Germinal center B-cell subgroups in the tumor microenvironment cannot be overlooked: their involvement in prognosis, immunotherapy response, and treatment resistance in head and neck squamous carcinoma

Germinal center B-cell subgroups in the tumor microenvironment cannot be overlooked: their involvement in prognosis, immunotherapy response, and treatment resistance in head and neck squamous carcinoma | 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 Germinal center B-cell subgroups in the tumor microenvironment cannot be overlooked: their involvement in prognosis, immunotherapy response, and treatment resistance in head and neck squamous carcinoma Li Lin, Jiani Zou, Shengbin Pei, Wenyi Huang, Yichi Zhang, Zhijie Zhao, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3804853/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background More than 60% of patients with head and neck squamous carcinoma (HNSCC) are diagnosed at advanced stages and miss radical treatment. This has prompted the need to find new biomarkers to achieve early diagnosis and predict early recurrence and metastasis of tumors. Methods Single-cell RNA sequencing (scRNA-seq) data from HNSCC tissues and peripheral blood samples were obtained through the Gene Expression Omnibus (GEO) database (GSE164690) to characterize the B-cell subgroups, differentiation trajectories, and intercellular communication networks in HNSCC and to construct a prognostic model of the associated risks. In addition, this study analyzed the differences in clinical features, immune cell infiltration, functional enrichment, tumor mutational burden (TMB), and drug sensitivity between the high- and low-risk groups. Results Using scRNA-seq of HNSCC, we classified B and plasma cells into a total of four subgroups for the first time: naive B cells (NBs), germinal center B cells (GCBs), memory B cells (MBs), and plasma cells (PCs) and analyzed each subgroup separately. Pseudotemporal trajectory analysis revealed that NBs and GCBs were at the early stage of B cell differentiation, while MBs and PCs were at the end stage. Cellular communication revealed that GCBs acted on tumor cells through the CD99 and SEMA4 signaling pathways. By employing univariate Cox regression, Lasso, and multivariate Cox regression analysis, we developed several risk scores based on marker genes of GCBs subgroups, especially the independent prognostic value of MEF2B + GCB score was validated. The results of immune cell infiltration, TMB, and drug sensitivity assays were significantly different in HNSCC samples from high and low- MEF2B + GCB score groups. Conclusions This study provides ideas to unravel the fundamental biological functions of B cells and their complex mechanistic roles. For the first time, we identified GCBs as B cell-specific prognostic biomarkers. The MEF2B + GCB score fills the research gap in the genetic prognostic prediction model of HNSCC and is expected to provide a theoretical basis for finding new therapeutic targets for HNSCC. Head and neck squamous carcinoma germinal center B cells single-cell RNA-seq immune cell infiltration drug sensitivity Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 1 Introduction Head and neck squamous carcinoma (HNSCC), an immunosuppressive disease characterized by molecular heterogeneity and tumor-host interactions, has been increasing in incidence and mortality year by year and has now become the sixth most common cancer and the eighth leading cause of cancer deaths worldwide 1 . About 90% of head and neck malignancies are squamous cell carcinomas, and early diagnosis and treatment can lead to a better prognosis 2–3 . The diagnosis and treatment of HNSCC have made significant progress in the past few decades. Surgical treatment, radiotherapy, and chemotherapy can significantly improve the disease control rate and survival of patients with HNSCC. However, the overall survival (OS) of patients with HNSCC is still relatively low, with the 5-year survival rate at all stages around 40–50%, mainly due to the recurrence of tumors and metastasis after treatment 4–6 . However, in the treatment of HNSCC, due to the specificity of the tumor site, surgical treatment may change the corresponding anatomical structure, and radiotherapy will form fibrous scars in the corresponding area, thus making it particularly difficult to achieve early detection of local recurrence directly through physical examination or re-biopsy 7 . The rest of the diagnostic tools, such as Computed Tomography (CT), ultrasound, and Magnetic Resonance Imaging (MRI), are not sensitive to early lesions, so we must rely on specific biomarkers to predict early recurrence and metastasis of tumors. Improved survival in patients with HNSCC depends partly on disease stage, lymph node involvement, patient response to treatment, and early diagnosis of local tumor recurrence. Nevertheless, it has been suggested that the clinical staging of tumors is not a good reflection of the behavior of the tumor and the outcome of the disease 8 . This is why a growing number of studies have concluded, by evaluating clinical and pathological indicators, that non-advanced tumors tend to be accompanied by early recurrence and that advanced tumors tend to have longer disease-free intervals 9 . These factors have prompted the need to search for essential tumor markers for preoperative assessment of the likelihood of regional metastasis, the significance of postoperative adjuvant therapy, assessment of the patient's response to treatment, and prognostic analysis of the patient. Tumor immune microenvironment (TIME), which consists of immune-related components such as immune cells and their secreted immunomodulatory molecules, has become a hot topic in oncology research in recent years due to its critical regulatory role in the growth of tumors as well as metastasis 10–11 . As an essential component of TIME, there is growing evidence that B cells can participate in the immune response to tumors as well as immune evasion through the secretion of specific antibodies, the processing of presenting antigens, and the maintenance of the structure and function of tertiary lymphatic structures, which are biologically important 12–13 . However, research on TIME has focused mainly on T cells, with relatively few B cell-related studies 14 . The role of B cells is dual. On the one hand, most B cells can inhibit tumor growth through antibody production, secretion of anti-tumor cytokines, and direct killing. On the other hand, some B cell subgroups may also play a tumor-promoting role. For example, a group of regulatory B cell subgroups (Bregs) may promote tumor growth and invasion through the secretion of specific tumor-promoting cytokines (e.g., IL-10, TGF-β, IL-35) and other mechanisms 15 . In addition, B cells are closely associated with the prognosis of various tumors, including skin melanoma, gastric cancer, and lung cancer 16–17 . An early study of 33 HNSCC patients demonstrated improved outcomes with increased peritumoral B-cells in lymph node metastases. Recent analyses of B-cell phenotypes and responses in HNSCC have delineated notable heterogeneity 18 . Nevertheless, the impact of B cells on survival or their influence on responses to immunotherapy or radiation remained unexplored 19–20 , possibly due to the lack of uniformity in the choice of B cell markers. In recent decades, the emergence of bulk RNA-sequencing (bulk RNA-seq) has revolutionized oncology research and has been widely used in diagnosing and prognostic assessing cancer patients. However, there are individual differences between different cell types in tumor tissues. Still, bulk sequencing can only reflect the overall average gene expression level in the whole sample and cannot accurately characterize the gene expression at the individual cell level. Single-cell RNA-sequencing (scRNA-seq) allows access to the genetic signature of a particular cell type by extracting, reverse transcribing, amplifying, and sequencing RNA at the level of a single cell, better revealing heterogeneity in the intra-tumor microenvironment and providing assistance in the development of individualized therapeutic regimens 21 . Although scRNA-seq technology can characterize genes in different cell types more accurately than traditional bulk RNA-seq, bulk RNA-seq datasets have larger sample sizes and more complete follow-up information. Therefore, there is a need to integrate scRNA-seq and bulk RNA-seq datasets to identify new tumor biomarkers and develop more accurate tumor prognosis-related models. In summary, the biological functions of B cells are complex, and their prognostic value in HNSCC remains unknown. Screening for appropriate B-cell markers in HNSCC and in-depth exploration of their biological functions and prognostic value could help to assess the prognosis. It may provide new targets for the treatment of patients with HNSCC. 2 Materials and methods 2.1 Data collection and processing The GPL18573 platform-based Head and neck squamous cell carcinomas (HNSCCs) -related single-cell RNA sequencing (scRNA-seq)data GSE164690 [PMID: 34921143] was downloaded from Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/ ). In addition, bulk RNA sequencing data of HNSCC tissue samples were downloaded from the Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/ ) database, along with clinical information including age, gender, TNM stage, race, and survival time. 2.2 Single-cell sequencing data analysis Using the Seurat R package (version 4.3.0), raw gene expression substrates were imported into R software (version 4.2.2). Using the DoubletFinder R package (version 2.0.3), the scRNA-seq data underwent rigorous quality control. The percentage of mitochondrial and erythrocyte-related genes in each cell was calculated using the "PercentageFeatureSet" function in the Seurat package, followed by cell quality control, with the following exclusion criteria: ( 1 ) the number of genes in the cell was 7500; ( 3 ) the percentage of mitochondrial genes in the cell was > 25. Then use the “NormalizeData” function for normalization (normalization.method = "LogNormalize", scale.factor = 10000). Use the “CellCycleScoring” function to determine cell cycle conditions. Find the top 2000 top variable genes for subsequent analysis using the “FindVariableFeatures” function. The data are standardized using the "ScaleData" function. The harmony R package (version 0.1.1) was used to remove batch effects between samples. The dim value was set to 30 and the resolution to 1.2. The "RunPCA" function was used to cluster the dataset and to downscale the uniform manifold approximation and projection (UMAP) visualisation 100 . Marker genes for each B-cell subgroup were compiled by searching previous literature, and the clusters were manually annotated after the downscaling of the subgroups. Next, the "FindAllMarkers" function was used to find differentially expressed genes in each B-cell subpopulation, where min.pct was set to 0.25 and logfc.threshold was set to 0.25. 2.3 Pseudotime analysis We performed pseudotime trajectory analysis of B and plasma cell subgroups using the Monocle R package (version 2.24.0) 101 . The UMAP method was employed to reduce dimensionality, and the functions "plot_cell_trajectory" was utilized for visualization. Different cell subpopulations were arranged in a proposed chronological order, and pseudotemporal heatmaps were used to identify and display genes that changed simultaneously with pseudotime. 2.4 Cell-cell communication analysis The CellChat R package (version 1.6.1) was applied to infer cell-to-cell interactions among cell subgroups and establish a regulatory network based on ligand-receptor crosstalk 102 . The “netVisual_circle”, “netVisual_aggregate”, “netVisual_individual”, “netVisual_heatmap”, “netVisual_bubble”, “netVisual_chord_gene”, “netAnalysis_river”, “netAnalysis_dot”, and “netAnalysis_signalingRole_heatmap” functions were used to visualize cell-to-cell interactions. 2.5 Construction of a risk score in HNSCC patients The expression values of signature genes of patients in the GDC TCGA Head and Neck Cancer (HNSC) cohort were extracted after obtaining the list of MBs or GCBs signature genes (top 100 genes). Signature genes associated with OS of HNSCC patients were screened using univariate Cox regression analysis. To avoid the multicollinearity of these genes, we further filtered them using LASSO regression (glmnet, version 4.1-6) to fileted these genes. The risk coefficients for each gene were then calculated using multivariate Cox regression. Risk scores of the prognostic risk score model (X: coefficient, Y: gene expression level) = \(\sum _{\text{i}}^{\text{n}}\text{X}\text{i}\times \text{Y}\text{i}\) . Patients were categorized into high and low-risk groups based on median of risk scores. Kaplan-Meier survival analysis (version 3.3-1) was performed in the TCGA-HNSC cohort to compare the prognosis of patients in different risk groups. The predictive performance of risk score was further evaluated by plotting the area under the curve of ROC (AUC) over time using the timeROC R package (version 0.4). Risk score, age, TNM stage, and molecular subtype were subjected to were subjected to multivariate Cox regression analysis to screen out independent prognostic factors in HNSCC patients. The nomogram column-line graph prediction model for predicting 1-, 3-, and 5-year overall survival of HNSCC patients was constructed using the "rms" package (version 6.7-0). The constructed nomogram prediction model was validated using the ROC curve and calibration curve. 2.6 Immune infiltration analysis Previous literature was queried to obtain marker genes for different immune cells, and ssGSEA enrichment analysis was performed to assess the level of immune cell infiltration for each sample in the TCGA-HNSC cohort 103 . The relative proportion of immune cells was quantified using the CIBERSORT R package (version 0.1.0) 104 . The Wilcox.test was used to compare the levels of immune cell infiltration and the expression levels of joint exempted checkpoints between different risk groups. Correlation analyses between immune cell infiltration levels of each type and between immune cell infiltration levels and risk genes were performed using the "corrplot" R package (version 0.92). Stromal and immune cells in HNSCC tissues were estimated using the ESTIMATE R package (version 1.0.13), and stromal score, immune score, ESTIMATE score, and tumor purity were calculated 105 . 2.7 Functional enrichment analysis The gene expression matrices of HNSCC patients in the high and low-risk subgroups were analyzed differently by the limma R package (version 3.54.0). The GO 106 and KEGG 107 were used to analyze the DEGs between different risk subgroups by using clusterProfiler R package (version 4.6.0). We downloaded the list of pathways gene sets from the MSIGDB database ( https://www.gsea-msigdb.org/ ) and performed GSEA 108 enrichment analyses on these pathways. 2.8 Tumor mutation burden analysis The maftools package was utilized to analyze the overall mutation status of the tumor, TMB, as well as the transformation status of SNPs. The Wilcox.test was used to compare the TMB levels between risk subgroups of patients, and the Spearman correlation test was performed to determine the relationship between risk score and TMB. 2.9 Prediction of drug sensitivity The Genomics of Drug Sensitivity in Cancer (GDSC)( https://www.cancerrxgene.org/)i s a public dataset containing information about drug sensitivity and molecular markers of drug response in cancer 109 . The dataset information was downloaded as reference files, and the pRRophetic R package (version 0.5) 110 was used to predict the sensitivity to chemotherapy and targeted therapies for each sample. The Wilcox.test was performed to determine whether there was a statistically significant difference in medicine sensitivity between the high and low risk groups. Pearson's correlation was utilized to examine the relationship between risk scores and drug sensitivity. 2.10 Multiplexed immunofluorescence assay Multicolor IHC staining was performed using a Multiple Immunofluorescence Kit (AiFang biological, AFIHC035) according to the manufacturer’s instructions. Briefly, the paraffin-embedded tissues were dewaxed, then the antigen retrieval was performed by sodium citrate at 98℃ for 20 minutes. Ater infiltrating in 3% H 2 O 2 for 15 minutes, the tissues were blocked by 5% BSA. Then the tissues were incubated by primary antibody at 4℃ overnight. Then the tissues were incubated by secondary antibody for 30 minutes and stained by TSA for 5 minutes. For each targe, the steps of antigen retrieval, blocking, incubation of primary antibody, incubation of secondary antibody, and TSA staining are required. Finally, the nuclei were stained with DAPI. PBS was used to wash tissues between each step. These images were obtained by the LEICA DMi8 system and analyzed by the ImageJ software. 2.11 Cell culture HN-5 cells and UMSCC-47 cells originated from the American Type Culture Collection (ATCC). HN-5 cells underwent cultivation utilizing RPMI 1640 medium, whereas UMSCC-47 cells were subjected to DMEM medium (Gibco BRL, United States). Both cell lines were nurtured within an incubation environment at 37°C, featuring a 5% CO2 concentration. 2.12 Cell transfection and Real-time fluorescent quantitative PCR (RT-qPCR) Two small interfering RNAs (siRNAs) targeting the PLXNB2 gene, along with their corresponding negative controls (si-NC), were synthesized by Ribobio Guangzhou (China). The transfection procedure adhered to the methodologies delineated in the Lipofectamine 3000 instructions provided by Invitrogen, USA. Thermo's TRIzol reagent (15596018) was employed to extract total RNA from the cellular lines, and RNA concentrations were standardized. Subsequently, cDNA synthesis was carried out using Vazyme's PrimeScript™ RT kit (R232-01). Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was executed utilizing the SYBR Green Kit from TaKaRa Biotechnology (Dalian, China), with GAPDH as the internal reference control. Further details regarding primer and siRNA sequences are available in supplemental Table 1. 2.13 Cell Counting A total of 6000 transfected cells were seeded into individual wells of 96-well plates (Corning, USA, 3599). Following cellular adhesion to the well surfaces, the cells underwent treatment with the CCK-8 labeling reagent (A311-01, Vazyme) on days 1, 2, 3, 4, and 5, respectively. Incubation of the cells occurred in conditions devoid of light for a duration of 2 hours, during which absorbance (OD) values were meticulously recorded. 2.14 Wound healing Transfected cells were nurtured in a six-well culture plate (Corning, USA, 3516). Upon achieving a cellular density of approximately 90–95%, a 200 µl sterile pipette was employed to delicately abrade the cell layer in a linear fashion. Subsequently, a gradual rinsing with PBS was conducted to eliminate unattached cells and debris. Following this, the serum-free cell medium was substituted to sustain optimal cell proliferation. Consistent photographic documentation was undertaken at identical locations, both at the initiation of the culture and 48 hours subsequent to the initiation. 2.15 Transwell In the Transwell experimental setup, the Transwell chamber (Corning, USA, 341201) was positioned within a 24-well plate. Subsequently, 8,000 cells were uniformly distributed in each chamber and cultured utilizing 200 µl of serum-free DMEM or RPMI 1640 medium. To the lower chamber, 600 µl of complete DMEM or RPMI 1640 medium, enriched with 10% serum, was added. The experimental plate was then placed in an incubation apparatus and cultured for a duration of 36 hours, allowing for the migration of cells from the upper chamber to the lower chamber. Following this incubation period, the Transwell chamber underwent meticulous cleaning, fixation, and staining procedures. 3 Results 3.1 Phenotypic clusters of B cells according to scRNA-seq To investigate the variations in the immune microenvironment across all cell subgroups in HNSCC, we collected scRNA-seq datasets of HNSCC. By utilizing these datasets, we aim to gain a more profound comprehension of the molecular mechanisms at the individual cell level (Figure S1 - S2 ). Through the standard workflow of single-cell analysis, we identified four cell types in the integrated data, including NBs, MBs, GCBs, and PCs. We plotted the distribution of HPV infection status and phase (Fig. 1 A). To further elucidate the heterogeneity and intricacy of HNSCC, we have subdivided them into 16 distinct subgroups. Among these subgroups, the population of NBs encompasses C0 TCL1A + Naive B cells, C5 TCL1A + Naive B cells, C7 CD83 + Naive B cells, C8 ISG15 + Naive B cells, C10 TCL1A + Naive B cells, and C15 HSPA1B + Naive B cells. Additionally, the category of MBs includes C1 CRIP1 + Memory B cells, C2 LINC01781 + Memory B cells, C3 PPP1R15A + Memory B cells, C4 IFI30 + Memory B cells, C9 CD1C + Memory B cells, and C13 TLE5 + Memory B cells. Furthermore, the subgroup of GCBs comprises C11 MEF2B + Germinal center B cells and C14 STMN1 + Germinal center B cells. Lastly, we have identified PCs represented by C6 FKBP11 + IgG Plasma cells and C12 DERL3 + IgA Plasma cells (Fig. 1 B). The distribution of CNVscore, nCount_RNA, S.Score, and G2M. The score for each subpopulation is shown in Fig. 1 C, in which GCBs on the right side of the UMAP plot have a high CNVscore, a high nCount_RNA, and a more active cell cycle state. In addition, we analyzed the proportion of subgroups derived from peripheral blood (PBL) and tumor tissues. The pie chart displayed the proportion of B and plasma cell subtypes (Fig. 1 D). The results showed that most subgroups belonged to malignant cells, with a small proportion of NBs and MBs originating from the PBL (Fig. 1 E). Bubble maps and volcano plots demonstrated the expression of differentially expressed genes (DEGs) in each subgroup and the associated features of different HPV infection statuses (HPV + vs HPV − ) (Fig. 1 F-G). 3.2 Pseudotime trajectories unveiled the differentiation status of B and plasma cell subgroups To understand the dynamic transitional processes of B and plasma cells in HNSCC, we further performed a pseudotime analysis to trace the transcriptional trajectory. As shown in Fig. 2 A-B, with the progression of pseudotime, the number of NBs gradually decreased, while GCBs and MBs gradually differentiated toward PCs. The decrease in GCBs and MBs may be due to their gradual transformation into PCs, most of PCs were at the end of the cell developmental trajectory. Cellular composition was identified by colouring the pseudotemporal map with tissue of origin. By analysing Fig. 1 A, B & Fig. 2 C, it was found that the pseudotime of NBs was significantly lower than that of other B cell subgroups, and the pseudotime of PCs represented by C6 FKBP11 + IgG Plasma cells and C12 DERL3 + IgA Plasma cells was the highest. An enhanced graphical representation, a bar plot, elucidated that the NBs subgroups predominantly comprised neoplastic cells originating from State1. Likewise, the MBs subgroups exhibited a preponderance of cells stemming from State2. Furthermore, the GCBs subgroups was characterized by a prevailing presence of neoplastic cells deriving from State4. Lastly, the PCs subgroups exhibited a notable predominance of cells originating from State5 (Fig. 2 D, E). These observations collectively unveil that the NBs subgroups is positioned within the incipient stage of cellular differentiation, contrasting with the PCs, which signifies the end of differentiation. Additionally, cells within the GCBs and MBs subgroups display an augmented propensity for proliferation and differentiation capacity, thereby highlighting their distinctive biological characteristics. 3.3 NBs are in the initial differentiation stage of pseudotime trajectories According to our research, B and plasma cells could be classified into NBs, GCBs, MBs, and PCs. Each of these subgroups manifests distinct phenotypic characteristics and undertakes specific roles within the tumor microenvironment (TME). NBs can be further classified into 6 subgroups (Fig. 3 A), and the expression of top5 marker genes in different subgroups is also shown (Fig. 3 B, C). By analyzing the differences in cell cycle, gene expression and CNV among samples with different NBs subgroups and HPV infection states, it was found that the CNV score of C0 TCL1A + Naive B cells and C15 HSPA1B + Naive B cells was higher, the proportion of G2M and S phase in the cell cycle was higher, and the expression of differential genes was higher, indicating that their cell proliferation was more active and genetic instability (Fig. 3 D, I). The DEGs of 6 NBs subgroups and their corresponding transcription factors and surface proteins were analyzed by GO-BP enrichment analysis, and the active biological process of each tumor cell was obtained. It was found that C0 TCL1A + Naive B cells was active in the biological process of Cytoplasmic translation, and C8 ISG15 + Naive B cells was active in the biological process of Negative regulation of b cell activation (Fig. 3 E, H). Immune landscape studies showed that HPV + HNSCC were enriched with more NBs, HPV − HNSCC were infiltrated by fewer NBs, and C0 TCL1A + Naive B cells had the highest percentage of HPV + cells (Fig. 3 F). In the analysis of NBs subgroups and cell cycle, it was found that the proportion of NBs within the G1 phase of cellular growth was the highest, which may be related to the fact that NBs is in the initial stage of B cell differentiation (Fig. 3 G). Through GSEA enrichment analysis, it was found that C0 TCL1A + Naive B cells subgroup was negatively correlated with Negative regulation of metabolic process and Negative regulation of programmed cell death pathway, indicating that cells of C0 TCL1A + Naive B cells subgroup had higher ability of cell metabolism and proliferation (Fig. 3 J, K). 3.4 GCBs subgroups may have a high immune response ability In the germinal center, follicular helper T cells (T fh ) stimulate the B cell receptor (BCR) and induce the differentiation of NBs into a variety of fates, mainly including GCBs, MBs and PCs 24 . Complete transcriptional characteristics at the single-cell level would be particularly helpful in determining transcriptional trajectories and heterogeneity in the process of differentiation from NBs to PCs. C11 MEF2B + Germinal center B cells (C11 GCBs) and C14 STMN1 + Germinal center B cells (C14 GCBs) were obtained by further subgroup analysis of GCBs (Fig. 4 A). The bubble chart and UMAP plots showed the expression of GCBs top marker genes (Fig. 4 B-C). Among the two GCBs subgroups, the CNVscore of C11 was higher, and the nCount-RNA, G2M.Scorce and S.Score of C14 were higher (P < 0.001) (Fig. 4 D-F). The GCBs in patients with HNSCC mainly comes from tumor tissues, and some from PBL (Fig. 4 G). The percentage of C11 GCBs in HPV + samples was higher than in C14 GCBs (Fig. 4 H). In C11 GCBs, there were more cells with G1 cell cycle, and most of C14 GCBs were in G2M and S phase (Fig. 4 I). The proportion of cells in G1 phase was more in HPV + group, and the proportion of GCBs cells in G2M and S phase was more in HPV − group (Fig. 4 J). The DEGs of C11 GCBs and C14 GCBs were screened (Fig. 4 L), and the enrichment results of GO-BP among C11 GCBs and C14 GCBs was shown by heatmap. It was found that GOBP terms of C11 GCBs were up-regulated in biological processes such as cell activation, leukocyte activation, immune response, lymphocyte activation, regulation of immune system process, regulation of immune response, indicating that the DEGs of C11 GCBs may play a promoting role in the regulation of immune system (Fig. 4 K). Similarly, in the GO-BP enrichment analysis results of C11 GCBs, C11 GCBs was more active in cytoplasmic translation, B cell activation, B cell proliferation, production of molecular mediator of immune response and immunoglobulin production pathways. The high immune response ability of C11 GCBs subgroup was further verified, which was consistent with the previous research results (Fig. 4 M-N). 