The expression of FCN2 messenger RNA (mRNA) in liver cancer was investigated using the Oncomine 4.5 database (esophageal carcinoma, ESCA). Oncomine.org is the world’s largest oncogene chip database and is a data mining platform in general. A number of ESCA studies involving Cuichard liver cell carcinoma, Roessler liver cancer, and The Cancer Genome Atlas (TCGA) liver cancer research were dissected and data obtained. The expression of FCN2 in HCC and in normal tissues was then investigated. A related inquiry was carried out based on the comparison of the transcription results. The common values and multiples of p-values were as below: p < 0.01; multiple of difference, 1.5; gene arrangement, 10%; data type, mRNA.

The recently established interactive web server Gene Expression Profiling Interactive Analysis (GEPIA) (9, 10), which examines tumor and natural swatch sequencing expression data, was used to analyze the expression of FCN2 in liver cancer and normal liver tissues. GEPIA was also utilized to obtain the mRNA expression data of FCN2 in the liver cancer tissues in this study. A network analysis was then implemented. Gene expression profiles were obtained from 369 liver cancers and 50 normal cases.

The UCSC (University of California, Santa Cruz) database (https://genome.edu/) is one of the most widely used in biology. We used this database to download a single standardized global cancer dataset, and then retrieved and screened the gene expression data of FCN2 in each sample. To determine the expression differences between normal and tumor samples, R software (version 3.6.4) was used for analysis according to previous studies. The significance of the differences was determined using the unpaired Wilcoxon rank-sum and signed-rank tests. The gene expression data of FCN2 in each sample, as well as the gene sequences retrieved, were examined using the embryonic stem cell-specific (ECSC) dataset.

TISIDB (Tumor Immune System Interactions and Drug Bank database; http://cis.hku.hk/TISIDB/) is a tumor and immunity-related database that collects a variety of data. It can carry out a large number of special screenings of genes and report genes related to T-cell-mediated killing or immunotherapy. It also includes data from TCGA on the links between genes and their immunological functions in 30 different tumor types. The TISIDB database was used to look for a link between FCN2 expression and the liver cancer stage.

UALCAN (University of Alabama at Birmingham Cancer data analysis portal) (11, 12) draws on the TCGA database and employs the capacities of cancer data mining and online examination to analyze and handle data effectively. Besides identifying relevant genetic biomarkers, it also includes expression profile and survival analyses and can directly query related data in other databases using relevant links. In summary, this website can be used to efficiently mine and analyze TCGA data with simple operations. UALCAN was utilized to analyze the expression of FCN2 in normal and liver cancer tissues in this study.

The Cancer Cell Line Encyclopedia (CCLE; https://sites.broadinstitute.org/ccle/) is a database of cancer cell lines maintained by the Broad Institute that includes relevant data on previous cell lines. By searching for cell lines, annotations, and genes on this website, the expression level of the specified gene in each tumor cell line can be determined.

The cBioPortal (http://www.cbioportal.org) currently contains 225 cancer studies. Changes in the levels of FCN2 in the HCC samples from TCGA were analyzed using cBioPortal. The search parameters included mutations and mRNA expression. The OncoPrint tab displays the genetic mutation profile of each sample. Furthermore, the network shows the biological interaction network of FCN2 from the common pathway database using color coding and screening options grounded on the periodicity of each gene’s genome change, including neighboring genes that change more frequently. Kaplan–Meier plots were drawn to show that FCN2 gene mutations are associated with the overall survival (OS) of patients with HCC. A logarithmic rank test was performed to explain the survival curve.

Proteomics, transcriptomics, and systems biology data are used to map tissues, cells, and organs in the Human Protein Atlas (HPA) database (https://www.proteinatlas.org). This database can be used to assess the protein expression in tumor and normal tissues, as well as the OS of patients with tumors. We used the HPA database to obtain the immunochemistry data for related genes.

