cem\u003eCD74\u003c/em\u003e were originally identified In KPC PDAC murine models [5] and have subsequently been reported in a several human cancers e.g., breast cancer [40]. However, other studies have failed to identify apCAF. In PDAC, these have been shown to arise from mesothelial cells acquiring fibroblastic features through IL1 and TGFβ signalling [47]; this may explain the absence of apCAF in this study where the curation of the PCFA strictly excluded cells expressing markers denoting alternative cell types, including mural cells and mesothelial cells. MHC-II-related genes were expressed in several fibroblast clusters, with FRC-like cells having highest expression (albeit far lower than professional antigen presenting cells). Dominguez and colleagues similarly found all human CAF to express \u003cem\u003eCD74/HLA-DRA\u003c/em\u003e [31]. In breast cancer, a \u003cem\u003eCD74\u003c/em\u003e-expressing CAF cluster (IFNγ-iCAF) has been depicted as apCAF; this population also expressed \u003cem\u003eCCL19\u003c/em\u003e [6].\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, single cell analysis of HNSCC identifies inflammatory fibroblast subsets that are associated with distinct immune cell niches: FRC-like with CD4\u0026thinsp;+\u0026thinsp;T-cells and B-cells; \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF with inflammatory monocytes and neutrophils. HPV\u0026thinsp;+\u0026thinsp;ve HNSCC contain significantly higher levels of FRC-like fibroblasts; their spatial location within TLS, and their positive association with immunotherapy response suggests that these cells support anti-tumour immunity. We also identify transcriptionally discrete iCAF phenotypes including a low activation/transition phenotype (\u003cem\u003eIGF1+\u003c/em\u003e), likely the predominant iCAF in the current literature, as well as a more highly inflammatory \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;CAF subtype found within cancers of the GI tract. Distinguishing between these phenotypes and dissecting functional differences will be important considerations going forwards. It is intriguing that immunological differences within tumours may be tied to fibroblast phenotypes, and the association of fibroblast subtypes with both negative and positive effects on anti-tumour immunity raises intriguing therapeutic possibilities.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e\u003cstrong\u003eCAF\u003c/strong\u003e: Cancer-associated fibroblast\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHNSCC\u003c/strong\u003e: Head and neck squamous cell carcinoma\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHPV\u003c/strong\u003e: Human papillomavirus\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eiCAF\u003c/strong\u003e: Inflammatory CAF\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFRC\u003c/strong\u003e: Fibroblastic reticular cell\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eECM\u003c/strong\u003e: Extracellular matrix\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003emyCAF\u003c/strong\u003e: Myofibroblastic CAF\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eapCAF\u003c/strong\u003e: Antigen-presenting CAF\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003emeCAF\u003c/strong\u003e: Metabolic CAF\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePDAC\u003c/strong\u003e: Pancreatic ductal adenocarcinoma\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTIL\u003c/strong\u003e: Tumour infiltrating lymphocytes\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGI\u003c/strong\u003e: Gastrointestinal\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUMAP\u003c/strong\u003e:\u0026nbsp;Uniform Manifold Approximation and Projection\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDEG\u003c/strong\u003e: Differentially expressed gene\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNSCLC\u003c/strong\u003e: Non-small cell lung cancer\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGSEA\u003c/strong\u003e: Gene set enrichment