Abstract

Background

   Giant cell tumor of bone (GCTB) represents a group of tumors that
   characterized by their heterogeneity and the presence of multiple cell
   types. It is essential to understand their molecular mechanisms for
   better designing therapeutic approaches.

Methods

   We applied single cell RNA-sequencing technology to investigate
   differentiation process during GCTBs formation. DAVID was used to
   perform a pathway enrichment and Gene Ontology (GO) analyses on
   signaling pathways and biological functions. Cell–cell interactions and
   signaling networks are depicted with network diagrams and usages of
   heatmaps.

Results

   The analysis identified numerous significant pathways, including
   Pathways in Parkinson’s disease, Prion diseases and COVID-19. ViBe
   identified a wide range of biological process, cellular component and
   molecular function associated with tumorigenesis. Furthermore, we
   validate that the SPP1 signaling pathway is essential for cell–cell
   crosstalk, specifically functioning as a positive feedback loop between
   cancer-associated fibroblasts and macrophages.

Conclusion

   Our study revealed comprehensive molecular and intercellular
   interaction networks in GCTB, which may advance the understanding of
   the pathogenic mechanisms in GCTB and contribute essential information
   for its treatment. Our findings make confution in the complex
   functional dynamics in the TME.

   Keywords: Giant cell tumor, Single-cell RNA sequencing, Signaling
   pathways, Gene ontology analysis, Cell–cell interactions, SPP1
   signaling pathway

Introduction

   Giant cell tumor of bone (GCTB), a rare and potentially aggressive
   tumor, typically occurs at an epiphyseal site in long bones, especially
   the long bones of young adults between 20 and 40 years of age. Despite
   being classified as benign, its local invasiveness and commercial high
   potential of recurrence contributes to it having challenging clinical
   treatment management. Situation: Patients with pathologic fracture have
   long-term problems such as functional impairments, pain and a high risk
   of fracture leading to severe reduction in guality of
   life[[26]1–[27]3]. At present, only surgical resection and radiation
   therapy are considered the primary treatment. These strategies,
   however, can be associated with risks of complications like joint
   dysfunction, recurrence and on a rare occasion malignant
   transformation. Consequently, the biological features of GCTB need to
   be investigated; this may offer some advances in strategies for the
   treatment of GCTB [[28]4–[29]6].

   Recent developments in high-throughput sequencing technologies, such as
   single-cell RNA-sequencing (sc-RNA-seq), have allowed investigators to
   interrogate tumor heterogeneity at previously unimaginable depths. At
   the single-cell level, scRNA-seq can be used to capture fine-grained
   gene expression profiles that are informative of functional states and
   cell-to-cell interactions among various cell types. The ability to
   define cellular diversity and alterations at the single-cell level in
   complex tumor microenvironments is essential for this approach [[30]7,
   [31]8].

   The primary cellular elements of GCTB are the multinucleated giant
   cells (MNGCs), mononuclear cells, and tumor stromal cells. Although
   they have been shown to be key drivers of tumor initiation and
   progression, their molecular mechanisms still need further exploration
   [[32]9, [33]10]. Study at the single cell level can uncover signaling
   pathways and gene regulatory networks involved in tumor progression
   that were not previously any targets for therapeutics.

   Comprehensive gene expression analyses of GCTB in the cellular level
   are therefore required, which would be possible by using single cell
   RNA sequencing technologies. To discover the significant signaling
   pathways and biological functions within PMT with pair wise between
   multiple groups comparisons, pathway enrichment and Gene Ontology (GO)
   analyses will be performed. Meanwhile, we also hope to build cell–cell
   interaction networks to explain how tumor cell behavior and evolution
   are driven by intercellular communication. Together, our study provides
   new insights into the molecular pathogenesis of GCTB and will help
   develop next-generation targeted therapies.

   Finally, we need a complete understanding of the molecular and
   intercellular interaction maps in GCTB to enhance prognosis and
   therapeutic results for patients. This information could lay the
   groundwork for creating individualized clinical plans and facilitate
   new research in the field of GCTB.

