The bioinfomatics analysis of the M1 macrophage-related gene CXCL9 signature in cervical cancer

Abstract

Background

The expression and function of coexpression genes of M1 macrophage in cervical cancer have not been identified. And the CXCL9-expressing tumour-associated macrophage has been poorly reported in cervical cancer.

Methods

To clarify the regulatory gene network of M1 macrophage in cervical cancer, we downloaded gene expression profiles of cervical cancer patients in TCGA database to identify M1 macrophage coexpression genes. Then we constructed the protein–protein interaction networks by STRING database and performed functional enrichment analysis to investigate the biological effects of the coexpression genes. Next, we used multiple bioinformatics databases and experiments to overall investigate coexpression gene CXCL9, including western blot assay and immunohistochemistry assay, GeneMANIA, Kaplan-Meier Plotter, Xenashiny, TISCH2, ACLBI, HPA, TISIDB, GSCA and cBioPortal databases.

Results

There were 77 positive coexpression genes and 5 negative coexpression genes in M1 macrophage. The coexpression genes in M1 macrophage participated in the production and function of chemokines and chemokine receptors. Especially, CXCL9 was positively correlated with M1 macrophage infiltration levels in cervical cancer. CXCL9 expression would significantly decrease and high CXCL9 levels were linked to good prognosis in the cervical cancer tumour patients, it manifestly expressed in blood immune cells, and was positively related to immune checkpoints. CXCL9 amplification was the most common type of mutation. The CXCL9 gene interaction network could regulate immune-related signalling pathways, and CXCL9 amplification was the most common mutation type in cervical cancer. Meanwhile, CXCL9 may had clinical significance for the drug response in cervical cancer, possibly mediating resistance to chemotherapy and targeted drug therapy.

Conclusion

Our findings may provide new insight into the M1 macrophage coexpression gene network and molecular mechanisms in cervical cancer, and indicated that M1 macrophage association gene CXCL9 may serve as a good prognostic gene and a potential therapeutic target for cervical cancer therapies.

PLAIN LANGUAGE SUMMARY

Cervical cancer is a common gynaecological malignancy, investigating the precise gene expression regulation of M1 macrophage is crucial for understanding the changes in the immune microenvironment of cervical cancer. In our study, a total of 82 coexpression genes with M1 macrophages were identified, and these genes were involved in the production and biological processes of chemokines and chemokine receptors. Especially, the chemokine CXCL9 was positively correlated with M1 macrophage infiltration levels in cervical cancer. CXCL9 as a protective factor, it manifestly expressed in blood immune cells, and was positively related to immune checkpoints. CXCL9 amplification was the most common type of mutation. And CXCL9 expression could have an effect on the sensitivity of some chemicals or targeted drugs against cervical cancer. These findings may provide new insight into the M1 macrophage coexpression gene network and molecular mechanisms, and shed light on the role of CXCL9 in cervical cancer.

Keywords:

• M1 macrophage
• coexpression gene
• chemokine
• CXCL9
• cervical cancer

Introduction

Cervical cancer ranks fourth globally in the incidence of malignancies in women, with more than 300,000 deaths around the world each year (Ward et al. 2020). The advanced and metastasis cervical cancer patients have a poor 5-year survival rate of 16.5% (Li et al. 2016). In recent years, tumour immune microenvironment (TME), especially immune cells, has attracted growing attention and become a research hotspot in cervical cancer (Zhang et al. 2021). Accurate immunotherapy and biomarkers also have emerged as clinical research priorities (Volkova et al. 2021). These works suggest the importance and necessity of exploring the precise expression genes of immune cells in cervical cancer.

Activated macrophages, as highly plastic cells, which can be divided into two polarised states: M1 phenotype and M2 phenotype (Yunna et al. 2020). At early stages, the ratio is more favourable for M1 macrophages (Najafi et al. 2019). As the pro-inflammatory M1 phenotype, M1 macrophage features antigen-presenting ability and promotes inflammatory factor secretion to play the anticancer effects (Lewis & Pollard 2006; Zhang et al. 2022). Nowadays, the mechanism of driving macrophage plasticity phenotype through the activation of many molecular pathways in cervical cancer remain unclear. Therefore, a better understanding of the genetic regulatory relationships responsible for the macrophage in the microenvironment of cervical cancer is important.

