Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Ovarian cancer is the common tumor in female, the prognostic of which is influenced by a series of factors. In this study, 4 genes relevant to pathological grade in ovarian cancer were screened out by the construction of weighted gene co-expression network analysis.

METHODS:

GSE9891 with 298 ovarian cancer cases had been used to construct co-expression networks. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses was used to analyze the possible mechanism of genes involved in the malignant process of ovarian cancer. Hub genes were validated in other independent datasets, such as GSE63885, GSE26193 and GSE30161. Survival analysis based on the hub genes was performed by website of Kaplan Meier-plotter.

RESULTS:

The result based on weighted gene co-expression network analysis indicated that turquoise module has the highest association with pathological grade. Gene Ontology enrichment analysis revealed that the genes in turquoise module main enrichment in inflammatory response and immune response. Kyoto Encyclopedia of Genes and Genomes enrichment analysis revealed that the genes in turquoise module main enrichment in cytokine-cytokine receptor interaction and chemokine signaling pathway. In turquoise module, a total of 4 hub genes (MS4A4A, CD163, CPR65, MS4A6A) were identified. Then, 4 hub genes were effectively verified in the test datasets (GSE63885, GSE26193 and GSE30161) and tissue samples from Shengjing Hospital of China Medical University. Survival analysis indicated that the 4 hub genes were associated with poor progression-free survival of ovarian cancer.

CONCLUSIONS:

In conclusion, 4 hub genes (MS4A4A, CD163, CPR65, MS4A6A) were verified associated with pathological grade of ovarian cancer. Moreover, MS4A4A, CD163, MS4A6A may serve as a surface marker for M2 macrophages. Targeting the 4 hub genes may can improve the prognosis of ovarian cancer.

Keywords

Ovarian cancer WGCNA pathological grade M2 macrophages prognosis

1. Introduction

Epithelial ovarian cancer (EOC) is one of the most common and lethal malignant tumors, and the 5-year survival rate is less than 30% [1]. Tumor grade is closely related to the prognosis of patients. High grade serous ovarian carcinoma (HGSOC) as the most common histological subtype of EOC has a worse prognosis than low-grade ovarian cancer [2]. Currently, the standard treatment for HGSOC is tumor cell destruction combined with postoperative taxol/platinum chemotherapy. However, the vast majority of patients eventually relapse for a variety of reasons [3]. Therefore, finding predictive markers related to ovarian cancer grade is very significant, which can not only predict the prognosis of ovarian cancer, but also serve as a therapeutic target for OC.

In recent years, bioinformatic analysis is commonly used in various diseases, especially tumors. Compared with other methods, bioinformatic analysis is more accurate, comprehensive and economical. Weighted gene co-expression network analysis (WGCNA) is a widely used method in systems biology, especially for studying biological networks based on paired correlations between variables. Based on the correlation between gene expression data and clinical phenotypes, the WGCNA network can be constructed to analyze the change of biological process integrally and identify hundreds of pathogenic genes and therapeutic targets simultaneously. WGCNA network replaced the hard threshold with soft threshold, and introduced the concept of scale-free distribution network. Through using the correlation coefficient of network weighted expression matrix, gene pairs in each network meet a certain biological significance. WGCNA technology, based on the concept of scale-free distribution, has been successfully applied to solve many biological problems and make great contribution to the finding of many important discoveries [4]. In cancer research, WGCNA had been used to find tumor markers for screening, diagnosis, and treatment [5]. Zhu et al. [6] screened out genes related to the prognosis of soft tissue sarcoma by constructing WGCNA. Several miRNAs were identified as prognostic biomarkers through co-expression network analysis [7]. Similarly, biomarkers serve as clear cell Renal cell Carcinoma were identified in the same way [8].

In our research, WGCNA was established based on gene expression data and clinical information download from GSE9891 dataset and the hub gene module closely related to the pathological grade of OC were screened. GO and KEGG pathways analysis related to genes in hub module were performed. Subsequently, hub genes, which were closely related to ovarian cancer pathological grade, were screened out and verified in other data sets. Through a series of analyses, we speculated that the hub genes may be served as surface markers of M2 macrophages.