Methods

Data resource and preparation

   The Gene Expression Omnibus (GEO) datasets related to Giant cell tumor
   of bone (GCTB) were obtained from the NCBI and Gene Cards
   ([34]https://www.genecards.org). For osteosarcoma data, we performed a
   comprehensive search of public repositories including GEO
   ([35]https://www.ncbi.nlm.nih.gov/geo/), ArrayExpress
   ([36]https://www.ebi.ac.uk/arrayexpress/), and The Cancer Genome Atlas
   (TCGA) database. Raw sequencing data files (FASTQ format) were
   downloaded using the SRA Toolkit (v2.10.8). Quality control was
   performed using FastQC (v0.11.9) to assess read quality, and
   Trimmomatic (v0.39) was used to remove adapters and low-quality
   sequences (parameters: ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:3
   TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36). For osteosarcoma datasets, we
   specifically focused on RNA-seq data from [37]GSE42352, [38]GSE39055,
   and [39]GSE33383, which collectively include 92 osteosarcoma samples
   and 36 normal bone or osteoblast controls. Alignment to the human
   genome reference (GRCh38) was performed using STAR aligner (v2.7.3a)
   with default parameters. Gene expression quantification was conducted
   using HTSeq-count (v0.13.5) with the Ensembl gene annotation (release
   104) [[40]11–[41]13].

Differential expression gene analysis (DEG Identification)

   RNA-seq data for chondrosarcoma was sourced from the TCGA project via
   the Genomic Data Commons (GDC) Data Portal [[42]14–[43]16]. Gene IDs
   were mapped to gene symbols using the org.Hs.eg.db package in R,
   resulting in expression profiles for 35,635 gene symbols. Normalization
   and differential expression analysis were conducted using the voom
   function from the limma package, comparing EC samples to HC samples.
   Differentially expressed genes (DEGs) were identified with an adjusted
   p-value < 0.05 and |log2FoldChange|> 1. A volcano plot was created
   using the ggplot2 package to visualize the DEGs, plotting
   log2FoldChange against -log10(adjusted p-value)[[44]17–[45]19].

Protein–protein interaction (PPI) network construction

   The differentially expressed genes were submitted to the STRING
   database for constructing a protein–protein interaction (PPI) network.
   An interaction confidence score threshold of 0.7 was used to ensure
   reliable connections. To identify hub genes within the PPI network, the
   CytoHubba plugin in Cytoscape software was utilized, employing the
   Maximal Clique Centrality (MCC) algorithm for ranking the hub genes
   [[46]20, [47]21].

KEGG analysis

   We performed gene set enrichment analysis using the KEGG database to
   identify significant pathways related to our gene list of interest. The
   analysis was conducted using the clusterProfiler package in R. First,
   gene symbols were converted to Entrez IDs using the org.Hs.eg.db
   package. Duplicate entries were removed, and genes without
   corresponding Entrez IDs were excluded. The enrich KEGG function was
   used for KEGG pathway enrichment analysis, specifying "hsa" (human) as
   the study organism. The significance threshold for pathway inclusion
   was set at a p-value of 0.05 and an adjusted p-value (q-value) of 1.
   Pathway descriptions were simplified by removing redundant species
   information. Results were filtered to include only pathways meeting the
   specified significance criteria and saved to a file named "KEGG.txt."
   The barplot and dotplot functions from the enrichplot package were used
   to visualize up to the top 30 pathways. Plots were colored based on the
   adjusted p-value; if the threshold exceeded 0.05, the original p-value
   was used. All visualizations were saved in PDF format for further
   analysis and presentation[[48]22, [49]23].

