Skip to main content
NCBI home page
IET Systems Biology logoLink to IET Systems Biology
. 2026 Jun 26;20(1):e70069. doi: 10.1049/syb2.70069

Identification of Chemokine‐Related Genes Derived From T and NK Cells in the Tumour Microenvironment of Ovarian Cancer Based on scRNA‐Seq

Liang Wang 1,2,3, Guifen Liu 1,2,3, Liying Wang 1, Yisen Cao 1, JiaYi Jiang 4, Qunhui Wang 5, Xite Lin 1, Zhenhong Wang 1,2,3,✉
PMCID: PMC13309255  PMID: 42362198

ABSTRACT

Ovarian cancer (OC) is characterised by high malignancy. Tumour immune microenvironment (TME) may serve as a breakthrough for therapy. Chemokines are responsible for the recruitment of diverse immune cells in TME. We identify the association between chemokines and T and NK cells. This study was conducted to screen the chemokine derived from T and NK cells in OC single‐cell RNA sequencing (scRNA‐seq) dataset was acquired from the GEO database. Total 64 genes were collected. The intersection of marker genes of T and NK cells and chemokine‐related genes was considered as key chemokine‐related genes. Then, cell‐cell communication analysis, pseudotime analysis and transcription factors (TFs)–target genes regulatory network construction, were performed. Immunohistochemical staining experiments and wound healing assays are performed. The results found 21 distinct cell subgroups and 8 core cell types. Total 667 and 611 marker genes were identified in T and NK cells, respectively. Eight key chemokine‐related genes were found. Also, 49 TFs‐related to cell subtypes were obtained, and the TFs with the highest activation degree was SPI1. Eight key chemokine‐related genes were mainly distributed in T cells, NK cells and monocyte cells. The pseudotime trajectory shown that key chemokine‐related genes were highly expressed in T and NK cells in the later stage of differentiation. CCL5 exhibits differences in ovarian malignant tumour tissues compared with normal ovarian tissues. These findings provide novel insights for subsequent exploration in OC.

Keywords: bioinformatics, cancer, network analysis

Ovarian cancer characterised by high malignancy is mainly because of its complex tumour immune microenvironment. Here, eight chemokine‐related genes were derived from T and NK cells in the tumour immune microenvironment of ovarian cancer, namely CCL5, CXCR6, CXCR3, CXCR4, CCL4, XCL1 and CCL3. These 8 chemokine‐related genes may function in the later stage of differentiation stage to strengthen cell communications, exerting significant functions in the tumourigenesis of ovarian cancer.

graphic file with name SYB2-20-e70069-g004.jpg

Abbreviations

GO

gene ontology

KEGG

Kyoto encyclopaedia of genes and genomes

OC

ovarian cancer

PCA

principal component analysis

PPI

protein‐protein interaction

QC

quality control

scRNA‐seq

single‐cell RNA sequencing

TF

transcription factor

TME

tumour microenvironment

1. Introduction

Ovarian cancer (OC), which seriously threatens women's health, has the highest mortality rate among gynaecological tumours [1, 2]. Despite continuous progress in diagnosis and treatment methods, the overall prognosis of patients with OC remains unsatisfactory [3, 4]. The main reasons lie in the fact that its early symptoms are insidious, most patients are diagnosed at an advanced stage and the tumours have the characteristics of high invasiveness and high metastasis [5]. Therefore, it is urgent to deeply explore the pathogenesis of OC and search for effective diagnostic and therapeutic targets.

Recently, the critical role of the tumour microenvironment (TME) in tumourigenesis and development has received increasing attention with the rapid development of tumour immunology [6, 7, 8]. Chemokines, as an important component of the TME, are a type of small molecule cytokines that can induce the directional migration of cells [9, 10]. They play an extremely important role in tumour occurrence, development, invasion and metastasis by interacting with specific receptors on the cell surface [11, 12]. Studies have shown that multiple chemokines and their receptors are abnormally expressed in OC tissues and cell lines, and closely related to the clinicopathological characteristics of OC [13, 14, 15].Chemokines and their receptors directly affect the functions of T cells and NK cells in the ovarian cancer microenvironment [16, 17]. Therefore, it is essential to identify the key chemokines shared by these two critical cytotoxic cell populations. For example, Guo et al. have indicated that the CXCL12/CXCR4 axis plays a crucial role in the invasion and metastasis of OC, and blocking this axis can obviously inhibit the migration and invasion abilities of OC cells [18]. In addition, chemokines can regulate the recruitment and function of immune cells in TME, affecting the body's anti‐tumour immune response, thereby promoting immune escape in OC [19, 20]. Therefore, in‐depth research on the role of chemokines in the pathogenesis of OC not only helps to reveal the molecular mechanisms of the occurrence and development of OC but also identifies new targets and strategies for the diagnostic and treatment of OC.

Single‐cell RNA sequencing (scRNA‐seq) is a potent tool, which can analyse gene expression at the single‐cell level and reveal cellular heterogeneity in the TME [21, 22, 23, 24]. T cells and NK cells are the major cytotoxic immune cells, and their functions are regulated by shared chemokines. Identifying these common chemokines and their functional crosstalk is valuable for understanding anti‐tumour immunity in ovarian cancer. Thus, in this study, scRNA‐seq was employed to screen the chemokine‐related genes derived from T and NK cells in the TME of OC. This study provides new references and targets for the diagnosis of OC patients. ScRNA‐seq data is from platform (GPL24676). The GSE184880 dataset is seven treatment‐naive patients with early or advanced high‐grade serous ovarian cancer (HGSOC) and five age‐matched non‐malignant ovarian samples.

2. Methods and Methods

2.1. Data Source and Preparation

The scRNA‐seq dataset GSE184880 was acquired from GEO database [25]. ‘Seurat’ package (version 4.3.0.1) [26] was employed for quality control (QC). Finally, total 7 OC samples were extracted for follow‐up analysis. Cells were filtered when mitochondrial gene expression percentages were more than 40%, and the low‐quality cells with the QC criteria of nFeature‐RNA ≤ 200 and nFeature‐RNA ≥ 6000 were filtered. Finally, a total of 30,496 cells and 21,785 genes were obtained. Data normalisation was conducted utilising ‘NormalizeData’. The ‘FindVariableFeatures’ function was employed to extract genes with high coefficient of variation between cells, and the Top 2000 highly variable genes were selected for principal component analysis (PCA) analysis. In addition, total 64 chemokine‐related genes (Table S1) were collected through literature search [27].

2.2. Cell Type Annotation

‘FindNeighbors’, ‘FindClusters’ and ‘RunUMAP’ functions were employed to classify the cells into distinct clusters with the parameter of resolution = 0.4. Cell type annotation was carried out by using ‘SingleR’ (version 2.0.0) [28] combined with ‘the HumanPrimaryCellAtlasData’ dataset in ‘celldex’ database (version 1.8.0) [29], which included 713 microarray samples and classified into 38 major cell types and further annotated as 169 subtypes. Additionally, ‘FindAllMarkers’ function was utilised to conducted differential expression analysis to identify the marker genes in each cell type with the cutoff value of P.adj < 0.05 and |logFC| > 0.263. Also, ‘scRNAtoolVis’ (version 0.1.0) [30] was employed to visualise the Top 5 marker genes of T and NK cells, and the proportion of different cell types was observed. Additionally, cell type annotation is supplemented with manual marker gene verification.