2. Materials and methods

2.1 Obtain the specimen of OC

From April 2018 to July 2019, 70 OC tissue samples were collected from Shengjing Hospital of China Medical University, including 25 in the low-grade group and 45 high-grade group. There was no chemotherapy or any other treatment before operation, and the clinicopathological data was complete. This study was approved by the medical research ethics committee of the hospital.

2.2 GEO data acquisition

Ovarian cancer experimental data set GSE9891 [9], validation data sets GSE63885 [10], GSE26193 [11] and GSE30161 [12] and their clinical data were download from GEO (https://www.ncbi.nlm.nih.gov/geo/). All of data sets were derived from the GPL570 platform. GSE9891, GSE63885, GSE26193 and GSE30161were respectively including 285 cases (Fig. 1), 101 cases, 107 cases and 58 cases of ovarian cancer in different pathological grade.

Figure 1.

Clustering dendrogram of 298 tumor samples and the clinical traits in GSE9891. Red color intensity was proportional to higher pathological stage and tumor grade. Gray color was proportional to missing information.

2.3 Construction of weighted gene co-expression network

The “WGCNA” package was used to establish co-expression network as the same way used in our previous article [13]. First, the Pearson correlation matrix is constructed by using the correlation coefficient of genes. Then, the pearson correlation matrix was transformed into a weighted adjacency matrix by using the formula of amn $=|$ Cmn $|\beta$ . Finally, the weighted adjacency matrix was transformed into topological overlap matrix. Here, $\beta=$ 3 (R2 $=$ 0.97) have been used to ensure scale-free network operation (Fig. 2). According to the gene tree with the minimum size of 30, the similar genes were divided into one module for the establishing of average linkage hierarchical clustering. Finally, 1,000 genes were selected randomly to perform visualization studies.

Figure 2.

Determination of soft threshold and inspection of scale-free network. (A) The correlation coefficients of log(K) and log(p(K)) corresponding to different soft-thresholding power ( $\beta$ ). The higher the coefficient is, the more consistent the network is with the distribution of scale-free networks. (B) The mean value of gene adjacency coefficient in the gene network corresponding to different soft-thresholding power, which reflects the average connection level of the network. (C) The distribution of connectedness of each node in the network. (D) The scatter plot of log(k) and log(p(k)), and the linear regression results show that the correlation coefficients are consistent with the characteristics of scale-free networks.

2.4 Hub module screening and functional analysis

Two methods were used to screen a hub module. First, gene significance (GS) was defined as the log10 transformation based on the $P$ value (GS $=$ logP) of linear regression between clinical features and gene expression. Modular significance (MS) is defined as the average GS of all genes in a module. The module with highest MS was selected as the hub module. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses was used to analyze the possible mechanism of genes involved in the malignant process of ovarian cancer through DAVID database FDR $ï ¼ œ$ 0.01 [14].

2.5 Hub gene screening and validation

In our research, hub genes were considered to be genes with high module connectivity (cor.GeneModule Membership $>$ 0.8) and high correlation of clinical characters (cor.GeneTraitSignificance $>$ 0.2). We used text datasets (GSE63885, GSE26193 and GSE30161) to validate the differential expression of hub genes in ovarian cancer with different pathological grade. Survival analysis based on the hub genes was performed by website of Kaplan Meier-plotter [13].

2.6 Real time fluorescence quantitative PCR experiment

The RNA of ovarian cancer tissues were extracted by Trizol according to the manufacturer’s instructions. Then, we used the ThermoScript reverse transcriptase (RT)-PCR System to synthesize single-stranded cDNA. The 20 $\mu$ L reaction system containing 10 $\mu$ L SYBR premix EX Taq (2 $\times$ ), 2 $\mu$ L primer, 2 $\mu$ L cDNA and 6 $\mu$ L ddH2O was used for RNA amplification. Calculating the relative expression of mRNA with 2 ${}^{-}$ Ct. MS4A4A forward primer 5’-ACCATGCAAGGAATG GAACAG-3’, reverse primer 5’-TTCCCATGCTAAG GCTCATCA-3’. CD163 forward primer 5’-TTTGTCA ACTTGAGTCCCTTCAC-3’, reverse primer 5’-TCCC GCTACACTTGTTTTCAC-3’. CPR65 forward primer 5’-GGGAGCCTACACCCTACACG-3’, reverse primer 5’-ATCCACCATATCCTCCCTGGT-3’. MS4A6A forward primer 5’-TGTGGCATGATGGTATTGAGC-3’, reverse primer 5’-AGGGTCCTATGAATGGGTAAGC-3’. GAPDH forward primer 5’-ACAACTTTGGTATCG TGGAAGG-3’, reverse primer 5’-GCCATCACGCCA CAGTTTC-3’.