analysis\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003essGSEA\u003c/strong\u003e: single sample gene set enrichment analysis\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003escRNA-Seq\u003c/strong\u003e: single cell RNA-Sequencing\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMxIHC\u003c/strong\u003e: Multiplexed immunohistochemistry\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTfh\u003c/strong\u003e: T follicular helper\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGC\u003c/strong\u003e: Geminal centre\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTF\u003c/strong\u003e: Transcription factor\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eROI\u003c/strong\u003e: Region of interest\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRCTD\u003c/strong\u003e: Robust cell type deconvolution\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTLS\u003c/strong\u003e: Tertiary lymphoid structure\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLT\u003c/strong\u003e\u003cstrong\u003ebR\u003c/strong\u003e: Lymphotoxin beta receptor\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLT\u003c/strong\u003e:\u0026nbsp;lymphotoxin \u0026alpha;1\u0026beta;2\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNOF\u003c/strong\u003e: Normal primary oral fibroblast\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePCFA\u003c/strong\u003e: Pan-cancer fibroblast atlas\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFib\u003c/strong\u003e: Fibroblast\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCRC\u003c/strong\u003e:\u0026nbsp;Colorectal carcinoma\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eESCC\u003c/strong\u003e: Oesophageal squamous cell carcinoma\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMALT\u003c/strong\u003e:\u0026nbsp;Mucosa-associated lymphoid tissues\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMHC-II\u003c/strong\u003e: Major histocompatibility complex class II\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSLO\u003c/strong\u003e: Secondary lymphoid organ\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthical Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical approval for the study was obtained through the UK National Research Ethics Service (South Central - Hampshire B Research Ethics Committee) and written informed consent was obtained from all subjects (REC No. 09/H0501/90).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBHJ, GJT, CJH, JGB, wrote the manuscript. GJT, CJH, JGB, project led the study. BHJ, IT, MFSR, MJS, BRM, SM, AR, HZ, KL, AA, LL, LD, experimental design. ME, EVK, patient selection and sample acquisition. BHJ, IT, MFSR, MJS, BRM, SM, AR, HZ, KL, AA, LL, LD, ME, EVK, performed the research. IT, BHJ, tissue processing and isolation of primary fibroblasts. IT, MFSR, MJS, BRM, performed in vitro experiments. SM, AR, HZ, KL, AA, LL, LD, generated spatial transcriptomics and MxIHC (Phenocycler) data. IT carried out single cell RNA Sequencing. BHJ, KL, CJH, performed bioinformatics analysis. All authors reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was funded through a Cancer Research UK Programme grant (DRVRPG-Jun\\100004; GJT/CJH), a Cancer Research UK Centres Network Accelerator Award Grant (A20256; GJT), the CRUK and NIHR Experimental Cancer Medicine Center (ECMC) Southampton (A15581 \u0026amp; ECMCQQR-2022/100018), a Pathological Society of Great Britain \u0026amp; Ireland Visiting Fellowship (grant no. VF 1002 03; GJT/MFSR), Sao Paulo Research Foundation (FAPESP) grant 2022/05364-0 (MFSR) and a Project grant from Gilead Sciences Inc. BHJ was funded through an AstraZeneca PhD studentship\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll research data will be submitted on review. Sequencing data generated in this study (scRNA-Seq and Spatial Transcriptomics) will be available in the Gene Expression Omnibus (GEO). Code will be available on Github. MxIHC (Pheno-cycler) data will be made available. The data analysed in this study obtained from GEO: GSE164690, GSE161529, GSE150290, GSE160269, GSE178341, GSE129455, GSE159067 and GSE161537. PRJCA001063 was retrieved from the Genome Sequence Archive. PRJEB23709 was accessed through http://tide.dfci.harvard.edu/download/.