Abstract Background The tumor microenvironment (TME) has received an increasing amount of attention. CXC chemokines can regulate immune cell transport and tumor cell activity to exert anti-tumor immunity. However, studies on the expression and prognosis of CXC chemokines in cervical cancer (CC) are more limited. Methods The study investigated the role of CXC chemokines in TME of CC by using public databases. Moreover, quantitative real-time PCR (qRT-PCR) and immunohistochemistry (IHC) of CXC chemokines were performed to further verify. Results The transcriptional levels of CXCL1/3/5/6/8/9/10/11/13/16/17 in CC tissues were significantly elevated while the transcriptional levels of CXCL12/14 were significantly reduced. We reached a consistent conclusion that the expression of CXCL9/10/11/13 was verified by quantitative real-time PCR and immunohistochemistry. Moreover, CC patients with low transcriptional levels of CXCL1/2/3/4/5/8 were significantly associated with longer overall survival (OS). The CCL family was related to CXC chemokines neighboring alteration. RELA, NFKB1, LCK and PAK2 were the key transcription factors and kinase targets of CXC chemokines, respectively. We also found there were significant correlations between the expression of CXCL9/10/11 and the infiltration of immune cells (CD8+ T cell, CD4+ T cell, neutrophils and dendritic cells). Conclusions In brief, we conducted a comprehensive analysis of CXC chemokines via clinical data and some online public databases. Our results may provide a new idea for the selection of immunotherapeutic targets and prognostic biomarkers for cervical cancer. Supplementary Information The online version contains supplementary material available at 10.1186/s12935-021-02101-9. Keywords: Cervical cancer, CXC chemokine, Quantitative real-time PCR, Immunohistochemistry, Prognosis, Tumor microenvironment Background Cervical cancer accounts for approximately 12% of female cancers and is one of the major causes of death among women worldwide [[39]1, [40]2]. Although the early detection and treatment of CC have improved, there are still about 604,127 new cases and 341,831 deaths in 2020, according to data from the World Health Organization [[41]3]. The actual situation shows that the majority of patients with cervical cancer are in the advanced stage of the disease, with limited access to appropriate treatment [[42]4]. As a result, the mortality rate is still high, and the median overall survival rate of advanced cervical cancer is only 16.8 months [[43]5]. Nowadays, an increasing number of studies have explored the kinase and immune checkpoint inhibitors in cancer [[44]6]. Moreover, epigenetics, the new regulation of specific genes, has also shown certain importance in the progression of cervical cancer [[45]7]. This provides a new idea for identifying the therapeutic targets and prognostic biomarkers of cervical cancer. Chemokines are a type of secreted proteins with small molecular weights. Their role in autoimmune diseases, chronic inflammations and tumors has been continuously revealed [[46]8], mainly mediating the migration of immune cells and the development of lymphoid tissues [[47]9]. In recent years, the role of chemokines in the tumor microenvironment has been continuously reported by researchers. As a major subfamily of the chemokines family, the function and mechanism of CXC chemokines have been discovered in tumors [[48]10]. CXC chemokines’ altered expression in cancers dictates immune cell recruitment, angiogenesis, tumorigenesis, cancer cell proliferation and metastasis [[49]11, [50]12]. Previous studies have shown that there is a correlation between CXC chemokines, tumor microenvironment and tumor immunotherapy [[51]13], which has been confirmed in some cancers [[52]14–[53]16]. This suggests that CXC chemokines may be potential therapeutic targets and prognostic biomarkers of cancer, by modulating tumor progression and immunotherapy efficacy. However, the function of the CXC chemokine family in CC has not been comprehensively described. CXC