Bioinformatics analysis reveals link between alternative complement cascade pathway and colorectal cancer liver metastasis

doi:10.21203/rs.3.rs-3691348/v1

Download PDF

Research Article

Bioinformatics analysis reveals link between alternative complement cascade pathway and colorectal cancer liver metastasis

https://doi.org/10.21203/rs.3.rs-3691348/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Colorectal cancer (CRC) has a high incidence mortality rate and is characterized by liver metastasis, which is the main cause of CRC patient death. In this study, a transcriptome sequencing dataset (GSE81558) from the integrated Gene Expression Omnibus database was evaluated to gain new insights into the pathogenesis of CRC and potential therapeutic targets.

Methods

All raw data were processed using R and screened for differentially expressed genes (DEGs) using LIMMA software. In-depth Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were conducted and visualized using R and Cytoscape software. Protein‒protein interactions (PPIs) associated with the DEGs were assessed using the Search Tool for the Retrieval of Interacting Genes/Proteins database. A mouse model of CRC liver metastasis of CRC was used to verify key associated signaling pathways.

Results

The GO biological processes (BPs) and KEGG pathway analyses revealed that DEGs between the normal colon and CRC samples were mainly involved in the cell cycle and the P53 signaling pathway, which regulate the cell cycle and alter tumor signaling pathways. The 10 hub genes identified by PPI were cell cycle-related. In CRC versus and CRC liver metastasis samples, the GO BPs were mainly associated with platelets and coagulation, and the KEGG pathways were mainly enriched in the complement and coagulation cascades and drug metabolism. The PPI hub genes were blood protein-related, such as ALB, AHSG, and APOH, or plasma protease inhibitors, such as SERPINC1. To confirm bioinformatics analysis results, we used wild-type (WT), C4 (an important molecule in the classical and lectin complement cascade pathways), and complement factor B (fB, an important molecule in the alternative complement cascade pathway) knockout (KO) mice to construct a CRC liver metastasis model. Compared with WT mice, fB-KO mice demonstrated significantly reduced liver metastasis and inflammation, while there was no difference in C4-KO mice.

Conclusion

Bioinformatics analyses revealed that the complement cascade is related to CRC liver metastasis and that the cell cycle is related to CRC. The role of the alternative complement pathway in CRC liver metastasis was confirmed in mice, indicating that this pathway is a potential therapeutic target in CRC liver metastasis and providing a theoretical basis for further research.

colorectal cancer

liver metastasis

GEO database

bioinformatics

differential genes

Colorectal cancer (CRC) is one of the most frequently occurring malignancies, with a high incidence and mortality rate. Over 1.8 million patients were newly diagnosed with CRC worldwide in 2020, and approximately 915,880 CRC-related deaths occurred ^[1]. Surgical removal of the primary tumor is considered the first choice for patients with CRC. However, the prognosis of CRC patients is relatively poor, with metastases present in approximately two-thirds of at the time of diagnosis. The liver is the most frequent location of CRC metastasis ^[2]. The median survival time of CRC patients with liver metastases without surgical removal is 7–15 months, and the 5-year survival rate of those with surgical removal is only 20–40%. Although chemotherapy and targeted therapy can eliminate tumor metastasis, these treatments cannot completely cure metastases due to drug resistance. Therefore, bioinformatic screening of the key signaling pathways and mechanisms that affect liver metastasis may reveal an effective treatment method to block the malignant progression of CRC ^[3–5].

CRC liver metastasis involves a complex cascade reaction with multiple factors and steps; thus, the underlying molecular mechanisms are intricate and varied. Therefore, further analysis is required to study the roles and mechanisms of various factors in this process. Bioinformatics is an important tool for analyzing large quantities of biological data that can greatly promote tumor research by mining tumor pathogenic genes and studying their roles in biological processes, molecular functions, and signaling pathways ^[6]. The recent development of bioinformatics has resulted in numerous studies on CRC using gene chip data to identify key associated pathways and genes. For example, some studies have found that S100P, CD133, LIP-C, APOBEC3G, and other genes affect CRC liver metastasis. However, because of the use of diverse data platforms and processing methods, various studies have yielded different results regarding the hub genes and associated signaling pathways. For example, Zhang et al. ^[7] analyzed differently expressed genes (DEGs) using the GEO2R online tool, which is widely used for data analysis because of its simplicity and convenience. However, the average gene expression values of each sample in the raw data were inconsistent, making it difficult to guarantee the accuracy of the analysis results. In the present study, we analyzed the raw data in CEL format and normalized it using the RMA algorithm to maintain the same gene expression level across samples, ensuring the reliability of the downstream analysis. In addition, some authors have systematized the bioinformatic and experimental identification of CRC-related hub genes ^[8] and found that the NF/KB signaling pathway may affect CRC liver metastasis. Different data processing may produce different outcomes. For example, Katipally et al ^[9] evaluated resected metastatic tumors by RNA sequencing and microRNA (miRNA) analyses and found an increased number of DEGs associated with DNA repair and cell cycle regulation/proliferation pathways (including E2F, G2M, and mitotic spindle pathways). Most bioinformatics analyses of CRC have screened the differential genes and molecular markers between normal colon tissue and CRC tissue or between CRC and CRC liver metastases, focusing on prognostic value. There is no clear distinction between liver metastatic and normal CRC. In our study, we compared CRC and normal colon tissue and CRC and CRC liver metastasis datasets. The results from the two dataset analyses were compared and analyzed, and it was found that the inflammatory complement pathway was enriched only in the liver metastasis dataset, indicating that the complement pathway plays a key role in CRC liver metastasis.