2.3. Screening the Common Genes Between Chemokine‐Related Genes and Marker Genes in T and NK Cells

The marker genes in T and NK cells were intersected with the 64 chemokine‐related genes, and the common genes were acquired as the key chemokine‐related genes in T and NK cells. Moreover, GO and KEGG enrichment analysis was performed on these key chemokine‐related genes using ‘clusterProfiler’ (version 4.6.2) [31] with P.adj < 0.05 and Count ≥ 2. Additionally, the GSEA analysis of the marker genes in T and NK cells was also applied by ‘clusterProfiler’ combined with ‘MSigDB’ database (version 2024.1) [32] corresponding to the subsets of c2.cp.kegg.v2024.1.Hs.symbols.gmt and c5.go.v2024.1.Hs.symbols.gmt as the background dataset with the threshold of |NES| > 1 and p < 0.05.

2.4. Protein‐Protein Interaction (PPI) Network Establishment

To predict the interactions among key chemokine‐related genes, ‘STRING’ database [33] was used to construct PPI network, with a threshold score set to 0.4.

2.5. Gene‐Gene Interation (GGI) Analysis

In order to further observe the potential relationship between key chemokine‐related genes and other genes, as well as potential pathways, ‘GeneMANIA’ database [34] was used for GGI analysis, to predict their colocalisation, shared protein domains, co‐expression, and correlations and pathways.

2.6. Clustering Analysis and Cell Type Annotation of T and NK Cells

To explore the different cell subtypes in T and NK cells, PCA clustering and cell type annotation were performed on both cells again. At the same time, differential expression analysis was performed on the annotated cell subgroups, followed by GO and KEGG enrichment analysis.

2.7. Cell‐Cell Communication Analysis

To explore the quantitative inference and analysis of communication networks between different cells, ‘CellChat’ (version 1.6.1) [35] was used to perform cell–cell communication analysis on cell annotated data.

2.8. Pseudotime Analysis

To understand the gene expression changes experienced by each cell during the cell state transition process, ‘monocle’ (version2.28.0) [36] was conducted to perform pseudotime analysis on T and NK cells. Subsequently, GO enrichment analysis was conducted on differentially expressed genes (DEGs) of each branch along the cell axis.

2.9. Transcription Factor (TF)‐Target Genes Regulatory Network Construction

Firstly, ‘GENIE3’ (version 1.20.0) [37] was used to infer the co‐expression module between TFs and candidate target genes based on co‐expression. ‘RcisTarget’ (version 1.18.2) [38] was employed to perform cis‐regulatory motif analysis on each co‐expressed module. Then TF‐motif enrichment analysis was performed to identify direct targets. Finally, the ‘AUCell’ algorithm (version 1.28.0) [39] was utilised to score the activity of each regulon in each cell. Cytoscape [40] was used to construct and visualise TF–target genes regulatory network.

2.10. Copy Number Variation Analysis

Using ‘infircnv’ package (version 1.14.2) [41], common human normal cells were used as controls to observe the changes in gene expression intensity at various positions on the tumour genome. The relative expression levels of genes on each chromosome were displayed in the heatmap, and the malignant tumour cells were identified based on the CNV score.

2.11. Metabolic Pathway Analysis

To observe the metabolic activity of different cell types in single cells, the ‘scMetabolism’ package (version 0.2.1) [42] was used to quantitatively score the metabolic activity of different cell types. KEGG was selected as the background pathway gene set to score each cell, and then the metabolic scores of different cell types were visualised.

2.12. Single‐Cell Expression Pattern of Key Chemokine‐Related Genes

Based on the key chemokine‐related genes and single‐cell data obtained above, the expression of key chemokine‐related genes among different cell types and subtypes of T and NK cells were analysed, and their pseudotime trajectory in single‐cell data was visualised.

2.13. Immunohistochemistry of Human Ovarian Malignant Tumour Tissues and Normal Ovarian Tissues

This study was approved by the Ethics Committee of Fujian Maternity and Child Health Hospital, and informed consent was obtained from all patients. A total of 21 tissue samples were collected from female patients diagnosed with ovarian malignant tumours at Fujian Maternity and Child Health Hospital, along with 21 normal ovarian tissue samples from female patients with benign disease. The inclusion criteria for the ovarian malignant tumour group were as follows: (1) Patients with ovarian malignant tumours who had not received chemotherapy, radiotherapy, or other neoadjuvant therapies; (2) no history of other malignant tumours and (3) clear pathological diagnosis. The collected tissue samples were excised, fixed in formalin, sectioned and then embedded in paraffin. For the normal ovarian tissue group, the first inclusion criterion was patients with pathologically confirmed benign disease and underwent ovarian resection; the other criteria were consistent with those of the ovarian malignant tumour group. Patient age and pathological information are available in Table S2.

The immunohistochemical procedure was performed according to routine protocol. The sections were incubated with the primary antibody (rabbit anti‐human CCL5 antibody [dilution ratio 1:500, catalogue number ab307712; purchased from Abcam Inc., UK]). For the secondary antibody, 100 μL of goat anti‐rabbit IgG (kit catalogue number Kit‐0038; purchased from Zhongshan Jinqiao Biotechnology Co. Ltd.) was used. The results were determined by two senior pathologists using a semi‐quantitative assessment method. According to the immunohistochemical staining results, weakly positive, positive and strongly positive were considered as positive expression of the protein in the tissue.

2.14. Quantitative Real‐Time Polymerase Chain Reaction

Total RNA was prepared from cultured cells using Trizol reagent (Invitrogen) following the manufacturer's instructions. RNA was reverse transcribed to cDNA using the TransScript All‐in‐One first‐strand cDNA synthesis SuperMix (Transgen Biotech, Beijing, China). Real‐time PCR was performed using the TransStart Top Green qPCR SuperMix (Transgen Biotech) and the ABI 7300 Real‐Time qPCR system. All PCR experiments were performed in triplicate. The primers were synthesised by Sangon Biotech (Shanghai) Co. Ltd.

2.15. In Vitro ‘Wound’ Closure Assay

A2780‐NC and A2780‐CCL5 cells were plated in 6‐well plates until cells reached confluence. A ‘wound’ was then created by scratching the surface of the well with a pipette tip. Following injury, cells were further cultured in serum‐free 1640 for 24h. Cells were then photographed using a phase‐contrast microscope (× 100) and the wound closure was measured.

2.16. Statistical Analysis

Data analyses were all processed in R software (version 4.3.0) and Linux. Experimental data were analysed by GraphPad Prism 9.0, and the results were expressed as the mean ± standard deviation. All the results were repeated at least three or more times. The t‐test or Mann–Whitney test were used for comparison between the two groups. All statistical studies used p < 0.05 or the corrected p value < 0.05 as the threshold for statistical significance.