3.5 MBs play a key role in the proliferation and differentiation of B cells MBs, the subgroups with the largest number of B cells (Fig. 5 A), could be further divided into six subgroups and the expression of top5 marker genes in different subgroups are also shown in Fig. 5 B-C. UMAP plots showed that the samples in the C4 IFI30 + Memory B cells (C4 MBs) had higher CNV score, higher nCount_RNA and more active cell cycle state. Among them, the chromosome stability, transcription level and cell cycle state of different subgroups of MBs are also different, which is of great significance for the study of B cell heterogeneity (Fig. 5 D, H). The bar charts revealed that the majority of MBs subgroups were associated with HPV + infection, whereas the cell cycle mostly occurred in the G1 and S phases (Fig. 5 E-F). The AUCell analysis of the GOBP-positive regulation of B cell proliferation in several subgroups of MBs revealed that the C4 MBs had a higher AUC value for this GOBP term compared to other subgroups (Fig. 5 G, left). The CNV score of C13 TLE5 + Memory B cells (C13 MBs) exhibited a lower value compared to the other subgroups (Fig. 5 G, right), while the nCount_RNA of C13 MBs was higher than other subgroups (Fig. 5 H, left). The G2M.Score of C4 MBs was lower than other subgroups of MBs (Fig. 5 H, right). In GO-BP enrichment analysis, C4 MBs were active in cytoplasmic translation, B cell receptor signaling pathway, immune response-regulating cell surface receptor signaling pathway, B cell differentiation, Fc receptor signaling pathway (Fig. 5 I). Similarly, in GSEA enrichment analysis, the expression of C4 MBs was up-regulated in positive regulation of cellular biosynthetic process, which was positively correlated with positive regulation of cell communication, positive regulation of cell death, positive regulation of cellular metabolic process (Fig. 5 J). These results suggest that the DEGs of C4 MBs plays a key role in the differentiation of B cells and promotes the regulation of immune system process, which may reflect the high immune response ability of tumor cells. 3.6 B cells eventually differentiate into mature PCs GCBs that have a strong attraction to tumor-associated antigens ultimately develop into fully developed PCs that generate immunoglobulins (IgG, IgA, IgM, and IgE). These immunoglobulins have distinct purposes within the given situation 25 . According to the difference in the expression of tumor-associated antibodies in PBs, we divided the subgroups into C6 FKBP11 + IgG Plasma cells (C6 IgG), C12 DERL3 + IgA Plasma cells (C12 IgA) (Fig. 6 A). The UMAP visualization showed the difference between C6 IgG and C12 IgA in terms of CNV score, nCount_RNA, S.Score, and G2M (Fig. 6 B). The bubble chart and UMAP plots displayed the top marker genes for C6 IgG and C12 IgA (Fig. 6 C-D). The bar charts depicted the distribution of PCs among various categories (Fig. 6 E-G). To investigate the molecular diversity among PCs subgroups and its functionality. The GO-BP enrichment analysis of the DEGs in PCs revealed that C6 IgG was involved in the biosynthesis of amyloid precursor protein and glycoprotein. Additionally, C12 IgA was found to be active in the biological processes of responding to endoplasmic reticulum stress, the ERAD pathway, and topologically incorrect protein (Fig. 6 H-I). 3.7 Exploring the intercellular communication between B cells, plasma cells, and tumor cells Numerous investigations have demonstrated the existence of a diverse spectrum of mutual influence between B cells, plasma cells, and tumor cells. This interplay exerts a profound influence on the progression of cancer 25 . To elucidate the intricate mechanism of cell-cell interaction between these cells, we established a separate intercellular communication network between B cells, plasma cells, and tumor cells in HNSCC (Fig. 7 A). 17 pairs of interactions and 20 significant signaling pathways were identified (Fig. 7 B). We generally describe the inferred strengths of incoming and outcoming interaction, in which C11 GCBs and C14 GCBs are the main cell type expressing ligands and receptors that actively participate in cell-cell interaction. These results suggest that GCBs is associated with tumors and plays a central role in TME (Fig. 7 C). As the observed C11 GCBs and C14 GCBs played a central role in the intercellular communications at HNSCC, we further explore the interactions between incoming and outgoing communication patterns of target cells and secreting cells. Notably, in outgoing communication patterns of secreting cells of GCBs, CD99 and SEMA4 are shown as the most predominant signaling in C11 GCBs compared to other signaling pathways (Fig. 7 D-F). Consistent with these observations, GCBs and MBs are the core of the interaction between B cells and tumor cells (Fig. 7 G-H). In summary, the intercellular interaction based on CD99 and SEMA4 signaling pathways is the molecular basis of GCBs-centered TME observed in HNSCC. 3.8 GCBs based regulatory network detect tumor-specific signaling pathways in HNSCC To unveil the intricate signaling pathways that contribute to the multifaceted intercellular communications, we computed the CD99 and SEMA4 signaling pathways (Fig. 8 A, F). Tumor-specific C11 MEF2B + Germinal center B cells could express CD99 and SEMA4 to interact with some receptors of tumor cells to promote tumor progression (Fig. 8 B, G). Regarding the CD99 signaling pathways, we observed that C11 MEF2B + Germinal center B cells and C14 STMN1 + Germinal center B cells highly expressed CD99 to interact with the tumor cell receptors, C4 and C13 of memory B cells (Fig. 8 C-E). In SEMA4 signaling pathway networks, the specific interaction networks of C11 MEF2B + Germinal center B cells and C14 STMN1 + Germinal center B cells with tumor cells were also observed (Fig. 8 H-J). These phenomena show that CD99 and SEMA4 could potentially serve as pivotal pathways mediating the interplay between GCBs and tumor cells. Furthermore, these signaling pathways possess the capacity to regulate B cell differentiation, consequently exercising control over tumor growth. Overall, our findings underscore that the C11 MEF2B + Germinal center B cells manifest intricate intercellular interplays that exacerbate the virulence of the tumor. Targeting tumor-specific cell interaction networks could be a promising strategy for treating HNSCC. 3.9 Screening and validation of prognostic genes in GCBs and MBs Considering the presence of tumor-specific signaling pathways in MBs and GCBs mentioned earlier, we will conduct an in-depth examination of these subgroups. A total of 499 HNSCC patients from the TCGA database were finally included in this study. Univariate Cox regression analysis was performed to identify the prognosis-related genes in MBs and GCBs (Fig. 9 A, 9 F, 10 A, 10 F). The candidate genes were recruited into the Least Absolute Shrinkage and Selection Operator (LASSO) Cox regression model to screen out the robust prognosticators (Fig. 9 B, 9 G, 10 B, 10 G). Ultimately, using multivariate Cox regression analysis, we were able to construct four prognostic risk scores consisting of prognostic genes specific to the MBs and GCBs subgroups (Fig. 9 C, 9 D, 9 H, 9 I, 10 C, 10 D, 10 H, 10 I). The individual genes that constituted risk scores, which showed an important association with overall survival in MBs and GCBs, were thoroughly shown (Fig. 9 E, 9 J, 10 E, 10 J, 10 K). The expression of CD38, SMIM14, and TCL1A in C11 MEF2B + Germinal center B cells varied considerably over distinct N stages (Fig. 10 L, P < 0.05). 3.10 MEF2B + GCB score based on the prognosis-related genes in C11 GCBs We then investigated the possible mechanisms of prognostic effects of 18 prognosis-related genes that comprised the MEF2B + GCB score. Bubble maps demonstrated the expression of 18 prognosis-related genes in each GCBs subgroup and the associated features of different HPV infection statuses (HPV + vs HPV − ), Phase (G1, S, G2M), G2M.Scorce, S.Score, nCount_RNA, CNVscore (Fig. 11 A ). The expression of 18 prognostic genes in C11 and C14 GCBs subgroups was further visualized in Fig. 11 B and 11 C. Risk scores were computed for individuals diagnosed with HNSCC within the TCGA datasets. The collected data were subsequently partitioned into cohorts of low and high-risk groups, demarcated by the median risk score (Fig. 11 D above). The heatmap specifically showed the expression of genes in high and low risk groups (Fig. 11 D below). More HNSCC patients had higher survival rates ( p < 0.05) observed in the low-risk group, while the distribution of stage and TNM phases was not statistically different (Fig. 11 E). The Kaplan-Meier survival curves showed that HNSCC patients in the high-risk group had a lower probability of survival and shorter survival time (Fig. 11 F). Moreover, the time-dependent ROC curve demonstrated the proficient prognostic capacity of this model concerning the OS of HNSCC patients. The calculated Area Under the ROC Curve (AUC) values stood at 0.690 (1 year), 0.711 (3 years) and 0.677 (5 years), as evidenced in Fig. 11 G. Then, we further assessed the correlation between gene expression and risk score in the prognostic model (Fig. 11 H). The expression of five genes, ACTB, BASP1, EZR, PRDX6, and RRAS2, was positively correlated with the risk scores (Fig. 11 I), which was in agreement with the results of the previous prognostic studies (Fig. 10 J), indicating that these five genes may be tumor-related risk genes in HNSCC. Meanwhile, the five genes of ACTB, BASP1, EZR, PRDX6, RRAS2 were highly expressed in the high-risk group, and the difference between the high and low risk groups was statistically significant (Fig. 11 J, P < 0.001). They were negatively correlated with OS (Fig. 11 K). Both EZR and RRAS2 expressions were all positively correlated with the other four genes (Fig. 11 L-M, P < 0.05)Figure. We further investigated the differences of EZR and RRAS2 in high- and low-risk groups, different ages, different races, and different TNM stages, and found that the gene expression was higher in all high-risk groups (Fig. 11 N). We observed that the MEF2B + GCB score had independent prognostic value and had stronger predictive power than other clinical features ( P < 0.001), which increases its potential for clinical applications (Fig. 12 A). A nomogram plot was established to get a better clinical application ability by integrating the risk score and clinical features (Fig. 12 B). The time-dependent ROC curve demonstrated the proficient prognostic capacity of this model. The AUC values were 0.738 (1 year), 0.769 (3 years) and 0.634 (5 years) (Fig. 12 C, D). For the 1-, 3-, and 5-year survival, a satisfactory fit was observed between the survival predicted by the nomogram and the actual survival (Fig. 12 E). 3.11 Immune cells play a dual role in the dynamic ecosystem of the HNSCC microenvironment The initiation, progression, and invasion of HNSCC are controlled by the complex network of signaling pathways in the TME as well as cell-cell interactions between the tumor stroma, immune cells, and tumor cells 26 . HNSCC microenvironment contained multiple types of immune cells, most of which are T cells and macrophages (Fig. 12 F). The high-risk group had lower immune scores and higher tumor purity than the low-risk group (both P < 0.001), which may be related to the fact that immune cells infiltrate into the tumor stroma within the HNSCC and interact with tumor cells (Fig. 12 G-H). By analyzing the principal components of immune cells in HNSCC, it was found that there were significant differences in the infiltration of many kinds of immune cells between the two risk subgroups ( P < 0.05). There was more CD8 T cell infiltration in the low-risk group and more macrophage M2 infiltration in the high-risk group (Fig. 12 I-J). In the correlation analysis between risk score and immune cells, mast cells, NK cells, and macrophages were significantly positively correlated with the risk score, whereas various T cells such as T cells regulatory (Tregs), T cells follicular helper, and CD8 T cells were significantly negatively correlated with the risk score (Fig. 12 K). Combined with the results of the previous survival analysis of the two subgroups (Fig. 11 F), this suggests that tumor-associated immunocompetent cells may be the key immunosuppressive population of TME cells. In addition, we also analyzed the correlation between the expression of 18 prognostic genes and the proportion of immune cells in HNSCC. The results showed that the risk score, the gene expression of ACTB, BASP1, PRDX6 was positively correlated with macrophage M2 (all P < 0.05), and the risk score, the gene expression of ACTB, BASP1, RRAS2 was negatively correlated with CD8 T cells (all P < 0.01) (Fig. 12 L-N). The gene expression of RRAS2 was negatively correlated with naive B cells, plasma cells, stromal score, and immunity score (all P < 0.05), and was positively correlated with tumor purity ( P < 0.001) (Fig. 12 O-P). Furthermore, RRAS2 was positively correlated with dendritic cells activated, macrophages M0, CD4 memory T cells resting, and NK cells resting (all P < 0.05). RRAS2 was negatively correlated with CD4 memory T cells activated, Mast cells resting, CD8 T cells, and Tregs (Fig. 12 Q). 3.12 Functional enrichment and GSEA analysis of differentially expressed genes in HNSCC patients To scrutinize the molecular heterogeneity and explore the biological mechanism distinguishing the high-risk from the low-risk subgroups, we discerned DEGs within the TCGA cohort [ \(\left|{\text{log}}_{2}\left(fold change\right)\right|\) >1, fdr < 0.05]. The results of Gene Ontology (GO) enrichment analysis revealed that the DEGs primarily participated in biological processes such as immunoglobulin complex, circulating immunoglobulin complex, and external side of plasma membrane. Their products were predominantly engaged in immune components, encompassing immunoglobulin production, production of molecular mediator of immune response, B cell activation, etc. Furthermore, these genes played a pivotal role in biological molecular functions, which covered antigen binding and immunoglobulin receptor binding, etc. (Fig. 13 A-C). The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways mainly included the B cell receptor signaling pathway, Ras signaling pathway, PI3K − Akt signaling pathway, Cytokine − cytokine receptor interaction, Primary immunodeficiency, Hematopoietic cell lineage, etc. (Fig. 13 D). Our GSEA results, presented in Fig. 13 E, revealed that Ribosomal large subunit biogenesis, Mitochondrial translation, Ncrna 3 End processing, Maturation of ssu Rrna from tricistronic Rrna transcript ssu Rrna 5 8s Rrna lsu Rrna, amino acid activation were activated, while antigen receptor mediated signaling, regulation of B cell activation, B cell mediated immunity, positive regulation of B cell activation, B cell receptor signaling pathway were suppressed in the GCBs subgroup. According to the results of GSEA, many B-cell-associated pathways were suppressed in GCBs, highlighting the presence of a B-cell response in patients with HNSCC. 3.13 Relationship of the OS in HNSCC patients among the TMB subgroups and drug sensitivity In the gene mutation spectrum of the risk and age groups, 466 (91.37%) of 510 HNSCC patients had the gene mutation (Fig. 13 F left). Among the 18 prognostic genes, 31 (6.08%) had gene mutations in 510 HNSCC patients (Fig. 13 F right). We further investigated the frequency of CNV in a total of 526 samples. As shown in Fig. 13 G, ACTB, RRAS2, SERPINA9, TCL1A, PKM, EZR, LCK, CD38, and BIK had significant CNV amplification frequencies. In addition, the probability of simultaneous mutations in two groups of genes, PKM and SERPINA9, TCL1A and STAG3, was higher (Fig. 13 H, P < 0.05). We selected the top 6 genes of somatic mutation rate for single nucleotide polymorphism (SNP) analysis and found that SERPINA9, STAG3, ACTB, CD38, LCK, and PKM were dominated by missense mutation (Fig. 13 I). In the two risk subgroups, the TMB of the high-risk group was higher than that of the low-risk group (Fig. 13 J, P < 0.001), and TMB was positively correlated with the risk score (Fig. 13 K, P < 0.001). Combined with the survival analysis in Fig. 13 L, the survival time of patients in high risk-high TMB group was the shortest, while that in low risk-low TMB group was the longest, and the OS of patients in each group was significantly different ( P < 0.0001). The AUC values were 0.541 (1 year), 0.532 (3 years), and 0.5I2 (5 years) (Fig. 13 M). This indicates that the TMB model has a discernible predictive ability for the prognosis of HNSCC patients. Furthermore, in drug sensitivity analysis, HNSCC patients in the low-risk subgroup demonstrated enhanced responsiveness to cisplatin and docetaxel in chemotherapy (Fig. 13 N). In contrast, the high-risk subgroup was more sensitive to Nilotinib and Rapamycin, both targeted drugs (Fig. 13 O). The IC50 values of ABT.263, AKT.inhibitor.Ⅷ, CCT007093 were higher in the high-risk subgroup, and Bosutinib, Epothilone. B and Pazopanib were higher in the low-risk subgroup (Fig. 13 P). These discoveries not only furnish invaluable insights into selecting suitable chemotherapy and targeted drugs following the risk score of HNSCC patients afflicted with HNSCC but also facilitate informed clinical therapeutic deliberations. 3.14 Immunofluorescence detection of MEF2B + B cells To validate the existence of MEF2B + B cells, immunofluorescence assays were executed on three human HNSCC tissues (Figure S3 - S5 ). In the multiple immunofluorescence assay, MEF2B and CD79a were subjected to staining. A conspicuous co-localization was discerned between MEF2B + cells and CD79a + cells, thereby affirming the presence of MEF2B + B cells. To further elucidate the interplay among MEF2B + B cells, CD8 + T cells, and tumor cells. Strikingly, spatial co-localization between CD79a and MEF2B in CD8 + and EPCAM + regions was identified, decisively establishing a substantial correlation between MEF2B + B cells and CD8 + T cells and tumor cells. 3.15 In vitro experiment Based on the cell communication analysis findings using cellchat between MEF2B + B cells and tumor cells, we conducted in vitro knockdown experiments targeting the PLXNB2 gene in the head and neck cancer cell line. The results, illustrated in Fig. 14 A and B, manifest a conspicuous reduction in the proliferation capacity of both cell lines, discernibly contrasting with the control group. Furthermore, plate cloning experiments unveiled a pronounced suppression in both the quantity and size of cell colonies after the knockdown of the PLXNB2 gene in both cell lines (Fig. 14 C, D). The inhibitory impact of PLXNB2 knockdown on the proliferation of the two head and neck cancer cell lines was further substantiated, indirectly suggesting that the interaction between B cells and tumor cells can modulate the biological functions of head and neck cancer cells. Subsequent wound healing experiments showcased a statistically significant decrease in the cell migration rate within the PLXNB2 knockdown group (Fig. 14 E, F). Moreover, Transwell experiments demonstrated a marked reduction in the number of cells penetrating the lower chamber following PLXNB2 gene knockout, indicative of a substantial inhibition of the migration and invasion capabilities of head and neck cancer cells (Fig. 14 G, H) 4 Discussion HNSCC is one of the most common malignant tumors in the world, with a high metastatic rate and low survival rate, often accompanied by dysfunctions such as articulation and eating and disruption of facial appearance 27 . It was found that HNSCC exhibited immunosuppressive features 28 . Tumor-infiltrated B cells are an essential component of the TME, and they play a fundamental role in the immunotherapeutic response as the primary effectors of humoral immunity 29 . However, in TME, B cells exert both pro- and anti-tumor effects. On the one hand, B cells exert their anti-tumor effects in three ways: 1. secreting tumor-specific antibodies, promoting T cell functions 30 ; 2. secreting Granzyme B, interferon-γ (IFNγ), and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) to kill tumor cells directly 31–33 ; 3. secreting immunoglobulins to mediate the processes associated with complement-dependent cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cell-mediated phagocytosis (ADCP) 34 . On the other hand, B cells can also promote tumor progression through several mechanisms, such as TIM-1 + Bregs. B cells can produce IL10, which inhibits the immune-killing effect of T cells 15 . B cells can also produce cytokines such as IL-1β to promote the growth and invasion of cancer cells 35 . CD20 + B cells in TME are mainly associated with a good prognosis in different solid tumors, including melanoma, colorectal, gastric, and non-small cell lung cancer 34–37 . Distel et al. showed that CD20 + tumor-infiltrating lymphocytes are prognostic indicators of favorable outcomes in early-stage HNSCC. Conversely, an inverse correlation emerged in advanced stages, implying a plasticity in B cells' functional and compositional attributes throughout the disease 40 . Consequently, B cells may be significant in the prognosis of HNSCC patients and may work as a predictive biomarker for immunotherapy 41 . Historically, investigations into B cells have predominantly relied on Immunohistochemistry (IHC) or flow cytometry techniques, which inherently constrain the precise categorization of B cell populations. Typically, based on their differentiation progression, B cells can be stratified into three primary subgroups: naïve B cells, MBCs, and Antibody-Secreting Cells (ASCs), each manifesting distinctive phenotypic profiles and functions in TME 42 . Due to the classification, functionality, and spatial presence of different B cell subgroups in different tissues and the low proportion of B cells in specific tumors, B cells are also commonly classified as CD19 + CD20 + B cells and CD138 + SDC1 + PCs. The application of scRNA-seq provides higher resolution for identifying B cell subgroups. In our study, we analyzed the intratumoral B cell profiles within HNSCC and investigated the functional information of tumor-responsive B cells. From a longitudinal perspective, we used cluster analysis to divide B cells into NBs, GCBs, MBs, and PCs for the first time. High-throughput scRNA-seq and pseudotime trajectory analysis showed that in the early stage of tumor-infiltrating B cell development in HNSCC, NBs and GCBs with high differentiation ability were primarily present. In contrast, in the later stage of development, a large number of MBs and a small number of PCs appeared. Nevertheless, the composition of B cells changes differently during the development of different tumors and promotes or inhibits tumor development. In advanced non-small cell lung cancer (NSCLC), a decrease in naive-like B cells is associated with a poor prognosis, whereas plasma-like B cells inhibit cancer cell growth in the early stage of NSCLC but promote cell growth in the advanced stage of NSCLC 43 . B cells isolated from early-stage colorectal cancer (CRC) exhibited a pre-B cell-like phenotype characterized by the expression of tumor suppressor molecules. In contrast, B cells derived from advanced CRC displayed a propensity for differentiation into plasma cells and predicted a poor prognosis 44 . In addition, the composition of tumor-infiltrating B cells varies in many ways across the transversal dimensions of the tumor. The objective diversity of B cell subgroups in different tumors and the subjective individualized approach of single-cell analysis contribute to the diversity of B cell subgroup definitions. For example, a higher proportion of IgG plasma cells were found at the tumor sites than in the adjacent normal mucosal tissues in CRC 45 ; however, in a study in NSCLC, there were more B cells in the adjacent tissues than in the tumor, which may be related to infiltration of inflammatory cells 43 . B cells show less infiltration in esophageal squamous cell carcinoma (ESCC), breast invasive carcinoma (BRCA), kidney cancer (KIRC), and hepatocellular carcinoma (HCC) than lung cancer (LC), colon adenocarcinoma (COAD), and skin cutaneous melanoma (SKCM) 46 . The above results can reveal the existence of specific B cell subgroups of TME in different tissues, thus revealing the immunogenic or immunosuppressive of TME. The germinal center (GC) is located in the follicular zone of the secondary lymphoid organs. This particular structure aggregates B cells during the humoral immune response 47 . Antigen-specific B cells accumulate in this remarkable structure and subsequently expand to produce high-frequency mutations in somatic cells, undergo antibody-type switching, and eventually turn into high-affinity B cells against the antigen, a specific means of targeting foreign invading antigens 48–49 . B cells undergo somatic high-frequency mutations along with high-affinity clones directed against antigens, and the clones produce immunoglobulin molecules with higher affinity for antigens. 