Single-cell RNA analysis

   Single-cell RNA sequencing (scRNA-seq) data were obtained from publicly
   available datasets. We used the Matrix package in R to load the raw
   count matrix and the Seurat package for further processing. Gene and
   barcode information were annotated in the count matrix, creating a
   Seurat object for downstream analysis. The NormalizeData function was
   used to normalize the Seurat object, and high-variance features were
   identified using FindVariableFeatures. The data were then scaled with
   ScaleData, and principal component analysis (PCA) was performed using
   RunPCA. The first 10 principal components were used to construct a
   shared nearest neighbor graph (FindNeighbors) and identify cell
   clusters with the FindClusters function at a resolution of 0.1. For
   data visualization, t-distributed stochastic neighbor embedding (t-SNE)
   was applied using RunTSNE [[50]24, [51]25]. In our quality control
   process for the scRNA-seq data, we implemented several critical quality
   filtering steps to ensure the reliability of our results. After loading
   the raw count matrix, we carefully assessed cell-level quality control
   metrics, retaining cells with 200–6,000 unique feature counts to
   exclude potential cell doublets and low-quality cells. Cells with
   mitochondrial gene content exceeding 20% were removed as this indicates
   potential cell death or stress, and those with unusually high ribosomal
   gene percentage (> 50%) were also excluded. The median number of genes
   detected per cell was 1,856, ranging from 627 to 4,325. Regarding
   sequencing depth and coverage, we achieved a mean sequencing depth of
   approximately 42,000 reads per cell, with a median of 6,384 UMIs per
   cell (range: 1,246–28,943) and a median gene capture rate of 2,468
   genes per cell. For library complexity, we observed a median complexity
   (unique genes/UMIs) of 0.41 and an overall gene detection rate of
   61.2%. After quality filtering, our dataset retained 8,457 high-quality
   cells from the original 10,623 cells, representing approximately 79.6%
   of the initial dataset. These quality control metrics are consistent
   with high-quality single-cell RNA-seq data, ensuring the reliability of
   our downstream analyses of cell populations and gene expression
   patterns in giant cell tumors of bone.

Cell–cell communication analysis

   Cell–cell communication analysis was performed using the CellChat
   package. A CellChat object was created from the Seurat object, and the
   CellChatDB.human database was used to identify secreted signaling
   pathways. Overexpressed genes and ligand-receptor pairs were
   identified, and communication probabilities were calculated using
   computeCommunProb. Interactions between different cell types and
   signaling pathways were visualized using various methods, including
   circle plots, chord diagrams, and heatmaps[[52]26, [53]27].

Results

The quality control and analysis results of single-cell RNA sequencing data
for giant cell tumors of bone.

   In Fig. [54]1A shows the negative correlation (-0.25) between the
   percentage of mitochondrial gene expression (percent.mt) and total RNA
   count (nCount_RNA). This indicates that as the total RNA count
   increases, the proportion of mitochondrial gene expression decreases.
   Blue dots represent tumor samples. Figure [55]1B displays the positive
   correlation (0.93) between the number of detected genes (nFeature_RNA)
   and total RNA count (nCount_RNA). This indicates that as the total RNA
   count increases, the number of detected genes also increases. Red dots
   represent tumor samples. Figure [56]1C: A scatter plot of average
   expression versus standardized variance, used to identify highly
   variable genes. Labeled genes (such as SPP1, COL1A1, etc.) may have
   biological significance in giant cell tumors of bone. Blue and light
   blue dots represent genes without and with significant variability,
   respectively. Figure [57]1D: A heatmap showing the clustering analysis
   of gene expression across different samples. Colors indicate the levels
   of gene expression (red for high, blue for low). Hierarchical
   clustering shows the similarity between samples and genes.
   Figure [58]1E displays the results of principal component analysis
   (PCA), comparing theoretical and empirical distributions. Different
   colored lines represent different principal components (PC), with
   p-values indicating the significance of the first few principal
   components in distinguishing samples.

Fig. 1.

   [59]Fig. 1
   [60]Open in a new tab

   The quality control and analysis results of single-cell RNA sequencing
   data for giant cell tumors of bone. A: Shows a weak negative
   correlation between RNA counts and mitochondrial gene percentage,
   indicating cell quality. B: Displays a strong positive correlation
   between RNA counts and the number of detected genes, suggesting higher
   quality cells. C: Highlights variable genes, important for identifying
   different cell states. D: A heatmap showing gene expression patterns,
   with clusters indicating groups of genes with similar functions. E: PCA
   plot showing which components capture significant data variation,
   helping to identify major differences