3. Results

3.1. Clustering Analysis and Cell Type Annotation

After scRNA‐seq data from GSE184880 dataset downloading, the stringent QC was conducted and highly variable genes were identified (Figure S1A,B). The samples showed no obvious batch separation, which may be attributed to the limited sample size. Subsequently, ‘JackStrawPlot’ and ‘ElbowPlot’ functions were utilised to pick the most appropriate number of PC, the results found that when the number of PC was 15, the cumulative standard deviation of PCs decreased slowly (Figure S1C,D). Then a total of 21 distinct cell subgroups were screened after descending clustering using ‘UMAP’ and ‘TSNE’ (Figure 1A,B), and total 8 cell types were annotated, containing T cells, monocyte, NK cells, B cells, epithelial cells, smooth muscle cells, endothelial cells and tissue stem cells (Figure 1A,B), and the marker genes expression in these cell types were shown in Figure 1C,D. Notably, 667, 611 marker genes were identified in T and NK cells, respectively. In addition, the percentage of the different cell types was shown in Figure 1E. Furthermore, scatter plot of the Top 5 marker genes distribution in each cell was drawn. Herein, the scatter plot of the Top 5 marker genes distribution in T cell was presented in Figure 1F.

FIGURE 1.

FIGURE 1

Open in a new tab

Clustering analysis and cell type annotation. Cell type classification and cluster annotation using ‘UMAP’ (A) and ‘TSNE’ (B). Heatmap (C) and volcano plot (D) of the marker genes expression in cell types. (E) The percentage of the different cell types. Left: overall proportion; Right: proportion in each sample. (F) The scatter plot of the top 5 marker genes distribution in T cells.

3.2. Screening of Key Chemokine‐Related Genes in T and NK Cells

After merging and deduplicating the obtained marker genes in T and NK cells, and 988 marker genes were obtained. Then the 988 marker genes were intersected with the 64 chemokine‐related genes, and 8 overlapped genes were screened as key chemokine‐related genes in T and NK cells (Figure 2A), namely CCL5, CXCR6, CXCR3, CXCR4, CCL4, XCL2, XCL1 and CCL3. Moreover, enrichment analysis revealed that these 8 genes enriched in 147 GO terms and 9 KEGG pathways, containing chemokine‐mediated signalling pathway, natural killer cell chemotaxis, T cell chemotaxis etc (Figure 2B,C). Additionally, the GSEA analysis of the marker genes in T and NK cells was also applied. The marker genes in NK cells enriched in 353 GO terms and 10 KEGG pathways (Figure 2D), and the marker genes in T cells involved in 445 GO terms and 12 KEGG pathways (Figure 2E).

FIGURE 2.

FIGURE 2

Open in a new tab

Screening key chemokine‐related genes functional enrichment analysis, as well as PPI and GGI networks. (A) Venn diagram of common genes between chemokine‐related genes and marker genes in T and NK cells. Gene ontology (GO) terms (B) and Kyoto encyclopaedia of genes and genomes (KEGG) pathways (C) that 8 key chemokine‐related genes enriched. Enrichment analysis of marker genes in NK (D) and T cells (E). F protein–protein interaction (PPI) network of 8 key chemokine‐related genes. (G) GGI analysis of 8 key chemokine‐related genes and 20 their interaction genes.

3.3. PPI and GGI Analysis

Based on the 8 key chemokine‐related genes, a PPI network was constructed (Figure 2F), and significant protein‐interaction relationships were found among the 8 genes. Also, GGI analysis uncovered that 20 genes were related to these genes, as well as some pathways, containing cytokine activity, chemokine receptor binding, cellular response to chemokine, response to chemokine, leucocyte chemotaxis, leucocyte migration and cell chemotaxis (Figure 2G).

3.4. Clustering Analysis and Cell Type Annotation of T and NK Cells

Total 12 and 10 distinct cell subgroups of T and NK cells were screened after clustering using ‘UMAP’ and ‘TSNE’, respectively (Figure 3A,B), and 3 (CD4+ effecto memory T cell, CD8+ T cell and gamma‐delta T cell) and 7 (CD56bright NK cell, KLRC1+ NK cell, general NK cell, trNK cell, CD3+ NK cell, CD56dim NK cell and CD62L + NK cell) cell types of T and NK cells were annotated, respectively (Figure 3C,D). Besides, differential expression analysis was performed on the annotated cell subgroups, and the marker genes in annotated cell subgroups of T and NK cells were acquired (Figure 3E,F). Moreover, enrichment analysis demonstrated that the marker genes of the annotated T cell subclusters were significantly enriched in 588 GO terms and 53 KEGG pathways, including response to virus, the T cell receptor signalling pathway and the NOD‐like receptor signalling pathway (Figure 3G,H). Similarly, 754 GO terms and 68 KEGG pathways, including natural killer cell mediated immunity, immune response‐activating signalling pathway and natural killer cell mediated cytotoxicity, were enriched by the marker genes in annotated cell subgroups of NK cells (Figure 3I,J).

FIGURE 3.

FIGURE 3

Open in a new tab

Clustering analysis and cell type annotation of T and NK cells. Cell type classification of T (A) and NK (B) cells using ‘UMAP’ and ‘TSNE’. Cell type annotation of T (C) and NK (D) cells using ‘UMAP’ and ‘TSNE’. The marker genes in annotated cell subgroups of T (E) and NK (F) cells. Left: Heatmap of the Top 10 marker genes; left: bubble plot of marker genes. Gene ontology (GO) terms (G) and Kyoto encyclopaedia of genes and genomes (KEGG) pathways (H) that the marker genes in annotated cell subgroups of T cells enriched. GO terms (I) and KEGG pathways (J) that the marker genes in annotated cell subgroups of NK cells enriched.

3.5. Cell–Cell Communication Analysis

Normalised single‐cell expression profiles were used as input data, cell types were annotated as cell information, and high‐expressed ligand receptor genes in each cell subtype were identified. Subsequently, overexpressed ligand–receptor pairs were identified based on the CellChatDB database, and cellular communication networks were inferred according to ligand–receptor expression levels. The number of interactions and interaction weights/strength between cells were visualised in Figure 4A,B. Subsequently, the ligand receptor pairs between T and NK cells and other cells were shown in Figure 4C.

FIGURE 4.

FIGURE 4

Open in a new tab

Cell–cell communication analysis. The number of interactions (A) and interaction weights/strength (B) between cells. Each colour represents a different cell, the arrow represents the order, and the thickness of the line represents the number/weight. (C) The ligand receptor pairs between T and NK cells and other cells. The horizontal axis represents the direction of interaction between different cells, and the vertical axis represents specific ligand–receptor pairs. The different colours of the dots represent communication probabilities, and the size of the dot represents the p‐value of the communication significance.

3.6. Pseudotime Analysis

‘Monocle’ was employed to observe the differentiation relationships of intracellular subtypes and infer relationships between cell trajectories and pseudotime. It was found that the density of gamma‐delta T cells was higher in the early stage of differentiation, whereas CD8+ T cells were mainly distributed in the middle stage of differentiation, and CD4+ effector memory T cells and gamma‐delta T cells were mainly in the late stage of differentiation in T cells (Figure 5A–D). In NK cells, KLRC1+ NK cells had a high density in the early stage of differentiation, while other subtypes of NK cells accounted for the majority in the late stage of differentiation (Figure 5E–H). Then the expression heatmaps of DEGs in T and NK cells with pseudotime changes were observed in branches of different pseudotime nodes (Figure 5I,J).