Chemokines are a family of small cytokines, dysregulation in the chemokine system has been linked to cancer, chemokine can both exert pro- and antitumorigenic effects (Märkl et al. 2022). The recruitment of immune cells into the TME depends heavily on different chemokines and their receptors (Propper & Balkwill 2022). Furthermore, in pathological circumstances, the accumulation and phenotype of the macrophages are regulated by chemokines secreted in the TME (Kitamura & Pollard 2015). It has been reported that CXCL9 inhibits tumour growth and drives anti-PD-L1 therapy in ovarian cancer (Seitz et al. 2022). And CXCL9-expressing tumour-associated macrophage plays a new role in the fight against cancer (Marcovecchio et al. 2021). Nevertheless, due to the lack of research on chemokine CXCL9 in cervical cancer, we performed a comprehensive analysis for chemokine CXCL9 and better understood its immune regulatory effects in cervical cancer.

Therefore, in our study, we firstly performed a comprehensive analysis to identify M1 macrophages coexpression genes by using RNA sequencing data from the TCGA database, in order to elucidate the functions of M1 macrophages coexpression genes in cervical cancer. Then, we sought to perform a comprehensive analysis of CXCL9 in cervical cancer using multiple databases, western blot assay and immunohistochemistry assay. These data were suggestive of the functions of CXCL9 and provide new insights into the pathogenesis of CXCL9 in cervical cancer.

Materials and methods

Identification of M1 macrophage coexpression genes

The RNA sequencing data of cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) was obtained from the cancer genome atlas (TCGA) data portal (https://portal.gdc.cancer.gov/). The CIBERSORT algorithm was used to calculate immune cell infiltration levels for each sample. And the coexpression genes of M1 macrophage were identified by analysing CIBERSORT results using limma, tidyverse, ggplot2, ggpubr, ggExtra R packages in R software (version:×64 4.1.2).

Construction of protein–protein and gene-gene interaction networks

STRING (https://cn.string-db.org/) database is a comprehensive gene expression profiling and network visual analytics websites (Szklarczyk et al. 2019). The expression of coexpression genes were uploaded to the database to construct the protein–protein interaction network. GeneMANIA database helps to predict the function of favourite genes and gene sets (https://genemania.org/) (Warde-Farley et al. 2010), we constructed the gene-gene interaction network of CXCL9 to make clear the gene interaction relationships.

Functional enrichment analysis

Gene Ontology (GO) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) functional enrichment analysis of coexpression genes were executed by using the R packages including clusterProfiler, enrichplot, ggplot2 and org.Hs.eg.db. Biological process (BP), cellular component (CC), and molecular function (MF) were the three categories included in the GO analysis.

Prognostic values and infiltration levels of CXCL9 in cervical cancer

Then Kaplan-Meier Plotter database and Xenashiny database were researched to analyse the clinical prognosis of CXCL9 in cervical cancer (http://kmplot.com/analysis/, https://shiny.hiplot.cn/ucsc-xena-shiny/) (Lánczky & Győrffy 2021; Wang et al. 2022). The prognostic value of CXCL9 in pancancer and the immune cell infiltration level at single-cell level in the TME was explored by analysing TISCH2 database (http://tisch.comp-genomics.org/home/) (Sun et al. 2021). The relationship between CXCL9 expression with immune-infiltrating cells was researched by the Assistant for Clinical Bioinformatic (ACLBI) database (https://www.aclbi.com/static/index.html#/). The expression levels of CXCL9 in various blood immune cells in cervical cancer were represented by using the Human Protein Atlas (HPA) database (https://www.proteinatlas.org/)(Uhlen et al. 2019).

Western blot assay and immunohistochemistry

All enrolled patients were reviewed and approved by the hospital’s ethics committee, and they all provided informed consent by signing a document (Ethics Number: WDRY2022-K083, 2022-06-30). As previous study (Liao et al. 2023), western blot and immunohistochemistry were performed to analyse the protein level of CXCL9 (anti-CXCL9, western blot diluted at 1:500, immunohistochemistry diluted at 1:1000, ABclonal) in our study. GAPDH was used to normalise protein expression levels. The western blot results were performed with ChemiDocTM imaging system (Bio-Rad Laboratories, Inc., USA). And the IHC tissue sections were examined with microscopic (Olympus).

The relationship of CXCL9 with markers of macrophages and immune checkpoints

TISIDB (http://cis.hku.hk/TISIDB/) is a web portal for tumour and immune system interaction (Ru et al. 2019), which was used in the present work to analyse the association between CXCL9 with markers of macrophages and immune checkpoints.