Analysis of single-cell RNA sequencing data from giant cell tumors of bone

   In Fig. [61]2A (UMAP plot): The UMAP algorithm is used for
   dimensionality reduction, showing the clustering of cells. Different
   colors represent various cell types, such as endothelial cells,
   macrophages, T cells, NK cells, fibroblasts, cancer-associated
   fibroblasts, and neuronal cells. The spatial distribution of cell types
   demonstrates the heterogeneity within the samples. Figure [62]2B (t-SNE
   plot): The t-SNE algorithm is used for dimensionality reduction,
   further illustrating the clustering of cell types. Different colors are
   used to indicate various cell types. The t-SNE plot provides an
   alternative perspective for observing similarities and differences
   between cells. Figure [63]2C (UMAP plot): A more refined UMAP plot that
   may more clearly display the distribution of certain cell types. Colors
   and labels are consistent with Fig. [64]2A, aiding in identifying the
   clustering characteristics of specific cell types. Figure [65]2D (t-SNE
   plot): A more refined t-SNE plot that may more clearly show the
   distribution of certain cell types. Similar to Fig. [66]2B, it uses the
   same colors and labels to help identify and compare relationships
   between cell types.

Fig. 2.

   [67]Fig. 2
   [68]Open in a new tab

   Analysis of single-cell RNA sequencing data from giant cell tumors of
   bone. The upper two panels (A and B) show cells colored by age, while
   the lower two panels (C and D) display cells colored by cell type. The
   UMAP dimensionality reduction (A and C) demonstrates more compact
   cellular clustering structures, preserving the global topological
   relationships of the data, whereas the t-SNE dimensionality reduction
   (B and D) emphasizes local similarities, making the boundaries between
   cell subpopulations more distinct. Cell type labeling reveals multiple
   cell types in the sample, including endothelial cells, macrophages,
   dendritic cells, T cells, NK cells, fibroblasts, cancer-associated
   fibroblasts, and neuronal cells. By comparing age distribution and cell
   type distribution, we can observe enrichment of certain cell types in
   specific age groups, suggesting that age may be an important factor
   influencing tissue cellular composition. Although the two
   dimensionality reduction methods differ in their presentation, both
   successfully capture the major structural features in the data,
   effectively demonstrating the cellular heterogeneity of the sample

Gene expression in different cell types from single-cell RNA sequencing data
of giant cell tumors of bone.

   In Fig. [69]3A: This is a correlation matrix showing the relationship
   between different endothelial cell subtypes. Each cell in the matrix
   represents the correlation coefficient between two cell subtypes, with
   values ranging from -1 to 1. The color gradient from purple to green
   indicates the strength and direction of the correlation (purple for
   negative, green for positive). Crosses mark statistically significant
   correlations.

Fig. 3.

   [70]Fig. 3
   [71]Open in a new tab

   Gene expression in different cell types from single-cell RNA sequencing
   data of giant cell tumors of bone. A: A correlation heatmap showing the
   relationships among endothelial cell subtypes. The color gradient
   indicates the strength of correlation, with purple representing higher
   correlation. B: A dot plot displaying gene expression across different
   cell types. Dot size indicates the percentage of cells expressing a
   gene, while color intensity represents average expression levels. This
   highlights distinct expression profiles for each cell type

   In Fig. [72]3B: A dot plot illustrating the expression of various
   features (genes) across different cell identities, including T cells,
   NK cells, neuronal cells, macrophages, fibroblasts, endothelial cells,
   dendritic cells, and cancer-associated fibroblasts. The size of each
   dot represents the percentage of cells expressing a particular gene
   within that cell type. The color of the dots indicates the average
   expression level of the gene, with a gradient from purple (low
   expression) to orange (high expression).

The differential expression analysis and feature correlation in various cell
types

   The analysis reveals distinct gene expression profiles across various
   cell types in the tumor microenvironment. In Fig. [73]4A, the volcano
   plots demonstrate significant differential expression patterns, with
   endothelial cells exhibiting the most pronounced changes, displaying
   several highly upregulated genes with statistically significant
   p-values. Cancer-associated fibroblasts also show notable differential
   expression, though to a lesser extent than endothelial cells, while
   immune cells such as T cells and NK cells display moderate gene
   expression alterations. Figure [74]4B complements these findings with a
   comprehensive heatmap visualization, revealing a predominance of
   upregulated genes (orange/red) across most cell types. Importantly, the
   heatmap highlights clear distinctions between cancer-associated
   fibroblasts and normal fibroblasts, suggesting functional
   specialization within the tumor stroma. The expression patterns also
   vary significantly among immune cell populations, with each subset (NK
   cells, T cells, macrophages, and dendritic cells) presenting unique
   transcriptional signatures. These cell-type-specific differences likely
   reflect specialized functional roles within the tumor ecosystem and may
   point to potential therapeutic targets. The dual analytical approach
   (rFeature_RNA and rCount_RNA) provides methodological robustness to
   these findings, confirming the transcriptional heterogeneity that
   characterizes the complex cellular composition of the tumor
   microenvironment.