FIGURE 5.

FIGURE 5

Open in a new tab

Pseudotime analysis. (A) The T cell type maps to the ordering of cell pseudotime values. (B) Sequencing differences in pseudotime values of T cell differentiation. Dark blue indicates early differentiation, and light blue indicates late differentiation. (C) The density change curves of different subtypes of T cells under different pseudotime. (D) Differentiation of T cells in different pseudotime sequences under different cell types. (E) The NK cell type maps to the ordering of cell pseudotime values. (F) Sequencing differences in pseudotime values of NK cell differentiation. Dark blue indicates early differentiation, and light blue indicates late differentiation. (G) The density change curves of different subtypes of NK cells under different pseudotime. (H) Differentiation of NK cells in different pseudotime sequences under different cell types. (I) Analysis of the T cell gene expression heatmap by ‘Beam’. (J) Analysis of the NK cell gene expression heatmap by ‘Beam’.

3.7. TF‐Target Genes Regulatory Network Construction

Total 49 TFs related to cell subtypes were obtained, and each of these processed TFs and their potential direct target genes were called as a regulon. The regulons activity score of each cell was analysed using ‘AUCell’. The basis of the score was the expression value of the gene, and the higher score with the higher activation degree of the gene set. The results showed that among the TFs directly related to cells, the TFs with the highest activation degree was SPI1, and the TFs with the highest activation degree was LYL1 among the TFs associated with cells including indirect correlations (Figure 6A). Additionally, the TF–target genes regulatory network was constructed (Figure 6B).

FIGURE 6.

FIGURE 6

Open in a new tab

Transcription factor (TF)–target genes regulatory network construction. (A) Scatter plots of the active regions of Top 5 TFs and the distribution of regulon AUC. Left: The distribution of regulon AUC. The horizontal axis represents the AUC value, and the vertical axis represents the number of cells corresponding to AUC. Middle: AUC binary activity. Right: AUC activity. The darker the colour with the higher activity. (B) TF–target genes regulatory network construction. The orange nodes indicate TFs, and the green nodes indicate genes.

3.8. Copy Number Variation and Metabolic Pathway Analysis

Copy number variation analysis found that compared with the control cells, epithelial cells showed abnormal expression on multiple chromosomes (Figure 7A). Besides, metabolic pathway analysis revealed that the Top 10 pathways with higher average cell type scores included glycolysis/gluconeogenesis, Pentose phosphate pathway, D‐Glutamine and D‐glutamate metabolism etc (Figure 7B).

FIGURE 7.

FIGURE 7

Open in a new tab

Copy number variation and metabolic pathway analysis. (A) InferCNV result of 8 cell types. The upper half indicates references (cells), and the lower half indicates observations (cells). (B) Bubble chart of Top 10 metabolic pathway scoring of 8 cell types.

3.9. Single‐Cell Expression Pattern of Key Chemokine‐Related Genes

The expression of key chemokine‐related genes among different cell types were analysed. The results showed that the expression levels of 8 key chemokine‐related genes were generally high in single‐cell data, and their expression was mainly distributed in T cells, NK cells and monocyte cells (Figure 8A). The pseudotime trajectory analysis showed that these 8 genes were highly expressed in T and NK cells in the later stage of differentiation with the development of cells (Figure 8B).

FIGURE 8.

FIGURE 8

Open in a new tab

Single‐cell expression pattern of key chemokine‐related genes. (A) The expression of key chemokine‐related genes among different cell types. (B) The pseudotime trajectory of 8 key chemokine‐related genes in single‐cell data.

3.10. Validation for Expression of CCL5 in Clinical Tissues and Ovarian Cancer Cells

Twenty‐one patients with pathologically confirmed ovarian malignant tumours were enroled, and 21 patients whose ovarian tissues were resected due to benign were collected as control. There was no statistically significant difference in the general data between the two groups. The expression of CCL5 was detected by immunohistochemical staining. The results showed that there was a statistically significant difference in the expression of CCL5 between ovarian malignant tumour tissues and normal ovarian tissues (p < 0.05, see Table 1). The immunohistochemical staining of CCL5‐positive ovarian malignant tumour tissues and CCL5‐negative ovarian tissues is shown in the figure. However, it should be noted that even in ovarian malignant tumour tissues, the positive rate of CCL5 is not high (Figure 9A). The ovarian cancer cell line A2780 was transfected with CCL5 overexpression plasmid, and qRT‐PCR confirmed the successful establishment of CCL5‐overexpressing cell lines (Figure 9B). The cell migration ability was compared between A2780‐NC and A2780‐CCL5 overexpression cell lines, and no statistically significant difference was observed in the 24‐h wound healing assay (Figure 9C).

TABLE 1.

Comparison of CCL5 expression between patients with ovarian malignant tumours and non‐ovarian malignant tumours.

Ovarian malignant tumours (N = 21) Non‐ovarian malignant tumours (N = 21) p value
CCL5 positive 7 (33.33%) 1 (4.76%) p < 0.05
CCL5 negative 14 (66.67%) 20 (95.24%)
Open in a new tab

FIGURE 9.

FIGURE 9

Open in a new tab

Validation for expression of CCL5 in clinical tissues and ovarian cancer cells. There is positive expression of CCL5 in ovarian malignant tumour tissues, and the expression of CCL5 in normal ovarian tissues is negative (A). Because A2780 is a CCL5‐negative cell line, we successfully constructed the CCL5‐positive AI780 cell line, and verified the successful construction by qRT‐PCR (B). However, in the in vitro environment where ovarian cancer cells are cultured alone, whether CCL5 is overexpressed has no effect on the metastatic ability (C).

4. Discussion

Ovarian cancer is the third most common gynaecological malignancy, yet it has the highest mortality rate [43, 44]. In this study, scRNA‐seq analysis was utilised to screen the chemokine‐related genes derived from T and NK cells in the TME of OC, revealing the heterogeneity of NK and T cell subsets in the microenvironment of OC. Finally, total 8 key chemokine‐related genes were acquired, namely CCL5, CXCR6, CXCR3, CXCR4, CCL4, XCL2, XCL1 and CCL3.

In this study, scRNA‐seq analysis was performed on matching 7 OC samples. The samples were successfully annotated 8 cell types, containing T cells, monocyte, NK cells, B cells, epithelial cells, smooth muscle cells, endothelial cells and tissue stem cells. Notably, the percentage of the T cells, NK cells and monocytes was relatively high. T cells are an important component of the immune system and play a key role in cancer, including recognising cancer cells, directly killing cancer cells and secreting cytokines. [45]. NK cells are the core effector cells of the innate immune system, possessing the ability to directly kill tumour cells and regulate the immune microenvironment [46, 47]. Besides, the GSEA analysis of the marker genes in T and NK cells showed that the marker genes in NK cells enriched in 353 GO terms and 10 KEGG pathways, and the marker genes in T cells involved in 445 GO terms and 12 KEGG pathways. Even within ovarian cancer, T cell and NK cell populations exhibit highly complex heterogeneity. Therefore, identifying common factors among different cell subsets is also critically important [48].Interestingly, the marker genes in T and NK cells were both enriched in antigen processing and presentation, cytokine–cytokine receptor interaction and natural killer cells mediated cytotoxicity pathways. Under normal circumstances, the body can recognise and eliminate cancer cells expressing abnormal antigens through the process of antigen processing and presentation [49]. It is reported that cytokine‐cytokine receptor interaction is involved in shaping the tumour microenvironment, influencing the interaction between tumour cells and surrounding cells [50]. Thus, these data indicated that T and NK cells might exert a significant function in the progression of OC, and the marker genes in T and NK cells might involve in the progression of OC via antigen processing and presentation, cytokine–cytokine receptor interaction and natural killer cells mediated cytotoxicity pathways.