Gene mutation status analysis

The cBioPortal for Cancer Genomics has large-scale cancer genomics data sets (https://www.cbioportal.org/) (Cerami et al. 2012). We explored the genomic alterations of CXCL9 in cervical cancer using cBioPortal tool. CXCL9 variants were represented using OncoPrint and Cancer Types Summary.

Drug sensitivity analysis

Gene Set Cancer Analysis (GSCA, http://bioinfo.life.hust.edu.cn/GSCA/#/) combines clinical information and small molecular drugs, which can be used to screen candidate biomarkers and valuable drugs (Liu et al. 2022). We explored the relevance between CXCL9 with drug sensitivity based on GDSC in the GSCA database.

Statistic analysis

The chemokines and immune cell expression were analysed using Spearman’s correlation statistics. The data were considered statistically significant with the P value < 0.05.

Results

M1 macrophage coexpression genes and protein-protein interaction(PPI) network in cervical cancer

In order to elucidate the coexpression genes of M1 macrophage in cervical cancer, the CIBERSORT results from the TCGA database were identified. The results revealed that there were 77 positive coexpression genes and 5 negative coexpression genes in M1 macrophage (Supplementary Material 1). The gene coexpression networks were presented in Figure 1A. According to the coexpression networks, we established a correlation graph for the top 30 significantly associated genes, and the results indicated that these genes were positively correlated with M1 macrophage (Figure 1B). Furthermore, the coexpression genes were uploaded to construct the protein-protein interaction (PPI) networks from STRING database (Figure 1C). Based on the PPI networks, we were able to better visualise the importance of the relationships between proteins. These results showed a detailed description of the M1 macrophage coexpression genes and PPI networks in cervical cancer.

Figure 1. The coexpression gene networks and protein-protein interaction networks correlated with M1 macrophage in cervical cancer. (A) The positive and negative coexpression genes associated with M1 macrophage. Red indicates positive correlations while blue indicates negative correlations. (B) The correlation graph of the top 30 genes positively associated with M1 macrophage. (C) PPI networks of M1 macrophage coexpressed genes constructed based on the STRING database.

The GO and KEGG functional enrichment analysis of coexpression genes in M1 macrophage

To further explore the functions of coexpression genes, functional enrichment analysis was performed. The Gene Ontology (GO) functional analysis results in M1 macrophage coexpression genes were markedly enriched in the defense response to virus, cytokine-mediated signalling pathway, response to interferon-gamma, cytokine activity, cytokine receptor binding, chemokine receptor binding, chemokine activity, and CXCR chemokine receptor binding (Figure 2A–B). And Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway enrichment analysis were markedly enriched in cytokine-mediated signalling pathway, chemokine-mediated signalling pathway, regulation of innate immune response and macrophage activation involved in immune response (Figure 2C–D). All these results suggested that coexpression genes in M1 macrophage participated in the production and function of chemokines and chemokine receptors.

Figure 2. The GO and KEGG functional enrichment analysis of coexpression genes in M1 macrophage. (A–B) The circos plot and bubble plot of GO functional analysis in M1 macrophage. (C-D) The bar plot and circos plot of KEGG pathway enrichment analysis in M1 macrophage.

Expression levels and prognostic values of CXCL9 in cervical cancer

Chemokines and chemokine receptors have exerted a conflicting role in tumour development (Gowhari et al. 2021). Considering the functional enrichment analysis results above, we further assessed the relevance between M1 macrophage infiltration levels with chemokines and chemokine receptors in cervical cancer. The findings indicated that M1 macrophage infiltration levels were significantly and positively correlated with the abundance of CXCL9 (R = 0.65, p < 2.2e-16) (Figure 3A). In order to better understand the underlying mechanisms of the chemotactic responses in M1 macrophage, the expression levels and prognostic values of chemokines and chemokine receptors were investigated. We detected the expression levels of CXCL9 in normal cervix patients and cervical cancer patients. CXCL9 expression would significantly decrease in the cervical cancer tumour patients (Figure 3B). Furthermore, to more comprehensive prognostic information, we explored the TISCH, GEPIA and Xenashiny databases, we found that high CXCL9 levels were linked to good prognosis for cervical cancer patients (Figure 3C), and the Overall survival (OS) (log-rank P = 0.021 and 0.048) (Figure 3D and 3E) for patients with high CXCL9 levels were evidently much longer. Therefore, CXCL9 may serve as a good prognostic gene in cervical cancer.