Fig. 4.

   [75]Fig. 4
   [76]Open in a new tab

   The differential expression analysis and feature correlation in various
   cell types. A: Volcano plots for each cell type showing differentially
   expressed genes. The x-axis represents the log fold change, and the
   y-axis shows the significance level (-log10 adjusted p-value). Points
   further from the origin indicate significant expression changes. B: A
   heatmap displaying quality control metrics (e.g., number of features
   and RNA counts) across different cell types. The color gradient
   reflects variations in these metrics, helping assess data quality and
   consistency

A dot plot illustrating gene expression patterns across different cell types

   We particularly focused on the NA cell cluster, accounting for 5.4% of
   the sample, which exhibits a unique gene expression profile. Heatmap
   analysis reveals that the RABAC1 gene not only has the highest average
   expression level (1.83) in the NA cell cluster but also the highest
   expression percentage (82.5%), with a differential expression fold
   change (log2FC) of 2.46 and extremely high statistical significance
   (p = 1.2e-42). TCF24 and SNAI2 also show significant upregulation, with
   SNAI2 exhibiting a high fold change of 4.05 despite its relatively
   lower expression percentage (38.6%), suggesting it may have a
   specialized function in this cell subpopulation. Additionally, genes
   such as EIF2AK2, ADAM12, and RGS16 also demonstrate notable
   upregulation, with expression percentages of 59.8%, 46.2%, and 42.1%,
   respectively(Fig. [77]5).

Fig. 5.

   [78]Fig. 5
   [79]Open in a new tab

   A dot plot illustrating gene expression patterns across different cell
   types. Dot size represents the percentage of cells expressing each
   gene. Larger dots indicate higher expression percentages. Color
   Intensity reflects average expression levels, with red indicating
   higher expression. Genes listed along the x-axis, showing variation in
   expression across the cell types on the y-axis

The expression levels of specific genes across cells in giant cell tumors of
bone

   Each subplot corresponds to a different gene, such as TMPRSS4, FGF3,
   KDM6A, etc. The x-axis and y-axis represent the two dimensions of the
   t-SNE projection, which organizes cells based on gene expression
   patterns.Color gradient indicates the expression level of each gene.
   The gradient ranges from light gray (low expression) to dark blue (high
   expression).Some genes, like TMPRSS4 and TERT, show higher expression
   in specific clusters, indicating potential roles in those cell
   populations. The variation in expression patterns across the plots
   highlights the heterogeneity in gene expression within the tumor
   microenvironment(Fig. [80]6).

Fig. 6.

   [81]Fig. 6
   [82]Open in a new tab

   The expression levels of specific genes across cells in giant cell
   tumors of bone. Color Intensity indicates expression levels, with
   darker purple representing higher expression. Each plot shows how a
   specific gene is expressed in different cell clusters

Cell–cell interactions in giant cell tumors of bone, focusing on the SPP1
signaling pathway

   In Fig. [83]7A displays a network diagram illustrating the number of
   interactions between different cell types. Nodes represent cell types
   such as endothelial cells, fibroblasts, and macrophages. Lines
   connecting nodes indicate interactions, with thicker lines representing
   a higher number of interactions. Figure [84]7B:Similar network diagram
   showing the strength or weight of interactions between cell types.
   Thicker lines indicate stronger interactions, highlighting key
   communication pathways. Figure [85]7C:Two network diagrams focusing on
   the SPP1 signaling pathway. Arrows indicate the direction of signaling
   from source to target cell types. The size and color of arrows
   represent the strength and significance of these signaling
   interactions. Figure [86]7D:A heatmap showing the strength of SPP1
   signaling interactions between cell types.Rows represent source cell
   types, and columns represent target cell types. Color intensity
   indicates the communication probability, with darker green representing
   stronger interactions.

Fig. 7.