Chemokine CCL5, is a critical mediator for immune cells to recruit tumours [51, 52]. Han et al. revealed that upregulation of CCL5 is associated with the immune profile of TME [53]. Clinical tissues and ovarian cancer cell lines were used to validate CCL5. Analysis of clinical tissues revealed a significant difference in CCL5 expression between ovarian malignant tumour tissues and normal ovarian tissues, suggesting that increased CCL5 expression may be associated with the occurrence and progression of OC. Previous studies have confirmed that CCL5 exhibits heterogeneous expression in OC and a variety of other tumours [54]. However, when ovarian cancer cell lines were cultured alone to verify the effect of CCL5 on cell metastatic capacity, no significant difference was observed. Tumour cells represent a major source of chemokines. Although secreted by cancer cells, these chemokines do not act on tumour cells directly but function primarily by recruiting immune cells (e.g., CD8+ T cells, NK cells and macrophages) into the tumour microenvironment via binding to their cognate receptors on immune cells [55]. This may be attributed to the inability of CCL5 to exert its function in an environment lacking immune cells. Notably, the negative results of this study still hold reference value for subsequent research.

The correlation between chemokines and OC progression has been continuously confirmed. OC cells can secrete and release chemokines, thereby mediating chemotherapy resistance by regulating the expression and activity of histone deacetylases (HDACs) [56]. It's reported that the interaction between CXCR6 and ligand CXCL16 is associated with cancer metastasis and invasion [57]. Mabrouk et al. have highlighted that the expression of CXCR6 can serve as a marker for T cell differentiation and may serve as a biomarker for cancer [17]. XCL2 is a structurally conserved chemokine participated in activating cytotoxic T cells, which exerts a significant function in the development and progression of various cancers, containing OC [58].Macrophage‐derived CCL23 is upregulated in ovarian cancer and drives CD8+ T‐cell exhaustion via GSK3β signalling to promote immune suppression [59].

In this study, total 988 marker genes were obtained after merging and deduplicating the obtained marker genes in T and NK cells. Then the 988 marker genes were intersected with the 64 chemokine‐related genes, and eight overlapped genes were identified as key chemokine‐related genes in T and NK cells, namely CCL5, CXCR6, CXCR3, CXCR4, CCL4, XCL2, XCL1 and CCL3. These findings emphasise that these 8 chemokine‐related genes derived from T and NK cells in the TME of OC and may exert function in OC‐related immune response. Moreover, the pseudotime trajectory analysis found that the 8 key chemokine‐related genes were highly expressed in T and NK cells in the later stage of differentiation with the development of cells. Therefore, these genes may function in the later stage of differentiation stage to strengthen cell communications, exerting significant functions in the tumourigenesis of OC.

TFs are a class that can bind to specific DNA sequences in the promoter region of genes [60], and they can regulate tumour cells to secrete various cytokines, chemokines and growth factors, thereby affecting the TME [61]. Herein, total 49 TFs‐related to cell subtypes were obtained, and the TFs with the highest activation degree was SPI1 among the TFs directly related to cells, and the TFs with the highest activation degree was LYL1 among the TFs associated with cells including indirect correlations. SPI1 belongs to the ETS TF family, and SPI1 has been reported to be linked to the development of OC, gastric cancer, cervical cancer etc. [62, 63, 64]. The LYL1 gene is believed to regulate cell proliferation and differentiation, and it is also associated to numerous cancer‐related properties [65, 66]. Thus, SPI1 and LYL1 might play a crucial role in OC immunity by regulating multiple aspects such as the development, differentiation and function of immune cells in OC.

However, there are numerous deficiencies that should be acknowledged. First, the chemokine‐related genes are collected through literature search, which is being updated ceaselessly, and more genes are yet to be acquired. Second, the in vitro culture of ovarian cancer cells alone lacks the involvement of immune cells, and this research model is not suitable for studying chemokine functions. In the future, using single‐cell sequencing data, we will further explore the relationship between chemokines and ovarian cancer, employing strategies such as analysing chromatin interaction networks and investigating the roles of non‐coding RNAs [67, 68, 69]. As for the eight genes that have been screened, further in vitro and in vivo experiments would be under a co‐culture model involving both immune cells and ovarian cancer cells.

5. Conclusion

In conclusion, total 8 chemokine‐related genes were derived from T and NK cells in the TME of OC, namely CCL5, CXCR6, CXCR3, CXCR4, CCL4, XCL2, XCL1 and CCL3. The pathways and transcription factors associated with these key chemokine‐related genes were identified. Analysing chemokines using clinical tissue samples is reasonable; however, in vitro tumour cell culture should ideally be performed within a model that incorporates immune cell components. This study may provide a theoretical foundation and novel therapeutic targets for the diagnosis and treatment of ovarian cancer.

Author Contributions

Liang Wang: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, writing – review and editing. Guifen Liu: resources, software, writing – original draft, writing – review and editing. Liying Wang: resources, software, supervision, validation, visualization, writing – original draft, writing – review and editing. Yisen Cao: funding acquisition, investigation, validation, visualization, writing – original draft, writing – review and editing. JiaYi Jiang: conceptualization, data curation, investigation, methodology, project administration, resources, software, writing – review and editing. Qunhui Wang: data curation, investigation, project administration, resources, software, supervision, validation, visualization, writing – review and editing. Xite Lin: resources, software. Zhenhong Wang: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration.

Funding

The work was supported by National Natural Science Foundation of China (NSFC) (Grant No. 82303734), Fujian Medical University Training Programme of Innovation and Entrepreneurship for Undergraduates (Grant No. C2025072), Science and Technology Innovation Joint Fund Project of Fujian Provincial Health Commission (Grant No. 2024Y9535), Fujian Provincial Natural Science Foundation of China (Grant Nos. 2025J01201, 2022J011027), and Fujian Provincial Health Technology Project (Grant No. 2021QNA048).

Ethics Statement

This study was approved by Ethics Committee of Fujian Provincial Maternity and Child Health Hospital (2022KYLLRD03063). The research was carried out according to the Declaration of Helsinki.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Preprocessing, quality control, highly variable gene identification and principal component selection of the GSE184880 scRNA‐seq dataset. The scRNA‐seq data of GSE184880 were downloaded and subjected to strict quality control (QC), and highly variable genes were screened out (A, B). The JackStrawPlot and ElbowPlot algorithms were applied to determine the optimal principal component (PC) number. The cumulative standard deviation tended to decline gently when the PC number was set to 15 (C, D).