48–49 . Increased immunoglobulin affinity also requires antibody-type conversion. This process that requires the involvement of an enzyme called activation-induced cytidine deaminase (AID), which is highly expressed in B cells in the germinal center and is an essential enzyme in regulating the immune response in the germinal center, as well as being involved in the process of high-frequency mutation in somatic cells described above 52–53 . AID-mediated high-frequency mutations and type conversions make B-cell mutation rates much more likely, implying the potential for generating uncontrolled cancer cell clones and that germinal centers are critical sites for lymphocytic cancer production 54 . The immune response to GC depends on T fh cell, found in the follicular zone of secondary lymphoid organs, and co-localizes with GCBs to assist in B cell differentiation and clonal selection 55–56 . T fh cells secrete cytokines and membrane surface molecules in direct contact with the GCBs, transmitting signals to the GCBs to help them differentiate into MBs or high-affinity PCs 57–58 . The recruitment of suitable T fh cells into the GC during the humoral immune response is vital for the maturation of antibody affinity and the generation of an effective humoral immune response. This paper finds that molecules on GCBs enable B-cell action on tumor cells by expressing the high-affinity ligands CD99 and SEMA4, which are involved in synaptic junctions to transmit signals for intercellular communication. SEMA4 is a member of the Semaphorin family of molecules, named for the fact that family members all have a sema structural domain consisting of 7 β-folded sheets in their molecular structures, consists of 8 subfamilies, and initially associated with axon guidance and neural development 59–61 . Because of its functional specificity, the immune system is most commonly characterized by migratory movements of cells. The most crucial function of Semaphorin signaling is its involvement in the migratory process of immune cells. Dendritic cells (DCs) require SEMA4A for T-cell activation, and Th1 cells also precisely express SEMA4A, promoting the differentiation of activated T cells into Th1 subgroups 62 . SEMA4D is also abundantly expressed on activated T cells and has been reported to stimulate the production of high-affinity antibodies by humoral immunity 63–64 . Semaphorin involves several physiological processes such as cardiac development 65 , angiogenesis and regeneration 66–67 , osteoclastogenesis 68 , and immune regulation 69 . It also engages in many pathological human disorders, including tumorigenesis and metastasis 70 , neurogenic disorders 71 , and immune disorders 72 . A growing amount of evidence suggests that Semaphorin plays an essential role in tumor development and holds promise as a potential therapeutic target in various tumors. CD99 is a 32 kDa transmembrane protein encoded by the pseudoautosomal gene MIC2 73–74 . CD99 is well known to participate in several crucial biological processes, including cell adhesion 75–76 , migration 77–78 , apoptosis 77,79 , T cell differentiation 80–81 , inflammation 82 , and immune responses 83 , which suggests that CD99 plays a vital role in the peripheral immune response and cell differentiation. CD99 exerts dual effects on the clinical outcomes of patients afflicted with solid tumors. On the one hand, CD99 may act as a tumor suppressor in certain tumors. In patients with NSCLC, increased levels of CD99 stromal expression are associated with better prognosis 84 . In osteosarcoma, CD99 expression is down-regulated in the majority of patient samples. In contrast, overexpression of CD99 significantly negatively impacts osteosarcoma cell growth and survival, eliminating tumorigenicity and metastatic capacity 85 . On the other hand, CD99 can also be an oncogene for tumor migration and invasion. CD99 is strongly correlated with low survival and poor prognosis in gastric adenocarcinoma 86 . Overexpression of CD99 significantly increased migration and invasiveness of glioma cells and decreased survival 87–88 . High expression of CD99 is a distinguishing feature of Ewing sarcoma (EWS). As a result, CD99 stands out as the most frequently used diagnostic marker and therapeutic target for EWS 89 . Our study found that tumor cells are the only target cells of CD99 and SEMA4 signaling pathway, and the source of CD99 and SEMA4 ligands is also only GCBs, so it is speculated that GCBs may also be able to promote the development of HNSCC through CD99 and SEMA4 signaling pathway. Hence, CD99 and SEMA4 might serve as robust biomarkers for HNSCC and a promising target for tumor therapy. By integrating risk scores and clinical characteristics, we created a risk score model based on 18 GCBs risk genes, and the independent prognostic value of this model was subsequently validated. We used CIBERSORT, ssGSEA to analyze immune cell infiltration in HNSCC samples and to distinguish between different immune cell subgroups in the immune microenvironment, which have different prognostic implications. Compared with patients in the high-risk group, the low-risk group had a higher immune score and a greater abundance of tumor-infiltrating immune cells (TIIC), which resulted in a better prognosis. The low-risk group had more CD8 + T cell infiltration, and the high-risk group had more macrophage M2 infiltration. CD8 + T cells are the central effector cells in anti-tumor immunity by directly lysing tumor cells by releasing multiple cytokines and enzymes 90–91 . In the breast cancer tumor microenvironment interacting with tumor cells, the higher the number of CD8 + T cells, the better the regulation of anti-tumor activity 92 . CD8 + T cells are closely linked to the survival of tumor patients, with patients with higher CD8 + T cell counts having a higher 5-year survival rate 93 and a better prognosis 94–95 . M2 macrophages can suppress tumor immunity by secreting the immunosuppressive factors IL-10 and TGF-β, and at the same time, secrete EGF to promote neovascularisation within the tumor, which promotes tumor invasion and metastasis 96–97 . Increased infiltration of M2 macrophages in breast cancers indicates a poor prognosis 98–99 . These findings are consistent with our results in HNSCC. In summary, the ratio of immune effector cells to suppressor cells, the location of immune cell infiltration, and the degree of immune cell activation in TME also affect the prognosis of patients and their clinical immunotherapy response. There are some limitations to this study. Firstly, the data in this study came from part of the open single-cell cohort, which may not be representative because of the small sample size of B-cell studies in HNSCC. Secondly, there is a lack of studies on the association of B cells with treatment efficacy in HNSCC, which needs to be analyzed and studied in the future by further expanding the sample size. Thirdly, some studies related to assessing B cells on prognosis and efficacy in HNSCC are still controversial. These differences may be because B cell subgroups are highly functionally plastic and heterogeneous, so the composition and localization of tumor-infiltrating B cells vary. Hence, future validation in clinical trials is necessary. In conclusion, we report that GCBs are B-cell-specific prognostic biomarkers for survival in HNSCC. The emergence and development of HNSCC are dependent on somatic cell gene mutations and are closely connected with the tumor microenvironment. All anti-tumor or pro-tumor factors form a network in the tumor microenvironment that determines tumor initiation, progression, and outcome. Therefore, we can further apply B cells in diagnosis and therapeutic evaluation and develop targeted therapies against the tumor immune microenvironment. These results underscore a compelling rationale for employing B-cell markers in patient stratification and exploring innovative therapeutics targeting B-cell subgroups. Abbreviations Head and neck squamous carcinoma (HNSCC) Overall survival (OS) Tumor immune microenvironment (TIME) Bulk RNA-sequencing (bulk RNA-seq) Single-cell RNA-sequencing (scRNA-seq) Head and Neck Cancer (HNSC) Follicular helper T cells (T fh cells) B cell receptor (BCR) Peripheral blood (PBL) Differentially expressed genes (DEGs) Naive B cells (NBs) Germinal center B cells (GCBs) Memory B cells (MBs) Plasma cells (PCs) Gene Expression Omnibus (GEO) Tumor mutational burden (TMB) Computed Tomography (CT) Magnetic Resonance Imaging (MRI) Tumor immune microenvironment (TIME) Regulatory B cell subgroups (Bregs) GCBs Least Absolute Shrinkage and Selection Operator (LASSO) Area Under the ROC Curve (AUC) Tumor microenvironment (TME) T cells regulatory (Tregs) Gene Ontology (GO) Kyoto Encyclopedia of Genes and Genomes (KEGG) Interferon-γ (IFNγ) Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) Complement-dependent cytotoxicity (CDC) Antibody-dependent cell-mediated cytotoxicity (ADCC) Antibody-dependent cell-mediated phagocytosis (ADCP) Immunohistochemistry (IHC) Antibody-Secreting Cells (ASCs) Non-small cell lung cancer (NSCLC) Colorectal cancer (CRC) Esophageal squamous cell carcinoma (ESCC) Breast invasive carcinoma (BRCA) Kidney cancer (KIRC hepatocellular carcinoma (HCC) Lung cancer (LC) Colon adenocarcinoma (COAD) Skin cutaneous melanoma (SKCM) Germinal center (GC) Activation-induced cytidine deaminase (AID) Dendritic cells (DCs) Ewing sarcoma (EWS) Declarations Ethics approval and consent to participate Ethical approval for this study was obtained from the Ethics Committee of the First Affiliated Hospital of Anhui Medical University. As the other data used in this study are publicly available, no ethical approval was required. Consent for publication All authors have read the manuscript and agree to publish. Availability of data and material The raw data for this study were generated at the corresponding archives, further inquiries can be directed to the corresponding authors. Authors’ contributions Li Lin, Zhijie Zhao, and Can Xiao conceived and designed the research. Zhijie Zhao, Jiani Zou, Wenyi Huang, and Yichi Zhang downloaded and collected the data. Zhijie Zhao, Jiani Zou, and Li Lin analyzed the data and wrote the article. Shengbin Pei and Yantao Ding performed the experimental part of the paper. Shengbin Pei and Yantao Ding conducted quality control on the articles. Zhijie Zhao and Jiani Zou guided the submission. All authors read and approved the final manuscript. Li Lin, Jiani Zou, and Shengbin Pei are co-first authors. Zhijie Zhao, Yantao Ding, and Can Xiao are co-correspondent author. Competing interests The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Funding This study was funded by the National Natural Science Foundation of China (82301158), Shanghai Post-doctoral Excellence Incentive Program (2023M742332), Foundation of National Center for Translational Medicine (Shanghai) SHU Branch (SUITM-202304), Clinical Research Plan of SHDC (SHDC2020CR3070B), Shanghai Jiao Tong University School of Medicine Two-hundred Talent (20161420), Clinical Research Program of 9th People's Hospital affiliated to Shanghai Jiao Tong university School of Medicine (JYLJ202108), Interdisciplinary Program of Shanghai Jiao Tong University (YG2022QN048), Shanghai Clinical Research Center of Plastic and Reconstructive Surgery supported by Science and Technology Commission of Shanghai Municipality (22MC1940300) and the First Affiliated Hospital of Soochow University Department of Stomatology Clinical Trial Institution Capacity Improvement(LGY2016035). Conflict of Interest: None. References C.R. Leemans, P. Snijders, R.H. Brakenhoff, The molecular landscape of head and neck cancer. Nat. Rev. Cancer. 18 (5), 269–282 (2018) A. Wyss et al., Cigarette, cigar, and pipe smoking and the risk of head and neck cancers: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. Am. J. Epidemiol. 178 (5), 679–690 (2013) J.W. Carlisle et al., An update on the immune landscape in lung and head and neck cancers. CA Cancer J Clin. 70 (6), 505–517 (2020) D.C. Travassos et al., Squamous cell carcinoma antigen as a prognostic marker and its correlation with clinicopathological features in head and neck squamous cell carcinoma: Systematic review and meta-analysis. J. Oral Pathol. Med. 47 (1), 3–10 (2018) S.S. Sawant, S.M. Zingde, M.M. Vaidya, Cytokeratin fragments in the serum: their utility for the management of oral cancer. Oral Oncol. 44 (8), 722–732 (2008) A. Kulasinghe, L. Kenny, C. Punyadeera, Circulating tumour cell PD-L1 test for head and neck cancers. Oral Oncol. 75 , 6–7 (2017) R.M. Nagler et al., Early diagnosis and treatment monitoring roles of tumor markers Cyfra 21 – 1 and TPS in oral squamous cell carcinoma. Cancer. 85 (5), 1018–1025 (1999) M.B. Amin et al., The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more personalized approach to cancer staging. CA Cancer J Clin. 67 (2), 93–99 (2017) DE D. Paz et al., Prognostic Roles of SCC Antigen, CRP and CYFRA 21 – 1 in Oral Cavity Squamous Cell Carcinoma. Anticancer Res. 39 (4), 2025–2033 (2019) T. Fu et al., Spatial architecture of the immune microenvironment orchestrates tumor immunity and therapeutic response. J. Hematol. Oncol. 14 (1), 98 (2021) T. Tang et al., Advantages of targeting the tumor immune microenvironment over blocking immune checkpoint in cancer immunotherapy. Signal. Transduct. Target. Ther. 6 (1), 72 (2021) C.M. Laumont et al., Tumour-infiltrating B cells: immunological mechanisms, clinical impact and therapeutic opportunities. Nat. Rev. Cancer. 22 (7), 414–430 (2022) S.S. Kim et al., Role of B Cells in Responses to Checkpoint Blockade Immunotherapy and Overall Survival of Cancer Patients. Clin. Cancer Res. 27 (22), 6075–6082 (2021) O. Demaria et al., Harnessing innate immunity in cancer therapy. Nature. 574 (7776), 45–56 (2019) S.S. Wang et al., Tumor-infiltrating B cells: their role and application in anti-tumor immunity in lung cancer. Cell. Mol. Immunol. 16 (1), 6–18 (2019) J.R. Conejo-Garcia et al., Neglected no more: B cell-mediated anti-tumor immunity. Semin Immunol. 65 , 101707 (2023) J. Tang et al., B Cells and Tertiary Lymphoid Structures Influence Survival in Lung Cancer Patients with Resectable Tumors. Cancers (Basel), 2020. 12(9). D. Pretscher et al., Distribution of immune cells in head and neck cancer: CD8 + T-cells and CD20 + B-cells in metastatic lymph nodes are associated with favourable outcome in patients with oro- and hypopharyngeal carcinoma. BMC Cancer. 9 , 292 (2009) A. Lechner et al., Tumor-associated B cells and humoral immune response in head and neck squamous cell carcinoma. Oncoimmunology. 8 (3), 1535293 (2019) O. Wood et al., Gene expression analysis of TIL rich HPV-driven head and neck tumors reveals a distinct B-cell signature when compared to HPV independent tumors. Oncotarget. 7 (35), 56781–56797 (2016) P.H. Li et al., Recent developments in application of single-cell RNA sequencing in the tumour immune microenvironment and cancer therapy. Mil Med. Res. 9 (1), 52 (2022) S. Sun et al., Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis. Genome Biol. 20 (1), 269 (2019) X. Qiu et al., Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods. 14 (10), 979–982 (2017) J. Qi et al., Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat. Commun. 13 (1), 1742 (2022) P. Charoentong et al., Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell. Rep. 18 (1), 248–262 (2017) A.M. Newman et al., Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods. 12 (5), 453–457 (2015) K. Yoshihara et al., Inferring tumour purity and stromal and immune cell admixture from expression data. Nat. Commun. 4 , 2612 (2013) M. Ashburner et al., Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 25 (1), 25–29 (2000) M. Kanehisa et al., KEGG: integrating viruses and cellular organisms. Nucleic Acids Res. 49 (D1), D545–D551 (2021) A. Subramanian et al., Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U S A 102 (43), 15545–15550 (2005) D. Maeser, R.F. Gruener, R.S. Huang, oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief. Bioinform, 2021. 22 (6) P. Geeleher, N. Cox, R.S. Huang, pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PLoS One. 9 (9), e107468 (2014) A. Cerutti, M. Cols, I. Puga, Marginal zone B cells: virtues of innate-like antibody-producing lymphocytes. Nat. Rev. Immunol. 13 (2), 118–132 (2013) S. Gu et al., Significance of intratumoral infiltration of B cells in cancer immunotherapy: From a single cell perspective. Biochim. Biophys. Acta Rev. Cancer. 1876 (2), 188632 (2021) K. Starska-Kowarska, The Role of Different Immunocompetent Cell Populations in the Pathogenesis of Head and Neck Cancer-Regulatory Mechanisms of Pro- and Anti-Cancer Activity and Their Impact on Immunotherapy. Cancers (Basel), 2023. 15 (6) F. Bray et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 68 (6), 394–424 (2018) R. Mandal et al., The head and neck cancer immune landscape and its immunotherapeutic implications. JCI Insight. 1 (17), e89829 (2016) Z.N. Willsmore et al., Cells in Patients With Melanoma: Implications for Treatment With Checkpoint Inhibitor Antibodies. Front. Immunol. 11 , 622442 (2020) D. Ghosh et al., New insights into B cells as antigen presenting cells. Curr. Opin. Immunol. 70 , 129–137 (2021) A.M. Gocher, C.J. Workman, D. Vignali, Interferon-gamma: teammate or opponent in the tumour microenvironment? Nat. Rev. Immunol. 22 (3), 158–172 (2022) W.L. Xu et al., Granzyme B-Producing B Cells Function as a Feedback Loop for T Helper Cells in Liver Transplant Recipients with Acute Rejection. Inflammation. 44 (6), 2270–2278 (2021) G.V. Sharonov et al., cells, plasma cells and antibody repertoires in the tumour microenvironment. Nat. Rev. Immunol. 20 (5), 294–307 (2020) S. Herter et al., Preclinical activity of the type II CD20 antibody GA101 (obinutuzumab) compared with rituximab and ofatumumab in vitro and in xenograft models. Mol. Cancer Ther. 12 (10), 2031–2042 (2013) G.J. Yuen, E. Demissie, S. Pillai, B lymphocytes and cancer: a love-hate relationship. Trends Cancer. 2 (12), 747–757 (2016) A. Hennequin et al., Tumor infiltration by Tbet + effector T cells and CD20 + B cells is associated with survival in gastric cancer patients. Oncoimmunology. 5 (2), e1054598 (2016) J. Berntsson et al., Prognostic impact of tumour-infiltrating B cells and plasma cells in colorectal cancer. Int. J. Cancer. 139 (5), 1129–1139 (2016) A. Ladanyi et al., Prognostic impact of B-cell density in cutaneous melanoma. Cancer Immunol. Immunother. 60 (12), 1729–1738 (2011) K.I. Al-Shibli et al., Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer. Clin. Cancer Res. 14 (16), 5220–5227 (2008) L.V. Distel et al., Tumour infiltrating lymphocytes in squamous cell carcinoma of the oro- and hypopharynx: prognostic impact may depend on type of treatment and stage of disease. Oral Oncol. 45 (10), e167–e174 (2009) M. Wouters, B.H. Nelson, Prognostic Significance of Tumor-Infiltrating B Cells and Plasma Cells in Human Cancer. Clin. Cancer Res. 24 (24), 6125–6135 (2018) S. Gu et al., Significance of intratumoral infiltration of B cells in cancer immunotherapy: From a single cell perspective. Biochim. Biophys. Acta Rev. Cancer. 1876 (2), 188632 (2021) J. Chen et al., Single-cell transcriptome and antigen-immunoglobin analysis reveals the diversity of B cells in non-small cell lung cancer. Genome Biol. 21 (1), 152 (2020) W. Wang et al., Multiregion single-cell sequencing reveals the transcriptional landscape of the immune microenvironment of colorectal cancer. Clin. Transl Med. 11 (1), e253 (2021) J. Xia et al., Single-cell landscape and clinical outcomes of infiltrating B cells in colorectal cancer. Immunology. 168 (1), 135–151 (2023) Y. Zheng et al., Immune suppressive landscape in the human esophageal squamous cell carcinoma microenvironment. Nat. Commun. 11 (1), 6268 (2020) G.D. Victora, M.C. Nussenzweig, Germinal Centers. Annu. Rev. Immunol. 40 , 413–442 (2022) C. Berek, A. Berger, M. Apel, Maturation of the immune response in germinal centers. Cell. 67 (6), 1121–1129 (1991) J. Jacob et al., Intraclonal generation of antibody mutants in germinal centres. Nature. 354 (6352), 389–392 (1991) M.G. Weigert et al., Variability in the lambda light chain sequences of mouse antibody. Nature. 228 (5276), 1045–1047 (1970) D. McKean et al., Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin. Proc. Natl. Acad. Sci. U S A 81 (10), 3180–3184 (1984) M. Muramatsu et al., Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell. 102 (5), 553–563 (2000) R. Pavri, M.C. Nussenzweig, AID targeting in antibody diversity. Adv. Immunol. 110 , 1–26 (2011) A. Nussenzweig, M.C. Nussenzweig, Origin of chromosomal translocations in lymphoid cancer. Cell. 141 (1), 27–38 (2010) D. Breitfeld et al., Follicular B helper T cells express CXC chemokine receptor 5, localize to B cell follicles, and support immunoglobulin production. J. Exp. Med. 192 (11), 1545–1552 (2000) P. Schaerli et al., CXC chemokine receptor 5 expression defines follicular homing T cells with B cell helper function. J. Exp. Med. 192 (11), 1553–1562 (2000) K. Luthje et al., The development and fate of follicular helper T cells defined by an IL-21 reporter mouse. Nat. Immunol. 13 (5), 491–498 (2012) D. Eto et al., IL-21 and IL-6 are critical for different aspects of B cell immunity and redundantly induce optimal follicular helper CD4 T cell (Tfh) differentiation. PLoS One. 6 (3), e17739 (2011) A.L. Kolodkin, D.J. Matthes, C.S. Goodman, The semaphorin genes encode a family of transmembrane and secreted growth cone guidance molecules. Cell. 75 (7), 1389–1399 (1993) P.K. Hota, M. Buck, Plexin structures are coming: opportunities for multilevel investigations of semaphorin guidance receptors, their cell signaling mechanisms, and functions. Cell. Mol. Life Sci. 69 (22), 3765–3805 (2012) Y. Luo, D. Raible, J.A. Raper, Collapsin: a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell. 75 (2), 217–227 (1993) A. Kumanogoh et al., Nonredundant roles of Sema4A in the immune system: defective T cell priming and Th1/Th2 regulation in Sema4A-deficient mice. Immunity. 22 (3), 305–316 (2005) A. Kumanogoh et al., Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for regulating B cell signaling. Immunity. 13 (5), 621–631 (2000) C. Gu et al., Semaphorin 3E and plexin-D1 control vascular pattern independently of neuropilins. Science. 307 (5707), 265–268 (2005) A. Casazza et al., Sema3E-Plexin D1 signaling drives human cancer cell invasiveness and metastatic spreading in mice. J. Clin. Invest. 120 (8), 2684–2698 (2010) F. Maione et al., Semaphorin 3A overcomes cancer hypoxia and metastatic dissemination induced by antiangiogenic treatment in mice. J. Clin. Invest. 122 (5), 1832–1848 (2012) F. Maione et al., Semaphorin 3A is an endogenous angiogenesis inhibitor that blocks tumor growth and normalizes tumor vasculature in transgenic mouse models. J. Clin. Invest. 119 (11), 3356–3372 (2009) M. Hayashi et al., Osteoprotection by semaphorin 3A. Nature. 485 (7396), 69–74 (2012) G.M. Delgoffe et al., Stability and function of regulatory T cells is maintained by a neuropilin-1-semaphorin-4a axis. Nature. 501 (7466), 252–256 (2013) L. Tamagnone, Emerging role of semaphorins as major regulatory signals and potential therapeutic targets in cancer. Cancer Cell. 22 (2), 145–152 (2012) S. Nojima et al., A point mutation in Semaphorin 4A associates with defective endosomal sorting and causes retinal degeneration. Nat. Commun. 4 , 1406 (2013) T. Okuno, Y. Nakatsuji, A. Kumanogoh, The role of immune semaphorins in multiple sclerosis. FEBS Lett. 585 (23), 3829–3835 (2011) P.N. Goodfellow et al., MIC2: a human pseudoautosomal gene. Philos. Trans. R Soc. Lond. B Biol. Sci. 322 (1208), 145–154 (1988) M. Pasello, M.C. Manara, K. Scotlandi, CD99 at the crossroads of physiology and pathology. J. Cell. Commun. Signal. 12 (1), 55–68 (2018) V. Cerisano et al., Molecular mechanisms of CD99-induced caspase-independent cell death and cell-cell adhesion in Ewing's sarcoma cells: actin and zyxin as key intracellular mediators. Oncogene. 23 (33), 5664–5674 (2004) W. Kasinrerk et al., CD99 monoclonal antibody induce homotypic adhesion of Jurkat cells through protein tyrosine kinase and protein kinase C-dependent pathway. Immunol. Lett. 71 (1), 33–41 (2000) A. Bremond et al., Regulation of HLA class I surface expression requires CD99 and p230/golgin-245 interaction. Blood. 113 (2), 347–357 (2009) S.S. Yoon et al., Engagement of CD99 triggers the exocytic transport of ganglioside GM1 and the reorganization of actin cytoskeleton. FEBS Lett. 540 (1–3), 217–222 (2003) G. Bernard et al., The E2 molecule (CD99) specifically triggers homotypic aggregation of CD4 + CD8 + thymocytes. J. Immunol. 154 (1), 26–32 (1995) G. Bernard et al., The E2 molecule (CD99) specifically triggers homotypic aggregation of CD4 + CD8 + thymocytes. J. Immunol. 154 (1), 26–32 (1995) G. Bernard et al., Apoptosis of immature thymocytes mediated by E2/CD99. J. Immunol. 158 (6), 2543–2550 (1997) E.M. Dufour et al., CD99 is essential for leukocyte diapedesis in vivo. Cell. Commun. Adhes. 15 (4), 351–363 (2008) R.L. Watson et al., Endothelial CD99 signals through soluble adenylyl cyclase and PKA to regulate leukocyte transendothelial migration. J. Exp. Med. 212 (7), 1021–1041 (2015) K. Edlund et al., CD99 is a novel prognostic stromal marker in non-small cell lung cancer. Int. J. Cancer. 131 (10), 2264–2273 (2012) M.C. Manara et al., CD99 acts as an oncosuppressor in osteosarcoma. Mol. Biol. Cell. 17 (4), 1910–1921 (2006) J.H. Lee et al., Clinical significance of CD99 down-regulation in gastric adenocarcinoma. Clin. Cancer Res. 13 (9), 2584–2591 (2007) H.J. Seol et al., Overexpression of CD99 Increases the Migration and Invasiveness of Human Malignant Glioma Cells. Genes Cancer. 3 (9–10), 535–549 (2012) L.C. Cardoso et al., CD99 Expression in Glioblastoma Molecular Subtypes and Role in Migration and Invasion. Int. J. Mol. Sci., 2019. 20 (5) N. Riggi, M.L. Suva, I. Stamenkovic, Ewing's Sarcoma. N Engl. J. Med. 384 (2), 154–164 (2021) A. Ocana et al., Tumor-infiltrating lymphocytes in breast cancer: ready for prime time? J. Clin. Oncol. 33 (11), 1298–1299 (2015) C. Gu-Trantien et al., CD4(+) follicular helper T cell infiltration predicts breast cancer survival. J. Clin. Invest. 123 (7), 2873–2892 (2013) W.H. Fridman et al., The immune contexture in human tumours: impact on clinical outcome. Nat. Rev. Cancer. 12 (4), 298–306 (2012) S.M. Mahmoud et al., Tumor-infiltrating CD8 + lymphocytes predict clinical outcome in breast cancer. J. Clin. Oncol. 29 (15), 1949–1955 (2011) A. Rody et al., T-cell metagene predicts a favorable prognosis in estrogen receptor-negative and HER2-positive breast cancers. Breast Cancer Res. 11 (2), R15 (2009) D.G. DeNardo et al., CD4(+) T cells regulate pulmonary metastasis of mammary carcinomas by enhancing protumor properties of macrophages. Cancer Cell. 16 (2), 91–102 (2009) I. Shihab et al., Understanding the Role of Innate Immune Cells and Identifying Genes in Breast Cancer Microenvironment. Cancers (Basel), 2020. 12(8). A. Mantovani, A. Sica, Macrophages, innate immunity and cancer: balance, tolerance, and diversity. Curr. Opin. Immunol. 22 (2), 231–237 (2010) S.D. Heys et al., Characterisation of tumour-infiltrating macrophages: impact on response and survival in patients receiving primary chemotherapy for breast cancer. Breast Cancer Res Treat. 135 (2), 539–548 (2012) I. Shabo et al., Breast cancer expression of CD163, a macrophage scavenger receptor, is related to early distant recurrence and reduced patient survival. Int. J. Cancer. 123 (4), 780–786 (2008) Additional Declarations No competing interests reported. Supplementary Files FigureS1.tif FigureS2.tif FigureS3.tif FigureS4.tif FigureS5.tif SupplementaryTableI.