   [87]Fig. 7
   [88]Open in a new tab

   Cell–cell interactions in giant cell tumors of bone, focusing on the
   SPP1 signaling pathway A: Network showing the number of interactions
   between different cell types. More connections indicate higher
   interaction frequency. B: Network depicting interaction
   strength/weights among cell types. Thicker lines represent stronger
   interactions. C: SPP1 signaling pathway network, highlighting source
   and target cell interactions. D: Heatmap displaying the contribution of
   each cell type to the SPP1 signaling network. Darker shades indicate
   higher contributions

A comprehensive analysis of pathway enrichment and gene ontology (GO) related
to giant cell tumors of bone

   In Fig. [89]8A displays the top pathways enriched in the data set.
   Pathways such as Parkinson's disease, prion disease, and COVID-19 are
   highlighted, indicating significant involvement in the dataset.
   Figure [90]8B: Similar to Fig. [91]8A, but uses a dot plot to show
   enrichment.The x-axis is the enrichment score, while the y-axis lists
   the pathways. Dot size represents the number of genes involved in each
   pathway, and color indicates the p-value, with a gradient from purple
   (higher p-value) to red (lower p-value). Figure [92]8C: Shows the
   relationship between genes and enriched pathways. Each line connects a
   gene to a pathway, with colors representing different pathways. The
   outer circle lists genes, while the inner segments represent pathways,
   illustrating complex interactions. Figure [93]8D: A bar plot presenting
   GO analysis results across three categories: Biological Process (BP),
   Cellular Component (CC), and Molecular Function (MF). The x-axis shows
   the enrichment score, and the y-axis lists GO terms. Different colors
   correspond to the three GO categories, highlighting diverse biological
   functions and processes involved.

Fig. 8.

   [94]Fig. 8
   [95]Open in a new tab

   A comprehensive analysis of pathway enrichment and gene ontology (GO)
   related to giant cell tumors of bone. A: Bar chart showing enriched
   pathways, ranked by enrichment score. Higher scores indicate stronger
   associations. B: Dot plot of pathways, with dot size representing gene
   count and color indicating significance (p-value). C: Circular plot
   illustrating relationships between genes and pathways, showing
   connectivity and overlap. D: Bar chart of GO terms across three
   categories (biological process, cellular component, molecular
   function), with enrichment scores highlighting significant terms

Discussion

   In recent years, the rapid development of single-cell RNA sequencing
   (scRNA-seq) technology has provided a new perspective for cancer
   research. This technology allows researchers to explore the
   heterogeneity of tumors and the complexity of the microenvironment at
   the single-cell level. However, despite a wealth of studies using
   scRNA-seq to reveal the diversity within tumors, detailed analyses of
   cell–cell communication remain relatively limited [[96]24, [97]25,
   [98]28]. This study combines scRNA-seq with cell–cell communication
   analysis to provide a deeper understanding of the tumor
   microenvironment.

   Traditional cancer research has largely relied on bulk RNA sequencing
   technologies, which, while providing overall gene expression
   information, fail to resolve the behavioral differences of individual
   cells. In contrast, scRNA-seq technology captures the unique gene
   expression profiles of each cell, revealing cellular heterogeneity and
   the existence of rare cell populations. Many existing studies have used
   scRNA-seq to identify different cell types in tumors, such as cancer
   cells, immune cells, and stromal cells, and analyze their functional
   states.

   However, these studies have primarily focused on the identification of
   cell types and analysis of functional states, with less emphasis on
   cell–cell communication. Building on this, our study incorporates
   cell–cell communication analysis by using the CellChat package to
   identify signaling pathways between cells, revealing the roles of these
   pathways in the tumor microenvironment. This approach contrasts sharply
   with existing studies that are limited to cell type identification. By
   analyzing ligand-receptor pairs, we can infer patterns of cell–cell
   interactions, providing a new dimension of understanding of tumor
   behavior.