SYB2-20-e70069-s001.docx (407.5KB, docx)

Table S1: 64 chemokine‐related genes.

SYB2-20-e70069-s002.docx (19.2KB, docx)

Table S2: Comparison of age and pathological data between patients with ovarian malignant tumors and normal ovarian tissue, those normal samples were obtained from patients who underwent oophorectomy for benign diseases or benign ovarian cysts and normal ovarian tissue was used after excluding cystic lesions during specimen processing.

SYB2-20-e70069-s003.docx (18KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  • 1. Sowamber R., Lukey A., Huntsman D., and Hanley G., “Ovarian Cancer: From Precursor Lesion Identification to Population‐Based Prevention Programs,” Current Oncology 30, no. 12 (2023): 10179–10194, 10.3390/curroncol30120741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Bray F., Laversanne M., Sung H., et al., “Global Cancer Statistics 2022: Globocan Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries,” CA: A Cancer Journal for Clinicians 74, no. 3 (2024): 229–263, 10.3322/caac.21834. [DOI] [PubMed] [Google Scholar]
  • 3. Morand S., Devanaboyina M., Staats H., Stanbery L., and Nemunaitis J., “Ovarian Cancer Immunotherapy and Personalized Medicine,” International Journal of Molecular Sciences 22, no. 12 (2021): 6532, 10.3390/ijms22126532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Kuroki L. and Guntupalli S. R., “Treatment of Epithelial Ovarian Cancer,” BMJ 371 (2020): m3773, 10.1136/bmj.m3773. [DOI] [PubMed] [Google Scholar]
  • 5. Konstantinopoulos P. A. and Matulonis U. A., “Clinical and Translational Advances in Ovarian Cancer Therapy,” Nature Cancer 4, no. 9 (2023): 1239–1257, 10.1038/s43018-023-00617-9. [DOI] [PubMed] [Google Scholar]
  • 6. Xiao Y. and Yu D., “Tumor Microenvironment as a Therapeutic Target in Cancer,” Pharmacology & Therapeutics 221 (2021): 107753, 10.1016/j.pharmthera.2020.107753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Jin M. Z. and Jin W. L., “The Updated Landscape of Tumor Microenvironment and Drug Repurposing,” Signal Transduction and Targeted Therapy 5, no. 1 (2020): 166, 10.1038/s41392-020-00280-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Bader J. E., Voss K., and Rathmell J. C., “Targeting Metabolism to Improve the Tumor Microenvironment for Cancer Immunotherapy,” Molecules and Cells 78, no. 6 (2020): 1019–1033, 10.1016/j.molcel.2020.05.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Masih M., Agarwal S., Kaur R., and Gautam P. K., “Role of Chemokines in Breast Cancer,” Cytokine 155 (2022): 155909, 10.1016/j.cyto.2022.155909. [DOI] [PubMed] [Google Scholar]
  • 10. Zhang Y., Tang X., Wang Y., et al., “Recent Advances Targeting Chemokines for Breast Cancer,” International Immunopharmacology 146 (2025): 113865, 10.1016/j.intimp.2024.113865. [DOI] [PubMed] [Google Scholar]
  • 11. Jung H. and Paust S., “Chemokines in the Tumor Microenvironment: Implications for Lung Cancer and Immunotherapy,” Frontiers in Immunology 15 (2024): 1443366, 10.3389/fimmu.2024.1443366. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Jia S. N., Han Y. B., Yang R., and Yang Z. C., “Chemokines in Colon Cancer Progression,” Seminars in Cancer Biology 86, no. Pt 3 (2022): 400–407, 10.1016/j.semcancer.2022.02.007. [DOI] [PubMed] [Google Scholar]
  • 13. Bose S., Saha P., Chatterjee B., and Srivastava A. K., “Chemokines Driven Ovarian Cancer Progression, Metastasis and Chemoresistance: Potential Pharmacological Targets for Cancer Therapy,” Seminars in Cancer Biology 86, no. Pt 2 (2022): 568–579, 10.1016/j.semcancer.2022.03.028. [DOI] [PubMed] [Google Scholar]
  • 14. Ullah A., Chen Y., Singla R. K., Cao D., and Shen B., “Pro‐Inflammatory Cytokines and Cxc Chemokines as Game‐Changer in Age‐Associated Prostate Cancer and Ovarian Cancer: Insights From Preclinical and Clinical Studies' Outcomes,” Pharmacological Research 204 (2024): 107213, 10.1016/j.phrs.2024.107213. [DOI] [PubMed] [Google Scholar]
  • 15. Li W., Ma J. A., Sheng X., and Xiao C., “Screening of Cxc Chemokines in the Microenvironment of Ovarian Cancer and the Biological Function of Cxcl10,” World Journal of Surgical Oncology 19, no. 1 (2021): 329, 10.1186/s12957-021-02440-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Zhuo L., Meng F., Sun K., Zhou M., and Sun J., “Integrated Immuno‐Transcriptomic Analysis of Ovarian Cancer Identifies a Four‐Chemokine‐Dominated Subtype With Antitumor Immune‐Active Phenotype and Favorable Prognosis,” British Journal of Cancer 131, no. 6 (2024): 1068–1079, 10.1038/s41416-024-02803-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Mabrouk N., Tran T., Sam I., et al., “Cxcr6 Expressing T Cells: Functions and Role in the Control of Tumors,” Frontiers in Immunology 13 (2022): 1022136, 10.3389/fimmu.2022.1022136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Guo Q., Gao B. L., Zhang X. J., et al., “Cxcl12‐Cxcr4 Axis Promotes Proliferation, Migration, Invasion, and Metastasis of Ovarian Cancer,” Oncology Research Featuring Preclinical and Clinical Cancer Therapeutics 22, no. 5–6 (2014): 247–258, 10.3727/096504015x14343704124430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. You Y., Li Y., Li M., et al., “Ovarian Cancer Stem Cells Promote Tumour Immune Privilege and Invasion via Ccl5 and Regulatory T Cells,” Clinical and Experimental Immunology 191, no. 1 (2018): 60–73, 10.1111/cei.13044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Shao D., Zhou H., Yu H., and Zhu X., “Cx3cr1 Is a Potential Biomarker of Immune Microenvironment and Prognosis in Epithelial Ovarian Cancer,” Medicine (Baltimore) 103, no. 3 (2024): e36891, 10.1097/md.0000000000036891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Zhou X. and Wu H., “Schiclassifier: A Deep Learning Framework for Cell Type Prediction by Fusing Multiple Feature Sets From Single‐Cell Hi‐C Data,” Briefings in Bioinformatics 26, no. 1 (2024): bbaf009, 10.1093/bib/bbaf009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Yang X., Mann K. K., Wu H., and Ding J., “Sccross: A Deep Generative Model for Unifying Single‐Cell Multi‐Omics With Seamless Integration, Cross‐Modal Generation, and in Silico Exploration,” Genome Biology 25, no. 1 (2024): 198, 10.1186/s13059-024-03338-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Shi Z., Cong L., and Wu H., “Momicpred: A Cell Cycle Prediction Framework Based on Dual‐Branch Multi‐Modal Feature Fusion for Single‐Cell Multi‐Omics Data,” IEEE Journal of Biomedical and Health Informatics 30, no. 3 (2026): 2242–2251, 10.1109/jbhi.2025.3595904. [DOI] [PubMed] [Google Scholar]