Metastasis indicates tumor deterioration and has long been linked to the tumor's evasion of the immune system and the unrestrained multiplication of transformed cells ^[10]. There is growing evidence that inflammatory immune cells influence metastasis. As a major part of innate immunity, many complement components play key regulatory roles in the tumor microenvironment during cancer progression. For example, complement anaphylactotoxins directly or indirectly inhibit anti-tumor t cell responses at primary and metastatic sites, promote tumor cell proliferation, accelerate tumor metastasis and angiogenesis, and reduce the immune system response to tumors ^[11]. Complement system activation within the tumor liberates complement components, such as C1q, C3a, and C5a, into the tumor microenvironment (TME), fostering tumorigenesis. These complement components attract immune cells comprising tumor-associated macrophages (TAMs), tumor-associated neutrophils (TANs), myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs), and dendritic cells (DCs) toward the neoplastic mass. These cells transform into a protumor phenotype and contribute to neoplastic progression at various levels, such as cell proliferation, angiogenesis, epithelial-mesenchymal transition (EMT), invasion and metastasis, and immunosuppression with antitumor activities. The release of immunosuppressive cytokines, including Arg-1, IL-10, and TGF-β, is also involved in this process. Additionally, unbalanced complement activation and inflammation can promote cancer metastasis to the brain, lung, and liver by degrading the extracellular matrix and disrupting tissue barriers. However, the specific pathway by which the complement cascade is activated in CRC metastasis remains unclear. In this study, we used bioinformatics methods to compare and analyze two different datasets, the normal colon and colorectal cancer contrast dataset and the colon cancer and colorectal liver metastasis contrast dataset. Next, GO and KEGG enrichment analyses revealed that CRC-associated DEGs are mainly related to the cell cycle, while liver metastatic CRC was mainly related to the inflammatory complement pathway. Finally, we used a mouse CRC liver metastasis model to verify that the complement pathway is a key signaling pathway in CRC liver metastasis.

2.1 Sources of Data

The CRC transcriptome sequencing dataset GSE81558 was downloaded from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo) and included 23 sporadic colon adenocarcinoma samples, 19 liver metastasis samples, and 9 normal colon tissue samples.

2.2 Data Handling

We obtained the unprocessed CEL file for GSE81558 from the GEO repository. Subsequently, we processed the data using the R software (version 4.2.2). The Affymetrix GeneChip probe-level data were extracted from the CEL files using the "AFFY" package ^[12].

We acquired the probe expression matrix after performing background correction, normalization, and aggregation using the robust multichip averaging (RMA) algorithm ^[13]. Next, we utilized the GPL15207 annotation package ("hugene 10st transcript ptcluster.db") to convert probe IDs into gene symbols (international standard gene names). We removed probe IDs that lacked corresponding gene symbols, and for genes with multiple probes, the highest value was considered the gene expression value. The gene expression matrix obtained was subjected to additional quality control through cluster and principal component analysis (PCA) ^[14].

2.3 Screening and Integration of DEGS

The DEG analysis was conducted using the "LIMMA" package ^[15]. The gene expression values were normalized using the RMA algorithm after log2 transformation, and a linear model was fitted using weighted least squares. The CRC and control groups were analyzed to assess differential expression. The Limma software package computed the log2-fold change rate (Log2FC) and P values for the adjusted false discovery rate (FDR) for each gene. The Benjamini‒Hochberg method was used to correct the FDR for P values to manage multiple false positivity tests. DEGs were determined based on adjusted P values < 0.05 and |log2 Fc| ≥1.0.

2.4 GO and KEGG Enrichment Analyses

GO ^[16] and KEGG ^[17] enrichment analyses were performed on core genes through the Database for Annotation, Visualization and Integrated Discovery (DAVID; http://https://david.ncifcrf.gov/), a large-scale molecular data resource for understanding biological systems and functions ^[18]. GO is a bioinformatics tool for analyzing gene biological processes (BP), molecular functions (MF), and cellular components (CC) ^[19]. P < 0.05 was considered to indicate a significant difference. Visualization of the analysis results was achieved using R (version 4.2.2) and Cytoscape software (version 3.8.2).

2.5 Specific PPI network analysis and Hub gene and micRNA recognition

The protein interaction network of the DEGs was constructed using the interactive gene search tool (STRING v11.5; http://string-db.org). The online platform mirwalk (http://mirwalk.umm.uni-heidelberg.de/) was used to create the interaction network between hub genes and micRNAs. Cytoscape (v3.8.2), an open-source bioinformatics tool, was employed to visualize protein‒protein interaction networks. MCODE, a Cytoscape plugin, clusters a given network to identify tightly connected regions. In addition, the cytoHubba plugin facilitates hub gene screening.

Cytoscape was used to map the protein interaction network, MCODE was used to determine the important modules in the protein interaction network, and cytoHubba was used to screen the key genes in the important pathways.