chemokine is an important component of TME. Although previous studies have confirmed the expression and role of some members of the CXC chemokine family in cervical cancer, there is still a lack of comprehensive and systematic research. Therefore, it is worth exploring CXC chemokines as therapeutic targets and prognostic markers of cervical cancer. In this study, public databases were used to investigate the mRNA expression, prognosis, and related targets or kinase pathways of the CXC chemokine family in CC. Immunohistochemistry and qRT-PCR further verified the conclusion. Taken together, this study complements the function of CXC chemokines in cervical cancer, suggesting that certain CXC chemokines can be used as potential therapeutic targets and prognostic biomarkers for CC. Materials and methods Study population In this study, clinical data and pathological specimens were collected retrospectively, and the patients’ informed consent was obtained before the pathological specimens were collected. Cervical cancer and para-carcinoma tissue samples for qRT-PCR and IHC analysis were taken from patients who underwent surgery from January 1, 2017 to December 31, 2018. We collected frozen tissues of 60 patients to analyze the mRNA levels of CXC chemokines by qRT-PCR and analyzed the protein expression levels of CXC chemokines by IHC staining in the paraffin tissues of 60 patients. This research work had been approved by the Academic Committee of The Third Clinical Medical College of Xinjiang Medical University (affiliated Tumor Hospital) and was carried out under the rules put forward in the Declaration of Helsinki. This study had the relevant informed consent exemption certificate issued by the Academic Committee. As for the public databases, neither ethics committee approval nor patient informed consent was needed for analyzing data. Quantitative real-time PCR Total RNA was isolated from tumor tissues and adjacent tissues using Trizol according to the manufacturer’s instructions. The extracted RNA was converted into cDNA with 5× primescript buffer, prime script RT enzyme mix I, oligo-dT primer and random 6 mers. The qRT-PCR was performed in the BioRad CFX96 Real-Time PCR Detection System machine in the presence of GAPDH, CXCL9, CXCL10, CXCL11, CXCL12 and CXCL13. We verified CXCL9/10/11/13 because these factors were expressed on the intersection of the three common databases including ONCOMINE, GEPIA and UALCAN. The detection of CXCL12 was because it was the only one that tended towards low expression in two databases. The transcription level of target genes was measured and normalized to GAPDH expression. The following primer sequences were used: GAPDH, 5′-GAAGGTGAAGGTCGGAGTC-3′ (forward) and 5′-GAAGATGGTGATGGGATTTC-3′ (reverse); CXCL9, 5′-TGAGAAAGGGTCGCTGTTCC-3′ (forward) and 5′-GGGCTTGGGGCAAATTGTTT-3′ (reverse); CXCL10, 5′-TGCCATTCTGATTTGCTGCC-3′ (forward) and 5′-TGCAGGTACAGCGTACAGTT-3′ (reverse); CXCL11, 5′-CCCTGGGGTAAAAGCAGTGA-3′ (forward) and 5′-TAAGCCTTGCTTGCTTCGAT-3′ (reverse); CXCL12, 5′-AGATGCCCATGCCGATTCTT-3′ (forward) and 5′-AGGGCACAGTTTGGAGTGTT-3′ (reverse); CXCL13, 5′-CGACATCTCTGCTTCTCATGC-3′ (forward) and 5′-ACTGAGCTCTCTTGGACACAT-3′ (reverse). Immunohistochemistry Formalin-fixed paraffin-embedded surgical specimens were used for immunohistochemical study. The sections were dried at 60 °C for 2 h, subsequently were dewaxed in xylene and graded alcohols, were hydrated and washed in phosphate-buffered saline. After antigen repair was treated in a microwave oven (15 min in citrate buffer, pH 6.0), the endogenous peroxidase was inhibited with 3% H[2]O[2] for 30 min, then the sections were incubated with 10% normal goat serum for 40 min. Primary antibodies composed of rabbit anti-CXCL9 antibody (bs-2551R [Bioss], 1:100), rabbit anti-CXCL10 antibody (bs-1502R [Bioss], 1:150), rabbit anti-CXCL11 antibody (DF9917 [Affinity], 1:150) and rabbit anti-CXCL13 antibody (bs-2553R [Bioss], 1:100) were applied overnight in a moist room at 4°C. Then the tissues were incubated with secondary