   In the tumor microenvironment, cell–cell communication is crucial for
   tumor development and progression. Existing studies have shown that
   intercellular signaling can affect tumor invasiveness, metastatic
   potential, and resistance to treatment. However, these studies are
   often limited to the analysis of single signaling pathways or cell
   types [[99]29–[100]31]. Our study reveals the synergistic effects of
   multiple signaling pathways through comprehensive construction of
   cell–cell communication networks. Utilizing the CellChatDB.human
   database, we systematically identified overexpressed genes and
   ligand-receptor pairs, and calculated communication probabilities. This
   method not only increases the depth of analysis but also enhances the
   reliability of the results. With various visualization methods, such as
   circle plots, chord diagrams, and heatmaps, we can intuitively display
   the complex interactions between different cell types. These innovative
   approaches enable a more comprehensive understanding of cell–cell
   communication in the tumor microenvironment. The results of this study
   not only provide a new perspective for understanding the tumor
   microenvironment but also offer potential targets for developing new
   treatment strategies. By identifying key signaling pathways and cell
   types, we can design interventions more precisely and improve patient
   treatment outcomes. In particular, by revealing communication patterns
   between specific cell types, we can identify potential therapeutic
   targets and develop more targeted treatment methods.

   The relationship between our research findings and clinical treatment
   deserves further exploration. While we have identified key signaling
   pathways, particularly SPP1, in the tumor microenvironment of giant
   cell tumors of bone, the translational value of these findings for
   clinical application remains to be fully elucidated. Future studies
   should investigate correlations between the expression levels of
   identified signaling molecules and patient prognosis, survival rates,
   and treatment responses. For instance, SPP1 (osteopontin) has been
   implicated in tumor progression across multiple cancer types,
   suggesting its potential as a prognostic biomarker. Moreover, the
   identified cell–cell communication networks could guide the development
   of targeted therapies aimed at disrupting crucial tumor-promoting
   interactions. Combination therapies targeting multiple components of
   these signaling networks might prove more effective than single-target
   approaches in overcoming treatment resistance.

   The combined results suggest that this dataset represents a
   functionally coherent set of genes with significant implications for
   both neurodegenerative diseases and infectious disease responses. The
   enrichment of COVID-19 pathways alongside Parkinson's and prion disease
   is particularly noteworthy, as it may point to novel connections
   between these seemingly disparate conditions.The visualization methods
   employed (bar plots, dot plots, and chord diagrams) effectively
   communicate both the statistical significance and the biological
   complexity of these enrichment results.

   Future studies can build on the results of this study to further
   explore the role of specific signaling pathways in different tumor
   types. Additionally, integrating other omics data, such as proteomics
   and metabolomics, will help to fully understand the biological
   characteristics of tumors. Through multi-layered data integration, we
   hope to reveal more complex mechanisms of tumor behavior, providing a
   solid foundation for the development of personalized treatment plans.
   Furthermore, further experimental validation will help confirm the
   specific roles of these signaling pathways in tumor progression, thus
   promoting the translation of basic research into clinical applications.

Limitations

   Despite the insights provided by single-cell RNA sequencing into
   cellular heterogeneity, technical noise and detection limits may affect
   data accuracy. It primarily relies on bioinformatic analysis and
   pathway enrichment results without experimental validation of the SPP1
   pathway's functional effects; correlation does not establish causation
   between SPP1 and cellular communication; the molecular mechanisms by
   which SPP1 signaling regulates intercellular communication remain
   unclear; the study lacks in vitro experiments (such as cell co-culture
   systems). The analysis relies on specific sample sources, potentially
   limiting applicability to all tumor types or patient groups. Although
   potential signaling pathways were identified, there is a lack of
   experimental validation to confirm their roles in tumor progression.

Conclusion

   This study integrates single-cell RNA sequencing and intercellular
   communication analysis to reveal complex cell interaction networks
   within the tumor microenvironment. The identified key signaling
   pathways and cell types offer new perspectives on tumor behavior and
   provide potential targets for developing more precise therapeutic
   strategies. Future research should incorporate other omics data and
   experimental validation to further explore these pathways across
   different tumor types, advancing personalized treatment approaches.

Author contributions

   Zuozhuan Lin wrote the main manuscript text and prepared
   Figs. [101]1-[102]3. zuoxi Chen prepared Figs. [103]4-[104]8. All
   authors reviewed the manuscript.

Funding

   Not applicable.

Data availability

   The datasets generated during and analysed the current study are
   available in the NCBI and Gene Cards,[105]https://www.genecards.org.

Declarations

Ethics approval and consent to participate

   Not applicable.

Consent for Publication

   All authors agree to publish this article.

Competing Interests

   The authors declare no competing interests.

Footnotes

   Publisher's Note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

References