Each node in this model represents a protein encoded by a gene, and the connections between the nodes depict the interactions between the proteins. The nodes with the highest connectivity represent crucial proteins or hub genes that play significant regulatory roles in CRC liver metastasis.

2.6 Kaplan–Meier Survival Analyses of the Hub Genes

The survival analysis of patients with metastatic CRC was conducted using the GEPIA tool (http://gepia.cancer-pku.cn/). Kaplan‒Meier survival analyses were used to divide the cancer patients into low or high-expression groups according to the median mRNA expression of hub genes. The significance level was set at P < 0.05.

2.7 Establishment of a Mouse Model of CRC Liver Metastasis

Six wild-type, factor 3 B knockout, and factor 3 C4 knockout C57 mice were used to establish the in vivo liver metastasis models with SL4, a murine CRC cell line. After mice were anesthetized with an intraperitoneal injection of pentobarbital (100 mg/kg), the left spleen was transected and exposed, and 5×10⁵ SL4 cells in 100 µL DMEM/F12 medium were injected into the spleen using a 26-gauge needle. The mice were sacrificed on day 12 after inoculation, and the spleens and livers were harvested and weighed to observe metastasis. All mice were housed in standard cages at a constant temperature range of 22 ± 2°C, 60% relative humidity, and an artificial 12-hour light/dark cycle.

2.8 Histology and Immunohistochemistry

Livers from control and experimental mice were collected, paraffinized, sectioned at 4 microns, and analyzed histologically and by hematoxylin-eosin (HE) staining according to standard procedures. For immunohistochemical staining, liver sections were dehydrated, cleared, and incubated with anti-MAC-3 primary antibody (1:100, Santa) overnight at 4°C and ZSGB-BIO kit (PV-6002) secondary antibody at room temperature for 30 minutes. The DBA dyeing observation, hematoxylin dyeing. Slides were stained and viewed using a Nikon Labophot 2 microscope equipped with a SONY CCD-IRIS/RGB color camera and connected to a computerized imaging system. Images were captured and analyzed with ImageJ software.

3.1 Transcriptome Data

We processed the GSE81558 dataset to identify CRC and CRC liver metastasis-associated DEGS. The raw data were downloaded and processed using the "AFFY" package. Gene expression was normalized using the RMA comprehensive preprocessing algorithm. Figure 1(A) shows the data before and after normalization. The cluster analysis and PCA showed that the data in the CRC liver metastasis and control groups were easily distinguishable (Fig. 1(B)). Therefore, we could conduct follow-up analyses of genetic differences.

3.2 DEGs in CRC Liver Metastases

DEGS between the normal colon and CRC datasets and the CRC and CRC liver metastasis datasets were analyzed using the "LIMMA" package. A total of 1908 DEGs were screened in the normal colon and CRC comparison dataset, including 771 upregulated and 1137 downregulated genes. Twenty-three upregulated DEGs and 100 downregulated DEGs were identified in the CRC versus CRC liver metastasis datasets (Fig. 2(A)). The top 100 genes are shown as a heatmap in Fig. 2(B).

3.3 GO Enrichment Analysis and KEGG Enrichment Map of DEGs

The "clusterProfiler" program was used for the GO enrichment analysis of the DEGs. We focused on analyzing BPs. To better find new targets for treating colorectal cancer liver metastasis, we conducted a comparative analysis of normal colon versus CRC tissue and CRC versus CRC liver metastasis tissue. In the normal colon and CRC datasets, DEGs were mainly concentrated in mitotic cell cycle process, cell division, response to hormone, mitotic nuclear division, response to lipid, and other biological processes related to the cell cycle and mitosis. However, in the CRC and CRC liver metastasis datasets, the DEGs were mainly enriched in coagulation reaction processes, such as thrombocyte degranulation, regulation of blood clotting, acute phase response, hemostasis regulation, and regulation of blood clotting (Fig. 3 (A)). The KEGG channels with the most significant DEGs were obtained after entering the DEG list into DAVID. The results showed that the main enriched signaling pathways in CRC were the cell cycle and the p53 signaling pathway. The liver metastasis dataset was mainly enriched in signaling pathways related to complement and coagulation cascade and metabolism. Different pathways and genes are represented by colored ellipses in Fig. 3(B). The signaling pathways enriched in the two datasets differed, indicating that the inflammatory complement plays a key role only in CRC liver metastasis.

3.4 Specific PPI Network Analysis and Hub Gene Identification

After constructing the PPI network of differential genes using STRING (Fig. 4(A)), the Cytoscape plugin MCODE and cytoHubba were applied to screen meaningful modules and hub genes (Fig. 4(B)). Comparative analysis of the two comparative datasets found that the top 10 HUB genes evaluated according to PPI network connectivity in the normal colon and CRC comparison datasets were CDK1, CCNA2, CCNB1, CDC20, BUB1B, BUB1, DLGAP5, CCNB2, KIF11, and KIF20A (Table 1). Among them, CDK1, CCNA2, CCNB1, BUB1B, BUB1, and CCNB2 were related to the cell cycle, and DLGAP5 was a potential cell cycle regulator, consistent with the enrichment of signaling pathways in the cell cycle. The top 10 genes in the CRC and CRC liver metastasis comparison dataset were ALB, FGG, SERPINC1, AHSG, APOA1, F2, AMBP, APOH, APOB, and KLF20A (Table 2), most of which are common proteins in human blood, such as ALB, AHSG, and APOH, or plasma protease inhibitors, such as SERPINC1, which are also enriched in complement and coagulation. These results indicate that complement is only related to CRC liver metastasis.