2.7 Statistical analyses

SPSS 23.0 and R 3.5.3 were used for analysis. All statistical tests were bilateral, and a P value less than 0.05 was studied statistically significant. Continuous variables having to be in conformity with the customary distribution were compared by independent t test, while continuous variables with skewed distribution were compared by Mann-Whitney U test. The Kaplan-Meier curve was utilized to analyze the relationship between gene and overall survival. log-rank test is employed for evaluation.

3. Results

3.1 Construction of WGCNA and screening hub module

We used the “WGCNA” to assign genes with similar expression patterns into one module, and 7 gene modules (black, yellow, brown, blue, red, green and turquoise) were obtained (Fig. 3A). Then, we explored the correlation between the 7 gene modules and clinical traits, such as age, FIGO stage and grade. The result showed that the correlation between turquoise module and pathological grade of ovarian cancer was the highest (Fig. 3B, $r=$ $-$ 0.15, $p=$ 0.01), which indicated that the genes in turquoise module were closely related to the pathological grade of ovarian cancer. To screen for genes associated with the pathological grade of ovarian cancer, the turquoise module was selected as a meaningful module for subsequent research. We plotted network heatmap based on the interactive relationships of the 7 modules (Fig. 3C). The results showed that the gene modules were independent of each other. Besides, eigengene adjacency heatmap was plotted to explore co-expression similarity of the 7 modules, and the results showed that the 7 modules were mainly divided into two clusters that one cluster includes black and yellow modules, another cluster includes brown, blue, red, green and turquoise modules (Fig. 3D).

Figure 3.

Identification of hub module associated with the pathological grade of OC. (A) Dendrogram of all expressed genes clustered based on a dissimilarity measure (1-TOM), Dynamic Tree Cut corresponds to the original module and Merged Dynamic corresponds to the final module. Since no modules need to be merged, the results are exactly the same. (B) Heatmap of the correlation between Modular significance (MS) and clinical traits of ovarian cancer (OC), each row corresponds to a trait module and each column corresponds to a gene module, the numbers in the square brackets represent P values, the numbers in the square represent correlation coefficients, the red represents positive correlation, the blue represents negative correlation. (C) The heat map of interconnection degree after power conversion of correlation coefficient between genes in each module based on an adjacency matrix. In this heatmap progressively more saturated yellow colors indicate the high co-expression interconnectedness. Modules correspond to highly interconnected genes blocks. Genes of high intramuscular connectivity are located at the tip of the module branches because they show the highest interconnectedness with the rest of the genes in the module. (D) Heatmap plot of the adjacencies in the hub gene network includes the trait weight. Each column and row correspond to one module hub gene (labeled by color) or weight. In the heatmap, red represents high adjacency (positive correlation), while blue color represents low adjacency (negative correlation). Squares of red color along the diagonal are the meta-module.

3.2 GO and KEGG enrichment analysis

GO and KEGG enrichment analysis was used to analyze the possible mechanism of genes in turquoise module involved in the malignant process of ovarian cancer. GO analysis revealed that the genes in turquoise module main enrichment in inflammatory response, immune response and lipoprotein metabolic process (Fig. 4A). KEGG analysis revealed that the genes in turquoise module main enrichment in cytokine-cytokine receptor interaction and chemokine signaling pathway (Fig. 4B). All of the results suggested that the genes in turquoise module may be involved in the malignant process of ovarian cancer by regulating the tumor microenvironment.

Figure 4.