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors gratefully acknowledge the support of the Faculty of Medicine Tissue Bank, University of Southampton.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBuechler MB, Pradhan RN, Krishnamurty AT, Cox C, Calviello AK, Wang AW, et al. Cross-tissue organization of the fibroblast lineage. Nature. 2021;593:575\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eDavidson S, Coles M, Thomas T, Kollias G, Ludewig B, Turley S, et al. Fibroblasts as immune regulators in infection, inflammation and cancer. Nat Rev Immunol. 2021;21:704\u0026ndash;17. \u003c/li\u003e\n\u003cli\u003eHanley CJ, Waise S, Ellis MJ, Lopez MA, Pun WY, Taylor J, et al. Single-cell analysis reveals prognostic fibroblast subpopulations linked to molecular and immunological subtypes of lung cancer. Nat Commun. 2023;14:387. \u003c/li\u003e\n\u003cli\u003eBiffi G, Oni TE, Spielman B, Hao Y, Elyada E, Park Y, et al. IL1-Induced JAK/STAT Signaling Is Antagonized by TGF\u0026beta; to Shape CAF Heterogeneity in Pancreatic Ductal Adenocarcinoma. Cancer Discov. 2019;9:282\u0026ndash;301. \u003c/li\u003e\n\u003cli\u003eElyada E, Bolisetty M, Laise P, Flynn WF, Courtois ET, Burkhart RA, et al. Cross-Species Single-Cell Analysis of Pancreatic Ductal Adenocarcinoma Reveals Antigen-Presenting Cancer-Associated Fibroblasts. Cancer Discov. 2019;9:1102\u0026ndash;23. \u003c/li\u003e\n\u003cli\u003eKieffer Y, Hocine HR, Gentric G, Pelon F, Bernard C, Bourachot B, et al. Single-Cell Analysis Reveals Fibroblast Clusters Linked to Immunotherapy Resistance in Cancer. Cancer Discov. 2020;10:1330\u0026ndash;51. \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;24:1277\u0026ndash;89. \u003c/li\u003e\n\u003cli\u003eKrausgruber T, Fortelny N, Fife-Gernedl V, Senekowitsch M, Schuster LC, Lercher A, et al. Structural cells are key regulators of organ-specific immune response. Nature. 2020;583:296\u0026ndash;302. \u003c/li\u003e\n\u003cli\u003eDammeijer F, Gulijk M van, Mulder EE, Lukkes M, Klaase L, Bosch T van den, et al. The PD-1/PD-L1-Checkpoint Restrains T cell Immunity in Tumor-Draining Lymph Nodes. Cancer Cell. 2020;38:685-700.e8. \u003c/li\u003e\n\u003cli\u003eFord K, Hanley CJ, Mellone M, Szyndralewiez C, Heitz F, Wiesel P, et al. NOX4 Inhibition Potentiates Immunotherapy by Overcoming Cancer-Associated Fibroblast-Mediated CD8 T-cell Exclusion from Tumors. Cancer Res. 2020;canres;0008-5472.CAN-19-3158v2. \u003c/li\u003e\n\u003cli\u003eMariathasan S, Turley SJ, Nickles D, Castiglioni A, Yuen K, Wang Y, et al. TGF\u0026beta; attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature. 2018;554:544\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eTsukamoto H, Fujieda K, Miyashita A, Fukushima S, Ikeda T, Kubo Y, et al. Combined Blockade of IL6 and PD-1/PD-L1 Signaling Abrogates Mutual Regulation of Their Immunosuppressive Effects in the Tumor Microenvironment. Cancer Research. 2018;78:5011\u0026ndash;22. \u003c/li\u003e\n\u003cli\u003eCroft AP, Campos J, Jansen K, Turner JD, Marshall J, Attar M, et al. Distinct fibroblast subsets drive inflammation and damage in arthritis. Nature. 2019;570:246\u0026ndash;51. \u003c/li\u003e\n\u003cli\u003eGrauel AL, Nguyen B, Ruddy D, Laszewski T, Schwartz S, Chang J, et al. TGF\u0026beta;-blockade uncovers stromal plasticity in tumors by revealing the existence of a subset of interferon-licensed fibroblasts. Nat Commun. 2020;11:6315. \u003c/li\u003e\n\u003cli\u003eGalon J, Bruni D. Approaches to treat immune hot, altered and cold tumours with combination immunotherapies. Nat Rev Drug Discov. 2019;18:197\u0026ndash;218. \u003c/li\u003e\n\u003cli\u003eWard MJ, Thirdborough SM, Mellows T, Riley C, Harris S, Suchak K, et al. Tumour-infiltrating lymphocytes predict for outcome in HPV-positive oropharyngeal cancer. Br J Cancer. 