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-3804853","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":263550094,"identity":"624d6180-7313-4e88-9f90-f9d025f81eec","order_by":0,"name":"Li Lin","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Lin","suffix":""},{"id":263550095,"identity":"aa4b3b74-91d7-4a56-bacd-aa3033d053cd","order_by":1,"name":"Jiani Zou","email":"","orcid":"","institution":"Shandong University of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Jiani","middleName":"","lastName":"Zou","suffix":""},{"id":263550098,"identity":"4d3e009c-34e6-45d1-8f44-1b80c89fa534","order_by":2,"name":"Shengbin Pei","email":"","orcid":"","institution":"National Cancer Center, Chinese Academy of Medical Sciences and Peking Union Medical College","correspondingAuthor":false,"prefix":"","firstName":"Shengbin","middleName":"","lastName":"Pei","suffix":""},{"id":263550102,"identity":"9b880a57-9478-40ee-bd15-4ccccf7d810e","order_by":3,"name":"Wenyi Huang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Wenyi","middleName":"","lastName":"Huang","suffix":""},{"id":263550104,"identity":"e3d26aac-3a7b-4c9f-a962-f81964523acf","order_by":4,"name":"Yichi Zhang","email":"","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":false,"prefix":"","firstName":"Yichi","middleName":"","lastName":"Zhang","suffix":""},{"id":263550106,"identity":"50b948f3-adbf-4421-803c-e687d5ec038d","order_by":5,"name":"Zhijie Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYBACNvnDhx98MJCo52dvPvggoaKGsBY+CbY0wxkVFgmSPceSDR6cOUZYi5wEj4E0x5mKBIMZPmaSD1uYiXCYdFuCMWObRJ6BBINZRWIDGwN/e3cCfi0yhw88LmyTKDaXbki7kbhDhkHizNkN+LUwpCUYz2yTYNw558CxG4ln2BgMJHIJackxkOYFatlwI7GtILGNmQgtEkAtPGckEjfcSGZjIE4LzzFQIEsYAwOZWSLhzDEegn6Rb28GRWWdHD97/8ePPypq5Pjbe/FrwQA8pCkfBaNgFIyCUYAVAACDg0wrHqb8BQAAAABJRU5ErkJggg==","orcid":"","institution":"Shanghai Jiao Tong University","correspondingAuthor":true,"prefix":"","firstName":"Zhijie","middleName":"","lastName":"Zhao","suffix":""},{"id":263550108,"identity":"be00b343-6d50-4abe-a8b2-27a60de133ad","order_by":6,"name":"Yantao Ding","email":"","orcid":"","institution":"Anhui Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yantao","middleName":"","lastName":"Ding","suffix":""},{"id":263550109,"identity":"2ed1ab65-b640-40d3-b9a0-9c4fb691588b","order_by":7,"name":"Can Xiao","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Can","middleName":"","lastName":"Xiao","suffix":""}],"badges":[],"createdAt":"2023-12-25 14:29:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3804853/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3804853/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":49080284,"identity":"8537bdbe-9b3d-41f4-979b-d6303cf588e1","added_by":"auto","created_at":"2024-01-02 19:55:07","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":761083,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIntegration and analysis of scRNA-seq data of B and plasma cells in HNSCC.\u003c/strong\u003e \u003cstrong\u003e(A-B)\u003c/strong\u003e The UMAP plots of the expression profiles of the 7082 B and plasma cells derived from tumor tissues and PBL of HNSCC. B cells are classified into 16 distinct transcriptional clusters. HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e), Phase (G1, S, G2M). \u003cstrong\u003e(C)\u003c/strong\u003e The distribution of CNVscore, nCount_RNA, S.Score, G2M.Scorce for each subgroup was visualized using UMAP. \u003cstrong\u003e(D)\u003c/strong\u003e Pie chart showed the proportions of each subgroup in the HNSCC samples. \u003cstrong\u003e(E)\u003c/strong\u003e The proportion of subgroups derived from PBL and tumor tissues. \u003cstrong\u003e(F)\u003c/strong\u003e Visualization of sample characteristics associated with different B and plasma cell subgroups and different HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e), Phase (G1, S, G2M), G2M.Score, S.Score, nCount_RNA, CNVscore, differential gene expression. \u003cstrong\u003e(G)\u003c/strong\u003e Volcano plots showed DEGs in subgroups.\u003c/p\u003e","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/7df89964cdc39b9eaba6581d.png"},{"id":49080285,"identity":"5f3f049f-9c9f-408f-be6d-08f0779f5226","added_by":"auto","created_at":"2024-01-02 19:55:07","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":528681,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTrajectory of B and plasma cells.\u003c/strong\u003e \u003cstrong\u003e(A)\u003c/strong\u003e The differentiation trajectory of B and plasma cells showed the derivation of HNSCC cells, based on the pseudotime of cells (from left to right). \u003cstrong\u003e(B)\u003c/strong\u003e Pseudotime differentiation trajectory plots of B and plasma cell subgroups. \u003cstrong\u003e(C)\u003c/strong\u003e The UMAP visualization delineates the pseudotime scores of the entire spectrum of cells. The violin plot illustrates the pseudotime scores attributed to each distinct subgroup. The ridge map provides an insight into the density distribution of pseudotime values. \u003cstrong\u003e(D-E)\u003c/strong\u003eBar charts of the changes in the proportion of B and plasma cell subgroups in the pseudotime state, respectively.\u003c/p\u003e","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/a1c2b4834e32aa65ca0db0d8.png"},{"id":49080286,"identity":"2f7db53c-999c-4e55-b888-7d38619676ba","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":174621,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLandscape of NBs subgroups in the HNSCC. (A)\u003c/strong\u003e The distribution of UMAP for the 6 subgroups in NBs. \u003cstrong\u003e(B)\u003c/strong\u003e Visualisation of sample characteristics associated with different NBs subgroups and different HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e), Phase (G1, S, G2M), G2M.Scorce, S.Score, nCount_RNA, CNVscore, differential gene expression. \u003cstrong\u003e(C)\u003c/strong\u003e Expression of top5 marker genes in different subgroups. \u003cstrong\u003e(D)\u003c/strong\u003e The distribution of CNVscore, nCount_RNA, S.Score, G2M.Scorce for each subgroup was visualised using UMAP. \u003cstrong\u003e(E)\u003c/strong\u003e Results of GO-BP enrichment analysis of differential genes and their corresponding transcription factors and surface proteins in 6 NBs subgroups. \u003cstrong\u003e(F)\u003c/strong\u003e Histogram of HPV infection status of NBs subgroups. \u003cstrong\u003e(G)\u003c/strong\u003e The cell cycle of NBs subgroups. \u003cstrong\u003e(H-I)\u003c/strong\u003e Violin plots of GO-BP Negative regulation of B cell activation and CNV score. \u003cstrong\u003e(J-L)\u003c/strong\u003e GSEA enrichment analysis among NBs subgroups.\u003c/p\u003e","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/086039f009ca094fe2a731d4.png"},{"id":49082104,"identity":"82f21836-068e-4480-89ed-afcfe735e6c4","added_by":"auto","created_at":"2024-01-02 20:11:08","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":78533,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLandscape of GCBs subgroups in the tumor microenvironment of HNSCC.\u003c/strong\u003e \u003cstrong\u003e(A)\u003c/strong\u003e 332 GCBs were subjected to UMAP dimension reduction, divided into two cell clusters including C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells (C11 GCBs) and C14 STMN1\u003csup\u003e+\u003c/sup\u003e Germinal center B cells (C14 GCBs). The contribution of pie chart to patients (13 patients), HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e) and Phase (G1, S, G2M) were visualized. \u003cstrong\u003e(B-C)\u003c/strong\u003e The bubble chart and UMAP showed the expression of GCBs top marker genes. \u003cstrong\u003e(D)\u003c/strong\u003e Visualization of relevant sample features for 332 GCBs using UMAP dimensionality reduction clustering: nCount_RNA, CNVscore, G2M.Scorce, S.Score. \u003cstrong\u003e(E)\u003c/strong\u003e The violin plot shows the nCount_RNA, CNVscore, G2M.Scorce, S.Score of C11 GCBs and C14 GCBs and their differences. *\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05, **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01, ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001, means significant difference, ns means no significant difference. The boxplots showed the median (middle line), upper and lower quantiles (boxplot), and range of the data (violin plot). \u003cstrong\u003e(F)\u003c/strong\u003e nCount_RNA, CNVscore, G2M.Scorce, S.Score of GCBs with different HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e) and their differences. **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05, **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01, ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001, means significant difference, ns means no significant difference. The boxplots showed the median (middle line), upper and lower quantiles (boxplot), and range of the data (violin plot). \u003cstrong\u003e(G)\u003c/strong\u003e The proportion of C11 GCBs and C14 GCBs in PBL or tumor group among patients (above). The ratio of C11 GCBs and C14 GCBs in different groups of HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e) among patients (below). \u003cstrong\u003e(H)\u003c/strong\u003e The C11 GCBs and C14 GCBs ratios between the HPV\u003csup\u003e+\u003c/sup\u003e and HPV\u003csup\u003e- \u003c/sup\u003egroups. \u003cstrong\u003e(I-J)\u003c/strong\u003e Bar graphs depicted the difference in cell proportion between subgroups and phases, as well as the difference in cell proportion between different HPV infection statuses and phases. \u003cstrong\u003e(K)\u003c/strong\u003e GSEA enrichment analysis of GCBs subgroups. The DEGs of C11 GCBs and C14 GCBs were screened, and the enrichment items of GOBP were scored by GSEA. NES \u0026gt; 0 was up-regulated and \u0026lt; 0 was down-regulated. NES, N stands for standardization and ES for enrichment scores. \u003cstrong\u003e(L)\u003c/strong\u003e Volcano plots showing expression of DEGs in GCBs subgroups. \u003cstrong\u003e(M)\u003c/strong\u003e The heat map showed the active biological processes of each subgroup of GCBs. \u003cstrong\u003e(N)\u003c/strong\u003e Word cloud for GO-BP enrichment analysis of GCBs DEGs.\u003c/p\u003e","description":"","filename":"OnlineFigure4.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/6a0c77a9aedb24eb5ad6af28.png"},{"id":49081502,"identity":"d6831879-f95e-4b6c-822a-882eb472ef44","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":163515,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLandscape of MBs subgroups in the HNSCC microenvironment.\u003c/strong\u003e \u003cstrong\u003e(A) \u003c/strong\u003e3253 MBs were subjected to UMAP dimension reduction, divided into 6 cell clusters including C1 CRIP1\u003csup\u003e+\u003c/sup\u003e Memory B cells(C1 MBs, 868)，C2 LINC01781\u003csup\u003e+\u003c/sup\u003e Memory B cells(C2 MBs, 738)，C3 PPP1R15A\u003csup\u003e+\u003c/sup\u003e Memory B cells(C3 MBs, 620)，C4 IFI30\u003csup\u003e+\u003c/sup\u003e Memory B cells(C4 MBs, 614) ，C9 CD1C\u003csup\u003e+\u003c/sup\u003e Memory B cells(C9 MBs, 258) ，C13 TLE5\u003csup\u003e+\u003c/sup\u003e Memory B cells(C13 MBs, 155). The contribution of pie chart to HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e) and Phase (G1, S, G2M) were visualized. \u003cstrong\u003e(B-C)\u003c/strong\u003e The bubble chart and UMAP showed the expression of MBs top5 marker genes. \u003cstrong\u003e(D) \u003c/strong\u003eVisualization of relevant sample features for 3253 MBs using UMAP dimensionality reduction clustering: nCount_RNA, CNVscore, G2M.Scorce, S.Score. \u003cstrong\u003e(E-F)\u003c/strong\u003e The histogram showed the distribution of HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e) and Phase (G1, G2M, S) in each subgroup of MBs intuitively. \u003cstrong\u003e(G-H)\u003c/strong\u003e The violin plot shows the nCount_RNA, CNVscore, G2M.Scorce, S.Score of MBs subgroups and their differences. **P \u0026lt; 0.05, **P \u0026lt; 0.01, ***P \u0026lt; 0.001, means significant difference, ns means no significant difference. The boxplots showed the median (middle line), upper and lower quantiles (boxplot), and range of the data (violin plot). \u003cstrong\u003e(I) \u003c/strong\u003eThe activity of different pathways in MBs subgroups was demonstrated by GO-BP enrichment analysis and word cloud. \u003cstrong\u003e(J)\u003c/strong\u003e The correlation of DEGs among MBs subgroups in positive regulation of cell communication，positive regulation of cell death，positive regulation of cellular metabolic process was analyzed by GSEA enrichment analysis.\u003c/p\u003e","description":"","filename":"OnlineFigure5.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/70bc8225a03a9e92fa5cf0e6.png"},{"id":49080292,"identity":"254fe388-c844-4da8-96bd-19d3d8551968","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":129000,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLandscape of PCs subgroups in the HNSCC microenvironment.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A)\u003c/strong\u003e The distribution of UMAP plots for the 2 subgroups in PCs. \u003cstrong\u003e(B) \u003c/strong\u003eUMAP plots were used to illustrate the distribution of CNVscore, nCount_RNA, S.Score, and G2M.Score for each subgroup. \u003cstrong\u003e(C) \u003c/strong\u003eVisualisation of sample characteristics associated with different PCs and different HPV infection status (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e-\u003c/sup\u003e): Phase (G1, S, G2M), G2M.Scorce, S.Score, nCount_RNA, CNVscore, differential gene expression. \u003cstrong\u003e(D)\u003c/strong\u003e Expression of marker genes in different PCs. \u003cstrong\u003e(E-G)\u003c/strong\u003e The histogram displayed the differences in cell percentage across different HPV groups and phase groups. \u003cstrong\u003e(H-I)\u003c/strong\u003e Results of GO-BP enrichment analysis of differential genes and their corresponding transcription factors and surface proteins in 2 PCs.\u003c/p\u003e","description":"","filename":"OnlineFigure6.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/c198d9876fb8f8425e2d2f93.png"},{"id":49081505,"identity":"44814ff4-c778-4dca-8a53-629bd7855c69","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":106102,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCell communication network.\u003c/strong\u003e \u003cstrong\u003e(A)\u003c/strong\u003eCircle visualization of interaction strength (left) and number (right) among the B and plasma cell subgroups and tumor cells. \u003cstrong\u003e(B) \u003c/strong\u003eCell and communication patterns. \u003cstrong\u003e(C) \u003c/strong\u003eIllustration of the incoming and outgoing interaction strengths for each of the B and plasma cell subgroups and tumor cells. \u003cstrong\u003e(D)\u003c/strong\u003e Incoming communication patterns of target cells (left). Outgoing communication patterns of secreting cells (right). \u003cstrong\u003e(E)\u003c/strong\u003e Heatmap shows an overview of cell-cell interactions in both outgoing and incoming signaling patterns. \u003cstrong\u003e(F)\u003c/strong\u003e Interaction bubble diagram among GCBs, MBs and tumor cells. \u003cstrong\u003e(G)\u003c/strong\u003e The circle plots show the interaction strength (left) and number (right) of B and plasma cell subgroups on tumor cells. \u003cstrong\u003e(H)\u003c/strong\u003eCircle visualization of interaction strength (left) and number (right) about tumor cells on B cell subgroups.\u003c/p\u003e","description":"","filename":"OnlineFigure7.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/cbec981f4da1bf9396611f9b.png"},{"id":49081504,"identity":"4a489d66-0bc8-42f9-8ba1-6d5469695cbf","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":70915,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGCBs based regulatory network detect tumor-specifc signaling pathways in HNSCC.\u003c/strong\u003e \u003cstrong\u003e(A, F)\u003c/strong\u003eIllustration of the incoming and outgoing interaction strengths for CD99 and SEMA4 signaling pathways. \u003cstrong\u003e(B, G) \u003c/strong\u003eHeatmaps show dominant senders, receivers, mediators and influencers in CD99 and SEMA4 signaling pathways of C11 GCBs inferred by network centrality score. \u003cstrong\u003e(C, H)\u003c/strong\u003e The violin diagram depicts the gene expression levels in the CD99 and SEMA4 signaling pathway networks including B and plasma cells as well as malignant cells. \u003cstrong\u003e(D)\u003c/strong\u003eThe circle diagram shows the interaction of CD99 ligand on MBs and tumor cells. \u003cstrong\u003e(I)\u003c/strong\u003e The circle plot shows the interaction of SEMA4 ligand on GCBs and tumor cells. \u003cstrong\u003e(E, J)\u003c/strong\u003e The hierarchy plots of signaling pathway networks indicate the regulation of CD99 and SEMA4 on B cells and tumor cells in HNSCC, as predicted by CellChat anlysis.\u003c/p\u003e","description":"","filename":"OnlineFigure8.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/42793482534486a71bf12e53.png"},{"id":49081507,"identity":"22180dd6-8b54-4bc1-b597-e80f3a0c0a29","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":288938,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScreening for prognosis-related memory B cells in HNSCC patients.\u003c/strong\u003e \u003cstrong\u003e(A, F)\u003c/strong\u003eUnivariate Cox regression analysis was performed to identify the prognosis-related genes. \u003cstrong\u003e(B, G) \u003c/strong\u003eLasso regression analysis of prognosis-related genes. \u003cstrong\u003e(C, H)\u003c/strong\u003e Multivariate COX regression analysis on OS in the TCGA cohort. (HR \u0026gt; 1 protective factor, \u0026lt; 1 risk factor.) \u003cstrong\u003e(D, I)\u003c/strong\u003e Kaplan–Meier curves were created to estimate OS for high- and low-score groups from TCGA cohort (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.0001). \u003cstrong\u003e(E, J)\u003c/strong\u003e Kaplan-Meier curve showed that the prognosis of HNSCC with high expression of ATP6V0E1, PGK1, TPM3, ADK, LGALS1, TKT was poor in OS, while that of HNSCC with high expression of MS4A1, TRAC, BLK, ITGB7 was better in OS.\u003c/p\u003e","description":"","filename":"OnlineFigure9.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/e8342d5627da75dc226e8955.png"},{"id":49080294,"identity":"584ddce4-14cd-4a28-8c0d-afccdf26017c","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":439854,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScreening for prognosis-related germinal center B cells in HNSCC patients. (A, F)\u003c/strong\u003e Univariate Cox regression analysis was performed to identify the prognosis-related genes. \u003cstrong\u003e(B, G)\u003c/strong\u003e Lasso regression analysis of prognosis-related genes. \u003cstrong\u003e(C, H) \u003c/strong\u003eMultivariate COX regression analysis on OS in the TCGA cohort. (HR \u0026gt; 1 protective factor, \u0026lt; 1 risk factor.). \u003cstrong\u003e(D, I)\u003c/strong\u003e Kaplan–Meier curves were created to estimate OS for high- and low-score groups from TCGA cohort (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.0001). \u003cstrong\u003e(E, J, K)\u003c/strong\u003eKaplan-Meier curve showed that the prognosis of HNSCC with high expression of HMMR, HSP90AA1, PRDX6, ACTB, BASP1, EZR, RRAS2 was poor in OS, while that of HNSCC with high expression of LRMP, RGS13, CD38, LCK, SMIM14, TCL1A was better in OS. \u003cstrong\u003e(L) \u003c/strong\u003eThe bar diagrams show the distribution of the expression of CD38, SMIM14 and TCL1A in N0, N1, N2, N3, NX.\u003c/p\u003e","description":"","filename":"OnlineFigure10.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/e75455ca7bea36dd06b9bd9a.png"},{"id":49080296,"identity":"860b2d58-ff82-40b9-81a3-b91775212fca","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":537485,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMEF2B\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e GCB score. (A)\u003c/strong\u003e Visualization of genes constituted the MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score in different GCBs subgroups and different HPV infection statuses (HPV\u003csup\u003e+\u003c/sup\u003e vs. HPV\u003csup\u003e-\u003c/sup\u003e), Phase (G1, S, G2M), and G2M.Scorce, S.Score, nCount_RNA, CNVscore, differential gene expression. \u003cstrong\u003e(B-C) \u003c/strong\u003eExpression of 18 prognosis-related genes in C11 and C14 GCBs subgroups. \u003cstrong\u003e(D)\u003c/strong\u003e Distribution and median value of risk scores of HNSCC patients. Distribution of OS status and risk scores of HNSCC patients. The heatmap showed the expression of 18 prognosis-related genes in high and low-risk groups. \u003cstrong\u003e(E)\u003c/strong\u003e Pie charts depict the differences in clinical characteristics (status, stage, TNM, phase) between high and low-risk groups in HNSCC. \u003cstrong\u003e(F)\u003c/strong\u003e Kaplan–Meier curve showing the OS rate differences between high and low-risk groups. \u003cstrong\u003e(G) \u003c/strong\u003eTime-dependent ROC curves and AUC values evaluate the prognostic performance of the risk score model at 1, 3, and 5 years. \u003cstrong\u003e(H) \u003c/strong\u003eAn overview of the correlation between 18 prognosis-related genes. Red indicates a positive correlation, and blue indicates a negative correlation. \u003cstrong\u003e(I) \u003c/strong\u003eACTB, BASP1, EZR, PRDX6, and RRAS2 were positively correlated with risk scores. CD38, LCK, RGS13, SMIM14, and TCL1A were negatively correlated with risk score (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). \u003cstrong\u003e(J)\u003c/strong\u003e The differential expression of ACTB, BASP1, EZR, PRDX6, and RRAS2 in high and low-risk groups. \u003cstrong\u003e(K)\u003c/strong\u003e ACTB, BASP1, EZR, PRDX6, RRAS2, and risk score were negatively correlated with OS. \u003cstrong\u003e(L-M) \u003c/strong\u003eBoth EZR and RRAS2 expressions were all positively associated with the other four genes (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05). \u003cstrong\u003e(N)\u003c/strong\u003e The differences of EZR and RRAS2 in high and low risk groups, different age groups, different races and different TNM stages.\u003c/p\u003e","description":"","filename":"OnlineFigure11.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/820aea0c6fc76ec5a9f7bbc5.png"},{"id":49080298,"identity":"4f319642-6010-4aa2-ba2c-ee3333b1e6ef","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":245131,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and verification of Nomogram and analysis of immune cell infiltration. (A)\u003c/strong\u003e Multivariate analyses of the prognosis model. \u003cstrong\u003e(B) \u003c/strong\u003eNomogram plot constructed by combining the risk score and clinical features. \u003cstrong\u003e(C-D) \u003c/strong\u003eAUC values evaluate the prognostic performance of the risk score model at 1, 3 and 5 years. \u003cstrong\u003e(E)\u003c/strong\u003e Deviations of actual 1-year, 3-year, and 5-year survival probabilities. \u003cstrong\u003e(F) \u003c/strong\u003eA heatmap of principal component analysis for the immune cell infiltration between high and low risk groups. \u003cstrong\u003e(G-H)\u003c/strong\u003e Differential analysis of estimate score, immune score, stromal score, and tumor purity between high and low risk groups. \u003cstrong\u003e(I) \u003c/strong\u003eA boxplot of immune cells in the two risk subgroups. \u003cstrong\u003e(J) \u003c/strong\u003eDifferential analysis of immune cell infiltration distribution between high and low risk groups. \u003cstrong\u003e(K) \u003c/strong\u003eThe correlation between risk score and immune cells. The horizontal axis represents the correlation with the risk score, the vertical axis represents the proportion of immune cells, and the size of the circle represents the absolute value of the correlation coefficient. The larger the circle, the higher the correlation. \u003cstrong\u003e(L) \u003c/strong\u003eThe correlation between prognostic genes and immune cells. Red indicates positive correlation, blue indicates negative correlation, *\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05, **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01. \u003cstrong\u003e(M) \u003c/strong\u003eAn overview of the correlation of immune cells. \u003cstrong\u003e(N) \u003c/strong\u003eThe distribution of prognostic genes, estimate score, immune score, stromal score, tumor purity and immune cells in high and low risk groups. \u003cstrong\u003e(O) \u003c/strong\u003eRRAS2 was positively correlated with NBs, MBs, PB. \u003cstrong\u003e(P) \u003c/strong\u003eRRAS2 was negatively correlated with stromal score, immune score and tumor purity. \u003cstrong\u003e(Q) \u003c/strong\u003eAnalysis of the correlation between RRAS2 and immune cells.\u003c/p\u003e","description":"","filename":"OnlineFigure12.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/b48028a7fb9530edc8e356b3.png"},{"id":49080291,"identity":"d17d2e6a-6cb6-4d38-b4ea-98258f5d0268","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":120999,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(A-B)\u003c/strong\u003e DEGs analysis between high and low risk groups. \u003cstrong\u003e(C-E) \u003c/strong\u003eFunctional enrichment analysis between the high and low risk groups of HNSCC patients. \u003cstrong\u003e(C)\u003c/strong\u003e Results of the GO enrichment analysis. \u003cstrong\u003e(D) \u003c/strong\u003eResults of the KEGG enrichment analysis. \u003cstrong\u003e(E)\u003c/strong\u003eEnrichment plots from the gene set enrichment analysis. Biological processes enriched in the high-risk group of HNSCC patients. Signaling pathways enriched in the low-risk HNSCC samples. \u003cstrong\u003e(F)\u003c/strong\u003e The waterfall map showed the distribution of the top 20 genes of TMB in the high and low risk groups. The horizontal coordinates represent the different samples and the vertical coordinates represent the genes. The different colors in the note below represent different types of mutations, the upper bars represent the probability of mutation in each sample, and the right bars represent the proportion of different types of mutations in each gene. \u003cstrong\u003e(G) \u003c/strong\u003eThe CNV amplifcation, deletion or absence of data of risk genes. \u003cstrong\u003e(H) \u003c/strong\u003eAnalysis of the correlation between gene and gene mutation. Mutually exclusive: if one gene is mutated, there is a high probability that no mutation will occur in the corresponding gene. Co−occurrence: two genes mutate at the same time. \u003cstrong\u003e(I) \u003c/strong\u003eThe sample numbers, type, and location of mutations in each gene. The figure notes below represent the different mutation types. The dots in the figure represent the samples in which the mutation occurred, and the connected vertical lines indicate the location of the mutation. \u003cstrong\u003e(J) \u003c/strong\u003eTMB in high and low risk groups. \u003cstrong\u003e(K) \u003c/strong\u003eCorrelation analysis between risk scores and TMB (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001). \u003cstrong\u003e(L) \u003c/strong\u003eSurvival analysis of high risk-high TMB, high risk-low TMB, low risk-high TMB, low risk-low TMB. \u003cstrong\u003e(M) \u003c/strong\u003eAUC values evaluate the prognostic performance of the TMB model at 1, 3 and 5 years. \u003cstrong\u003e(N-P)\u003c/strong\u003e Chemotherapy and targeted drugs sensitivity of high- and low-risk subgroups.\u003c/p\u003e","description":"","filename":"OnlineFigure13.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/f97ba49e0488d409c3d3c482.png"},{"id":49081506,"identity":"70915e60-7c5a-4fcb-b349-b6dab1a4d527","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"png","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":115570,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePLXNB2 significantly affects the proliferation and migration of head and neck cancer cell lines.