3.5 MicRNA Network Analysis based on Hub Gene Identification

After screening the top 10 hub genes in the CRC and CRC liver metastasis comparison dataset, the miRWalk online database was used for an enrichment analysis of the miRNAs corresponding to the hub genes Fig. 5(A). A total of 337 micRNAs were enriched, of which AHSG-related micRNA was the most enriched. Cytoscape software was used for further screening and mapping, as shown in Fig. 5(B). The top ten micRNAs were miR-10b-5p, miR-20a-5p, miR-378g, miR-122-5p, miR-143-3p, miR-22-3p, miR-142-5p, miR-153-3p, miR-27b-3p, and miR-345-5p. Then, we intersected the enriched micRNA with the original data and found only miR-378g. MiR-378g is one of the most downregulated miRNAs in primary CRC tumors and can directly target PMEPA1, a TGF-β signaling regulator, promoting CRC liver metastasis.

3.6 Kaplan–Meier Survival Analyses According to Hub Genes

Kaplan‒Meier survival analyses were performed according to the 10 hub genes in the CRC versus CRC liver metastasis comparison dataset using the GEPIA tool. Overexpression of the hub gene KIF20A was associated with improved overall survival outcomes (Fig. 6). The P values of other genes were > 0.05, i.e., not significant. This finding suggests that KLF20A helps prevent the development of cancer.

3.7 Establishment of a Colon Cancer Liver Metastasis Model and Subsequent Experimental Verification

The signaling pathway enrichment analysis (3.3) comparing CRC with CRC liver revealed that the CRC liver metastasis dataset was mainly enriched in the complement pathway. Therefore, in the subsequent verification experiment, the complement pathway was primarily assessed for verification. Factor B is a classical constituent of the alternate complement pathway, and factor C4 is a constituent of both the classical and the lectin complement pathways. To verify the effect of the complement pathway on CRC liver metastasis in mice, we first established factor B-deficient and factor C4-deficient mouse models of CRC liver metastasis and performed HE staining and immunohistochemical analyses.

The extent of liver metastasis was significantly decreased in factor B null mice and did not change significantly in C4 null mice compared with controls (Fig. 7(A)). The HE staining and immunohistochemistry results were consistent with the phenotypic results (Fig. 7(B)). These results indicate that factor B deficiency could attenuate CRC liver metastasis in mice.

Despite our current understanding of the occurrence and progression of CRC, its incidence continues to increase. Therefore, it is critical to identify new therapeutic objectives to prevent CRC liver metastasis and significantly improve patient survival rates and quality of life ^[20]. Here, we used the GSE81558 CRC gene chip dataset from the GEO database to perform bioinformatics analyses to identify the key signaling pathways and hub genes associated with CRC liver metastasis. Among these signaling pathways, the alternative complement cascade was found to be crucial in CRC liver metastasis, which was verified in a mouse liver metastasis model.

In the present study, 771 upregulated and 1137 downregulated DEGs were identified in the normal colon versus CRC dataset. Similar to previous studies, compared with normal colon tissue, GO enrichment analysis revealed that BPs in CRC mainly involved mitotic cell cycle processes, cell division, response to hormones, mitotic nuclear division, and response to lipids, which are related to the tumor cell cycle. Many experimental and clinical studies have demonstrated that disordered cell cycle regulation is among the features of tumors, and tumors may result from abnormal cell cycle regulation. For example, Han et al. ^[21] found that betulinol could inhibit tumor metastasis in metastatic colon cancer cells by inducing cell cycle arrest, autophagy, and apoptosis. The KEGG signaling pathways of CRC were enriched in the cell cycle and P53 signaling pathway, consistent with the GO analyses. In tumor cells, the expression and regulation of various key cell cycle proteins are abnormal, resulting in the uncontrolled proliferation and division of tumor cells. For example, the mutation or inactivation of key proteins, such as p53, is often found in CRC, leading to an uncontrolled cell cycle ^[22]. More than 1/3 of CRC tumors have been reported to be associated with mutations in the TP53 gene. Colorectal cancer patients carrying TP53 mutations usually have a poorer prognostic outcome than that of those without mutations in this gene ^[23]. In addition, the abnormal expression or mutation of other proteins, such as CDK family proteins and retinoblastoma protein, can lead to an imbalance in cell cycle regulation. Miao et al ^[24] found that FERM domain-containing 8 (FRMD8) obstructs CDK4 activation and supports RB stabilization, leading to cell cycle arrest and CRC cell growth restriction. The top 10 hub genes in CRC in the PPI network interaction map were CDK1, CCNA2, CCNB1, CDC20, BUB1B, BUB1, DLGAP5, CCNB2, KIF11, and KIF20A. CDK1 is significantly differentially expressed in CRC with lymph node metastasis compared with that in normal mucosa ^[25]. CCNB1 is associated with survival in stage II CRC ^[26]. Overexpression of cell division cycle protein 20 (CDC20) predicts poor prognosis in CRC patients ^[27]. Among the remaining hub genes, the mitotic checkpoint gene BUB1 may be a specific driver of tumor metastasis and progression ^[28]. Additionally, a potential cell cycle regulator, DLGAP5, may reduce cell proliferation through cell cycle regulation ^[29]. In summary, compared with those in normal colon samples, most of the hub genes related to CRC were associated with the cell cycle, consistent with the GO and KEGG enrichment analysis results.