2014;110:489\u0026ndash;500. \u003c/li\u003e\n\u003cli\u003eK\u0026uuml;rten CHL, Kulkarni A, Cillo AR, Santos PM, Roble AK, Onkar S, et al. Investigating immune and non-immune cell interactions in head and neck tumors by single-cell RNA sequencing. Nat Commun. 2021;12:7338. \u003c/li\u003e\n\u003cli\u003eHynes RO, Naba A. Overview of the matrisome--an inventory of extracellular matrix constituents and functions. Cold Spring Harb Perspect Biol. 2012;4:a004903. \u003c/li\u003e\n\u003cli\u003eBonizzi G, Karin M. The two NF-\u0026kappa;B activation pathways and their role in innate and adaptive immunity. Trends in Immunology. 2004;25:280\u0026ndash;8. \u003c/li\u003e\n\u003cli\u003eGarcia-Alonso L, Holland CH, Ibrahim MM, Turei D, Saez-Rodriguez J. Benchmark and integration of resources for the estimation of human transcription factor activities. Genome Res. 2019;29:1363\u0026ndash;75. \u003c/li\u003e\n\u003cli\u003eCable DM, Murray E, Zou LS, Goeva A, Macosko EZ, Chen F, et al. Robust decomposition of cell type mixtures in spatial transcriptomics. Nat Biotechnol. 2022;40:517\u0026ndash;26. \u003c/li\u003e\n\u003cli\u003eDong R, Yuan G-C. SpatialDWLS: accurate deconvolution of spatial transcriptomic data. Genome Biol. 2021;22:145. \u003c/li\u003e\n\u003cli\u003eFletcher AL, Acton SE, Knoblich K. Lymph node fibroblastic reticular cells in health and disease. Nat Rev Immunol. 2015;15:350\u0026ndash;61. \u003c/li\u003e\n\u003cli\u003eCabrita R, Lauss M, Sanna A, Donia M, Skaarup Larsen M, Mitra S, et al. Tertiary lymphoid structures improve immunotherapy and survival in melanoma. Nature. 2020;577:561\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eSaut\u0026egrave;s-Fridman C, Petitprez F, Calderaro J, Fridman WH. Tertiary lymphoid structures in the era of cancer immunotherapy. Nat Rev Cancer. 2019;19:307\u0026ndash;25. \u003c/li\u003e\n\u003cli\u003eCoppola D, Nebozhyn M, Khalil F, Dai H, Yeatman T, Loboda A, et al. Unique ectopic lymph node-like structures present in human primary colorectal carcinoma are identified by immune gene array profiling. Am J Pathol. 2011;179:37\u0026ndash;45. \u003c/li\u003e\n\u003cli\u003eGu-Trantien C, Loi S, Garaud S, Equeter C, Libin M, de Wind A, et al. CD4+ follicular helper T cell infiltration predicts breast cancer survival. J Clin Invest. 2013;123:2873\u0026ndash;92. \u003c/li\u003e\n\u003cli\u003eHennequin A, Derang\u0026egrave;re V, Boidot R, Apetoh L, Vincent J, Orry D, et al. Tumor infiltration by Tbet+ effector T cells and CD20+ B cells is associated with survival in gastric cancer patients. Oncoimmunology. 2016;5:e1054598. \u003c/li\u003e\n\u003cli\u003eBrowaeys R, Saelens W, Saeys Y. NicheNet: modeling intercellular communication by linking ligands to target genes. Nat Methods. 2020;17:159\u0026ndash;62. \u003c/li\u003e\n\u003cli\u003eMa C, Yang C, Peng A, Sun T, Ji X, Mi J, et al. Pan-cancer spatially resolved single-cell analysis reveals the crosstalk between cancer-associated fibroblasts and tumor microenvironment. Mol Cancer. 2023;22:170. \u003c/li\u003e\n\u003cli\u003eDominguez CX, M\u0026uuml;ller S, Keerthivasan S, Koeppen H, Hung J, Gierke S, et al. Single-Cell RNA Sequencing Reveals Stromal Evolution into LRRC15+ Myofibroblasts as a Determinant of Patient Response to Cancer Immunotherapy. Cancer Discov. 2020;10:232\u0026ndash;53. \u003c/li\u003e\n\u003cli\u003eGrout JA, Sirven P, Leader AM, Maskey S, Hector E, Puisieux I, et al. Spatial positioning and matrix programs of cancer-associated fibroblasts promote T cell exclusion in human lung tumors. Cancer Discov. 2022;12:2606\u0026ndash;25. \u003c/li\u003e\n\u003cli\u003eHelmink BA, Reddy SM, Gao J, Zhang S, Basar R, Thakur R, et al. B cells and tertiary lymphoid structures promote immunotherapy response. Nature. 2020;577:549\u0026ndash;55. \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;171:1611-1624.e24. \u003c/li\u003e\n\u003cli\u003eNayar S, Campos J, Smith CG, Iannizzotto V, Gardner DH, Mourcin F, et al. Immunofibroblasts are pivotal drivers of tertiary lymphoid structure formation and local pathology. Proc Natl Acad Sci USA. 2019;116:13490\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eRemouchamps C, Boutaffala L, Ganeff C, Dejardin E. Biology and signal transduction pathways of the Lymphotoxin-\u0026alpha;\u0026beta;/LT\u0026beta;R system. Cytokine \u0026amp; Growth Factor Reviews. 