  • 24. Jia F., Sun S., Li J., et al., “Neoadjuvant Chemotherapy‐Induced Remodeling of Human Hormonal Receptor‐Positive Breast Cancer Revealed by Single‐Cell RNA Sequencing,” Cancer Letters 585 (2024): 216656, 10.1016/j.canlet.2024.216656. [DOI] [PubMed] [Google Scholar]
  • 25. Clough E., Barrett T., Wilhite S. E., et al., “Ncbi Geo: Archive for Gene Expression and Epigenomics Data Sets: 23‐Year Update,” Nucleic Acids Research 52, no. D1 (2024): D138–d144, 10.1093/nar/gkad965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Stuart T., Butler A., Hoffman P., et al., “Comprehensive Integration of Single‐Cell Data,” Cell 177, no. 7 (2019): 1888–1902.e1821, 10.1016/j.cell.2019.05.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Hughes C. E. and Nibbs R. J. B., “A Guide to Chemokines and Their Receptors,” FEBS Journal 285, no. 16 (2018): 2944–2971, 10.1111/febs.14466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Cao G., Xuan X., Li Y., et al., “Single‐Cell RNA Sequencing Reveals the Vascular Smooth Muscle Cell Phenotypic Landscape in Aortic Aneurysm,” Cell Communication and Signaling 21, no. 1 (2023): 113, 10.1186/s12964-023-01120-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Aran D., Looney A. P., Liu L., et al., “Reference‐Based Analysis of Lung Single‐Cell Sequencing Reveals a Transitional Profibrotic Macrophage,” Nature Immunology 20, no. 2 (2019): 163–172, 10.1038/s41590-018-0276-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Zappia L., Phipson B., and Oshlack A., “Exploring the Single‐Cell RNA‐Seq Analysis Landscape With the scRNA‐Tools Database,” PLoS Computational Biology 14, no. 6 (2018): e1006245, 10.1371/journal.pcbi.1006245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Yu G., Wang L. G., Han Y., and He Q. Y., “Clusterprofiler: An R Package for Comparing Biological Themes Among Gene Clusters,” OMICS 16, no. 5 (2012): 284–287, 10.1089/omi.2011.0118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Liberzon A., Subramanian A., Pinchback R., Thorvaldsdóttir H., Tamayo P., and Mesirov J. P., “Molecular Signatures Database (Msigdb) 3.0,” Bioinformatics 27, no. 12 (2011): 1739–1740, 10.1093/bioinformatics/btr260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. von Mering C., Huynen M., Jaeggi D., Schmidt S., Bork P., and Snel B., “String: A Database of Predicted Functional Associations Between Proteins,” Nucleic Acids Research 31, no. 1 (2003): 258–261, 10.1093/nar/gkg034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Franz M., Rodriguez H., Lopes C., et al., “Genemania Update 2018,” Nucleic Acids Research 46, no. W1 (2018): W60–w64, 10.1093/nar/gky311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Jin S., Guerrero‐Juarez C. F., Zhang L., et al., “Inference and Analysis of Cell‐Cell Communication Using Cellchat,” Nature Communications 12, no. 1 (2021): 1088, 10.1038/s41467-021-21246-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Trapnell C., Cacchiarelli D., Grimsby J., et al., “The Dynamics and Regulators of Cell Fate Decisions Are Revealed by Pseudotemporal Ordering of Single Cells,” Nature Biotechnology 32, no. 4 (2014): 381–386, 10.1038/nbt.2859. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Ruyssinck J., Huynh‐Thu V. A., Geurts P., Dhaene T., Demeester P., and Saeys Y., “Nimefi: Gene Regulatory Network Inference Using Multiple Ensemble Feature Importance Algorithms,” PLoS One 9, no. 3 (2014): e92709, 10.1371/journal.pone.0092709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Wang Y., Li B., and Zhao Y., “Inflammation in Preeclampsia: Genetic Biomarkers, Mechanisms, and Therapeutic Strategies,” Frontiers in Immunology 13 (2022): 883404, 10.3389/fimmu.2022.883404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Aibar S., González‐Blas C. B., Moerman T., et al., “Scenic: Single‐Cell Regulatory Network Inference and Clustering,” Nature Methods 14, no. 11 (2017): 1083–1086, 10.1038/nmeth.4463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Saito R., Smoot M. E., Ono K., et al., “A Travel Guide to Cytoscape Plugins,” Nature Methods 9, no. 11 (2012): 1069–1076, 10.1038/nmeth.2212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Chen K., Wang Y., Hou Y., et al., “Single Cell RNA‐Seq Reveals the Ccl5/Sdc1 Receptor‐Ligand Interaction Between T Cells and Tumor Cells in Pancreatic Cancer,” Cancer Letters 545 (2022): 215834, 10.1016/j.canlet.2022.215834. [DOI] [PubMed] [Google Scholar]
  • 42. Wu Y., Yang S., Ma J., et al., “Spatiotemporal Immune Landscape of Colorectal Cancer Liver Metastasis at Single‐Cell Level,” Cancer Discovery 12, no. 1 (2022): 134–153, 10.1158/2159-8290.cd-21-0316. [DOI] [PubMed] [Google Scholar]
  • 43. Pullen R. L. Jr., “Ovarian Cancer,” Nursing 54, no. 6 (2024): 17–28, 10.1097/nsg.0000000000000002. [DOI] [PubMed] [Google Scholar]
  • 44. Thull T. and Kempton D., “Ovarian Cancer: A Review for Primary Care Providers,” Jaapa 37, no. 7 (2024): 32–36, 10.1097/01.jaa.0000000000000042. [DOI] [PubMed] [Google Scholar]
  • 45. Thommen D. S. and Schumacher T. N., “T Cell Dysfunction in Cancer,” Cancer Cell 33, no. 4 (2018): 547–562, 10.1016/j.ccell.2018.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Portale F. and Di Mitri D., “Nk Cells in Cancer: Mechanisms of Dysfunction and Therapeutic Potential,” International Journal of Molecular Sciences 24, no. 11 (2023): 9521, 10.3390/ijms24119521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Cantoni C., Falco M., Vitale M., et al., “Human Nk Cells and Cancer,” Oncoimmunology 13, no. 1 (2024): 2378520, 10.1080/2162402x.2024.2378520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Meng L. and Sun S., “Single‐Cell RNA Sequencing Reveals the Change in Cytotoxic Nk/T Cells, Epithelial Cells and Myeloid Cells of the Tumor Microenvironment of High‐Grade Serous Ovarian Carcinoma,” Discover Oncology 15, no. 1 (2024): 417, 10.1007/s12672-024-01290-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Jhunjhunwala S., Hammer C., and Delamarre L., “Antigen Presentation in Cancer: Insights into Tumour Immunogenicity and Immune Evasion,” Nature Reviews Cancer 21, no. 5 (2021): 298–312, 10.1038/s41568-021-00339-z. [DOI] [PubMed] [Google Scholar]