The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database (https://www.string-db.org/) is an online database that investigates protein interactions that have already occurred. This website can assist in identifying key regulatory genes. This website’s supremacy stems from the fact that it contains the most species and interaction data. We created a protein–protein interaction (PPI) network for FCN2 using the STRING database. From the PPI network diagram, we found 10 fCN2-related genes. Furthermore, analyses using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were conducted for FCN2 and the 10 genes discovered.

The expression of FCN2 was examined using immunohistochemistry. Samples were collected from 30 patients with liver cancer at the Guangxi Zhuang Autonomous Region’s People’s Hospital. The samples were paraffin-fixed, cut into serial sections, and incubated with rabbit ficolin-2/ficolin-B polyclonal antibody (bs-13162R; Bioss, Woburn, MA, USA) at 4°C overnight. The sections were then dyed with hematoxylin and eosin (HE). Every procedure was executed as per the instructions in the SP kit. The number of positively stained cells during immunohistochemical labeling was determined using the antigen content, distribution density, tag method, and susceptibility. The distribution density of the positive results increased with the increase of antibody content, and the color development was brighter. Blue was the negative marker, light yellow indicated a slightly positive mark, brownish yellow was the moderately positive marker, and dark brown was the highly positive marker. In general, the more strongly positive regions should be photographed.

We took several representative images using the OlymbusX21 microscope. ImageJ was then used to execute a general morphometric analysis of each image. The photic density and positive area data for normal and cancer tissues were procured by measuring the chosen dyeing region using the ImageJ parameter. The higher the optical density and the more positive the expression, the higher the average level. Finally, statistical methods were used to determine whether there was a difference in the FCN2 expression between the normal and cancer groups.

The Kaplan–Meier plotter (13) is capable of assessing the correlation between the expression of all genes (mRNA, miRNA, and protein) and survival in more than 30,000 samples from 21 tumor types, including breast, ovarian, lung, and gastric cancer. Sources for the databases include the Gene Expression Omnibus (GEO), the European Genome–Phenome Archive (EGA), and TCGA. The primary purpose of the tool is the meta-analysis-based discovery and validation of survival biomarkers.

The Wilcoxon rank-sum test was used because normality monitoring made it clear that the specimen failed (p < 0.05). According to the Wilcoxon rank-sum test, the tumor was lower than normal, with a median difference of −0.084 between the two groups (−0.11 to 0.054). The diversity (p < 0.001) was statistically significant. Statistical analysis was executed with the R program (primarily ggplot2), with the following levels of significance: ns, p < 0.05, *p < 0.05, **p < 0.01, and ***p < 0.001.

In this study, we wanted to determine whether there were any differences in the gene expression between liver cancer and robust liver tissues. The GSE45267, GSE87630, and GSE54236 datasets from the National Center for Biotechnology Information (NCBI) were chosen as being related to liver cancer. Differentially expressed genes (DEGs) were filtered out using bioinformatics and R analysis, and the RobustRankAggreg (RRA) algorithm was employed to determine 20 genes with unique downregulated expressions. In this study, the 20 downregulated genes were selected for further analysis ( Figure 1 ). It was discovered that liver cancer tissues had lower levels of ANGPTL6, CFP, CLEC1B, CLEC4G, CLEC4M, COLEC10, CRHBP, CXCL12, DNASE1L3, FCN2, FCN3, GSTZ1, LCAT, NAT2, OIT3, RSPO3, VIPR1, STAB2, ECM1, and GPR128 expression compared to normal liver cells.

The 17 extracted genes (ANGPTL6, CFP, CLEC1B, CLEC4G, CLEC4M, COLEC10, CRHBP, CXCL12, DNASE1L3, FCN2, FCN3, GSTZ1, LCAT, NAT2, OIT3, RSPO3, and VIPR1) were validated using the GEPIA2 (Gene Expression Profiling Interactive Analysis) database. STAB2 was not found, and the levels of GPR128 and ECM1 were not significant. In contrast to their expression in normal liver tissues, the retrieved genes were only slightly expressed in HCC tissue samples ( Figure 2 ). This indicates that both the GEO and the TCGA database have confirmed that these genes are indeed significantly underexpressed in liver cancer.