Next, we identified 23 upregulated and 100 downregulated DEGs in the CRC versus CRC liver metastasis dataset ^[30]. The GO enrichment analysis of the BPs associated with the DEGs revealed that DEGs in the CRC versus CRC liver metastasis dataset were mainly enriched in platelet degranulation, regulation of blood coagulation, acute-phase response, hemostasis regulation, and regulation of coagulation, mainly related to hemostasis and platelet activation. The first sign of malignancy may be coagulopathy ^[31]. The coagulation system is hyperactive in patients with malignant tumors due to the procoagulant characteristics of cancer cells, and the activity of the fibrinolytic system is decreased, which affects various related factors involved in the coagulation process and eventually leads to abnormal coagulation. In this process, the interaction of substances involved in blood coagulation not only directly leads to the hypercoagulation state but also promotes the proliferation and infiltration of tumor cells by activating related pathways and provides a scaffold for tumor cells to adhere to and metastasize while helping them evade normal immune surveillance. Extrahepatic coagulation factor VII (FVII) produced by CRC cells can promote tumor invasion and metastasis by upregulating the expression of MMP ^[32]. Compared with the CRC dataset, ALB, FGG, SERPINC1, AHSG, APOA1, F2, AMBP, APOH, APOB, and KLF20A were the top 10 hub genes in the CRC liver metastasis dataset. ALB is an important transport protein in serum, and decreased ALB expression worsens the prognosis of CRC patients ^[33]. AHSG is a serum glycoprotein synthesized by hepatocytes and adipocytes that can promote tumor proliferation, migration, and invasion ^[34]. FGG is a fibrinogen component that plays an instrumental role in blood coagulation, fibrinolysis, and cell-matrix interactions. Some researchers ^[35] have found that upregulated FGG activates the epithelial-mesenchymal transition pathway and promotes the migration and invasion of liver cancer cells by regulating the expression of Slug and ZEB1. Kinesin-kif20a is an important regulator of cell mitosis. Studies have demonstrated that KIF20A is highly expressed in many tumor tissues, consistent with our survival curve results, indicating that KIF20A overexpression is highly correlated with cell proliferation, tumor progression, tumor invasion and inferior overall survival in various tumors ^[36]. Although the role of KIF20A in CRC metastasis has not been investigated, the above evidence suggests that KIF20A is a potential key factor in CRC metastasis.

Next, we performed micRNA mining of hub genes in the CRC and CRC liver metastasis comparison dataset to construct a micRNA-protein interaction network. The intersection of our enriched micRNA and original transcriptome data showed that miR-378g was expressed in both datasets. Zhang et al identified a regulatory interaction network driven by the miR-378 gene family associated with metastatic CRC by assessing miR-378 levels in CRC cell lines and normal adjacent mucosal tissues paired with CRC cancers ^[37]. miR-378 was found to impede cell proliferation and invasion in CRC, suggesting that it might be an independent prognostic factor. Consistent with the above results, the authors of the aforementioned original article found that miRNA-378c, -378d, -378f, -378i, -378g, and − 378e were the most significantly downregulated miRNAs in primary CRC tumors. Another gene network frequently modified in liver metastasis of colorectal cancer is the TGFβ signaling pathway, which regulates cell growth and the development and differentiation of multiple tumor types ^[38]. MiR-378g can directly target PMEPA1, a TGF-β signaling regulator, inhibiting prostate cancer bone metastasis ^[39].

Furthermore, the KEGG pathway analysis revealed the involvement of the complement pathway in CRC metastasis for the first time. This pathway ranked first in the comparison of the CRC and CRC liver metastasis datasets in the present study. Thus, we further investigated the complement cascade. Complement can be activated by three canonical pathways, namely, the classical pathway, the lectin pathway, and the alternative pathway, all of which share a common terminal pathway ^[40]. Activated complement molecules in the tumor microenvironment can impact various immunomodulatory pathways, which have a crucial effect on the incidence and progression of tumors and ultimately determine their fate. On the one hand, the classical complement pathway can be activated by complement through the binding of specific antibodies to corresponding antigens found on the surface of the cell membrane, forming complexes, thus exerting a lytic effect on tumor cells. Complement-dependent cytotoxicity can also activate the complement system and kill tumor cells through alternative pathways. These pathways are essential for promoting immunological surveillance and inhibiting tumor cells. On the other hand, excessive complement activation in the tumor microenvironment plays a crucial role in sustaining chronic local inflammation and facilitating tumorigenesis. The interference of complement in the signaling pathway of the immune response also greatly promotes tumor immune escape and immune suppression ^[41]. Mice injected with lung cancer TC1 cells demonstrated that complement can be activated by the classical pathway ^[42]. In another syngeneic lung cancer mouse model using the KRAS-mutant CMT167 lung cancer cell line, complement was found to be activated through an alternative pathway to promote tumor development ^[43–44]. Other studies have shown that pathogenic fungi activate complement through the MBL pathway and promote the progression of pancreatic ductal adenocarcinoma ^[45]. Here, we found that compared with WT mice, factor B knockout mice demonstrated significantly reduced CRC liver metastasis, while there was no difference in C4 knockout mice. These findings imply that the alternative pathway of the complement cascade might constitute a new therapeutic target. As the most significantly enriched signaling pathway, the complement pathway is also related to other signaling pathways. Components of the extracellular matrix (ECM) can induce local complement activation and inflammation upon exposure to bodily fluids. The attachment of soluble complement inhibitors to the extracellular matrix, such as factor H (FH), is important to prevent excessive complement activity. Papp et al. ^[46] found that the complement factor H-related proteins FHR1 and FHR5 can interact with extracellular matrix ligands, decreasing factor H regulatory activity and increasing complement activation. Ficolin-3 (FCN3) is a constituent of the human Ficolin family and initiates complement activation through a pathway linked with serine proteases of mannose-binding lectin. Ma et al. ^[47] found that the overexpression of FCN3 stimulates apoptosis and suppresses cell proliferation via the p53 signaling pathway. Given the above findings, the complement pathway may be a new pathwayinvolved in CRC liver metastasis.