antibody (37 °C 50 min), stained with diaminobenzidine, and counterstained with hematoxylin. Positive staining was evaluated using computer-aided image analysis and Image J software. The average CXC chemokines infiltration was determined from three randomized fields by two independent pathologists who were unaware of the patients’ pathological and clinical status. Transcription-related databases of CXC chemokines in patients of cervical cancer We used the public databases ONCOMINE ([54]http://www.oncomine.org) [[55]17], GEPIA ([56]http://gepia.cancer-pku.cn/index.html) [[57]18] and UALCAN ([58]http://ualcan.path.uab.edu) [[59]19] that could provide cancer RNA sequence data and clinical data to analyze the differential expression of CXC chemokines in cervical cancer and adjacent cancer or normal tissues by using Student’s t-test. The cut-off of the p-value was 0.05. In ONCOMINE, the fold change was 2.0, and the gene rank was in the top 10%. In brief, we entered the target genes in the input box of the database, and then searched for them. In addition, we conducted a prognostic study of CXC chemokines in cervical cancer by Kaplan–Meier curve in the GEPIA database. cBioPortal cBioPortal ([60]http://www.cbioportal.org) is an online open-access website, which involves the exploration, visualization, and analysis of multidimensional cancer genomics data [[61]20]. Genetic alterations of CXC chemokines were obtained from cBioPortal based on The Cancer Genome Atlas (TCGA) database. In this study, 293 cervical squamous cell carcinoma samples were analyzed (TCGA, PanCancer Atlas). The z-score of mRNA expression (log RNA Seq V2 RSEM) was obtained using the threshold of ± 2.0. CXC chemokines-related networks We studied the gene-related networks and protein–protein interaction (PPI) of CXC chemokines by GeneMANIA ([62]http://www.genemania.org) [[63]21], a website which could provide information on the protein and genetic interactions, pathways and co-expression of submitted genes, and the STRING database ([64]https://string-db.org/) [[65]22], respectively. In our research, we entered the searched species and gene names in the input box of databases. Moreover, on the right side of the GeneMANIA website, we also set up bioinformatics methods such as co-expression, physical interaction, gene enrichment analysis, predictive interaction and pathway, etc. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis DAVID 6.8 ([66]https://david.ncifcrf.gov/home.jsp), a website that can clarify the biological functions of genes [[67]23], was used to perform GO and KEGG pathway enrichment analysis of CXC chemokines and their neighboring 50 genes. Analysis of GO function mainly included biological process (BP), cell component (CC), molecular function (MF). Then, the “ggplot2” package of R software (4.0.2) was used for visualization. P-value < 0.05 was considered to be statistically different. Target analysis of kinases and transcription factors The TRRUST ([68]https://www.grnpedia.org/trrust/) database can provide transcription factor (TF)–target regulatory relationship [[69]24]. The module of LinkInterpreter in the LinkedOmics ([70]http://www.linkedomics.org/) [[71]25] was used to obtain biological insights on the enrichment of kinase targets of CXC chemokines. In the LinkedOmics database, Gene Set Enrichment Analysis (GSEA) was investigated with a minimum number of genes (size) of 3 and a simulation of 500. The results were analyzed by the Spearman correlation test. The p-value cut-off was 0.05. TIMER database TIMER ([72]https://cistrome.shinyapps.io/timer/) could provide a systematic assessment of the infiltration of different immune cells and their clinical effects [[73]26]. We conducted the module of correlation in the TIMER database to evaluate the interrelation between immune cell infiltration and CXC chemokines level (or CXCRs) by the purity-corrected partial Spearman method in the tumor microenvironment. Moreover, the correlations between cell markers of CD8+ T cells (CD8A and CD8B), natural killer (NK) cells (KLRK1, KIR2DL4, KIR3DL2, NCR1, and NCR3), T helper 1 (Th1) cells (TBX21, STAT1) and CXCL9–11/CXCR3 were explored. Statistical analysis The data of clinical patients were presented as mean ± SD, the Chi-square test and Wilcoxon rank-sum test was used to compare the data between the tumor group and the para-cancerous group (SPSS 26.0). The qRT-PCR and immunohistochemical data of 60 patients were analyzed using Student’s t-test and Mann–Whitney U test (Graphpad Prism 8.0). Differences with p values < 0.05 were considered statistically significant. Results The clinicopathological characteristics of the patients The clinicopathological features of 60 patients were studied. The results are shown in Table [74]1. There were no differences in ethnicity (p = 0.566), tumor size (p = 0.756), differentiation (p = 1.000), or FIGO stage (p = 0.378) between the tumor tissue group and the adjacent tissue group, except for first diagnostic age (p = 0.000). Table 1. Clinical characteristics of patients with cervical cancer Clinical or pathologic feature N (%) or mean (s.d.) p value Tumor tissue Adjacent tissue Total number of patients enrolled 40 20 Age at first diagnosis (years) 49.90 ± 10.20 48.30 ± 8.68 0.000 Ethnicity 0.566 Han 25 (62.5) 14 (70.0) Others 15 (37.5) 6 (30.0) Tumor size (cm) 0.756 ≤ 4 30 (75.0) 16 (80.0) > 4 10 (25.0) 4 (20.0) Differentiation 1.000 Poor 10 (25.0) 5 (25.0) Middle/high 30 (75.0) 15 (75.0) FIGO stage 0.378 IA 4 (10.0) 0 (0.0) IB 15 (37.5) 8 (40.0) IIA 20 (50.0) 11 (55.0) IIB 1 (2.5) 1 (5.0) [75]Open in a new tab Abnormal expression of CXC chemokines in CC patients To investigate the transcription level of CXC chemokines between tumor and adjacent or normal tissues in CC, we performed an analysis using the ONCOMINE, GEPIA and UALCAN database. Sixteen CXC chemokines were retrieved using the ONCOMINE databases. The results are presented in Fig. [76]1 and Table [77]2. The transcriptional level of CXCL1/3/5/6/8/9/10/11/13/16 in cervical cancer tissues was significantly elevated, while the transcriptional level of CXCL12/14 was significantly lower than that of adjacent cancer tissue. In Zhai cervix statistics [[78]27], CXCL1/3/5/6/8/13 were overexpressed in cervical tumor tissues versus tumor-adjacent tissues. In Scotto Cervix 2 Statistics [[79]28], CXCL1/8/9/10/11/13 were elevated in CC tissues relative to the adjacent tissues of the tumor. Similarly, CXCL8/9/10/11 were overexpressed in CC tissues instead of para-carcinoma tissue in Biewenga Cervix Statistics [[80]29]. In the GEPIA database, the results indicated that the expressional levels of CXCL1/8/9/10/11/13/16/17 were increased in tumor tissues rather than normal tissues, while CXCL12 was reduced (Fig. [81]2a). In the UALCAN database, as expected, the transcriptional levels of CXCL6 (p = 1.18e−4), CXCL9 (p = 1.63e−12), CXCL10 (p = 7.77e−16), CXCL11 (p = 8.36e−13), CXCL13 (p = 2.81e−7) and CXCL17 (p = 1.74e−3) in cervical tissues were significantly elevated (Fig. [82]2b–g). Since the normal control of the study in the ONCOMINE database came from para-carcinoma tissues rather than normal tissues, it was rational to acquire diverse results from the three databases. In addition, the sample size was not sufficient to capture variability as there were only three normal cervical patients in the UALCAN database. This part of the data has yet to be further confirmed in clinical practice. Taken together, these data suggest that these CXC chemokines play a significant role in the tumorigenesis and progression of cervical cancer. Fig. 1. [83]Fig. 1 [84]Open in a new tab mRNA levels of CXC chemokines in diverse types of cancers (ONCOMINE). It shows the numbers of datasets with statistically significant mRNA over-expression (red) or down-regulated expression (blue) of CXC chemokines by Students’ t-test. The parameters are as follows, p-value: 0.01, fold change: 2.0, gene rank: 10% Table 2. The significant changes of CXC chemokines expression in transcription level between different types of CC TLR Type Fold change p-value t-test References