The Kaplan–Meier plotter database was used to determine and evaluate the prognostic value of 20 genes (p < 0.05). To ensure the simplicity of the experiment, genes that showed no statistical significance were eliminated. Finally, the genes ANGPTL6, CFP, CLEC1B, CLEC4G, CLEC4M, COLEC10, CRHBP, CXCL12, DNASE1L3, FCN2, FCN3, GSTZ1, LCAT, NAT2, OIT3, RSPO3, and VIPR1 were obtained ( Figures 3A–Q ). The PPI network of DEGs was affected and was analyzed using the STRING database ( Figure 3R ). Using the cell type Minimal Common Oncology Data Elements (mCODE), we identified the most important nodes and retrieved six central nodes ( Figure 3S ). Among them, FCN2 was a central node.

The expression of FCN2 mRNA in various tumor tissues was investigated. The Oncomine database was used to examine the expression of FCN2 mRNA in various cancers and normal clinical samples ( Figure 4A ). A total of 639 datasets with a total of 78,661 samples were selected. The results showed that, in contrast to normal tissues, the mRNA expression of FCN2 is downregulated in leukemia and liver cancer, demonstrating that the transcription of FCN2 is tumor-specific.

The expression data of FCN2 in pan-cancer can be procured from the BioGPS database ( Figure 4B ). The expression data of the FCN2 gene in each sample ( Figure 4C ) were derived from a unified and standardized dataset from the UCSC. We also screened the following tissue sources: protopathic solid tumors, protopathic tumors, normal tissue, protopathic blood-derived cancer—bone marrow, and protopathic blood-derived cancer—peripheral blood samples. We then utilized R software to calculate the expression differences between normal and tumor tissues in each tumor (version 3.6.4). It was discovered that the expression levels in 31 malignancies were significantly downregulated, including glioblastoma (GBM), breast cancer (BRCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), lung adenocarcinoma (LUAD), kidney renal papillary cell carcinoma (KIRP), colon adenocarcinoma (COAD), prostate adenocarcinoma (PRAD), liver hepatocellular carcinoma (LIHC), tenosynovial giant cell tumors (TGCT), and adenoid cystic carcinoma (ACC), utilizing unopposed Wilcoxon rank-sum and signed-rank monitoring. A log2(x + 0.001) transformation was also applied for each expression value ( Figure 4D ), and the statistical significance was determined using paired Wilcoxon signed-rank test.

An Oncomine database search of the expression of FCN2 in diverse liver cancer studies produced three studies showing a lower expression of FCN2 in LIHC tissues compared to normal liver tissues (p = 1E−4, multiple of difference: 2) ( Figures 5A–C ).

The transcription and translation results of FCN2 mRNA in 369 liver cancer and 50 normal liver tissues were acquired using the GEPIA2 database. The findings revealed that the mRNA and protein expressions of FCN2 were lower in liver cancer tissues than those in normal liver tissues in all cases (p < 0.05) ( Figure 5D ). To confirm these results in depth, we used the UALCAN online analysis tool to further analyze the FCN2 gene, the result of which was similar in that the expression level of FCN2 in HCC cells was markedly lessened compared to that in normal cells ( Figures 5E, F ).

Liver cancer samples from 30 patients were taken from the People’s Hospital of Guangxi Zhuang Autonomous Region and immunohistochemistry staining was performed. Each staining procedure followed a rigorous set of guidelines. Five fields of view were randomly chosen from each sample after staining, and the average optical density value was determined using ImageJ. SPSS 19.0 was used for statistical analysis (IBM Inc., Armonk, NY, USA).