2011;22:301\u0026ndash;10. \u003c/li\u003e\n\u003cli\u003eKang W, Feng Z, Luo J, He Z, Liu J, Wu J, et al. Tertiary Lymphoid Structures in Cancer: The Double-Edged Sword Role in Antitumor Immunity and Potential Therapeutic Induction Strategies. Front Immunol. 2021;12:689270. \u003c/li\u003e\n\u003cli\u003eRuffin AT, Cillo AR, Tabib T, Liu A, Onkar S, Kunning SR, et al. B cell signatures and tertiary lymphoid structures contribute to outcome in head and neck squamous cell carcinoma. Nat Commun. 2021;12:3349. \u003c/li\u003e\n\u003cli\u003eSchumacher TN, Thommen DS. Tertiary lymphoid structures in cancer. Science. 2022;375:eabf9419. \u003c/li\u003e\n\u003cli\u003eCords L, Tietscher S, Anzeneder T, Langwieder C, Rees M, de Souza N, et al. Cancer-associated fibroblast classification in single-cell and spatial proteomics data. Nat Commun. 2023;14:4294. \u003c/li\u003e\n\u003cli\u003eKumar T, Nee K, Wei R, He S, Nguyen QH, Bai S, et al. A spatially resolved single-cell genomic atlas of the adult human breast. Nature. 2023;620:181\u0026ndash;91. \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;11:5077. \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;214:579\u0026ndash;96. \u003c/li\u003e\n\u003cli\u003ePelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, et al. Spatially organized multicellular immune hubs in human colorectal cancer. Cell. 2021;184:4734-4752.e20. \u003c/li\u003e\n\u003cli\u003eSmillie CS, Biton M, Ordovas-Monta\u0026ntilde;es J, Sullivan KM, Burgin G, Graham DB, et al. Cellular and inter-cellular rewiring of the human colon during ulcerative colitis. Cell. 2019;178:714-730.e22. \u003c/li\u003e\n\u003cli\u003eWilliams DW, Greenwell-Wild T, Brenchley L, Dutzan N, Overmiller A, Sawaya AP, et al. Human oral mucosa cell atlas reveals a stromal-neutrophil axis regulating tissue immunity. Cell. 2021;184:4090-4104.e15. \u003c/li\u003e\n\u003cli\u003eHuang H, Wang Z, Zhang Y, Pradhan RN, Ganguly D, Chandra R, et al. Mesothelial cell-derived antigen-presenting cancer-associated fibroblasts induce expansion of regulatory T cells in pancreatic cancer. Cancer Cell. 2022;40:656-673.e7. \u003c/li\u003e\n\u003cli\u003eKuleshov MV, Jones MR, Rouillard AD, Fernandez NF, Duan Q, Wang Z, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 2016;44:W90\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eWu T, Hu E, Xu S, Chen M, Guo P, Dai Z, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb). 2021;2:100141. \u003c/li\u003e\n\u003cli\u003eSchubert M, Klinger B, Kl\u0026uuml;nemann M, Sieber A, Uhlitz F, Sauer S, et al. Perturbation-response genes reveal signaling footprints in cancer gene expression. Nat Commun. 2018;9:20. \u003c/li\u003e\n\u003cli\u003eDann E, Henderson NC, Teichmann SA, Morgan MD, Marioni JC. Differential abundance testing on single-cell data using k-nearest neighbor graphs. Nat Biotechnol. 2022;40:245\u0026ndash;53. \u003c/li\u003e\n\u003cli\u003eCerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401\u0026ndash;4. \u003c/li\u003e\n\u003cli\u003eBecht E, Giraldo NA, Lacroix L, Buttard B, Elarouci N, Petitprez F, et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 2016;17:218.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"molecular-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"molc","sideBox":"Learn more about [Molecular Cancer](http://gsejournal.biomedcentral.com/)","snPcode":"12943","submissionUrl":"https://submission.nature.com/new-submission/12943/3","title":"Molecular Cancer","twitterHandle":"@SN_Oncology","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5125055/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5125055/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCancer-associated Fibroblasts (CAFs) have emerged as critical regulators of anti-tumour immunity, with both beneficial and detrimental properties that remain poorly characterised. To investigate