GO functional enrichment analysis and KEGG analysis. The size of the circle represents the number of genes. The y-axis shows the GO or KEGG pathway terms. The -log 10(pvalue) of each term is colored according to the legend. The redder the color the higher the value of -log 10 (pvalue) (A) GO functional enrichment analysis, the results revealed that the genes in turquoise module main enrichment in inflammatory response, immune response and lipoprotein metabolic process. (B) KEGG analysis, the results revealed that the genes in turquoise module were mainly enriched in cytokine-cytokine receptor interaction and chemokine signaling pathway.

Figure 5.

Hub gene screening. (A) Scatter plot for correlation between gene module membership in the turquoise module and gene significance. (B) PPI network analysis of top 25 genes in the turquoise module, the circle is the gene name, and the line represents the degree of connection between genes. The higher the connectivity, the stronger the interaction between genes.

Figure 6.

Hub gene validation. Differential expression of MS4A4A, CD163, CPR65, MS4A6A between Grade1/2 and Grade3/4 of OC tissues based on datasets GSE63885, GSE26193 and GSE30161. (A) Differential expression of MS4A4A in GSE63885 ( $P=$ 0.03). (B) Differential expression of CD163 in GSE63885 ( $P=$ 0.03). (C) Differential expression of CPR65 in GSE63885 ( $P=$ 0.04). (D) Differential expression of MS4A6A in GSE63885 ( $P=$ 0.01). (E) Differential expression of MS4A4A in GSE26193 ( $P=$ 0.01). (F) Differential expression of CD163 in GSE26193 ( $P=$ 0.014). (G) Differential expression of CPR65 in GSE26193 ( $P=$ 0.005). (H) Differential expression of MS4A6A in GSE26193 ( $P=$ 0.03). (I) Differential expression of MS4A4A in GSE30161 ( $P=$ 0.03). (J) Differential expression of CD163 in GSE30161 ( $P=$ 0.03). (K) Differential expression of CPR65 in GSE30161 ( $P=$ 0.04). (L) Differential expression of MS4A6A in GSE30161 ( $P=$ 0.01).

Figure 7.

Survival analysis. Kaplan-Meier plotter database prognostic analysis of MS4A4A, CD163, CPR65, MS4A6A [progression-free survival (PFS)]. (A) High expression of MS4A4A indicated shorter PFS of OC ( $P=$ 0.015). (B) High expression of CD163 indicated shorter PFS of OC ( $P=$ 0.0092). (C) High expression of CPR65 indicated shorter PFS of OC ( $P=$ 0.003). (D) High expression of MS4A6A indicated shorter PFS of OC ( $P=$ 0.0033).

3.3 Hub gene screening and validation

In this study, the hub genes refers to the turquoise module were the genes most related to the pathological grade of ovarian cancer. WGCNA screened 4 genes (MS4A4A, CD163, CPR65, MS4A6A) as hub genes based on hub gene screening method (cor.ModuleMembership in turquoise module $>$ 0.8, cor.GeneSignificance for Grade $>$ 0.2) (Fig. 5A). To show the interaction between the 25 genes most associated with pathological grade in turquoise module, we mapped the PPI network in Fig. 5B. The results suggested that these genes are involved in the malignant progression of ovarian cancer through complex interconnections. To verify the correlation between 4 hub genes and the pathological grade of ovarian cancer, the 4 genes were verified in three separate datasets GSE63885, GSE26193 and GSE30161through differential analysis method. The results of datasets in GSE63885, GSE26193 and GSE30161showed that the 4 genes were more highly expressed in high-grade ovarian cancer than in low-grade ovarian cancer, which verified that the 4 genes were closely related to the pathological grade of OC (Fig. 6, $p<$ 0.05). Different from the WGCNA method used to screen hub genes related to pathological grade in the training set, the differential analysis method was used to verify the correlation between the 4 screened hub genes and pathological grade of ovarian cancer in the verification dataset. Although the methods were different, the results all showed that the 4 hub genes were closely related to the pathological grade of ovarian cancer.