In summary, we analyzed the DEGs between CRC tissue and CRC liver metastasis foci using the GSE81558 dataset from the GEO database and found that the complement pathway was highly enriched in the CRC liver metastasis dataset. Two hub genes, BUB1 and KIF20A were found to be potential key factors in colorectal cancer metastasis. In addition, we demonstrated that the alternative pathway of the complement cascade is associated with CRC liver metastasis in a mouse. Therefore, we suggest that the alternative pathway of the complement cascade may be an effective therapeutic target in CRC liver metastasis.

Acknowledgements

The authors thank the contributors of the GEO database for sharing their data for uploading their meaningful datasets.

Funding

This study was supported by the National Natural Science Foundation of China (No. 82060537, No. 81301797, No. 81260304), and Beijing Municipal Natural Science Foundation (7122026).

Conflicts of interest

The authors declare no conflict of interest.

Author Contributions

Study design:PCM; Writing the draft:LY; Data analysis:WRQ and DJ;

Data collection and Literature search:YSQand ZQL.

Data availability statement

Raw data for this study is available on the Gene Expression omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/, NO. GSE81558).

Ethics approval

The study was conducted in accordance with the Declaration of Helsinki and approved by the Medical Review Committee of Yanbian University, with a report number of 2310120 and an approval date of November 30, 2023