this, we performed single-cell and spatial transcriptomic analysis, comparing immune-hot and immune-cold HNSCC subgroups (human papillomavirus [HPV]\u0026thinsp;+\u0026thinsp;ve and HPV-ve tumours respectively). This identified six fibroblast subpopulations, including two with immunomodulatory gene expression profiles (\u003cem\u003eIL-11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;inflammatory [i]CAF and fibroblastic reticular cell [FRC]-like). \u003cem\u003eIL-11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;iCAF were spatially associated with inflammatory monocytes and regulated \u003cem\u003ein vitro\u003c/em\u003e through synergistic activation of canonical NF-κB signalling by IL-1β and TNF-α. FRC-like were enriched in HPV\u0026thinsp;+\u0026thinsp;ve tumours, associated with CD4 T-cells and B-cells in tertiary lymphoid structures and regulated through non-canonical NF-κB signalling via lymphotoxin. Pan-cancer analysis revealed several 'iCAF\u0026rsquo; subgroups present in both normal and cancer tissues; \u003cem\u003eIL11\u0026thinsp;+\u003c/em\u003e\u0026thinsp;iCAF were found in cancers from the gastrointestinal tract and transcriptomically distinct from iCAFs previously described in pancreatic and breast cancers with greater inflammatory properties; FRC-like fibroblasts, a rare phenotype but present in all tumour types, were associated with significantly better survival in patients receiving checkpoint immunotherapy. This work clarifies and expands current literature on immunomodulatory CAFs, highlighting links with important immunological niches.\u003c/p\u003e","manuscriptTitle":"Single cell and spatial analysis of immune-hot and immune-cold tumours identifies fibroblast subtypes associated with distinct immunological niches and positive immunotherapy response","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-09 13:09:51","doi":"10.21203/rs.3.rs-5125055/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-11-14T22:41:51+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-14T17:03:36+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-08T16:07:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"250500850139360097190970141124224401624","date":"2024-11-06T09:31:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"90409034337110944556557424327211940714","date":"2024-11-06T08:47:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"170478814712860799728960215857277839349","date":"2024-11-04T00:46:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"38181991887835725092557204855964854679","date":"2024-11-04T00:43:06+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-11-04T00:22:28+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-09-23T23:11:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-09-23T23:09:36+00:00","index":"","fulltext":""},{"type":"submitted","content":"Molecular Cancer","date":"2024-09-20T16:50:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"molecular-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"molc","sideBox":"Learn more about [Molecular Cancer](http://gsejournal.biomedcentral.com/)","snPcode":"12943","submissionUrl":"https://submission.nature.com/new-submission/12943/3","title":"Molecular Cancer","twitterHandle":"@SN_Oncology","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"40e51a9a-a8a3-4698-bbcf-2f43d58ae11a","owner":[],"postedDate":"December 9th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-01-13T16:05:24+00:00","versionOfRecord":{"articleIdentity":"rs-5125055","link":"https://doi.org/10.1186/s12943-024-02191-9","journal":{"identity":"molecular-cancer","isVorOnly":false,"title":"Molecular Cancer"},"publishedOn":"2025-01-06 15:57:08","publishedOnDateReadable":"January 6th, 2025"},"versionCreatedAt":"2024-12-09 13:09:51","video":"","vorDoi":"10.1186/s12943-024-02191-9","vorDoiUrl":"https://doi.org/10.1186/s12943-024-02191-9","workflowStages":[]},"version":"v1","identity":"rs-5125055","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5125055","identity":"rs-5125055","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}