4. The prognostic value of the 4 genes

To investigate the influence of screened hub genes on the prognosis of ovarian cancer patients, survival analysis based on the 4 genes (MS4A4A, CD163, CPR65, MS4A6A) were performed by website of Kaplan Meier-plotter. The results showed that high expression of MS4A4A, CD163, CPR65, MS4A6A indicated shorter progression-free survival (PFS) of OC (Fig. 7, $p<$ 0.05).

5. Real time fluorescence quantitative PCR verification of the expression of the 4 genes

The results of Real time fluorescence quantitative PCR revealed that MS4A4A, CD163, CPR65, MS4A6A were overexpressed in high-grade group than the low-grade group. The difference was statistically significant (Fig. 8).

6. Discussion

Tumor pathological grade is considered to be closely related to tumor prognosis. High-grade serous ovarian cancer (HGSOC) has a poorer prognosis than low-grade ovarian cancer because it is more likely to metastasize into the peritoneal cavity [15]. Therefore, identification of biomarkers related to the pathological grade of ovarian cancer and intervening in patients at an early stage is crucial to improve the prognosis of patients.

In this study, in order to screen the biomarkers most related to the pathological grade of OC, we used WGCNA package to establish the relationship matrix between genes and phenotypes. Finally, 4 genes (MS4A4A, CD163, CPR65, MS4A6A) most associated with pathological grade of OC in turquoise module were screened for further study. To improve the reliability of the experiment, the 4 genes were verified in three other independent datasets (GSE63885, GSE26193 and GSE30161) and tissue samples from Shengjing Hospital of China Medical University. The results further proved that all of the 4 genes were related to the pathological grade of OC. Survival analysis showed that high expression of the 4 genes predicted a shorter PFS.

Figure 8.

MRNA levels verified by Real time fluorescence quantitative PCR experiment. (A) MS4A4A; (B) CD163; (C) CPR65; (D) MS4A6A.

MS4A4A and MS4A6A are a member of the tetracycline membrane, 4 domain families, subfamily A (MS4A), the expression of which was mainly confined to hematopoietic tissues [16]. At present, there are few related studies on MS4A4A and MS4A6A. The major research areas were limited to Alzheimer’s disease and blood diseases. Sanyal et al [17] reported that MS4A4A can be used as a surface marker for M2 macrophages, which suggest that targeting MS4A4A can inhibit the polarization of M2 macrophages. Wang et al. [18] revealed that MS4A4A correlates with stromal immune score and prognosis of gastric cancer patients. High expression of MS4A4A was associated with poor prognosis in gastric cancer, which was consistent with our findings. In addition, the role of MS4A4A in carcinogenesis and tumor progression has not been clearly found. Buddingh et al. [19] found that MS4A6A was specifically expressed on the surface of M2 macrophages in the process of screening differential genes related to osteosarcoma metastasis. Ma et al. [20] indicated that the hypermethylated and silenced of MS4A6A was associated with short-term survival in GBM patients, which was contrary to our research results. In addition, there is no more research on MS4A6A in tumor. CD163 is an important member of the scavenger receptor superfamily and mainly expressed on the surface of monocytes and macrophages. It is also served as a highly specific type marker of M2 tumor-associated macrophage [21]. CD163 not only acts as an immunomodulator against inflammation, but also plays an important role as a tumor-associated macrophage family member on tumor proliferation and metastasis. The expression of CD163 in malignant tumors has drawn more and more attention. Some studies have shown that CD163 is closely related to malignant cancers such as breast cancer [23], bladder cancer [24], lung cancer [25] and colorectal cancer [26]. A research also showed that CD163 is high-expressed in tumor-associated macrophages of OC and relevant to shorter PFS in OC, which was consistent with our findings. Besides, the research also showed that the number of CD163+ cells in OC was clearly associated with early relapse after first-line therapy [27]. However, there have not been any research associated with CPR65.