\u003c/strong\u003e \u003cstrong\u003e(A, B) \u003c/strong\u003eCCK-8 experiment. After PLXNB2 knockdown, the proliferation ability of HN-5 and UMSCC-47 cell lines decreased significantly. \u003cstrong\u003e(C, D)\u003c/strong\u003e Plate cloning experiment. After PLXNB2 knockdown, the colony-forming ability of HN-5 and UMSCC-47 cell lines was significantly decreased. \u003cstrong\u003e(E, F)\u003c/strong\u003e wound healing test. After PLXNB2 knockdown, the migration ability of HN-5 and UMSCC-47 cell lines decreased significantly. \u003cstrong\u003e(G, H)\u003c/strong\u003e Transwell experiment. After PLXNB2 knockdown, the migration and invasion ability of HN-5 and UMSCC-47 cell lines were significantly reduced. (*\u003cem\u003eP\u003c/em\u003e\u0026lt;0.05; **\u003cem\u003eP\u003c/em\u003e\u0026lt;0.01; ***\u003cem\u003eP\u003c/em\u003e\u0026lt;0.001)\u003c/p\u003e","description":"","filename":"OnlineFigure14.png","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/1d258cdf622f1746aa7c4765.png"},{"id":49083364,"identity":"024219cf-eca1-4763-8d1d-d405fa1ebc6f","added_by":"auto","created_at":"2024-01-02 20:27:13","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6700208,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/f3b4dfce-2e03-444f-b3f7-f4e08fbb1f9b.pdf"},{"id":49080302,"identity":"5fa3e6bb-f69d-4aa9-8384-c31b9a8c8a41","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":7954384,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS1.tif","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/0fe16be594936ba12ff8fb11.tif"},{"id":49080301,"identity":"10f3853a-0be0-4240-9ac9-d1e6c3b718d5","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":10145312,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS2.tif","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/d42a165d08eac9ac89978689.tif"},{"id":49081508,"identity":"fcd0981e-7ec5-4d29-bc13-0d5b3d633057","added_by":"auto","created_at":"2024-01-02 20:03:08","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":17308984,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS3.tif","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/5749faf7da490229fa17fbe7.tif"},{"id":49080300,"identity":"68af1560-536e-47b6-96c0-02d029298917","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":8714496,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS4.tif","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/afb4953feb9e883170e47bdd.tif"},{"id":49080297,"identity":"a460ae09-1328-4fbc-b333-6c18f8f72d01","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":5209372,"visible":true,"origin":"","legend":"","description":"","filename":"FigureS5.tif","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/546ea90f3fd5b1ac20c46a25.tif"},{"id":49080289,"identity":"6cbe33d7-7017-4b93-9c80-b4a2e4714785","added_by":"auto","created_at":"2024-01-02 19:55:08","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":16054,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTableI.docx","url":"https://assets-eu.researchsquare.com/files/rs-3804853/v1/c4fd1fb0892ee7fd8f91a862.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eGerminal center B-cell subgroups in the tumor microenvironment cannot be overlooked: their involvement in prognosis, immunotherapy response, and treatment resistance in head and neck squamous carcinoma\u003c/p\u003e","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eHead and neck squamous carcinoma (HNSCC), an immunosuppressive disease characterized by molecular heterogeneity and tumor-host interactions, has been increasing in incidence and mortality year by year and has now become the sixth most common cancer and the eighth leading cause of cancer deaths worldwide\u003csup\u003e1\u003c/sup\u003e. About 90% of head and neck malignancies are squamous cell carcinomas, and early diagnosis and treatment can lead to a better prognosis\u003csup\u003e2\u0026ndash;3\u003c/sup\u003e. The diagnosis and treatment of HNSCC have made significant progress in the past few decades. Surgical treatment, radiotherapy, and chemotherapy can significantly improve the disease control rate and survival of patients with HNSCC. However, the overall survival (OS) of patients with HNSCC is still relatively low, with the 5-year survival rate at all stages around 40\u0026ndash;50%, mainly due to the recurrence of tumors and metastasis after treatment\u003csup\u003e4\u0026ndash;6\u003c/sup\u003e. However, in the treatment of HNSCC, due to the specificity of the tumor site, surgical treatment may change the corresponding anatomical structure, and radiotherapy will form fibrous scars in the corresponding area, thus making it particularly difficult to achieve early detection of local recurrence directly through physical examination or re-biopsy\u003csup\u003e7\u003c/sup\u003e. The rest of the diagnostic tools, such as Computed Tomography (CT), ultrasound, and Magnetic Resonance Imaging (MRI), are not sensitive to early lesions, so we must rely on specific biomarkers to predict early recurrence and metastasis of tumors. Improved survival in patients with HNSCC depends partly on disease stage, lymph node involvement, patient response to treatment, and early diagnosis of local tumor recurrence. Nevertheless, it has been suggested that the clinical staging of tumors is not a good reflection of the behavior of the tumor and the outcome of the disease\u003csup\u003e8\u003c/sup\u003e. This is why a growing number of studies have concluded, by evaluating clinical and pathological indicators, that non-advanced tumors tend to be accompanied by early recurrence and that advanced tumors tend to have longer disease-free intervals\u003csup\u003e9\u003c/sup\u003e. These factors have prompted the need to search for essential tumor markers for preoperative assessment of the likelihood of regional metastasis, the significance of postoperative adjuvant therapy, assessment of the patient's response to treatment, and prognostic analysis of the patient.\u003c/p\u003e \u003cp\u003eTumor immune microenvironment (TIME), which consists of immune-related components such as immune cells and their secreted immunomodulatory molecules, has become a hot topic in oncology research in recent years due to its critical regulatory role in the growth of tumors as well as metastasis\u003csup\u003e10\u0026ndash;11\u003c/sup\u003e. As an essential component of TIME, there is growing evidence that B cells can participate in the immune response to tumors as well as immune evasion through the secretion of specific antibodies, the processing of presenting antigens, and the maintenance of the structure and function of tertiary lymphatic structures, which are biologically important\u003csup\u003e12\u0026ndash;13\u003c/sup\u003e. However, research on TIME has focused mainly on T cells, with relatively few B cell-related studies\u003csup\u003e14\u003c/sup\u003e. The role of B cells is dual. On the one hand, most B cells can inhibit tumor growth through antibody production, secretion of anti-tumor cytokines, and direct killing. On the other hand, some B cell subgroups may also play a tumor-promoting role. For example, a group of regulatory B cell subgroups (Bregs) may promote tumor growth and invasion through the secretion of specific tumor-promoting cytokines (e.g., IL-10, TGF-β, IL-35) and other mechanisms\u003csup\u003e15\u003c/sup\u003e. In addition, B cells are closely associated with the prognosis of various tumors, including skin melanoma, gastric cancer, and lung cancer\u003csup\u003e16\u0026ndash;17\u003c/sup\u003e. An early study of 33 HNSCC patients demonstrated improved outcomes with increased peritumoral B-cells in lymph node metastases. Recent analyses of B-cell phenotypes and responses in HNSCC have delineated notable heterogeneity\u003csup\u003e18\u003c/sup\u003e. Nevertheless, the impact of B cells on survival or their influence on responses to immunotherapy or radiation remained unexplored\u003csup\u003e19\u0026ndash;20\u003c/sup\u003e, possibly due to the lack of uniformity in the choice of B cell markers.\u003c/p\u003e \u003cp\u003eIn recent decades, the emergence of bulk RNA-sequencing (bulk RNA-seq) has revolutionized oncology research and has been widely used in diagnosing and prognostic assessing cancer patients. However, there are individual differences between different cell types in tumor tissues. Still, bulk sequencing can only reflect the overall average gene expression level in the whole sample and cannot accurately characterize the gene expression at the individual cell level. Single-cell RNA-sequencing (scRNA-seq) allows access to the genetic signature of a particular cell type by extracting, reverse transcribing, amplifying, and sequencing RNA at the level of a single cell, better revealing heterogeneity in the intra-tumor microenvironment and providing assistance in the development of individualized therapeutic regimens\u003csup\u003e21\u003c/sup\u003e. Although scRNA-seq technology can characterize genes in different cell types more accurately than traditional bulk RNA-seq, bulk RNA-seq datasets have larger sample sizes and more complete follow-up information. Therefore, there is a need to integrate scRNA-seq and bulk RNA-seq datasets to identify new tumor biomarkers and develop more accurate tumor prognosis-related models.\u003c/p\u003e \u003cp\u003eIn summary, the biological functions of B cells are complex, and their prognostic value in HNSCC remains unknown. Screening for appropriate B-cell markers in HNSCC and in-depth exploration of their biological functions and prognostic value could help to assess the prognosis. It may provide new targets for the treatment of patients with HNSCC.\u003c/p\u003e"},{"header":"2 Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data collection and processing\u003c/h2\u003e \u003cp\u003eThe GPL18573 platform-based Head and neck squamous cell carcinomas (HNSCCs) -related single-cell RNA sequencing (scRNA-seq)data GSE164690 [PMID: 34921143] was downloaded from Gene Expression Omnibus (GEO, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). In addition, bulk RNA sequencing data of HNSCC tissue samples were downloaded from the Cancer Genome Atlas (TCGA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://portal.gdc.cancer.gov/\u003c/span\u003e\u003cspan address=\"https://portal.gdc.cancer.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) database, along with clinical information including age, gender, TNM stage, race, and survival time.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Single-cell sequencing data analysis\u003c/h2\u003e \u003cp\u003eUsing the Seurat R package (version 4.3.0), raw gene expression substrates were imported into R software (version 4.2.2). Using the DoubletFinder R package (version 2.0.3), the scRNA-seq data underwent rigorous quality control.\u003c/p\u003e \u003cp\u003eThe percentage of mitochondrial and erythrocyte-related genes in each cell was calculated using the \"PercentageFeatureSet\" function in the Seurat package, followed by cell quality control, with the following exclusion criteria: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) the number of genes in the cell was \u0026lt;\u0026thinsp;300; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) the number of genes in the cell was \u0026gt;\u0026thinsp;7500; (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) the percentage of mitochondrial genes in the cell was \u0026gt;\u0026thinsp;25.\u003c/p\u003e \u003cp\u003eThen use the \u0026ldquo;NormalizeData\u0026rdquo; function for normalization (normalization.method = \"LogNormalize\", scale.factor\u0026thinsp;=\u0026thinsp;10000). Use the \u0026ldquo;CellCycleScoring\u0026rdquo; function to determine cell cycle conditions. Find the top 2000 top variable genes for subsequent analysis using the \u0026ldquo;FindVariableFeatures\u0026rdquo; function. The data are standardized using the \"ScaleData\" function. The harmony R package (version 0.1.1) was used to remove batch effects between samples. The dim value was set to 30 and the resolution to 1.2. The \"RunPCA\" function was used to cluster the dataset and to downscale the uniform manifold approximation and projection (UMAP) visualisation\u003csup\u003e100\u003c/sup\u003e. Marker genes for each B-cell subgroup were compiled by searching previous literature, and the clusters were manually annotated after the downscaling of the subgroups. Next, the \"FindAllMarkers\" function was used to find differentially expressed genes in each B-cell subpopulation, where min.pct was set to 0.25 and logfc.threshold was set to 0.25.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Pseudotime analysis\u003c/h2\u003e \u003cp\u003eWe performed pseudotime trajectory analysis of B and plasma cell subgroups using the Monocle R package (version 2.24.0)\u003csup\u003e101\u003c/sup\u003e. The UMAP method was employed to reduce dimensionality, and the functions \"plot_cell_trajectory\" was utilized for visualization. Different cell subpopulations were arranged in a proposed chronological order, and pseudotemporal heatmaps were used to identify and display genes that changed simultaneously with pseudotime.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Cell-cell communication analysis\u003c/h2\u003e \u003cp\u003eThe CellChat R package (version 1.6.1) was applied to infer cell-to-cell interactions among cell subgroups and establish a regulatory network based on ligand-receptor crosstalk\u003csup\u003e102\u003c/sup\u003e. The \u0026ldquo;netVisual_circle\u0026rdquo;, \u0026ldquo;netVisual_aggregate\u0026rdquo;, \u0026ldquo;netVisual_individual\u0026rdquo;, \u0026ldquo;netVisual_heatmap\u0026rdquo;, \u0026ldquo;netVisual_bubble\u0026rdquo;, \u0026ldquo;netVisual_chord_gene\u0026rdquo;, \u0026ldquo;netAnalysis_river\u0026rdquo;, \u0026ldquo;netAnalysis_dot\u0026rdquo;, and \u0026ldquo;netAnalysis_signalingRole_heatmap\u0026rdquo; functions were used to visualize cell-to-cell interactions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Construction of a risk score in HNSCC patients\u003c/h2\u003e \u003cp\u003eThe expression values of signature genes of patients in the GDC TCGA Head and Neck Cancer (HNSC) cohort were extracted after obtaining the list of MBs or GCBs signature genes (top 100 genes). Signature genes associated with OS of HNSCC patients were screened using univariate Cox regression analysis. To avoid the multicollinearity of these genes, we further filtered them using LASSO regression (glmnet, version 4.1-6) to fileted these genes. The risk coefficients for each gene were then calculated using multivariate Cox regression. Risk scores of the prognostic risk score model (X: coefficient, Y: gene expression level) = \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\sum _{\\text{i}}^{\\text{n}}\\text{X}\\text{i}\\times \\text{Y}\\text{i}\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003ePatients were categorized into high and low-risk groups based on median of risk scores. Kaplan-Meier survival analysis (version 3.3-1) was performed in the TCGA-HNSC cohort to compare the prognosis of patients in different risk groups. The predictive performance of risk score was further evaluated by plotting the area under the curve of ROC (AUC) over time using the timeROC R package (version 0.4).\u003c/p\u003e \u003cp\u003eRisk score, age, TNM stage, and molecular subtype were subjected to were subjected to multivariate Cox regression analysis to screen out independent prognostic factors in HNSCC patients. The nomogram column-line graph prediction model for predicting 1-, 3-, and 5-year overall survival of HNSCC patients was constructed using the \"rms\" package (version 6.7-0). The constructed nomogram prediction model was validated using the ROC curve and calibration curve.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Immune infiltration analysis\u003c/h2\u003e \u003cp\u003ePrevious literature was queried to obtain marker genes for different immune cells, and ssGSEA enrichment analysis was performed to assess the level of immune cell infiltration for each sample in the TCGA-HNSC cohort\u003csup\u003e103\u003c/sup\u003e. The relative proportion of immune cells was quantified using the CIBERSORT R package (version 0.1.0) \u003csup\u003e104\u003c/sup\u003e. The Wilcox.test was used to compare the levels of immune cell infiltration and the expression levels of joint exempted checkpoints between different risk groups. Correlation analyses between immune cell infiltration levels of each type and between immune cell infiltration levels and risk genes were performed using the \"corrplot\" R package (version 0.92). Stromal and immune cells in HNSCC tissues were estimated using the ESTIMATE R package (version 1.0.13), and stromal score, immune score, ESTIMATE score, and tumor purity were calculated\u003csup\u003e105\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Functional enrichment analysis\u003c/h2\u003e \u003cp\u003eThe gene expression matrices of HNSCC patients in the high and low-risk subgroups were analyzed differently by the limma R package (version 3.54.0). The GO\u003csup\u003e106\u003c/sup\u003e and KEGG\u003csup\u003e107\u003c/sup\u003e were used to analyze the DEGs between different risk subgroups by using clusterProfiler R package (version 4.6.0). We downloaded the list of pathways gene sets from the MSIGDB database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and performed GSEA\u003csup\u003e108\u003c/sup\u003e enrichment analyses on these pathways.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Tumor mutation burden analysis\u003c/h2\u003e \u003cp\u003eThe maftools package was utilized to analyze the overall mutation status of the tumor, TMB, as well as the transformation status of SNPs. The Wilcox.test was used to compare the TMB levels between risk subgroups of patients, and the Spearman correlation test was performed to determine the relationship between risk score and TMB.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Prediction of drug sensitivity\u003c/h2\u003e \u003cp\u003eThe Genomics of Drug Sensitivity in Cancer (GDSC)(\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cancerrxgene.org/)i\u003c/span\u003e\u003cspan address=\"https://www.cancerrxgene.org/)i\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003es a public dataset containing information about drug sensitivity and molecular markers of drug response in cancer\u003csup\u003e109\u003c/sup\u003e. The dataset information was downloaded as reference files, and the pRRophetic R package (version 0.5)\u003csup\u003e110\u003c/sup\u003e was used to predict the sensitivity to chemotherapy and targeted therapies for each sample. The Wilcox.test was performed to determine whether there was a statistically significant difference in medicine sensitivity between the high and low risk groups. Pearson's correlation was utilized to examine the relationship between risk scores and drug sensitivity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Multiplexed immunofluorescence assay\u003c/h2\u003e \u003cp\u003eMulticolor IHC staining was performed using a Multiple Immunofluorescence Kit (AiFang biological, AFIHC035) according to the manufacturer\u0026rsquo;s instructions. Briefly, the paraffin-embedded tissues were dewaxed, then the antigen retrieval was performed by sodium citrate at 98℃ for 20 minutes. Ater infiltrating in 3% H\u003csub\u003e2\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e for 15 minutes, the tissues were blocked by 5% BSA. Then the tissues were incubated by primary antibody at 4℃ overnight. Then the tissues were incubated by secondary antibody for 30 minutes and stained by TSA for 5 minutes. For each targe, the steps of antigen retrieval, blocking, incubation of primary antibody, incubation of secondary antibody, and TSA staining are required. Finally, the nuclei were stained with DAPI. PBS was used to wash tissues between each step. These images were obtained by the LEICA DMi8 system and analyzed by the ImageJ software.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Cell culture\u003c/h2\u003e \u003cp\u003eHN-5 cells and UMSCC-47 cells originated from the American Type Culture Collection (ATCC). HN-5 cells underwent cultivation utilizing RPMI 1640 medium, whereas UMSCC-47 cells were subjected to DMEM medium (Gibco BRL, United States). Both cell lines were nurtured within an incubation environment at 37\u0026deg;C, featuring a 5% CO2 concentration.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 Cell transfection and Real-time fluorescent quantitative PCR (RT-qPCR)\u003c/h2\u003e \u003cp\u003eTwo small interfering RNAs (siRNAs) targeting the PLXNB2 gene, along with their corresponding negative controls (si-NC), were synthesized by Ribobio Guangzhou (China). The transfection procedure adhered to the methodologies delineated in the Lipofectamine 3000 instructions provided by Invitrogen, USA. Thermo's TRIzol reagent (15596018) was employed to extract total RNA from the cellular lines, and RNA concentrations were standardized. Subsequently, cDNA synthesis was carried out using Vazyme's PrimeScript\u0026trade; RT kit (R232-01). Quantitative reverse transcription-polymerase chain reaction (qRT-PCR) was executed utilizing the SYBR Green Kit from TaKaRa Biotechnology (Dalian, China), with GAPDH as the internal reference control. Further details regarding primer and siRNA sequences are available in supplemental Table\u0026nbsp;1.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 Cell Counting\u003c/h2\u003e \u003cp\u003eA total of 6000 transfected cells were seeded into individual wells of 96-well plates (Corning, USA, 3599). Following cellular adhesion to the well surfaces, the cells underwent treatment with the CCK-8 labeling reagent (A311-01, Vazyme) on days 1, 2, 3, 4, and 5, respectively. Incubation of the cells occurred in conditions devoid of light for a duration of 2 hours, during which absorbance (OD) values were meticulously recorded.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 Wound healing\u003c/h2\u003e \u003cp\u003eTransfected cells were nurtured in a six-well culture plate (Corning, USA, 3516). Upon achieving a cellular density of approximately 90\u0026ndash;95%, a 200 \u0026micro;l sterile pipette was employed to delicately abrade the cell layer in a linear fashion. Subsequently, a gradual rinsing with PBS was conducted to eliminate unattached cells and debris. Following this, the serum-free cell medium was substituted to sustain optimal cell proliferation. Consistent photographic documentation was undertaken at identical locations, both at the initiation of the culture and 48 hours subsequent to the initiation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.15 Transwell\u003c/h2\u003e \u003cp\u003eIn the Transwell experimental setup, the Transwell chamber (Corning, USA, 341201) was positioned within a 24-well plate. Subsequently, 8,000 cells were uniformly distributed in each chamber and cultured utilizing 200 \u0026micro;l of serum-free DMEM or RPMI 1640 medium. To the lower chamber, 600 \u0026micro;l of complete DMEM or RPMI 1640 medium, enriched with 10% serum, was added. The experimental plate was then placed in an incubation apparatus and cultured for a duration of 36 hours, allowing for the migration of cells from the upper chamber to the lower chamber. Following this incubation period, the Transwell chamber underwent meticulous cleaning, fixation, and staining procedures.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Phenotypic clusters of B cells according to scRNA-seq\u003c/h2\u003e \u003cp\u003eTo investigate the variations in the immune microenvironment across all cell subgroups in HNSCC, we collected scRNA-seq datasets of HNSCC. By utilizing these datasets, we aim to gain a more profound comprehension of the molecular mechanisms at the individual cell level (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e-\u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThrough the standard workflow of single-cell analysis, we identified four cell types in the integrated data, including NBs, MBs, GCBs, and PCs. We plotted the distribution of HPV infection status and phase (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). To further elucidate the heterogeneity and intricacy of HNSCC, we have subdivided them into 16 distinct subgroups. Among these subgroups, the population of NBs encompasses C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells, C5 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells, C7 CD83\u003csup\u003e+\u003c/sup\u003e Naive B cells, C8 ISG15\u003csup\u003e+\u003c/sup\u003e Naive B cells, C10 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells, and C15 HSPA1B\u003csup\u003e+\u003c/sup\u003e Naive B cells. Additionally, the category of MBs includes C1 CRIP1\u003csup\u003e+\u003c/sup\u003e Memory B cells, C2 LINC01781\u003csup\u003e+\u003c/sup\u003e Memory B cells, C3 PPP1R15A\u003csup\u003e+\u003c/sup\u003e Memory B cells, C4 IFI30\u003csup\u003e+\u003c/sup\u003e Memory B cells, C9 CD1C\u003csup\u003e+\u003c/sup\u003e Memory B cells, and C13 TLE5\u003csup\u003e+\u003c/sup\u003e Memory B cells. Furthermore, the subgroup of GCBs comprises C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells and C14 STMN1\u003csup\u003e+\u003c/sup\u003e Germinal center B cells. Lastly, we have identified PCs represented by C6 FKBP11\u003csup\u003e+\u003c/sup\u003e IgG Plasma cells and C12 DERL3\u003csup\u003e+\u003c/sup\u003e IgA Plasma cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). The distribution of CNVscore, nCount_RNA, S.Score, and G2M. The score for each subpopulation is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, in which GCBs on the right side of the UMAP plot have a high CNVscore, a high nCount_RNA, and a more active cell cycle state. In addition, we analyzed the proportion of subgroups derived from peripheral blood (PBL) and tumor tissues. The pie chart displayed the proportion of B and plasma cell subtypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). The results showed that most subgroups belonged to malignant cells, with a small proportion of NBs and MBs originating from the PBL (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). Bubble maps and volcano plots demonstrated the expression of differentially expressed genes (DEGs) in each subgroup and the associated features of different HPV infection statuses (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e\u0026minus;\u003c/sup\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF-G).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Pseudotime trajectories unveiled the differentiation status of B and plasma cell subgroups\u003c/h2\u003e \u003cp\u003eTo understand the dynamic transitional processes of B and plasma cells in HNSCC, we further performed a pseudotime analysis to trace the transcriptional trajectory. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-B, with the progression of pseudotime, the number of NBs gradually decreased, while GCBs and MBs gradually differentiated toward PCs. The decrease in GCBs and MBs may be due to their gradual transformation into PCs, most of PCs were at the end of the cell developmental trajectory. Cellular composition was identified by colouring the pseudotemporal map with tissue of origin. By analysing Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, B \u0026amp; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC, it was found that the pseudotime of NBs was significantly lower than that of other B cell subgroups, and the pseudotime of PCs represented by C6 FKBP11\u003csup\u003e+\u003c/sup\u003e IgG Plasma cells and C12 DERL3\u003csup\u003e+\u003c/sup\u003e IgA Plasma cells was the highest. An enhanced graphical representation, a bar plot, elucidated that the NBs subgroups predominantly comprised neoplastic cells originating from State1. Likewise, the MBs subgroups exhibited a preponderance of cells stemming from State2. Furthermore, the GCBs subgroups was characterized by a prevailing presence of neoplastic cells deriving from State4. Lastly, the PCs subgroups exhibited a notable predominance of cells originating from State5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, E). These observations collectively unveil that the NBs subgroups is positioned within the incipient stage of cellular differentiation, contrasting with the PCs, which signifies the end of differentiation. Additionally, cells within the GCBs and MBs subgroups display an augmented propensity for proliferation and differentiation capacity, thereby highlighting their distinctive biological characteristics.