The areas of positive staining were brown, and the microscopy results revealed that normal liver tissues had bigger areas of positive staining than did liver cancer tissues ( Figure 5G ). Normal liver tissues had an average positive area of 1,783,320.5318 m² and an average optical density of 0.25662176264, while tumor tissues had an average positive area of 1,357,014.3320 m² and an average optical density of 0.16741577631. The normal and tumor groups had significance values of 0.200 and 0.086, respectively, suggesting that the positive staining area followed a normal distribution. Analysis of the t-test results revealed a significance of 0.036 (0.05), suggesting that the two sets of data are quite dissimilar and that the positive staining zone of the cancer group is smaller than that of the normal group ( Figure 5H ). The expression of FCN2 mRNA in liver tumor cell lines was shown to be at a low level using RNA sequencing (RNAseq) ( Attachment 1 ). Affymetrix analysis exhibited the expression of FCN2 mRNA in all tumor cell lines to be above 3.2. These studies have confirmed the significantly low expression of FCN2 in liver cancer tissues from different databases, as well as mRNA, protein, and clinical specimens.

As shown in ( Attachment 2 ), the TCGA dataset yielded 1,374 clinical and gene expression data on primary cancers. The sample included data from 253 men and 121 women, with a median age of 61 years. The protein expression of FCN2 was found to be strongly associated with height (p = 0.024) and the AFP level (p = 0.004) in the correlation analysis.Other clinicopathological features showed no association with FCN2 expression. The downregulation of FCN2 in HCC was linked to height (p = 0.018) and the AFP level (p = 0.003) in the univariate analysis ( Attachment 3 ).

To further verify the association between FCN2 and liver cancer, we analyzed the expression of the FCN2 gene in several markers using the R tool. The expression levels of FCN2 varied between the low- and high-AFP groups (p < 0.001) ( Figures 6A–K ). This indicates that FCN2 has certain clinical importance, whether from the high and low FCN2 expression groups (median) or the expression analysis of normal liver tissue and different liver cancer subgroups, and could be a complementary gene for AFP detection in liver cancer.

The survival rates of the high and low FCN2 expression level groups were compared to establish the predictive value of the expression of FCN2 in HCC. According to the Kaplan–Meier survival analysis, the OS of patients with HCC with a low FCN2 expression was poor [HR = 2.26 (1.58–3.24), p = 0.015], as was the DFS, compared to patients with a high FCN2 expression [HR = 0.69 (0.5–0.97), p = 0.03] ( Figures 7A–D ).

The nomogram included data on the tumor status, T stage, and FCN2 expression and was based on a Cox proportional hazards regression model ( Figure 7E ). The C-index of the prognostic model ( Figure 7F ) was 0.984 (95% CI = 0.970–0.997). We built a calibration plot to examine the similarity between the OS predicted by the prognostic model and the actual OS. The findings revealed that the nomogram’s prediction results were accurate ( Figure 7G ). These results imply that FCN2 could be a useful prognostic indicator for HCC.

STRING was used to complete an online analysis of the FCN2 protein, producing a network diagram of the PPI between 10 genes and the FCN2 gene ( Figures 8A, B ). The GO and KEGG pathway enrichment analyses for FCN2 and the other 10 genes ( Figures 8C–G ) showed that FCN2 is primarily involved in body immune processes, such as detecting and eliminating the molecular function of attaching to cellular constituents such as collagen trimer, mannose, complement, and opsonin and apoptotic cells, complement activation, and the lectin pathway. Signal pathway findings suggested that the FCN2 gene participates in S. aureus infection, the coagulation cascade, and the lectin pathway of complement activation.