Through the review of the above literature, we found that MS4A4A, CD163 and MS4A6A are mainly related to immunity. Our functional analysis results also revealed that genes in turquoise module play a role in the malignant progression of ovarian cancer mainly through immune regulation. Tumor immune infiltration microenvironment is an important part of tumor genesis and treatment [28]. Many theoretical and experimental evidences had illustrated that tumor-associated macrophages (TAMs) as a component of immune infiltrate have a crucial effect on tumor immune-infiltrating microenvironment [29]. TAMs have two opposite phenotypes: M1 subtype macrophages have anti-tumor effect, while M2 type macrophages have tumorigenic activity. Numerous studies had shown that M2 TAMs promote tumor metastasis by communicating with tumor cells [32]. Chen et al. [33] reported that the communication between cancer cells and M2 TAMs can promote migration of EOC cell. In our research, MS4A4A, CD163, CPR65, MS4A6A were associated with the pathological grade and PFS of OC. In further research, we can extend the PFS of OC by inhibiting the 4 genes.

In conclusion, our research used a series of bioinformatic analysis methods to identify and validate the relationship between the 4 genes and ovarian cancer pathological grade. The MS4A4A, CD163, MS4A6A may serve as a surface marker for M2 macrophages. Targeting the 4 hub genes may be can extend the PFS of OC.

Footnotes

Acknowledgments

We thank the authors who provided the GEO public datasets.

Conflict of interest

The authors declare no competing interests.

References

Siegel

R.L.

Miller

K.D.

and Jemal

, Cancer statistics, 2016. CA: A Cancer Journal for Clinicians 66 (2016), 7–30.

Bowtell

D.D.

Bohm

Ahmed

A.A.

et al., Rethinking ovarian cancer II: reducing mortality from high-grade serous ovarian cancer, Nature Reviews Cancer 15 (2015), 668–79.

Cooke

S.L.

and Brenton

J.D.

, Evolution of platinum resistance in high-grade serous ovarian cancer, The Lancet Oncology 12 (2011), 1169–74.

Langfelder

and Horvath

, WGCNA: An R package for weighted correlation network analysis, BMC Bioinformatic 9 (2008), 559.

Liu

Zhang

Liu

Z.Q.

et al., Associating transcriptional modules with colon cancer survival through weighted gene co-expression network analysis, BMC Genomics 18 (2017), 361.

Zhu

Jin

Deng

et al., Co-expression network analysis identifies four hub genes associated with prognosis in soft tissue sarcoma, Frontiers in Genetics 10 (2019), 37.

Zhu

E.C.

Y.L.

et al., Serum level of co-expressed hub miRNAs as diagnostic and prognostic biomarkers for pancreatic ductal adenocarcinoma, Journal of Cancer 9 (2018), 3991–99.

Chen

Yuan

Qian

et al., Identification of biomarkers associated with pathological stage and prognosis of clear cell renal cell carcinoma by co-expression network analysis, Frontiers in Physiology 9 (2018), 399.

Tothill

R.W.

Tinker

A.V.

George

et al., Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research 14 (2008), 5198–208.

10.

Lisowska

K.M.

Olbryt

Dudaladava

et al., Gene expression analysis in ovarian cancer – faults and hints from DNA microarray study, Frontiers in Oncology 4 (2014), 6.

11.

Gentric

Kieffer

Mieulet

et al., PML-regulated mitochondrial metabolism enhances chemosensitivity in human ovarian cancers, Cell Metabolism 29 (2019), 156–73e10..

12.

Ferriss

J.S.

Kim

Duska

et al., Multi-gene expression predictors of single drug responses to adjuvant chemotherapy in ovarian carcinoma: Predicting platinum resistance, PloS One 7 (2012), e30550.

13.

Chen

et al., Coexpression network analysis identified Kruppel-like factor 6 (KLF6) association with chemosensitivity in ovarian cancer, J Cell Biochem (2018).

14.

Chen

Yuan

Wang

et al., Co-expression network analysis identified FCER1G in association with progression and prognosis in human clear cell renal cell carcinoma, International Journal of Biological Sciences 13 (2017), 1361–72.

15.

Le Page

Rahimi

Kobel

et al., Characteristics and outcome of the COEUR Canadian validation cohort for ovarian cancer biomarkers, BMC Cancer 18 (2018), 347.

16.

Eon Kuek

Leffler

Mackay

G.A.

et al., The MS4A family: counting past 1, 2 and 3, Immunol Cell Biol 94 (2016), 11–23.

17.

Sanyal

Polyak

M.J.