  • 50. He K., Meng X., Su J., Jiang S., Chu M., and Huang B., “Oleanolic Acid Inhibits the Tumor Progression by Regulating Lactobacillus Through the Cytokine‐Cytokine Receptor Interaction Pathway in 4t1‐Induced Mice Breast Cancer Model,” Heliyon 10, no. 5 (2024): e27028, 10.1016/j.heliyon.2024.e27028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Dong Z., Wu J., He L., et al., “Egcg Inhibits Tobacco Smoke‐Promoted Proliferation of Lung Cancer Cells Through Targeting Ccl5,” Phytomedicine 139 (2025): 156512, 10.1016/j.phymed.2025.156512. [DOI] [PubMed] [Google Scholar]
  • 52. Schlechter B. L. and Stebbing J., “Ccr5 and Ccl5 in Metastatic Colorectal Cancer,” Journal for Immunotherapy of Cancer 12, no. 5 (2024): e008722, 10.1136/jitc-2023-008722. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53. Han Y., Guo Z., Jiang L., et al., “Cxcl10 and Ccl5 as Feasible Biomarkers for Immunotherapy of Homologous Recombination Deficient Ovarian Cancer,” American Journal of Cancer Research 13, no. 5 (2023): 1904–1922, https://pmc.ncbi.nlm.nih.gov/articles/PMC10244116/. [PMC free article] [PubMed] [Google Scholar]
  • 54. Shan J., Xu Y., and Lun Y., “Comprehensive Analysis of the Potential Biological Significance of Ccl5 in Pan‐Cancer Prognosis and Immunotherapy,” Scientific Reports 14, no. 1 (2024): 22138, 10.1038/s41598-024-73251-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Vilgelm A. E. and Richmond A., “Chemokines Modulate Immune Surveillance in Tumorigenesis, Metastasis, and Response to Immunotherapy,” Frontiers in Immunology 10 (2019): 333, 10.3389/fimmu.2019.00333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56. Singha B., Gatla H. R., and Vancurova I., “Transcriptional Regulation of Chemokine Expression in Ovarian Cancer,” Biomolecules 5, no. 1 (2015): 223–243, 10.3390/biom5010223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. La Porta C. A., “Cxcr6: The Role of Environment in Tumor Progression. Challenges for Therapy,” Stem Cell Reviews and Reports 8, no. 4 (2012): 1282–1285, 10.1007/s12015-012-9383-6. [DOI] [PubMed] [Google Scholar]
  • 58. Chen W., Zou F., Song T., et al., “Comprehensive Analysis Reveals Xcl2 as a Cancer Prognosis and Immune Infiltration‐Related Biomarker,” Aging (Albany NY) 15, no. 21 (2023): 11891–11917, 10.18632/aging.205156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. Kamat K., Krishnan V., and Dorigo O., “Macrophage‐Derived Ccl23 Upregulates Expression of T‐Cell Exhaustion Markers in Ovarian Cancer,” British Journal of Cancer 127, no. 6 (2022): 1026–1033, 10.1038/s41416-022-01887-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60. Koch S., “The Transcription Factor Foxq1 in Cancer,” Cancer & Metastasis Reviews 44, no. 1 (2025): 22, 10.1007/s10555-025-10240-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61. Bushweller J. H., “Targeting Transcription Factors in Cancer—From Undruggable to Reality,” Nature Reviews Cancer 19, no. 11 (2019): 611–624, 10.1038/s41568-019-0196-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62. Qiu P., Li X., Gong M., et al., “Spi1 Mediates N‐Myristoyltransferase 1 to Advance Gastric Cancer Progression via Pi3k/Akt/Mtor Pathway,” Chinese Journal of Gastroenterology and Hepatology 2023 (2023): 2021515, 10.1155/2023/2021515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63. Tian J., Yang L., Wang Z., and Yan H., “Mir503hg Impeded Ovarian Cancer Progression by Interacting With Spi1 and Preventing Tmeff1 Transcription,” Aging (Albany NY) 14, no. 13 (2022): 5390–5405, 10.18632/aging.204147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64. Han X., Xia L., Wu Y., Chen X., and Wu X., “M6a‐Modified Circstx6 as a Key Regulator of Cervical Cancer Malignancy via Spi1 and Il6/Jak2/Stat3 Pathways,” Oncogene 44, no. 14 (2025): 968–982, 10.1038/s41388-024-03260-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65. Kim S. I., Lee J. W., Lee N., et al., “Lyl1 Gene Amplification Predicts Poor Survival of Patients With Uterine Corpus Endometrial Carcinoma: Analysis of the Cancer Genome Atlas Data,” BMC Cancer 18, no. 1 (2018): 494, 10.1186/s12885-018-4429-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66. Heidari Horestani M., Schindler K., and Baniahmad A., “Functional Circuits of Lyl1 Controlled by Supraphysiological Androgen in Prostate Cancer Cells to Regulate Cell Senescence,” Cell Communication and Signaling 22, no. 1 (2024): 590, 10.1186/s12964-024-01970-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67. Zhang Y., Dai L., Dou Y., Li X., Lu C., and Wu H., “Eap‐Lstm: A Bi‐Lstm Based Deep Learning Framework for Quantitatively Predicting Enhancer Activity in Drosophila and Human Cell Lines,” IEEE Journal of Biomedical and Health Informatics 30, no. 4 (2026): 3257–3266, 10.1109/jbhi.2025.3620986. [DOI] [PubMed] [Google Scholar]
  • 68. Li L., Li X., and Wu H., “A Novel Deep Learning Framework With Dynamic Tokenization for Identifying Chromatin Interactions Along With Motif Importance Investigation,” Briefings in Bioinformatics 26, no. 3 (2025): bbaf289, 10.1093/bib/bbaf289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69. Mei Z., Bi X., Li D., Xia W., Yang F., and Wu H., “Dhhnn: A Dynamic Hypergraph Hyperbolic Neural Network Based on Variational Autoencoder for Multimodal Data Integration and Node Classification,” Information Fusion 119 (2025): 103016, 10.1016/j.inffus.2025.103016. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1: Preprocessing, quality control, highly variable gene identification and principal component selection of the GSE184880 scRNA‐seq dataset. The scRNA‐seq data of GSE184880 were downloaded and subjected to strict quality control (QC), and highly variable genes were screened out (A, B). The JackStrawPlot and ElbowPlot algorithms were applied to determine the optimal principal component (PC) number. The cumulative standard deviation tended to decline gently when the PC number was set to 15 (C, D).

SYB2-20-e70069-s001.docx (407.5KB, docx)

Table S1: 64 chemokine‐related genes.

SYB2-20-e70069-s002.docx (19.2KB, docx)

Table S2: Comparison of age and pathological data between patients with ovarian malignant tumors and normal ovarian tissue, those normal samples were obtained from patients who underwent oophorectomy for benign diseases or benign ovarian cysts and normal ovarian tissue was used after excluding cystic lesions during specimen processing.

SYB2-20-e70069-s003.docx (18KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from IET Systems Biology are provided here courtesy of Wiley

RESOURCES