To explore whether there is a link between FCN2 expression and tumor immune response, we conducted a single-sample gene set enrichment analysis (ssGSEA) to evaluate the immune cell infiltration in HCC tissues with various FCN2 expression levels. According to the findings, the infiltration levels of T follicular helper (TFH) and T helper 2 (Th2) cells were considerably lower in patients with HCC who had low FCN2 expression than those in patients who had high FCN2 expression. Neutrophil infiltration was higher in patients with HCC who had low FCN2 expression than in those who had high FCN2 expression. The infiltration levels of T cells, activated dendritic cells (aDCs), B cells, dendritic cells (DCs), interstitial dendritic cells (iDCs), macrophages, mast cells, CD56^dim natural killer (NK) cells, plasmacytoid dendritic (pDC) cells, T helper cells, effector memory T (Tem) cells, Th1 cells, Th17 cells, and regulatory T cells (Tregs) in patients with high and low FCN2 expressions were not substantially different ( Figure 9A ). We then examined the relationship between FCN2 expression levels and immune cell infiltration in HCC and discovered that the expression of FCN2 was positively linked to the number of neutrophils invading the tumor (r = 0.228, p < 0.001), eosinophils (r = 0.199, p < 0.001), NK cells (r = 0.165, p < 0.001), central memory T (Tcm) cells (r = 0.163, p < 0.001), and DCs (r = 0.161, p < 0.001) ( Figures 9B–F , respectively). The number of Th2 cells entering the body showed a negative correlation (r = −0.180, p < 0.001) ( Figure 9G ).

We used the GEPIA database to investigate the relationship between FCN2 expression and immune cell biomarkers in HCC in order to learn more about the role of FCN2 in tumor immunity. As seen in Table 1 , the expression of FCN2 was negatively correlated with the M1 macrophage biomarker (IRF5), neutrophil biomarker (ITGAM), and DC biomarker (ITGAX) in liver cancer. The B-cell biomarker (CD79A), CD8⁺ T-cell biomarker (CD8A), CD4⁺ T-cell biomarker (CD4), M1 macrophage biomarker (PTGS2), M2 macrophage biomarkers (CD163, VSIG4, and MS4A4A), neutrophil biomarker (CEACAM8), and DC biomarker (CD1C) were positively correlated with FCN2 expression. The above findings contribute to a better understanding of the link between FCN2 and immune cell infiltration.

To confirm the role of FCN2 in regulating LIHC immunity, we analyzed the correlation between FCN2 and the tumor microenvironment (TME) immunological signatures.

FCN2 was found to be favorably linked to the majority of immunomodulators, chemokines, and other factors, as well as the chemokine receptors in LIHC and diffuse large B-cell lymphoma (DLBCL), and negatively correlated with most of the immunomodulators, chemokines, and chemokine receptors in THYM and TGCT. FCN2 was positively correlated with most major histocompatibility complexes (MHCs) in LIHC, mesothelioma (MESO), and acute myeloid leukemia (LAML) and negatively correlated with most MHCs in TGCT ( Figure 10A ).

Analysis using the TISIDB database showed that three chemokines, CCL14 (ρ = 0.432, p < 2.2E−16), CCL16 (ρ = 0.232, p = 6.21E−06), and CCL23 (ρ = 0.402, p = 2.2E−16), were positively correlated with the expression of FCN2. The immunosuppressant kinase insert domain receptor (KDR) (ρ = 0.294, p = 8.34E−09) was also positively linked to the expression of FCN2. However, FCN2 expression was found to be inversely correlated with the expression of four chemokine receptors, namely, CD276 (ρ = −0.239, p = 3.19E−06), TNFRS4 (ρ = −0.204, p = 7.72E−05), CD276 (ρ = −0.239, p = 3.19E−06), and TNFSF4 (ρ = 0.211, p = 4.27E−05) ( Figures 10B–H ).

The HPA database confirmed that the expression levels of the immune-related markers CCL14, CCL23, KDR, CD276, TNFSF4, and TNFRSF4 in liver cancer are closely linked to FCN2. Apart from the lack of significant differences in the levels of positive rates, the immunohistochemistry results of the remaining molecules revealed that the positive rates in normal liver tissues were greater than the high specificity in liver cancer tissues, meaning that the expressions of these genes were lower in liver cancer tissues than those in normal liver tissues. The evidence shown in Figure 11 backs this up. These findings show that the expression of FCN2 is TME-specific, implying that FCN2 could be a potential immunotherapy target in the treatment of liver cancer.