Zuccolo

et al., MS4A4A: a novel cell surface marker for M2 macrophages and plasma cells, Immunol Cell Biol 95 (2017), 611–19.

18.

Wang

and Chen

, Stromal-immune score-based gene signature: A prognosis stratification tool in gastric cancer, Frontiers in Oncology 9 (2019), 1212.

19.

Buddingh

E.P.

Kuijjer

M.L.

Duim

R.A.

et al., Tumor-infiltrating macrophages are associated with metastasis suppression in high-grade osteosarcoma: a rationale for treatment with macrophage activating agents, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research 17 (2011), 2110–9.

20.

Hou

et al., Genome-wide methylation profiling reveals new biomarkers for prognosis prediction of glioblastoma, Journal of Cancer Research and Therapeutics 11(Suppl 2) (2015), C212–5.

21.

Martinez

V.G.

Moestrup

S.K.

Holmskov

et al., The conserved scavenger receptor cysteine-rich superfamily in therapy and diagnosis, Pharmacol Rev 63 (2011), 967–1000.

22.

Van Gorp

Delputte

P.L.

and Nauwynck

H.J.

, Scavenger receptor CD163, a Jack-of-all-trades and potential target for cell-directed therapy, Mol Immunol 47 (2010), 1650–60.

23.

Baxevanis

C.N.

Sofopoulos

Fortis

S.P.

et al., The role of immune infiltrates as prognostic biomarkers in patients with breast cancer, Cancer Immunology, Immunotherapy: CII (2019).

24.

Maniecki

M.B.

Etzerodt

Ulhoi

B.P.

et al., Tumor-promoting macrophages induce the expression of the macrophage-specific receptor CD163 in malignant cells, International Journal of Cancer 131 (2012), 2320–31.

25.

La Fleur

Boura

V.F.

Alexeyenko

et al., Expression of scavenger receptor MARCO defines a targetable tumor-associated macrophage subset in non-small cell lung cancer, International Journal of Cancer (2018).

26.

Wang

Sun

X.Q.

Lin

H.C.

et al., Correlation between immune signature and high-density lipoprotein cholesterol level in stage II/III colorectal cancer, Cancer Medicine (2019).

27.

Reinartz

Schumann

Finkernagel

et al., Mixed-polarization phenotype of ascites-associated macrophages in human ovarian carcinoma: Correlation of CD163 expression, cytokine levels and early relapse, International Journal of Cancer 134 (2014), 32–42.

28.

Pollard

J.W.

, Macrophages define the invasive microenvironment in breast cancer, J Leukoc Biol 84 (2008), 623–30.

29.

Bao

Zhao

Zhou

et al., Mitochondrial fission-induced mtDNA stress promotes tumor-associated macrophage infiltration and HCC progression, Oncogene, 2019.

30.

Zhang

Zhao

et al., Cancer-associated fibroblasts enhance tumor-associated macrophages enrichment and suppress NK cells function in colorectal cancer, Cell Death & Disease 10 (2019), 273.

31.

Pathria

Louis

T.L.

and Varner

J.A.

, Targeting tumor-associated macrophages in cancer, Trends Immunol, 2019.

32.

Zheng

Luo

Wang

et al., Tumor-associated macrophages-derived exosomes promote the migration of gastric cancer cells by transfer of functional Apolipoprotein E, Cell Death & Disease 9 (2018), 434.

33.

Chen

Zhou

et al., Exosomes derived from hypoxic epithelial ovarian cancer cells deliver microRNAs to macrophages and elicit a tumor-promoted phenotype, Cancer Lett 435 (2018), 80–91.

Four genes relevant to pathological grade and prognosis in ovarian cancer

Abstract

BACKGROUND:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Obtain the specimen of OC

2.2 GEO data acquisition

2.5 Hub gene screening and validation

2.6 Real time fluorescence quantitative PCR experiment

2.7 Statistical analyses

3. Results

3.1 Construction of WGCNA and screening hub module

4. The prognostic value of the 4 genes

5. Real time fluorescence quantitative PCR verification of the expression of the 4 genes

6. Discussion

Footnotes

Acknowledgments

Conflict of interest

References