Benson AB, Venook AP, Al-Hawary MM, et al. Colon Cancer, Version 2.2021, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2021;19(3):329–59. 10.6004/jnccn.2021.0012. Published 2021 Mar 2.
Tsai MS, Su YH, Ho MC, et al. Clinicopathological features and prognosis in resectable synchronous and metachronous colorectalcocln liver metastasis. Ann Surg Oncol. 2007;14(2):786–94. 10.1245/s10434-006-9215-5.
Mauri G, Bonazzina E, Amatu A et al. The Evolutionary Landscape of Treatment for BRAFV600E Mutant Metastatic Colorectalcocln Cancer. Cancers (Basel). 2021;13(1):137. Published 2021 Jan 4. 10.3390/cancers13010137.
Ruffinelli JC, Santos Vivas C, Sanz-Pamplona R, Moreno V. New advances in the clinical management of RAS and BRAF mutant colorectalcocln cancer patients. Expert Rev Gastroenterol Hepatol. 2021;15(1):65–79. 10.1080/17474124.2021.1826305.
Luo Q, Chen D, Fan X, Fu X, Ma T, Chen D, KRAS. PIK3CA bi-mutations predict a poor prognosis in colorectalcocln cancer patients: A single-site report. Transl Oncol. 2020;13(12):100874. 10.1016/j.tranon.2020.100874.
Wang DW, Su F, Yang LJ, et al. Bioinformatics Analysis and Identification of Potential Genes Associated with Pathogenesis and Prognosis of Gastric Cancer. Curr Med Sci. 2022;42(2):357–72. 10.1007/s11596-022-2515-6.
Zhang T, Guo J, Gu J, et al. Identifying the key genes and microRNAs in colorectalcocln cancer liver metastasis by bioinformatics analysis and in vitro experiments. Oncol Rep. 2019;41(1):279–91. 10.3892/or.2018.6840.
Ershov P, Poyarkov S, Konstantinova Y, Veselovsky E, Makarova A. Transcriptomic Signatures in Colorectal Cancer Progression. Curr Mol Med. 2023;23(3):239–49. 10.2174/1566524022666220427102048.
Katipally RR, Martinez CA, Pugh SA, et al. Integrated Clinical-Molecular Classification of Colorectal Liver Metastases: A Biomarker Analysis of the Phase 3 New EPOC Randomized Clinical Trial. JAMA Oncol Published online July. 2023;20. 10.1001/jamaoncol.2023.2535.
Ping Lu S, Wei X, Liang. Advances in research on the relationship between the complement system and antitumor immune response[J]. Chin J Clin Oncol. 2021;48(13):681–5. 10.3969/j.issn.1000-8179.2021.13.256.
Baldessari C, Pipitone S, Molinaro E, et al. Bone Metastases and Health in Prostate Cancer: From Pathophysiology to Clinical Implications. Cancers (Basel). 2023;15(5):1518. 10.3390/cancers15051518. Published 2023 Feb 28.
Gautier L, Cope L, Bolstad BM, Irizarry RA. affy–analysis of Affymetrix GeneChip data at the probe level. Bioinformatics. 2004;20(3):307–15. 10.1093/bioinformatics/btg405.
Irizarry RA, Hobbs B, Collin F, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249–64. 10.1093/biostatistics/4.2.249.
Wei RQ, Zhang WM, Liang Z, Piao C, Zhu G. Identification of Signal Pathways and Hub Genes of Pulmonary Arterial Hypertension by Bioinformatic Analysis. Can Respir J. 2022;2022:1394088. Published 2022 Aug 29. 10.1155/2022/1394088.
Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. 10.1093/nar/gkv007.
Young A, Whitehouse N, Cho J, Shaw C. OntologyTraverser: an R package for GO analysis. Bioinformatics. 2005;21(2):275–6. 10.1093/bioinformatics/bth495.
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16(5):284–7. 10.1089/omi.2011.0118.
Kanehisa M. The KEGG database. Novartis Found Symp. 2002;247:91–252.
Xia J, Benner MJ, Hancock RE. NetworkAnalyst–integrative approaches for protein-protein interaction network analysis and visual exploration. Nucleic Acids Res. 2014;W167–74. 10.1093/nar/gku443. 42(Web Server issue).
Wang D, Zhou Y, Hua L, Li J, Zhu N, Liu Y. CDK3, CDK5 and CDK8 Proteins as Prognostic and Potential Biomarkers in Colorectal Cancer Patients. Int J Gen Med. 2022;15:2233–45. 10.2147/IJGM.S349576. Published 2022 Feb 27.
Han YH, Mun JG, Jeon HD, Kee JY, Hong SH. Betulin Inhibits Lung Metastasis by Inducing Cell Cycle Arrest, Autophagy, and Apoptosis of Metastatic Colorectalcocln Cancer Cells. Nutrients. 2019;12(1):66. 10.3390/nu12010066. Published 2019 Dec 26.
Marei HE, Althani A, Afifi N, et al. p53 signaling in cancer progression and therapy. Cancer Cell Int. 2021;21(1):703. 10.1186/s12935-021-02396-8. Published 2021 Dec 24.
Dolma L, Muller PAJ. GOF Mutant p53 in Cancers: A Therapeutic Challenge. Cancers (Basel). 2022;14(20):5091. Published 2022 Oct 18. 10.3390/cancers14205091.
Yu M, Wu W, Sun Y, et al. FRMD8 targets both CDK4 activation and RB degradation to suppress colon cancer growth. Cell Rep. 2023;42(8):112886. 10.1016/j.celrep.2023.112886.
Urh K, Zidar N, Boštjančič E. Bioinformatics Analysis of RNA-seq Data Reveals Genes Related to Cancer Stem Cells in Colorectalcocln Cancerogenesis. Int J Mol Sci. 2022;23(21):13252. 10.3390/ijms232113252. Published 2022 Oct 31.
Hashemi R, Peymani M, Ghaedi K, Saffar H. In silico identification of the specific pathways in each stage of colorectalcocln cancer and the association of their top genes with drug resistance and sensitivity. Med Oncol. 2022;39(5):57. 10.1007/s12032-022-01661-8. Published 2022 Feb 12.
Wang L, Hu XD, Li SY, Liang XY, Ren L, Lv SX. ASPM facilitates colorectalcocln cancer cells migration and invasion by enhancing β-catenin expression and nuclear translocation. Kaohsiung J Med Sci. 2022;38(2):129–38. 10.1002/kjm2.12464.
Shichiri M, Yoshinaga K, Hisatomi H, Sugihara K, Hirata Y. Genetic and epigenetic inactivation of mitotic checkpoint genes hBUB1 and hBUBR1 and their relationship to survival. Cancer Res. 2002;62(1):13–7.