Figure 3. Expression levels and prognostic values of CXCL9 in cervical cancer. (A) M1 macrophages infiltration levels were positively associated with CXCL9. (B) CXCL9 was upregulated in the normal cervix group (NC) than in cervical cancer (CC) group by western blot assay and immunohistochemistry. (C) CXCL9 reduced the risk of cervical cancer patients in the TISCH database. (D-E) Overall survival of patients in the Kaplan–Meier database and Xenashiny database.

Infiltration levels and association with some immune checkpoints of CXCL9 in cervical cancer

Then we evaluated CXCL9 expression in various blood immune cells in HPA database, we found that CXCL9 expression was enhanced in naive CD8 T-cell activated and naive CD4 T-cell activated (Supplementary Figure S1A). Afterwards, to determine the connection between CXCL9 and various immune cell fractions, the heatmap in Supplementary Figure S1B was examined. We indeed confirmed that CXCL9 was significantly positively correlated with M1 macrophage. Similarly, it is worth noting that there was stronger association between CXCL9 expression and macrophages in CESC at the single-cell level (Supplementary Figure S1C). Additionally, in the TISIDB data portal, we found that CXCL9 was associated with some immune checkpoints. The expression of CXCL9 was correlated with CTLA-4 (r = 0.767, p < 2.2e-16), PD-L1 (CD274) (r = 0.596, p < 2.2e-16), and PD-1 (PDCD1) (r = 0.749, p < 2.2e-16) (Supplementary Figure S1D, E–G). These data suggested that alterations in CXCL9 expression may play a specific role in immunetherapy of cervical cancer. Based on the above analysis, we explicitly illustrated that higher expression of CXCL9 represented a better prognosis and was positively correlated with M1 macrophage infiltration levels in cervical cancer.

The gene–gene interaction network of CXCL9

Based on the above studies, it has been already known that CXCL9 may have impact on M1 macrophage. Molecular interaction networks and associated signalling pathways between genes are the fundamental to many biological processes. We next constructed the gene–gene interaction network of CXCL9 to define their functions by using GeneMANIA database. We observed a strong association between CXCL9 with other chemokines and their receptors genes, including CXCL1, CXCL3, CXCL5, CXCL6, CXCL8, CXCL10, CXCL11, CXCL13, CCL2, CCL5, CCL6, CCL11, CCL18, CCL19, CCL21 and CCL28. The gene interaction network could regulate immune-related signalling pathways, such as Chemokine_CXC_CS, Interleukin_8-like_sf, IL-23-mediated signalling events and TypeII interferon signalling (Supplementary Figure S2).

Mutation status and effect of CXCL9 expression level on drug sensitivity in cervical cancer

Analysis of genomic alterations of CXCL9 in cervical cancer was investigated in the cBioPortal database. It was found that CXCL9 genetic total mutation rate occurred 2% across 605 cervical cancer patients (Supplementary Figure S3A). Overall, there were gene altered in 2.02% of 297 cases in Cervical Squamous Cell Carcinoma (TCGA, PanCancer Atlas) and gene altered in 2.01% of 298 cases in Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma (TCGA, Firehose Legacy), including 2 cases of mutation, 1 case of structural variant, 6 cases of amplification, and 3 cases of deep deletion (Supplementary Figure S3B). Above all, the results revealed that CXCL9 amplification was the most common mutation type in cervical cancer. Finally, the GSCA database was utilised to illustrate the effect of CXCL9 expression on the sensitivity of the top 30 chemical or targeted drugs against cervical cancer. As shown in Supplementary Figure S3C and Supplementary Material 2, CXCL9 expression was positively associated with 17-AAG, but negatively associated with the sensitivity of 29 drugs, including 5-Fluorouracil, Methotrexate, PHA-793887 and PIK-93 (p < 0.05). Interestingly, M1 macrophage biomarkers CD80 and CD86 showed a synergisticly pharmacological effect with CXCL9. These results would have clinical significance for the drug response in cervical cancer, possibly mediating resistance to chemotherapy and targeted drug therapy.