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 NBs are in the initial differentiation stage of pseudotime trajectories\u003c/h2\u003e \u003cp\u003eAccording to our research, B and plasma cells could be classified into NBs, GCBs, MBs, and PCs. Each of these subgroups manifests distinct phenotypic characteristics and undertakes specific roles within the tumor microenvironment (TME). NBs can be further classified into 6 subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA), and the expression of top5 marker genes in different subgroups is also shown (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, C). By analyzing the differences in cell cycle, gene expression and CNV among samples with different NBs subgroups and HPV infection states, it was found that the CNV score of C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells and C15 HSPA1B\u003csup\u003e+\u003c/sup\u003e Naive B cells was higher, the proportion of G2M and S phase in the cell cycle was higher, and the expression of differential genes was higher, indicating that their cell proliferation was more active and genetic instability (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, I). The DEGs of 6 NBs subgroups and their corresponding transcription factors and surface proteins were analyzed by GO-BP enrichment analysis, and the active biological process of each tumor cell was obtained. It was found that C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells was active in the biological process of Cytoplasmic translation, and C8 ISG15\u003csup\u003e+\u003c/sup\u003e Naive B cells was active in the biological process of Negative regulation of b cell activation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE, H). Immune landscape studies showed that HPV\u003csup\u003e+\u003c/sup\u003e HNSCC were enriched with more NBs, HPV\u003csup\u003e\u0026minus;\u003c/sup\u003e HNSCC were infiltrated by fewer NBs, and C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells had the highest percentage of HPV\u003csup\u003e+\u003c/sup\u003e cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF). In the analysis of NBs subgroups and cell cycle, it was found that the proportion of NBs within the G1 phase of cellular growth was the highest, which may be related to the fact that NBs is in the initial stage of B cell differentiation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG). Through GSEA enrichment analysis, it was found that C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells subgroup was negatively correlated with Negative regulation of metabolic process and Negative regulation of programmed cell death pathway, indicating that cells of C0 TCL1A\u003csup\u003e+\u003c/sup\u003e Naive B cells subgroup had higher ability of cell metabolism and proliferation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ, K).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 GCBs subgroups may have a high immune response ability\u003c/h2\u003e \u003cp\u003eIn the germinal center, follicular helper T cells (T\u003csub\u003efh\u003c/sub\u003e) stimulate the B cell receptor (BCR) and induce the differentiation of NBs into a variety of fates, mainly including GCBs, MBs and PCs\u003csup\u003e24\u003c/sup\u003e. Complete transcriptional characteristics at the single-cell level would be particularly helpful in determining transcriptional trajectories and heterogeneity in the process of differentiation from NBs to PCs. C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells (C11 GCBs) and C14 STMN1\u003csup\u003e+\u003c/sup\u003e Germinal center B cells (C14 GCBs) were obtained by further subgroup analysis of GCBs (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). The bubble chart and UMAP plots showed the expression of GCBs top marker genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB-C). Among the two GCBs subgroups, the CNVscore of C11 was higher, and the nCount-RNA, G2M.Scorce and S.Score of C14 were higher (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD-F). The GCBs in patients with HNSCC mainly comes from tumor tissues, and some from PBL (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG). The percentage of C11 GCBs in HPV\u003csup\u003e+\u003c/sup\u003e samples was higher than in C14 GCBs (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eH). In C11 GCBs, there were more cells with G1 cell cycle, and most of C14 GCBs were in G2M and S phase (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eI). The proportion of cells in G1 phase was more in HPV\u003csup\u003e+\u003c/sup\u003e group, and the proportion of GCBs cells in G2M and S phase was more in HPV\u003csup\u003e\u0026minus;\u003c/sup\u003e group (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eJ). The DEGs of C11 GCBs and C14 GCBs were screened (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eL), and the enrichment results of GO-BP among C11 GCBs and C14 GCBs was shown by heatmap. It was found that GOBP terms of C11 GCBs were up-regulated in biological processes such as cell activation, leukocyte activation, immune response, lymphocyte activation, regulation of immune system process, regulation of immune response, indicating that the DEGs of C11 GCBs may play a promoting role in the regulation of immune system (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eK). Similarly, in the GO-BP enrichment analysis results of C11 GCBs, C11 GCBs was more active in cytoplasmic translation, B cell activation, B cell proliferation, production of molecular mediator of immune response and immunoglobulin production pathways. The high immune response ability of C11 GCBs subgroup was further verified, which was consistent with the previous research results (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eM-N).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.5 MBs play a key role in the proliferation and differentiation of B cells\u003c/h2\u003e \u003cp\u003eMBs, the subgroups with the largest number of B cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA), could be further divided into six subgroups and the expression of top5 marker genes in different subgroups are also shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB-C. UMAP plots showed that the samples in the C4 IFI30\u003csup\u003e+\u003c/sup\u003e Memory B cells (C4 MBs) had higher CNV score, higher nCount_RNA and more active cell cycle state. Among them, the chromosome stability, transcription level and cell cycle state of different subgroups of MBs are also different, which is of great significance for the study of B cell heterogeneity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD, H). The bar charts revealed that the majority of MBs subgroups were associated with HPV\u0026thinsp;+\u0026thinsp;infection, whereas the cell cycle mostly occurred in the G1 and S phases (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE-F). The AUCell analysis of the GOBP-positive regulation of B cell proliferation in several subgroups of MBs revealed that the C4 MBs had a higher AUC value for this GOBP term compared to other subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG, left). The CNV score of C13 TLE5\u003csup\u003e+\u003c/sup\u003e Memory B cells (C13 MBs) exhibited a lower value compared to the other subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG, right), while the nCount_RNA of C13 MBs was higher than other subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH, left). The G2M.Score of C4 MBs was lower than other subgroups of MBs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH, right). In GO-BP enrichment analysis, C4 MBs were active in cytoplasmic translation, B cell receptor signaling pathway, immune response-regulating cell surface receptor signaling pathway, B cell differentiation, Fc receptor signaling pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI). Similarly, in GSEA enrichment analysis, the expression of C4 MBs was up-regulated in positive regulation of cellular biosynthetic process, which was positively correlated with positive regulation of cell communication, positive regulation of cell death, positive regulation of cellular metabolic process (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ). These results suggest that the DEGs of C4 MBs plays a key role in the differentiation of B cells and promotes the regulation of immune system process, which may reflect the high immune response ability of tumor cells.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.6 B cells eventually differentiate into mature PCs\u003c/h2\u003e \u003cp\u003eGCBs that have a strong attraction to tumor-associated antigens ultimately develop into fully developed PCs that generate immunoglobulins (IgG, IgA, IgM, and IgE). These immunoglobulins have distinct purposes within the given situation\u003csup\u003e25\u003c/sup\u003e. According to the difference in the expression of tumor-associated antibodies in PBs, we divided the subgroups into C6 FKBP11\u003csup\u003e+\u003c/sup\u003e IgG Plasma cells (C6 IgG), C12 DERL3\u003csup\u003e+\u003c/sup\u003e IgA Plasma cells (C12 IgA) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). The UMAP visualization showed the difference between C6 IgG and C12 IgA in terms of CNV score, nCount_RNA, S.Score, and G2M (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). The bubble chart and UMAP plots displayed the top marker genes for C6 IgG and C12 IgA (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC-D). The bar charts depicted the distribution of PCs among various categories (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE-G). To investigate the molecular diversity among PCs subgroups and its functionality. The GO-BP enrichment analysis of the DEGs in PCs revealed that C6 IgG was involved in the biosynthesis of amyloid precursor protein and glycoprotein. Additionally, C12 IgA was found to be active in the biological processes of responding to endoplasmic reticulum stress, the ERAD pathway, and topologically incorrect protein (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH-I).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Exploring the intercellular communication between B cells, plasma cells, and tumor cells\u003c/h2\u003e \u003cp\u003eNumerous investigations have demonstrated the existence of a diverse spectrum of mutual influence between B cells, plasma cells, and tumor cells. This interplay exerts a profound influence on the progression of cancer\u003csup\u003e25\u003c/sup\u003e. To elucidate the intricate mechanism of cell-cell interaction between these cells, we established a separate intercellular communication network between B cells, plasma cells, and tumor cells in HNSCC (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). 17 pairs of interactions and 20 significant signaling pathways were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB). We generally describe the inferred strengths of incoming and outcoming interaction, in which C11 GCBs and C14 GCBs are the main cell type expressing ligands and receptors that actively participate in cell-cell interaction. These results suggest that GCBs is associated with tumors and plays a central role in TME (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). As the observed C11 GCBs and C14 GCBs played a central role in the intercellular communications at HNSCC, we further explore the interactions between incoming and outgoing communication patterns of target cells and secreting cells. Notably, in outgoing communication patterns of secreting cells of GCBs, CD99 and SEMA4 are shown as the most predominant signaling in C11 GCBs compared to other signaling pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD-F). Consistent with these observations, GCBs and MBs are the core of the interaction between B cells and tumor cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eG-H). In summary, the intercellular interaction based on CD99 and SEMA4 signaling pathways is the molecular basis of GCBs-centered TME observed in HNSCC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.8 GCBs based regulatory network detect tumor-specific signaling pathways in HNSCC\u003c/h2\u003e \u003cp\u003eTo unveil the intricate signaling pathways that contribute to the multifaceted intercellular communications, we computed the CD99 and SEMA4 signaling pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA, F). Tumor-specific C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells could express CD99 and SEMA4 to interact with some receptors of tumor cells to promote tumor progression (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB, G). Regarding the CD99 signaling pathways, we observed that C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells and C14 STMN1\u003csup\u003e+\u003c/sup\u003e Germinal center B cells highly expressed CD99 to interact with the tumor cell receptors, C4 and C13 of memory B cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC-E). In SEMA4 signaling pathway networks, the specific interaction networks of C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells and C14 STMN1\u003csup\u003e+\u003c/sup\u003e Germinal center B cells with tumor cells were also observed (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eH-J). These phenomena show that CD99 and SEMA4 could potentially serve as pivotal pathways mediating the interplay between GCBs and tumor cells. Furthermore, these signaling pathways possess the capacity to regulate B cell differentiation, consequently exercising control over tumor growth. Overall, our findings underscore that the C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells manifest intricate intercellular interplays that exacerbate the virulence of the tumor. Targeting tumor-specific cell interaction networks could be a promising strategy for treating HNSCC.\u003c/p\u003e\u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.9 Screening and validation of prognostic genes in GCBs and MBs\u003c/h2\u003e \u003cp\u003eConsidering the presence of tumor-specific signaling pathways in MBs and GCBs mentioned earlier, we will conduct an in-depth examination of these subgroups. A total of 499 HNSCC patients from the TCGA database were finally included in this study. Univariate Cox regression analysis was performed to identify the prognosis-related genes in MBs and GCBs (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eF, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eA, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eF). The candidate genes were recruited into the Least Absolute Shrinkage and Selection Operator (LASSO) Cox regression model to screen out the robust prognosticators (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eG, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eB, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eG). Ultimately, using multivariate Cox regression analysis, we were able to construct four prognostic risk scores consisting of prognostic genes specific to the MBs and GCBs subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eC, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eD, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eH, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eI, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eC, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eD, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eH, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eI). The individual genes that constituted risk scores, which showed an important association with overall survival in MBs and GCBs, were thoroughly shown (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eE, \u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eJ, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eE, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eJ, \u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eK). The expression of CD38, SMIM14, and TCL1A in C11 MEF2B\u003csup\u003e+\u003c/sup\u003e Germinal center B cells varied considerably over distinct N stages (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eL, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e3.10 MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score based on the prognosis-related genes in C11 GCBs\u003c/h2\u003e \u003cp\u003eWe then investigated the possible mechanisms of prognostic effects of 18 prognosis-related genes that comprised the MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score. Bubble maps demonstrated the expression of 18 prognosis-related genes in each GCBs subgroup and the associated features of different HPV infection statuses (HPV\u003csup\u003e+\u003c/sup\u003e vs HPV\u003csup\u003e\u0026minus;\u003c/sup\u003e), Phase (G1, S, G2M), G2M.Scorce, S.Score, nCount_RNA, CNVscore (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eA ). The expression of 18 prognostic genes in C11 and C14 GCBs subgroups was further visualized in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eB and \u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eC. Risk scores were computed for individuals diagnosed with HNSCC within the TCGA datasets. The collected data were subsequently partitioned into cohorts of low and high-risk groups, demarcated by the median risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eD above). The heatmap specifically showed the expression of genes in high and low risk groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eD below). More HNSCC patients had higher survival rates (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) observed in the low-risk group, while the distribution of stage and TNM phases was not statistically different (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eE). The Kaplan-Meier survival curves showed that HNSCC patients in the high-risk group had a lower probability of survival and shorter survival time (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eF). Moreover, the time-dependent ROC curve demonstrated the proficient prognostic capacity of this model concerning the OS of HNSCC patients. The calculated Area Under the ROC Curve (AUC) values stood at 0.690 (1 year), 0.711 (3 years) and 0.677 (5 years), as evidenced in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eG. Then, we further assessed the correlation between gene expression and risk score in the prognostic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eH). The expression of five genes, ACTB, BASP1, EZR, PRDX6, and RRAS2, was positively correlated with the risk scores (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eI), which was in agreement with the results of the previous prognostic studies (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eJ), indicating that these five genes may be tumor-related risk genes in HNSCC. Meanwhile, the five genes of ACTB, BASP1, EZR, PRDX6, RRAS2 were highly expressed in the high-risk group, and the difference between the high and low risk groups was statistically significant (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eJ, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). They were negatively correlated with OS (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eK). Both EZR and RRAS2 expressions were all positively correlated with the other four genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eL-M, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05)Figure. We further investigated the differences of EZR and RRAS2 in high- and low-risk groups, different ages, different races, and different TNM stages, and found that the gene expression was higher in all high-risk groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eN).\u003c/p\u003e \u003cp\u003eWe observed that the MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score had independent prognostic value and had stronger predictive power than other clinical features (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which increases its potential for clinical applications (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eA). A nomogram plot was established to get a better clinical application ability by integrating the risk score and clinical features (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eB). The time-dependent ROC curve demonstrated the proficient prognostic capacity of this model. The AUC values were 0.738 (1 year), 0.769 (3 years) and 0.634 (5 years) (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eC, D). For the 1-, 3-, and 5-year survival, a satisfactory fit was observed between the survival predicted by the nomogram and the actual survival (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eE).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e3.11 Immune cells play a dual role in the dynamic ecosystem of the HNSCC microenvironment\u003c/h2\u003e \u003cp\u003eThe initiation, progression, and invasion of HNSCC are controlled by the complex network of signaling pathways in the TME as well as cell-cell interactions between the tumor stroma, immune cells, and tumor cells\u003csup\u003e26\u003c/sup\u003e. HNSCC microenvironment contained multiple types of immune cells, most of which are T cells and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eF). The high-risk group had lower immune scores and higher tumor purity than the low-risk group (both \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), which may be related to the fact that immune cells infiltrate into the tumor stroma within the HNSCC and interact with tumor cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eG-H). By analyzing the principal components of immune cells in HNSCC, it was found that there were significant differences in the infiltration of many kinds of immune cells between the two risk subgroups (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). There was more CD8 T cell infiltration in the low-risk group and more macrophage M2 infiltration in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eI-J). In the correlation analysis between risk score and immune cells, mast cells, NK cells, and macrophages were significantly positively correlated with the risk score, whereas various T cells such as T cells regulatory (Tregs), T cells follicular helper, and CD8 T cells were significantly negatively correlated with the risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eK). Combined with the results of the previous survival analysis of the two subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eF), this suggests that tumor-associated immunocompetent cells may be the key immunosuppressive population of TME cells. In addition, we also analyzed the correlation between the expression of 18 prognostic genes and the proportion of immune cells in HNSCC. The results showed that the risk score, the gene expression of ACTB, BASP1, PRDX6 was positively correlated with macrophage M2 (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and the risk score, the gene expression of ACTB, BASP1, RRAS2 was negatively correlated with CD8 T cells (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eL-N). The gene expression of RRAS2 was negatively correlated with naive B cells, plasma cells, stromal score, and immunity score (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and was positively correlated with tumor purity (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eO-P). Furthermore, RRAS2 was positively correlated with dendritic cells activated, macrophages M0, CD4 memory T cells resting, and NK cells resting (all \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). RRAS2 was negatively correlated with CD4 memory T cells activated, Mast cells resting, CD8 T cells, and Tregs (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003eQ).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e3.12 Functional enrichment and GSEA analysis of differentially expressed genes in HNSCC patients\u003c/h2\u003e \u003cp\u003eTo scrutinize the molecular heterogeneity and explore the biological mechanism distinguishing the high-risk from the low-risk subgroups, we discerned DEGs within the TCGA cohort [\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\left|{\\text{log}}_{2}\\left(fold change\\right)\\right|\\)\u003c/span\u003e\u003c/span\u003e\u0026gt;1, fdr\u0026thinsp;\u0026lt;\u0026thinsp;0.05]. The results of Gene Ontology (GO) enrichment analysis revealed that the DEGs primarily participated in biological processes such as immunoglobulin complex, circulating immunoglobulin complex, and external side of plasma membrane. Their products were predominantly engaged in immune components, encompassing immunoglobulin production, production of molecular mediator of immune response, B cell activation, etc. Furthermore, these genes played a pivotal role in biological molecular functions, which covered antigen binding and immunoglobulin receptor binding, etc. (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eA-C). The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways mainly included the B cell receptor signaling pathway, Ras signaling pathway, PI3K\u0026thinsp;\u0026minus;\u0026thinsp;Akt signaling pathway, Cytokine\u0026thinsp;\u0026minus;\u0026thinsp;cytokine receptor interaction, Primary immunodeficiency, Hematopoietic cell lineage, etc. (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eD). Our GSEA results, presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eE, revealed that Ribosomal large subunit biogenesis, Mitochondrial translation, Ncrna 3 End processing, Maturation of ssu Rrna from tricistronic Rrna transcript ssu Rrna 5 8s Rrna lsu Rrna, amino acid activation were activated, while antigen receptor mediated signaling, regulation of B cell activation, B cell mediated immunity, positive regulation of B cell activation, B cell receptor signaling pathway were suppressed in the GCBs subgroup. According to the results of GSEA, many B-cell-associated pathways were suppressed in GCBs, highlighting the presence of a B-cell response in patients with HNSCC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003e3.13 Relationship of the OS in HNSCC patients among the TMB subgroups and drug sensitivity\u003c/h2\u003e \u003cp\u003eIn the gene mutation spectrum of the risk and age groups, 466 (91.37%) of 510 HNSCC patients had the gene mutation (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eF left). Among the 18 prognostic genes, 31 (6.08%) had gene mutations in 510 HNSCC patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eF right). We further investigated the frequency of CNV in a total of 526 samples. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eG, ACTB, RRAS2, SERPINA9, TCL1A, PKM, EZR, LCK, CD38, and BIK had significant CNV amplification frequencies. In addition, the probability of simultaneous mutations in two groups of genes, PKM and SERPINA9, TCL1A and STAG3, was higher (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eH, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). We selected the top 6 genes of somatic mutation rate for single nucleotide polymorphism (SNP) analysis and found that SERPINA9, STAG3, ACTB, CD38, LCK, and PKM were dominated by missense mutation (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eI). In the two risk subgroups, the TMB of the high-risk group was higher than that of the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eJ, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and TMB was positively correlated with the risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eK, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Combined with the survival analysis in Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eL, the survival time of patients in high risk-high TMB group was the shortest, while that in low risk-low TMB group was the longest, and the OS of patients in each group was significantly different (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001). The AUC values were 0.541 (1 year), 0.532 (3 years), and 0.5I2 (5 years) (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eM). This indicates that the TMB model has a discernible predictive ability for the prognosis of HNSCC patients. Furthermore, in drug sensitivity analysis, HNSCC patients in the low-risk subgroup demonstrated enhanced responsiveness to cisplatin and docetaxel in chemotherapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eN). In contrast, the high-risk subgroup was more sensitive to Nilotinib and Rapamycin, both targeted drugs (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eO). The IC50 values of ABT.263, AKT.inhibitor.Ⅷ, CCT007093 were higher in the high-risk subgroup, and Bosutinib, Epothilone. B and Pazopanib were higher in the low-risk subgroup (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e13\u003c/span\u003eP). These discoveries not only furnish invaluable insights into selecting suitable chemotherapy and targeted drugs following the risk score of HNSCC patients afflicted with HNSCC but also facilitate informed clinical therapeutic deliberations.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec32\" class=\"Section2\"\u003e \u003ch2\u003e3.14 Immunofluorescence detection of MEF2B\u003csup\u003e+\u003c/sup\u003e B cells\u003c/h2\u003e \u003cp\u003eTo validate the existence of MEF2B\u0026thinsp;+\u0026thinsp;B cells, immunofluorescence assays were executed on three human HNSCC tissues (Figure \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e-\u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e). In the multiple immunofluorescence assay, MEF2B and CD79a were subjected to staining. A conspicuous co-localization was discerned between MEF2B\u003csup\u003e+\u003c/sup\u003e cells and CD79a\u003csup\u003e+\u003c/sup\u003e cells, thereby affirming the presence of MEF2B\u003csup\u003e+\u003c/sup\u003e B cells. To further elucidate the interplay among MEF2B\u003csup\u003e+\u003c/sup\u003e B cells, CD8\u003csup\u003e+\u003c/sup\u003e T cells, and tumor cells. Strikingly, spatial co-localization between CD79a and MEF2B in CD8\u003csup\u003e+\u003c/sup\u003e and EPCAM\u003csup\u003e+\u003c/sup\u003e regions was identified, decisively establishing a substantial correlation between MEF2B\u003csup\u003e+\u003c/sup\u003e B cells and CD8\u003csup\u003e+\u003c/sup\u003e T cells and tumor cells.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec33\" class=\"Section2\"\u003e \u003ch2\u003e3.15 In vitro experiment\u003c/h2\u003e \u003cp\u003eBased on the cell communication analysis findings using cellchat between MEF2B\u003csup\u003e+\u003c/sup\u003e B cells and tumor cells, we conducted in vitro knockdown experiments targeting the PLXNB2 gene in the head and neck cancer cell line. The results, illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e14\u003c/span\u003eA and B, manifest a conspicuous reduction in the proliferation capacity of both cell lines, discernibly contrasting with the control group. Furthermore, plate cloning experiments unveiled a pronounced suppression in both the quantity and size of cell colonies after the knockdown of the PLXNB2 gene in both cell lines (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e14\u003c/span\u003eC, D). The inhibitory impact of PLXNB2 knockdown on the proliferation of the two head and neck cancer cell lines was further substantiated, indirectly suggesting that the interaction between B cells and tumor cells can modulate the biological functions of head and neck cancer cells.\u003c/p\u003e \u003cp\u003eSubsequent wound healing experiments showcased a statistically significant decrease in the cell migration rate within the PLXNB2 knockdown group (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e14\u003c/span\u003eE, F). Moreover, Transwell experiments demonstrated a marked reduction in the number of cells penetrating the lower chamber following PLXNB2 gene knockout, indicative of a substantial inhibition of the migration and invasion capabilities of head and neck cancer cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e14\u003c/span\u003eG, H)\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eHNSCC is one of the most common malignant tumors in the world, with a high metastatic rate and low survival rate, often accompanied by dysfunctions such as articulation and eating and disruption of facial appearance\u003csup\u003e27\u003c/sup\u003e. It was found that HNSCC exhibited immunosuppressive features\u003csup\u003e28\u003c/sup\u003e. Tumor-infiltrated B cells are an essential component of the TME, and they play a fundamental role in the immunotherapeutic response as the primary effectors of humoral immunity\u003csup\u003e29\u003c/sup\u003e. However, in TME, B cells exert both pro- and anti-tumor effects. On the one hand, B cells exert their anti-tumor effects in three ways: 1. secreting tumor-specific antibodies, promoting T cell functions\u003csup\u003e30\u003c/sup\u003e; 2. secreting Granzyme B, interferon-γ (IFNγ), and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) to kill tumor cells directly\u003csup\u003e31\u0026ndash;33\u003c/sup\u003e; 3. secreting immunoglobulins to mediate the processes associated with complement-dependent cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cell-mediated phagocytosis (ADCP)\u003csup\u003e34\u003c/sup\u003e. On the other hand, B cells can also promote tumor progression through several mechanisms, such as TIM-1\u003csup\u003e+\u003c/sup\u003e Bregs. B cells can produce IL10, which inhibits the immune-killing effect of T cells\u003csup\u003e15\u003c/sup\u003e. B cells can also produce cytokines such as IL-1β to promote the growth and invasion of cancer cells\u003csup\u003e35\u003c/sup\u003e. CD20\u003csup\u003e+\u003c/sup\u003e B cells in TME are mainly associated with a good prognosis in different solid tumors, including melanoma, colorectal, gastric, and non-small cell lung cancer\u003csup\u003e34\u0026ndash;37\u003c/sup\u003e. Distel et al. showed that CD20\u003csup\u003e+\u003c/sup\u003e tumor-infiltrating lymphocytes are prognostic indicators of favorable outcomes in early-stage HNSCC. Conversely, an inverse correlation emerged in advanced stages, implying a plasticity in B cells' functional and compositional attributes throughout the disease\u003csup\u003e40\u003c/sup\u003e. Consequently, B cells may be significant in the prognosis of HNSCC patients and may work as a predictive biomarker for immunotherapy\u003csup\u003e41\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHistorically, investigations into B cells have predominantly relied on Immunohistochemistry (IHC) or flow cytometry techniques, which inherently constrain the precise categorization of B cell populations. Typically, based on their differentiation progression, B cells can be stratified into three primary subgroups: na\u0026iuml;ve B cells, MBCs, and Antibody-Secreting Cells (ASCs), each manifesting distinctive phenotypic profiles and functions in TME\u003csup\u003e42\u003c/sup\u003e. Due to the classification, functionality, and spatial presence of different B cell subgroups in different tissues and the low proportion of B cells in specific tumors, B cells are also commonly classified as CD19\u003csup\u003e+\u003c/sup\u003eCD20\u003csup\u003e+\u003c/sup\u003e B cells and CD138\u003csup\u003e+\u003c/sup\u003eSDC1\u003csup\u003e+\u003c/sup\u003e PCs. The application of scRNA-seq provides higher resolution for identifying B cell subgroups. In our study, we analyzed the intratumoral B cell profiles within HNSCC and investigated the functional information of tumor-responsive B cells. From a longitudinal perspective, we used cluster analysis to divide B cells into NBs, GCBs, MBs, and PCs for the first time. High-throughput scRNA-seq and pseudotime trajectory analysis showed that in the early stage of tumor-infiltrating B cell development in HNSCC, NBs and GCBs with high differentiation ability were primarily present. In contrast, in the later stage of development, a large number of MBs and a small number of PCs appeared. Nevertheless, the composition of B cells changes differently during the development of different tumors and promotes or inhibits tumor development. In advanced non-small cell lung cancer (NSCLC), a decrease in naive-like B cells is associated with a poor prognosis, whereas plasma-like B cells inhibit cancer cell growth in the early stage of NSCLC but promote cell growth in the advanced stage of NSCLC\u003csup\u003e43\u003c/sup\u003e. B cells isolated from early-stage colorectal cancer (CRC) exhibited a pre-B cell-like phenotype characterized by the expression of tumor suppressor molecules. In contrast, B cells derived from advanced CRC displayed a propensity for differentiation into plasma cells and predicted a poor prognosis\u003csup\u003e44\u003c/sup\u003e. In addition, the composition of tumor-infiltrating B cells varies in many ways across the transversal dimensions of the tumor. The objective diversity of B cell subgroups in different tumors and the subjective individualized approach of single-cell analysis contribute to the diversity of B cell subgroup definitions. For example, a higher proportion of IgG plasma cells were found at the tumor sites than in the adjacent normal mucosal tissues in CRC\u003csup\u003e45\u003c/sup\u003e; however, in a study in NSCLC, there were more B cells in the adjacent tissues than in the tumor, which may be related to infiltration of inflammatory cells\u003csup\u003e43\u003c/sup\u003e. B cells show less infiltration in esophageal squamous cell carcinoma (ESCC), breast invasive carcinoma (BRCA), kidney cancer (KIRC), and hepatocellular carcinoma (HCC) than lung cancer (LC), colon adenocarcinoma (COAD), and skin cutaneous melanoma (SKCM)\u003csup\u003e46\u003c/sup\u003e. The above results can reveal the existence of specific B cell subgroups of TME in different tissues, thus revealing the immunogenic or immunosuppressive of TME.\u003c/p\u003e \u003cp\u003eThe germinal center (GC) is located in the follicular zone of the secondary lymphoid organs. This particular structure aggregates B cells during the humoral immune response\u003csup\u003e47\u003c/sup\u003e. Antigen-specific B cells accumulate in this remarkable structure and subsequently expand to produce high-frequency mutations in somatic cells, undergo antibody-type switching, and eventually turn into high-affinity B cells against the antigen, a specific means of targeting foreign invading antigens\u003csup\u003e48\u0026ndash;49\u003c/sup\u003e. B cells undergo somatic high-frequency mutations along with high-affinity clones directed against antigens, and the clones produce immunoglobulin molecules with higher affinity for antigens.\u003csup\u003e48\u0026ndash;49\u003c/sup\u003e. Increased immunoglobulin affinity also requires antibody-type conversion. This process that requires the involvement of an enzyme called activation-induced cytidine deaminase (AID), which is highly expressed in B cells in the germinal center and is an essential enzyme in regulating the immune response in the germinal center, as well as being involved in the process of high-frequency mutation in somatic cells described above\u003csup\u003e52\u0026ndash;53\u003c/sup\u003e. AID-mediated high-frequency mutations and type conversions make B-cell mutation rates much more likely, implying the potential for generating uncontrolled cancer cell clones and that germinal centers are critical sites for lymphocytic cancer production\u003csup\u003e54\u003c/sup\u003e. The immune response to GC depends on T\u003csub\u003efh\u003c/sub\u003e cell, found in the follicular zone of secondary lymphoid organs, and co-localizes with GCBs to assist in B cell differentiation and clonal selection\u003csup\u003e55\u0026ndash;56\u003c/sup\u003e. T\u003csub\u003efh\u003c/sub\u003e cells secrete cytokines and membrane surface molecules in direct contact with the GCBs, transmitting signals to the GCBs to help them differentiate into MBs or high-affinity PCs\u003csup\u003e57\u0026ndash;58\u003c/sup\u003e. The recruitment of suitable T\u003csub\u003efh\u003c/sub\u003e cells into the GC during the humoral immune response is vital for the maturation of antibody affinity and the generation of an effective humoral immune response.\u003c/p\u003e \u003cp\u003eThis paper finds that molecules on GCBs enable B-cell action on tumor cells by expressing the high-affinity ligands CD99 and SEMA4, which are involved in synaptic junctions to transmit signals for intercellular communication. SEMA4 is a member of the Semaphorin family of molecules, named for the fact that family members all have a sema structural domain consisting of 7 β-folded sheets in their molecular structures, consists of 8 subfamilies, and initially associated with axon guidance and neural development\u003csup\u003e59\u0026ndash;61\u003c/sup\u003e. Because of its functional specificity, the immune system is most commonly characterized by migratory movements of cells. The most crucial function of Semaphorin signaling is its involvement in the migratory process of immune cells.\u003c/p\u003e \u003cp\u003eDendritic cells (DCs) require SEMA4A for T-cell activation, and Th1 cells also precisely express SEMA4A, promoting the differentiation of activated T cells into Th1 subgroups\u003csup\u003e62\u003c/sup\u003e. SEMA4D is also abundantly expressed on activated T cells and has been reported to stimulate the production of high-affinity antibodies by humoral immunity\u003csup\u003e63\u0026ndash;64\u003c/sup\u003e. Semaphorin involves several physiological processes such as cardiac development\u003csup\u003e65\u003c/sup\u003e, angiogenesis and regeneration\u003csup\u003e66\u0026ndash;67\u003c/sup\u003e, osteoclastogenesis\u003csup\u003e68\u003c/sup\u003e, and immune regulation\u003csup\u003e69\u003c/sup\u003e. It also engages in many pathological human disorders, including tumorigenesis and metastasis\u003csup\u003e70\u003c/sup\u003e, neurogenic disorders\u003csup\u003e71\u003c/sup\u003e, and immune disorders\u003csup\u003e72\u003c/sup\u003e. A growing amount of evidence suggests that Semaphorin plays an essential role in tumor development and holds promise as a potential therapeutic target in various tumors.\u003c/p\u003e \u003cp\u003eCD99 is a 32 kDa transmembrane protein encoded by the pseudoautosomal gene \u003cem\u003eMIC2\u003c/em\u003e\u003csup\u003e73\u0026ndash;74\u003c/sup\u003e. CD99 is well known to participate in several crucial biological processes, including cell adhesion\u003csup\u003e75\u0026ndash;76\u003c/sup\u003e, migration\u003csup\u003e77\u0026ndash;78\u003c/sup\u003e, apoptosis\u003csup\u003e77,79\u003c/sup\u003e, T cell differentiation\u003csup\u003e80\u0026ndash;81\u003c/sup\u003e, inflammation\u003csup\u003e82\u003c/sup\u003e, and immune responses\u003csup\u003e83\u003c/sup\u003e, which suggests that CD99 plays a vital role in the peripheral immune response and cell differentiation. CD99 exerts dual effects on the clinical outcomes of patients afflicted with solid tumors. On the one hand, CD99 may act as a tumor suppressor in certain tumors. In patients with NSCLC, increased levels of CD99 stromal expression are associated with better prognosis\u003csup\u003e84\u003c/sup\u003e. In osteosarcoma, CD99 expression is down-regulated in the majority of patient samples. In contrast, overexpression of CD99 significantly negatively impacts osteosarcoma cell growth and survival, eliminating tumorigenicity and metastatic capacity\u003csup\u003e85\u003c/sup\u003e. On the other hand, CD99 can also be an oncogene for tumor migration and invasion. CD99 is strongly correlated with low survival and poor prognosis in gastric adenocarcinoma\u003csup\u003e86\u003c/sup\u003e. Overexpression of CD99 significantly increased migration and invasiveness of glioma cells and decreased survival\u003csup\u003e87\u0026ndash;88\u003c/sup\u003e. High expression of CD99 is a distinguishing feature of Ewing sarcoma (EWS). As a result, CD99 stands out as the most frequently used diagnostic marker and therapeutic target for EWS\u003csup\u003e89\u003c/sup\u003e. Our study found that tumor cells are the only target cells of CD99 and SEMA4 signaling pathway, and the source of CD99 and SEMA4 ligands is also only GCBs, so it is speculated that GCBs may also be able to promote the development of HNSCC through CD99 and SEMA4 signaling pathway. Hence, CD99 and SEMA4 might serve as robust biomarkers for HNSCC and a promising target for tumor therapy.\u003c/p\u003e \u003cp\u003eBy integrating risk scores and clinical characteristics, we created a risk score model based on 18 GCBs risk genes, and the independent prognostic value of this model was subsequently validated. We used CIBERSORT, ssGSEA to analyze immune cell infiltration in HNSCC samples and to distinguish between different immune cell subgroups in the immune microenvironment, which have different prognostic implications. Compared with patients in the high-risk group, the low-risk group had a higher immune score and a greater abundance of tumor-infiltrating immune cells (TIIC), which resulted in a better prognosis. The low-risk group had more CD8\u003csup\u003e+\u003c/sup\u003e T cell infiltration, and the high-risk group had more macrophage M2 infiltration. CD8\u003csup\u003e+\u003c/sup\u003e T cells are the central effector cells in anti-tumor immunity by directly lysing tumor cells by releasing multiple cytokines and enzymes\u003csup\u003e90\u0026ndash;91\u003c/sup\u003e. In the breast cancer tumor microenvironment interacting with tumor cells, the higher the number of CD8\u003csup\u003e+\u003c/sup\u003e T cells, the better the regulation of anti-tumor activity\u003csup\u003e92\u003c/sup\u003e. CD8\u003csup\u003e+\u003c/sup\u003e T cells are closely linked to the survival of tumor patients, with patients with higher CD8\u003csup\u003e+\u003c/sup\u003e T cell counts having a higher 5-year survival rate\u003csup\u003e93\u003c/sup\u003e and a better prognosis\u003csup\u003e94\u0026ndash;95\u003c/sup\u003e. M2 macrophages can suppress tumor immunity by secreting the immunosuppressive factors IL-10 and TGF-β, and at the same time, secrete EGF to promote neovascularisation within the tumor, which promotes tumor invasion and metastasis\u003csup\u003e96\u0026ndash;97\u003c/sup\u003e. Increased infiltration of M2 macrophages in breast cancers indicates a poor prognosis\u003csup\u003e98\u0026ndash;99\u003c/sup\u003e. These findings are consistent with our results in HNSCC. In summary, the ratio of immune effector cells to suppressor cells, the location of immune cell infiltration, and the degree of immune cell activation in TME also affect the prognosis of patients and their clinical immunotherapy response.\u003c/p\u003e \u003cp\u003eThere are some limitations to this study. Firstly, the data in this study came from part of the open single-cell cohort, which may not be representative because of the small sample size of B-cell studies in HNSCC. Secondly, there is a lack of studies on the association of B cells with treatment efficacy in HNSCC, which needs to be analyzed and studied in the future by further expanding the sample size. Thirdly, some studies related to assessing B cells on prognosis and efficacy in HNSCC are still controversial. These differences may be because B cell subgroups are highly functionally plastic and heterogeneous, so the composition and localization of tumor-infiltrating B cells vary. Hence, future validation in clinical trials is necessary.\u003c/p\u003e \u003cp\u003eIn conclusion, we report that GCBs are B-cell-specific prognostic biomarkers for survival in HNSCC. The emergence and development of HNSCC are dependent on somatic cell gene mutations and are closely connected with the tumor microenvironment. All anti-tumor or pro-tumor factors form a network in the tumor microenvironment that determines tumor initiation, progression, and outcome. Therefore, we can further apply B cells in diagnosis and therapeutic evaluation and develop targeted therapies against the tumor immune microenvironment. These results underscore a compelling rationale for employing B-cell markers in patient stratification and exploring innovative therapeutics targeting B-cell subgroups.\u003c/p\u003e"},{"header":"Abbreviations","content":" \u003cp\u003eHead and neck squamous carcinoma (HNSCC)\u003c/p\u003e \u003cp\u003eOverall survival (OS)\u003c/p\u003e \u003cp\u003eTumor immune microenvironment (TIME)\u003c/p\u003e \u003cp\u003eBulk RNA-sequencing (bulk RNA-seq)\u003c/p\u003e \u003cp\u003eSingle-cell RNA-sequencing (scRNA-seq)\u003c/p\u003e \u003cp\u003eHead and Neck Cancer (HNSC)\u003c/p\u003e \u003cp\u003eFollicular helper T cells (T\u003csub\u003efh\u003c/sub\u003e cells)\u003c/p\u003e \u003cp\u003eB cell receptor (BCR)\u003c/p\u003e \u003cp\u003ePeripheral blood (PBL)\u003c/p\u003e \u003cp\u003eDifferentially expressed genes (DEGs)\u003c/p\u003e \u003cp\u003eNaive B cells (NBs)\u003c/p\u003e \u003cp\u003eGerminal center B cells (GCBs)\u003c/p\u003e \u003cp\u003eMemory B cells (MBs)\u003c/p\u003e \u003cp\u003ePlasma cells (PCs)\u003c/p\u003e \u003cp\u003eGene Expression Omnibus (GEO)\u003c/p\u003e \u003cp\u003eTumor mutational burden (TMB)\u003c/p\u003e \u003cp\u003eComputed Tomography (CT)\u003c/p\u003e \u003cp\u003eMagnetic Resonance Imaging (MRI)\u003c/p\u003e \u003cp\u003eTumor immune microenvironment (TIME)\u003c/p\u003e \u003cp\u003eRegulatory B cell subgroups (Bregs)\u003c/p\u003e \u003cp\u003eGCBs\u003c/p\u003e \u003cp\u003eLeast Absolute Shrinkage and Selection Operator (LASSO)\u003c/p\u003e \u003cp\u003eArea Under the ROC Curve (AUC)\u003c/p\u003e \u003cp\u003eTumor microenvironment (TME)\u003c/p\u003e \u003cp\u003eT cells regulatory (Tregs)\u003c/p\u003e \u003cp\u003eGene Ontology (GO)\u003c/p\u003e \u003cp\u003eKyoto Encyclopedia of Genes and Genomes (KEGG)\u003c/p\u003e \u003cp\u003eInterferon-γ (IFNγ)\u003c/p\u003e \u003cp\u003eTumor necrosis factor-related apoptosis-inducing ligand (TRAIL)\u003c/p\u003e \u003cp\u003eComplement-dependent cytotoxicity (CDC)\u003c/p\u003e \u003cp\u003eAntibody-dependent cell-mediated cytotoxicity (ADCC)\u003c/p\u003e \u003cp\u003eAntibody-dependent cell-mediated phagocytosis (ADCP)\u003c/p\u003e \u003cp\u003eImmunohistochemistry (IHC)\u003c/p\u003e \u003cp\u003eAntibody-Secreting Cells (ASCs)\u003c/p\u003e \u003cp\u003eNon-small cell lung cancer (NSCLC)\u003c/p\u003e \u003cp\u003eColorectal cancer (CRC)\u003c/p\u003e \u003cp\u003eEsophageal squamous cell carcinoma (ESCC)\u003c/p\u003e \u003cp\u003eBreast invasive carcinoma (BRCA)\u003c/p\u003e \u003cp\u003eKidney cancer (KIRC\u003c/p\u003e \u003cp\u003ehepatocellular carcinoma (HCC)\u003c/p\u003e \u003cp\u003eLung cancer (LC)\u003c/p\u003e \u003cp\u003eColon adenocarcinoma (COAD)\u003c/p\u003e \u003cp\u003eSkin cutaneous melanoma (SKCM)\u003c/p\u003e \u003cp\u003eGerminal center (GC)\u003c/p\u003e \u003cp\u003eActivation-induced cytidine deaminase (AID)\u003c/p\u003e \u003cp\u003eDendritic cells (DCs)\u003c/p\u003e \u003cp\u003eEwing sarcoma (EWS)\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eEthics approval and consent to participate\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical approval for this study was obtained from the Ethics Committee of the First Affiliated Hospital of Anhui Medical University. As the other data used in this study are publicly available, no ethical approval was required.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eConsent for publication\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors have read the manuscript and agree to publish.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAvailability of data and material\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe raw data for this study were generated at the corresponding archives, further inquiries can be directed to the corresponding authors.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAuthors\u0026rsquo; contributions\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLi Lin, Zhijie Zhao, and Can Xiao conceived and designed the research. Zhijie Zhao, Jiani Zou, Wenyi Huang, and Yichi Zhang downloaded and collected the data. Zhijie Zhao, Jiani Zou, and Li Lin analyzed the data and wrote the article. Shengbin Pei and Yantao Ding performed the experimental part of the paper. Shengbin Pei and Yantao Ding conducted quality control on the articles. Zhijie Zhao and Jiani Zou guided the submission. All authors read and approved the final manuscript. Li Lin, Jiani Zou, and Shengbin Pei are co-first authors. Zhijie Zhao, Yantao Ding, and Can Xiao are co-correspondent author.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eCompeting interests\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eFunding\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was funded by the National Natural Science Foundation of China (82301158), Shanghai Post-doctoral Excellence Incentive Program\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e(2023M742332),\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eFoundation of National Center for Translational Medicine (Shanghai) SHU Branch (SUITM-202304), Clinical Research Plan of SHDC (SHDC2020CR3070B), Shanghai Jiao Tong University School of Medicine Two-hundred Talent (20161420), Clinical Research Program of 9th People\u0026apos;s Hospital affiliated to Shanghai Jiao Tong university School of Medicine (JYLJ202108), Interdisciplinary Program of Shanghai Jiao Tong University (YG2022QN048), Shanghai Clinical Research Center of Plastic and Reconstructive Surgery supported by Science and Technology Commission of Shanghai Municipality (22MC1940300) and the First Affiliated Hospital of Soochow University \u0026nbsp;Department of Stomatology Clinical Trial Institution Capacity Improvement(LGY2016035).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest: None.