Zhang S, Gong T, Nan Y, Feng R, Liu Z, Chen H. MAFB promotes the malignant phenotypes by IGFBP6 in esophageal squamous cell carcinomas. Exp Cell Res. 2022;416(1):113158. 10.1016/j.yexcr.2022.113158.
Sharma A, Yadav D, Rao P, et al. Identification of potential therapeutic targets associated with diagnosis and prognosis of colorectal cancer patients based on integrated bioinformatics analysis. Comput Biol Med. 2022;146:105688. 10.1016/j.compbiomed.2022.105688.
Admasu FT, Dejenie TA, Ayehu GW et al. Evaluation of thromboembolic event, basic coagulation parameters, and associated factors in patients with colorectal cancer: a multicenter study. Front Oncol. 2023;13:1143122. Published 2023 May 2. 10.3389/fonc.2023.1143122.
Kroone C, Tieken C, Kocatürk B, et al. Tumor-expressed factor VII is associated with survival and regulates tumor progression in breast cancer. Blood Adv. 2023;7(11):2388–400. 10.1182/bloodadvances.2022008455.
Wang B, Yin C, Yang X, et al. Six Genes Associated with Lymphatic Metastasis in Colon Adenocarcinoma Linked to Prognostic Value and Tumor Immune Cell Infiltration. Evid Based Complement Alternat Med. 2022;2022:4304361. 10.1155/2022/4304361. Published 2022 Aug 29.
Xing X, Cao F, Gao L, Song M. AHSG, a Gene Promoting Tumour Proliferation, Migration and Invasion, is an Independent Prognostic Factor for Poor Overall Survival in Lung Adenocarcinoma. Mol Biol Rep. 2023;50(9):7659–66. 10.1007/s11033-023-08623-x.
Qian X, Cai J, Qi Q et al. Preoperative fibrinogen is associated with the clinical survival of primary liver cancer patients and promotes hepatoma metastasis via the PTEN/AKT/mTOR pathway [published correction appears in Heliyon. 2023;9(9):e19952]. Heliyon. 2023;9(6):e16696. Published 2023 May 25. 10.1016/j.heliyon.2023.e16696.
Jin Z, Peng F, Zhang C, Tao S, Xu D, Zhu Z. Expression, regulating mechanism and therapeutic target of KIF20A in multiple cancer. Heliyon. 2023;9(2):e13195. 10.1016/j.heliyon.2023.e13195. Published 2023 Jan 25.
Zhang GJ, Zhou H, Xiao HX, Li Y, Zhou T. MiR-378 is an independent prognostic factor and inhibits cell growth and invasion in colorectal cancer. BMC Cancer. 2014;14:109. Published 2014 Feb 20. 10.1186/1471-2407-14-109.
Li X, Wu Y, Tian T. TGF-β Signaling in Metastatic Colorectal Cancer (mCRC): From Underlying Mechanism to Potential Applications in Clinical Development. Int J Mol Sci. 2022;23(22):14436. 10.3390/ijms232214436. Published 2022 Nov 20.
Baldessari C, Pipitone S, Molinaro E, et al. Bone Metastases and Health in Prostate Cancer: From Pathophysiology to Clinical Implications. Cancers (Basel). 2023;15(5):1518. 10.3390/cancers15051518. Published 2023 Feb 28.
Daha MR, Seelen M. Novel Approaches to Control of the Alternative Complement Pathway for the Treatment of C3 Glomerulopathies. J Am Soc Nephrol. 2018;29(8):2032–3. 10.1681/ASN.2018050554.
Coss SL, Zhou D, Chua GT, et al. The complement system and human autoimmune diseases. J Autoimmun. 2023;137:102979.
Roumenina LT, Daugan MV, Noé R, et al. Tumor Cells Hijack Macrophage-Produced Complement C1q to Promote Tumor Growth. Cancer Immunol Res. 2019;7(7):1091–105. 10.1158/2326-6066.CIR-18-0891.
Lu P, Ma Y, Wei S, Liang X. The dual role of complement in cancers, from destroying tumors to promoting tumor development. Cytokine. 2021;143:155522. 10.1016/j.cyto.2021.155522.
Kleczko EK, Poczobutt JM, Navarro AC et al. Upregulation of complement proteins in lung cancer cells mediates tumor progression. Front Oncol. 2023;12:1045690. Published 2023 Jan 5. 10.3389/fonc.2022.1045690.
Aykut B, Pushalkar S, Chen R, et al. The fungal mycobiome promotes pancreatic oncogenesis via activation of MBL. Nature. 2019;574(7777):264–7. 10.1038/s41586-019-1608-2.
Papp A, Papp K, Uzonyi B, et al. Complement Factor H-Related Proteins FHR1 and FHR5 Interact With Extracellular Matrix Ligands, Reduce Factor H Regulatory Activity and Enhance Complement Activation. Front Immunol. 2022;13:845953. 10.3389/fimmu.2022.845953. Published 2022 Mar 22.
Ma D, Liu P, Wen J, et al. FCN3 inhibits the progression of hepatocellular carcinoma by suppressing SBDS-mediated blockade of the p53 pathway. Int J Biol Sci. 2023;19(2):362–76. 10.7150/ijbs.69784. Published 2023 Jan 1.

Tables 1 to 2 are available in the Supplementary Files section

No competing interests reported.

Tables.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Bioinformatics analysis reveals link between alternative complement cascade pathway and colorectal cancer liver metastasis

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1. Introduction

2. Materials and Methods

2.1 Sources of Data

2.2 Data Handling

2.3 Screening and Integration of DEGS

2.4 GO and KEGG Enrichment Analyses

2.5 Specific PPI network analysis and Hub gene and micRNA recognition

2.6 Kaplan–Meier Survival Analyses of the Hub Genes

2.7 Establishment of a Mouse Model of CRC Liver Metastasis

2.8 Histology and Immunohistochemistry

3. Results

3.1 Transcriptome Data

3.2 DEGs in CRC Liver Metastases

3.3 GO Enrichment Analysis and KEGG Enrichment Map of DEGs

3.4 Specific PPI Network Analysis and Hub Gene Identification

3.5 MicRNA Network Analysis based on Hub Gene Identification

3.6 Kaplan–Meier Survival Analyses According to Hub Genes

3.7 Establishment of a Colon Cancer Liver Metastasis Model and Subsequent Experimental Verification

4. Discussion

5. Conclusion

Statements and Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1