Discussion

Macrophages are highly plastic cells that serve multiple functions in the developing of cancers (Anderson et al. 2021). Current studies on TAMs suggest that M1 macrophages are regarded as the anticancer effects, however, an increasing number of evidence has manifested that the polarisation of macrophages play crucial role in malignant progression of tumors(Qian & Pollard 2010). Therefore, the gene coexpression networks and regulatory mechanisms of M1 macrophages in cervical cancer should be further examined in order to obtain more precise and efficient anti-tumour effects. In our study, we downloaded the RNA-seq data of CESC from the TCGA database and constructed the coexpression networks of M1 macrophages. The results showed that there were 77 positive coexpression genes and 5 negative coexpression genes with M1 macrophages. Gene functional enrichment analysis manifested that coexpression genes were enriched in the cytokine-mediated signalling pathways, cytokine activity, cytokine receptor binding, chemokine receptor binding, chemokine activity, and macrophage activation involved in immune responses. Our study is critical to elucidate the underlying functional mechanisms of coexpression genes in M1 macrophage.

Chemokines have central roles in both anti-tumour and pro-tumour immune responses, the balance of chemokines may lead to the response and resistance among the immunological therapies (Ozga et al. 2021). The immune effect of macrophages requires specific chemokines to perform their functions, and these chemicals may represent valuable prognostic biomarkers of this process. Recent research showed that Macrophage-Derived CXCL9 may be a new player in the fight against cancer, which can strengthen the potential of patient responses after immune checkpoint therapy (House et al. 2020; Marcovecchio et al. 2021). However, further exploration of macrophage-derived CXCL9 in the tumour microenvironment of cervical cancer is needed. Therefore, we found as the macrophage coexpression gene, CXCL9 expression was lower in cervical cancer and represents a better prognosis for patients. It was also significantly positively correlated with M1 macrophage and was expressed in various blood immune cells. As the core of gene–gene interaction network, CXCL9 had strong associations with other chemokines and their receptors genes. Our research also demonstrated that the CXCL9 expression levels were positively correlated with immune checkpoints including PD-1, PD-L1 and CTLA4, which may suggest that alterations in CXCL9 expression may play a specific role in cervical cancer immunetherapy.

Next, we investigated the genetic mutation status of CXCL9 in cervical cancer. Our results indicated that CXCL9 amplification was the most common type of mutation in cervical cancer. Studies have verified that oncogene amplification plays a central role in tumorigenesis(Verhaak et al. 2019). Recently, the combination of chemical or targeted drugs with other chemotherapy drugs has been proved to improve the treatment of malignancies. Therefore, we predicted the relationship between CXCL9 and drug sensitivity. Our results suggested that 5-Fluorouracil, Methotrexate, PHA-793887 and PIK-93 were strongly linked with CXCL9 mRNA expression. These results manifested that the drug sensitivity to CXCL9 expression may suggest new strategies for clinical treatment of cervical cancer patients.

There are some limitations to our study. Further identification of the M1 macrophage coexpression genes is required. Follow-up experiments and prospective studies are needed to confirm the function of these genes.

Conclusion

In summary, our study identified multiple coexpression genes of M1 macrophage in cervical cancer. The M1 macrophage association gene CXCL9 may serve as a good prognostic gene and a potential therapeutic target for cervical cancer therapies.

Ethical approval

This study was approved by the Ethics Committee of Renmin Hospital of Wuhan University (Ethics Number: WDRY2022-K083, 2022-06-30). All participants were provided informed consent and signed a document.

Authors’ contributions

Wenxin Liao designed, analysed the main manuscript text. Tingting Liu and Yang Li wrote the manuscript text. Hua Liang prepared figures. Juexiao Deng did the experiments. Fujin Shen contributed to the study conception and design. All authors read and approved the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Public datasets were obtained from TCGA (https://portal.gdc.cancer.gov/), STRING (https://cn.string-db.org/), ACLBI (https://www.aclbi.com/static/index.html#/), HPA (https://www.proteinatlas.org/), TISCH2 (http://tisch.comp-genomics.org/home/), Xenashiny (https://shiny.hiplot.cn/ucsc-xena-shiny/), GeneMANIA

(https://genemania.org/),TISIDB(http://cis.hku.hk/TISIDB/), GSCA (http://bioinfo.life.hust.edu.cn/GSCA/#/), cBioPortal (https://www.cbioportal.org/).

Additional information

Funding

This work was supported by the Natural Science Foundation of Hubei Province of China (2021CFB430 and 2022CFB196).