\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eC.R. Leemans, P. Snijders, R.H. Brakenhoff, The molecular landscape of head and neck cancer. Nat. Rev. Cancer. \u003cb\u003e18\u003c/b\u003e(5), 269\u0026ndash;282 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Wyss et al., Cigarette, cigar, and pipe smoking and the risk of head and neck cancers: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. Am. J. Epidemiol. \u003cb\u003e178\u003c/b\u003e(5), 679\u0026ndash;690 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ.W. Carlisle et al., An update on the immune landscape in lung and head and neck cancers. CA Cancer J Clin. \u003cb\u003e70\u003c/b\u003e(6), 505\u0026ndash;517 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD.C. Travassos et al., Squamous cell carcinoma antigen as a prognostic marker and its correlation with clinicopathological features in head and neck squamous cell carcinoma: Systematic review and meta-analysis. J. Oral Pathol. Med. \u003cb\u003e47\u003c/b\u003e(1), 3\u0026ndash;10 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.S. Sawant, S.M. Zingde, M.M. Vaidya, Cytokeratin fragments in the serum: their utility for the management of oral cancer. Oral Oncol. \u003cb\u003e44\u003c/b\u003e(8), 722\u0026ndash;732 (2008)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Kulasinghe, L. Kenny, C. Punyadeera, Circulating tumour cell PD-L1 test for head and neck cancers. Oral Oncol. \u003cb\u003e75\u003c/b\u003e, 6\u0026ndash;7 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR.M. Nagler et al., Early diagnosis and treatment monitoring roles of tumor markers Cyfra 21\u0026thinsp;\u0026ndash;\u0026thinsp;1 and TPS in oral squamous cell carcinoma. Cancer. \u003cb\u003e85\u003c/b\u003e(5), 1018\u0026ndash;1025 (1999)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM.B. Amin et al., The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more personalized approach to cancer staging. CA Cancer J Clin. \u003cb\u003e67\u003c/b\u003e(2), 93\u0026ndash;99 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDE D. Paz et al., Prognostic Roles of SCC Antigen, CRP and CYFRA 21\u0026thinsp;\u0026ndash;\u0026thinsp;1 in Oral Cavity Squamous Cell Carcinoma. Anticancer Res. \u003cb\u003e39\u003c/b\u003e(4), 2025\u0026ndash;2033 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eT. Fu et al., Spatial architecture of the immune microenvironment orchestrates tumor immunity and therapeutic response. J. Hematol. Oncol. \u003cb\u003e14\u003c/b\u003e(1), 98 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eT. Tang et al., Advantages of targeting the tumor immune microenvironment over blocking immune checkpoint in cancer immunotherapy. Signal. Transduct. Target. Ther. \u003cb\u003e6\u003c/b\u003e(1), 72 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC.M. Laumont et al., Tumour-infiltrating B cells: immunological mechanisms, clinical impact and therapeutic opportunities. Nat. Rev. Cancer. \u003cb\u003e22\u003c/b\u003e(7), 414\u0026ndash;430 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.S. Kim et al., Role of B Cells in Responses to Checkpoint Blockade Immunotherapy and Overall Survival of Cancer Patients. Clin. Cancer Res. \u003cb\u003e27\u003c/b\u003e(22), 6075\u0026ndash;6082 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO. Demaria et al., Harnessing innate immunity in cancer therapy. Nature. \u003cb\u003e574\u003c/b\u003e(7776), 45\u0026ndash;56 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.S. Wang et al., Tumor-infiltrating B cells: their role and application in anti-tumor immunity in lung cancer. Cell. Mol. Immunol. \u003cb\u003e16\u003c/b\u003e(1), 6\u0026ndash;18 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ.R. Conejo-Garcia et al., Neglected no more: B cell-mediated anti-tumor immunity. Semin Immunol. \u003cb\u003e65\u003c/b\u003e, 101707 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Tang et al., B Cells and Tertiary Lymphoid Structures Influence Survival in Lung Cancer Patients with Resectable Tumors. Cancers (Basel), 2020. 12(9).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. Pretscher et al., Distribution of immune cells in head and neck cancer: CD8\u003csup\u003e+\u003c/sup\u003e T-cells and CD20\u003csup\u003e+\u003c/sup\u003e B-cells in metastatic lymph nodes are associated with favourable outcome in patients with oro- and hypopharyngeal carcinoma. BMC Cancer. \u003cb\u003e9\u003c/b\u003e, 292 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Lechner et al., Tumor-associated B cells and humoral immune response in head and neck squamous cell carcinoma. Oncoimmunology. \u003cb\u003e8\u003c/b\u003e(3), 1535293 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO. Wood et al., Gene expression analysis of TIL rich HPV-driven head and neck tumors reveals a distinct B-cell signature when compared to HPV independent tumors. Oncotarget. \u003cb\u003e7\u003c/b\u003e(35), 56781\u0026ndash;56797 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP.H. Li et al., Recent developments in application of single-cell RNA sequencing in the tumour immune microenvironment and cancer therapy. Mil Med. Res. \u003cb\u003e9\u003c/b\u003e(1), 52 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS. Sun et al., Accuracy, robustness and scalability of dimensionality reduction methods for single-cell RNA-seq analysis. Genome Biol. \u003cb\u003e20\u003c/b\u003e(1), 269 (2019)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eX. Qiu et al., Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods. \u003cb\u003e14\u003c/b\u003e(10), 979\u0026ndash;982 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Qi et al., Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat. Commun. \u003cb\u003e13\u003c/b\u003e(1), 1742 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP. Charoentong et al., Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell. Rep. \u003cb\u003e18\u003c/b\u003e(1), 248\u0026ndash;262 (2017)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA.M. Newman et al., Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods. \u003cb\u003e12\u003c/b\u003e(5), 453\u0026ndash;457 (2015)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK. Yoshihara et al., Inferring tumour purity and stromal and immune cell admixture from expression data. Nat. Commun. \u003cb\u003e4\u003c/b\u003e, 2612 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Ashburner et al., Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. \u003cb\u003e25\u003c/b\u003e(1), 25\u0026ndash;29 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Kanehisa et al., KEGG: integrating viruses and cellular organisms. Nucleic Acids Res. \u003cb\u003e49\u003c/b\u003e(D1), D545\u0026ndash;D551 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Subramanian et al., Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U S A \u003cb\u003e102\u003c/b\u003e(43), 15545\u0026ndash;15550 (2005)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. Maeser, R.F. Gruener, R.S. Huang, oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief. Bioinform, 2021. \u003cb\u003e22\u003c/b\u003e(6)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP. Geeleher, N. Cox, R.S. Huang, pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PLoS One. \u003cb\u003e9\u003c/b\u003e(9), e107468 (2014)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Cerutti, M. Cols, I. Puga, Marginal zone B cells: virtues of innate-like antibody-producing lymphocytes. Nat. Rev. Immunol. \u003cb\u003e13\u003c/b\u003e(2), 118\u0026ndash;132 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS. Gu et al., Significance of intratumoral infiltration of B cells in cancer immunotherapy: From a single cell perspective. Biochim. Biophys. Acta Rev. Cancer. \u003cb\u003e1876\u003c/b\u003e(2), 188632 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK. Starska-Kowarska, The Role of Different Immunocompetent Cell Populations in the Pathogenesis of Head and Neck Cancer-Regulatory Mechanisms of Pro- and Anti-Cancer Activity and Their Impact on Immunotherapy. Cancers (Basel), 2023. \u003cb\u003e15\u003c/b\u003e(6)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF. Bray et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. \u003cb\u003e68\u003c/b\u003e(6), 394\u0026ndash;424 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR. Mandal et al., The head and neck cancer immune landscape and its immunotherapeutic implications. JCI Insight. \u003cb\u003e1\u003c/b\u003e(17), e89829 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZ.N. Willsmore et al., Cells in Patients With Melanoma: Implications for Treatment With Checkpoint Inhibitor Antibodies. Front. Immunol. \u003cb\u003e11\u003c/b\u003e, 622442 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. Ghosh et al., New insights into B cells as antigen presenting cells. Curr. Opin. Immunol. \u003cb\u003e70\u003c/b\u003e, 129\u0026ndash;137 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA.M. Gocher, C.J. Workman, D. Vignali, Interferon-gamma: teammate or opponent in the tumour microenvironment? Nat. Rev. Immunol. \u003cb\u003e22\u003c/b\u003e(3), 158\u0026ndash;172 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW.L. Xu et al., Granzyme B-Producing B Cells Function as a Feedback Loop for T Helper Cells in Liver Transplant Recipients with Acute Rejection. Inflammation. \u003cb\u003e44\u003c/b\u003e(6), 2270\u0026ndash;2278 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG.V. Sharonov et al., cells, plasma cells and antibody repertoires in the tumour microenvironment. Nat. Rev. Immunol. \u003cb\u003e20\u003c/b\u003e(5), 294\u0026ndash;307 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS. Herter et al., Preclinical activity of the type II CD20 antibody GA101 (obinutuzumab) compared with rituximab and ofatumumab in vitro and in xenograft models. Mol. Cancer Ther. \u003cb\u003e12\u003c/b\u003e(10), 2031\u0026ndash;2042 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG.J. Yuen, E. Demissie, S. Pillai, B lymphocytes and cancer: a love-hate relationship. Trends Cancer. \u003cb\u003e2\u003c/b\u003e(12), 747\u0026ndash;757 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Hennequin et al., Tumor infiltration by Tbet\u003csup\u003e+\u003c/sup\u003e effector T cells and CD20\u003csup\u003e+\u003c/sup\u003e B cells is associated with survival in gastric cancer patients. Oncoimmunology. \u003cb\u003e5\u003c/b\u003e(2), e1054598 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Berntsson et al., Prognostic impact of tumour-infiltrating B cells and plasma cells in colorectal cancer. Int. J. Cancer. \u003cb\u003e139\u003c/b\u003e(5), 1129\u0026ndash;1139 (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Ladanyi et al., Prognostic impact of B-cell density in cutaneous melanoma. Cancer Immunol. Immunother. \u003cb\u003e60\u003c/b\u003e(12), 1729\u0026ndash;1738 (2011)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK.I. Al-Shibli et al., Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer. Clin. Cancer Res. \u003cb\u003e14\u003c/b\u003e(16), 5220\u0026ndash;5227 (2008)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL.V. Distel et al., Tumour infiltrating lymphocytes in squamous cell carcinoma of the oro- and hypopharynx: prognostic impact may depend on type of treatment and stage of disease. Oral Oncol. \u003cb\u003e45\u003c/b\u003e(10), e167\u0026ndash;e174 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Wouters, B.H. Nelson, Prognostic Significance of Tumor-Infiltrating B Cells and Plasma Cells in Human Cancer. Clin. Cancer Res. \u003cb\u003e24\u003c/b\u003e(24), 6125\u0026ndash;6135 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS. Gu et al., Significance of intratumoral infiltration of B cells in cancer immunotherapy: From a single cell perspective. Biochim. Biophys. Acta Rev. Cancer. \u003cb\u003e1876\u003c/b\u003e(2), 188632 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Chen et al., Single-cell transcriptome and antigen-immunoglobin analysis reveals the diversity of B cells in non-small cell lung cancer. Genome Biol. \u003cb\u003e21\u003c/b\u003e(1), 152 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW. Wang et al., Multiregion single-cell sequencing reveals the transcriptional landscape of the immune microenvironment of colorectal cancer. Clin. Transl Med. \u003cb\u003e11\u003c/b\u003e(1), e253 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Xia et al., Single-cell landscape and clinical outcomes of infiltrating B cells in colorectal cancer. Immunology. \u003cb\u003e168\u003c/b\u003e(1), 135\u0026ndash;151 (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY. Zheng et al., Immune suppressive landscape in the human esophageal squamous cell carcinoma microenvironment. Nat. Commun. \u003cb\u003e11\u003c/b\u003e(1), 6268 (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG.D. Victora, M.C. Nussenzweig, Germinal Centers. Annu. Rev. Immunol. \u003cb\u003e40\u003c/b\u003e, 413\u0026ndash;442 (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC. Berek, A. Berger, M. Apel, Maturation of the immune response in germinal centers. Cell. \u003cb\u003e67\u003c/b\u003e(6), 1121\u0026ndash;1129 (1991)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ. Jacob et al., Intraclonal generation of antibody mutants in germinal centres. Nature. \u003cb\u003e354\u003c/b\u003e(6352), 389\u0026ndash;392 (1991)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM.G. Weigert et al., Variability in the lambda light chain sequences of mouse antibody. Nature. \u003cb\u003e228\u003c/b\u003e(5276), 1045\u0026ndash;1047 (1970)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. McKean et al., Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin. Proc. Natl. Acad. Sci. U S A \u003cb\u003e81\u003c/b\u003e(10), 3180\u0026ndash;3184 (1984)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Muramatsu et al., Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell. \u003cb\u003e102\u003c/b\u003e(5), 553\u0026ndash;563 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR. Pavri, M.C. Nussenzweig, AID targeting in antibody diversity. Adv. Immunol. \u003cb\u003e110\u003c/b\u003e, 1\u0026ndash;26 (2011)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Nussenzweig, M.C. Nussenzweig, Origin of chromosomal translocations in lymphoid cancer. Cell. \u003cb\u003e141\u003c/b\u003e(1), 27\u0026ndash;38 (2010)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. Breitfeld et al., Follicular B helper T cells express CXC chemokine receptor 5, localize to B cell follicles, and support immunoglobulin production. J. Exp. Med. \u003cb\u003e192\u003c/b\u003e(11), 1545\u0026ndash;1552 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP. Schaerli et al., CXC chemokine receptor 5 expression defines follicular homing T cells with B cell helper function. J. Exp. Med. \u003cb\u003e192\u003c/b\u003e(11), 1553\u0026ndash;1562 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK. Luthje et al., The development and fate of follicular helper T cells defined by an IL-21 reporter mouse. Nat. Immunol. \u003cb\u003e13\u003c/b\u003e(5), 491\u0026ndash;498 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD. Eto et al., IL-21 and IL-6 are critical for different aspects of B cell immunity and redundantly induce optimal follicular helper CD4 T cell (Tfh) differentiation. PLoS One. \u003cb\u003e6\u003c/b\u003e(3), e17739 (2011)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA.L. Kolodkin, D.J. Matthes, C.S. Goodman, The semaphorin genes encode a family of transmembrane and secreted growth cone guidance molecules. Cell. \u003cb\u003e75\u003c/b\u003e(7), 1389\u0026ndash;1399 (1993)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP.K. Hota, M. Buck, Plexin structures are coming: opportunities for multilevel investigations of semaphorin guidance receptors, their cell signaling mechanisms, and functions. Cell. Mol. Life Sci. \u003cb\u003e69\u003c/b\u003e(22), 3765\u0026ndash;3805 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY. Luo, D. Raible, J.A. Raper, Collapsin: a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell. \u003cb\u003e75\u003c/b\u003e(2), 217\u0026ndash;227 (1993)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Kumanogoh et al., Nonredundant roles of Sema4A in the immune system: defective T cell priming and Th1/Th2 regulation in Sema4A-deficient mice. Immunity. \u003cb\u003e22\u003c/b\u003e(3), 305\u0026ndash;316 (2005)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Kumanogoh et al., Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for regulating B cell signaling. Immunity. \u003cb\u003e13\u003c/b\u003e(5), 621\u0026ndash;631 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC. Gu et al., Semaphorin 3E and plexin-D1 control vascular pattern independently of neuropilins. Science. \u003cb\u003e307\u003c/b\u003e(5707), 265\u0026ndash;268 (2005)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Casazza et al., Sema3E-Plexin D1 signaling drives human cancer cell invasiveness and metastatic spreading in mice. J. Clin. Invest. \u003cb\u003e120\u003c/b\u003e(8), 2684\u0026ndash;2698 (2010)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF. Maione et al., Semaphorin 3A overcomes cancer hypoxia and metastatic dissemination induced by antiangiogenic treatment in mice. J. Clin. Invest. \u003cb\u003e122\u003c/b\u003e(5), 1832\u0026ndash;1848 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF. Maione et al., Semaphorin 3A is an endogenous angiogenesis inhibitor that blocks tumor growth and normalizes tumor vasculature in transgenic mouse models. J. Clin. Invest. \u003cb\u003e119\u003c/b\u003e(11), 3356\u0026ndash;3372 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Hayashi et al., Osteoprotection by semaphorin 3A. Nature. \u003cb\u003e485\u003c/b\u003e(7396), 69\u0026ndash;74 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG.M. Delgoffe et al., Stability and function of regulatory T cells is maintained by a neuropilin-1-semaphorin-4a axis. Nature. \u003cb\u003e501\u003c/b\u003e(7466), 252\u0026ndash;256 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL. Tamagnone, Emerging role of semaphorins as major regulatory signals and potential therapeutic targets in cancer. Cancer Cell. \u003cb\u003e22\u003c/b\u003e(2), 145\u0026ndash;152 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS. Nojima et al., A point mutation in Semaphorin 4A associates with defective endosomal sorting and causes retinal degeneration. Nat. Commun. \u003cb\u003e4\u003c/b\u003e, 1406 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eT. Okuno, Y. Nakatsuji, A. Kumanogoh, The role of immune semaphorins in multiple sclerosis. FEBS Lett. \u003cb\u003e585\u003c/b\u003e(23), 3829\u0026ndash;3835 (2011)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP.N. Goodfellow et al., MIC2: a human pseudoautosomal gene. Philos. Trans. R Soc. Lond. B Biol. Sci. \u003cb\u003e322\u003c/b\u003e(1208), 145\u0026ndash;154 (1988)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM. Pasello, M.C. Manara, K. Scotlandi, CD99 at the crossroads of physiology and pathology. J. Cell. Commun. Signal. \u003cb\u003e12\u003c/b\u003e(1), 55\u0026ndash;68 (2018)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eV. Cerisano et al., Molecular mechanisms of CD99-induced caspase-independent cell death and cell-cell adhesion in Ewing's sarcoma cells: actin and zyxin as key intracellular mediators. Oncogene. \u003cb\u003e23\u003c/b\u003e(33), 5664\u0026ndash;5674 (2004)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW. Kasinrerk et al., CD99 monoclonal antibody induce homotypic adhesion of Jurkat cells through protein tyrosine kinase and protein kinase C-dependent pathway. Immunol. Lett. \u003cb\u003e71\u003c/b\u003e(1), 33\u0026ndash;41 (2000)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Bremond et al., Regulation of HLA class I surface expression requires CD99 and p230/golgin-245 interaction. Blood. \u003cb\u003e113\u003c/b\u003e(2), 347\u0026ndash;357 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.S. Yoon et al., Engagement of CD99 triggers the exocytic transport of ganglioside GM1 and the reorganization of actin cytoskeleton. FEBS Lett. \u003cb\u003e540\u003c/b\u003e(1\u0026ndash;3), 217\u0026ndash;222 (2003)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG. Bernard et al., The E2 molecule (CD99) specifically triggers homotypic aggregation of CD4\u003csup\u003e+\u003c/sup\u003e CD8\u003csup\u003e+\u003c/sup\u003e thymocytes. J. Immunol. \u003cb\u003e154\u003c/b\u003e(1), 26\u0026ndash;32 (1995)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG. Bernard et al., The E2 molecule (CD99) specifically triggers homotypic aggregation of CD4\u003csup\u003e+\u003c/sup\u003e CD8\u003csup\u003e+\u003c/sup\u003e thymocytes. J. Immunol. \u003cb\u003e154\u003c/b\u003e(1), 26\u0026ndash;32 (1995)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG. Bernard et al., Apoptosis of immature thymocytes mediated by E2/CD99. J. Immunol. \u003cb\u003e158\u003c/b\u003e(6), 2543\u0026ndash;2550 (1997)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE.M. Dufour et al., CD99 is essential for leukocyte diapedesis in vivo. Cell. Commun. Adhes. \u003cb\u003e15\u003c/b\u003e(4), 351\u0026ndash;363 (2008)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR.L. Watson et al., Endothelial CD99 signals through soluble adenylyl cyclase and PKA to regulate leukocyte transendothelial migration. J. Exp. Med. \u003cb\u003e212\u003c/b\u003e(7), 1021\u0026ndash;1041 (2015)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK. Edlund et al., CD99 is a novel prognostic stromal marker in non-small cell lung cancer. Int. J. Cancer. \u003cb\u003e131\u003c/b\u003e(10), 2264\u0026ndash;2273 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM.C. Manara et al., CD99 acts as an oncosuppressor in osteosarcoma. Mol. Biol. Cell. \u003cb\u003e17\u003c/b\u003e(4), 1910\u0026ndash;1921 (2006)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ.H. Lee et al., Clinical significance of CD99 down-regulation in gastric adenocarcinoma. Clin. Cancer Res. \u003cb\u003e13\u003c/b\u003e(9), 2584\u0026ndash;2591 (2007)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH.J. Seol et al., Overexpression of CD99 Increases the Migration and Invasiveness of Human Malignant Glioma Cells. Genes Cancer. \u003cb\u003e3\u003c/b\u003e(9\u0026ndash;10), 535\u0026ndash;549 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL.C. Cardoso et al., CD99 Expression in Glioblastoma Molecular Subtypes and Role in Migration and Invasion. Int. J. Mol. Sci., 2019. \u003cb\u003e20\u003c/b\u003e(5)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eN. Riggi, M.L. Suva, I. Stamenkovic, Ewing's Sarcoma. N Engl. J. Med. \u003cb\u003e384\u003c/b\u003e(2), 154\u0026ndash;164 (2021)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Ocana et al., Tumor-infiltrating lymphocytes in breast cancer: ready for prime time? J. Clin. Oncol. \u003cb\u003e33\u003c/b\u003e(11), 1298\u0026ndash;1299 (2015)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC. Gu-Trantien et al., CD4(+) follicular helper T cell infiltration predicts breast cancer survival. J. Clin. Invest. \u003cb\u003e123\u003c/b\u003e(7), 2873\u0026ndash;2892 (2013)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW.H. Fridman et al., The immune contexture in human tumours: impact on clinical outcome. Nat. Rev. Cancer. \u003cb\u003e12\u003c/b\u003e(4), 298\u0026ndash;306 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.M. Mahmoud et al., Tumor-infiltrating CD8\u003csup\u003e+\u003c/sup\u003e lymphocytes predict clinical outcome in breast cancer. J. Clin. Oncol. \u003cb\u003e29\u003c/b\u003e(15), 1949\u0026ndash;1955 (2011)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Rody et al., T-cell metagene predicts a favorable prognosis in estrogen receptor-negative and HER2-positive breast cancers. Breast Cancer Res. \u003cb\u003e11\u003c/b\u003e(2), R15 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD.G. DeNardo et al., CD4(+) T cells regulate pulmonary metastasis of mammary carcinomas by enhancing protumor properties of macrophages. Cancer Cell. \u003cb\u003e16\u003c/b\u003e(2), 91\u0026ndash;102 (2009)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eI. Shihab et al., Understanding the Role of Innate Immune Cells and Identifying Genes in Breast Cancer Microenvironment. Cancers (Basel), 2020. 12(8).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. Mantovani, A. Sica, Macrophages, innate immunity and cancer: balance, tolerance, and diversity. Curr. Opin. Immunol. \u003cb\u003e22\u003c/b\u003e(2), 231\u0026ndash;237 (2010)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eS.D. Heys et al., Characterisation of tumour-infiltrating macrophages: impact on response and survival in patients receiving primary chemotherapy for breast cancer. Breast Cancer Res Treat. \u003cb\u003e135\u003c/b\u003e(2), 539\u0026ndash;548 (2012)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eI. Shabo et al., Breast cancer expression of CD163, a macrophage scavenger receptor, is related to early distant recurrence and reduced patient survival. Int. J. Cancer. \u003cb\u003e123\u003c/b\u003e(4), 780\u0026ndash;786 (2008)\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Head and neck squamous carcinoma, germinal center B cells, single-cell RNA-seq, immune cell infiltration, drug sensitivity","lastPublishedDoi":"10.21203/rs.3.rs-3804853/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3804853/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eMore than 60% of patients with head and neck squamous carcinoma (HNSCC) are diagnosed at advanced stages and miss radical treatment. This has prompted the need to find new biomarkers to achieve early diagnosis and predict early recurrence and metastasis of tumors.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eSingle-cell RNA sequencing (scRNA-seq) data from HNSCC tissues and peripheral blood samples were obtained through the Gene Expression Omnibus (GEO) database (GSE164690) to characterize the B-cell subgroups, differentiation trajectories, and intercellular communication networks in HNSCC and to construct a prognostic model of the associated risks. In addition, this study analyzed the differences in clinical features, immune cell infiltration, functional enrichment, tumor mutational burden (TMB), and drug sensitivity between the high- and low-risk groups.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eUsing scRNA-seq of HNSCC, we classified B and plasma cells into a total of four subgroups for the first time: naive B cells (NBs), germinal center B cells (GCBs), memory B cells (MBs), and plasma cells (PCs) and analyzed each subgroup separately. Pseudotemporal trajectory analysis revealed that NBs and GCBs were at the early stage of B cell differentiation, while MBs and PCs were at the end stage. Cellular communication revealed that GCBs acted on tumor cells through the CD99 and SEMA4 signaling pathways. By employing univariate Cox regression, Lasso, and multivariate Cox regression analysis, we developed several risk scores based on marker genes of GCBs subgroups, especially the independent prognostic value of MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score was validated. The results of immune cell infiltration, TMB, and drug sensitivity assays were significantly different in HNSCC samples from high and low- MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score groups.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis study provides ideas to unravel the fundamental biological functions of B cells and their complex mechanistic roles. For the first time, we identified GCBs as B cell-specific prognostic biomarkers. The MEF2B\u003csup\u003e+\u003c/sup\u003e GCB score fills the research gap in the genetic prognostic prediction model of HNSCC and is expected to provide a theoretical basis for finding new therapeutic targets for HNSCC.\u003c/p\u003e","manuscriptTitle":"Germinal center B-cell subgroups in the tumor microenvironment cannot be overlooked: their involvement in prognosis, immunotherapy response, and treatment resistance in head and neck squamous carcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-02 19:55:02","doi":"10.21203/rs.3.rs-3804853/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"30b4f006-ce64-4241-b10b-3771a1aae1ea","owner":[],"postedDate":"January 2nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-01-02T19:55:05+00:00","versionOfRecord":[],"versionCreatedAt":"2024-01-02 19:55:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-3804853","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3804853","identity":"rs-3804853","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}