Immunohistochemistry (IHC)
The IHC analysis was conducted to evaluate the expression of FBXO1 in different clinical molecular subtypes of BC tissues. In brief, following 4% formalin fixation and paraffin-embedding of specimens, 3-µm thick sections were incubated with primary Rabbit anti-FBXO1 antibody (1:200, Sigma, HPA071600 ) overnight at 4 °C, washed 3 times with PBS, and incubated with the secondary antibody for 1 h at 37 °C and streptavidin-HRP. The DAB kit was purchased from Zhongshan Goldenbridge Biotechnology Company (Beijing, China). The sections were stained with hematoxylin. The breast tumor specimens of patients were obtained from department of pathology, the Second Affiliated Hospital of Dalian Medical University (Dalian, China). The research protocol was approved and recorded by the Ethics Committee of The Second Affiliated Hospital of Dalian Medical University. All procedures are carried out in accordance with the Helsinki Declaration.
Oncomine Database
The human mRNA expression levels of FBXO gene family members in BC were compared with normal tissues by using the Oncomine gene expression array database (http://www.oncomine.org), an integrated data-mining platform. Students’ t test was adopted and transcriptional data of FBXOs were represented as log2-transformed form. We conducted the selection criteria as follows: Statistically significant P-values threshold < 1E-2, fold change > 2 and the gene ranks in the top 10%. All statistical methods and data source were acquired directly from the online database.
The Gene Expression Profiling Interactive Analysis (GEPIA) Dataset
The transcriptional levels of FBXOs in breast invasive carcinoma (BRCA) and normal breast tissue were obtained from the GEPIA database(http://gepia.cancer-pku.cn), and a public dataset assembles varieties of gene expression profiling functional modules, which was developed by scientists of Peking University [28]. We focused on the analytical results among intrinsic subtypes of BC and normal tissue. The correlation analysis of FBXO1 and related genes in BRCA tumor and normal tissue datasets was based on the GTEx and TCGA data. By using one-way ANOVA test, we defined the absolute value of Log2(FC) cutoff is 1; statistically significant p-value Cutoff is 1E-3. The linear dependence (correlation) between FBXO1 and hub genes was measured using Spearman's correlation coefficient. The results were used the non-log scale for calculation and used the log-scale axis for visualization.
UALCAN database
UALCAN database(http://ualcan.path.uab.edu/) is a publicly accessible dataset for analyzing 31 cancer types’ OMICS data, which is built on PERL-CGI with high quality graphics using JavaScript and CSS. These resources allow researchers to understand the impact of gene expression levels and gather relative clinicopathological parameters of various individual cancer types from The Cancer Genome Atlas (TCGA) [29]. We acquired the FBXOs’ transcriptional data from TCGA pan-cancer view and major subclasses and stages of BRCA by using UALCAN database. The mRNA information was unified as transcripts per million (TPM) reads for data comparison from different sources. P-value < 0.05 was considered statistically significant.
The Human Protein Atlas (HPA) database
The Human Protein Atlas (HPA) (https://www.proteinatlas.org) aims to provide 24000 kinds of human protein distribution information in different tissues and cells, and it displays for more than 20 kind of cancer types’ immunohistochemical staining results. In this work, for comparing the expression difference of the FBXO protein, we showed the immunohistochemical staining images between breast tumor and normal tissues from the HPA database to observe the tissue location of the target protein directly.
TCGA dataset and cBioPortal Online Tools
CBioPortal for cancer genomics is an open source resource for interactive exploration of multiple cancer genomic datasets. It allows researchers to visualize and analyze multidimensional genetic changes in different samples, genes and pathways [30]. The Breast Invasive Carcinoma of the cancer genome atlas (TCGA, Firehose Legacy, 1108 total samples) was selected for genomics analysis. By using the cBioPortal online tool(http://www.cbioportal.org), we investigated FBXO gene family’s predicted copy number alterations, mRNA expression (RNA sequencing [RNA-seq] version (v.)2 RSEM), gene correlations and Mutations situation, the results were automatically calculated using a Z-score ± 2.0, Pearson’s correction was considered.
Bc-GenExMiner (v4.4) online tool
Breast Cancer Gene-Expression Miner (bc-GenExMiner v4.4) online tool (http://bcgenex.centregauducheau.fr/BC-GEM/GEM-Accueil.php) is a statistical mining tool of published BC transcriptomic data (DNA microarrays [n = 10001] and RNA-seq [n = 4712]). It incorporates three classical mining functions: correlation, expression and prognosis [31]. According to common clinical parameters, we analyzed the FBXO gene family’s expression data in different patient groups. The subtypes of parameter include age, nodal status, ER, PR, HER-2, Basal-like statues, Triple-negative statues (IHC) and P53 status (sequence-based). Scarff-Bloom-Richardson (SBR) grade, and Nottingham prognostic index (NPI). The correlative heatmap of FBXO1 and the cell cycle pathway related hub genes was drawn by using the correlation module.
Kaplan-Meier plotter (KM plotter) database for survival analysis
We evaluated the prognostic significance of FBXO family members in KM plotter online database (http://kmplot.com/). The KM plotter was utilized to estimate the effect of 54 k genes (mRNA, miRNA, protein) on survival in 21 cancer types based on the gene arrays, RNA-sequence or next generation sequencing (for mutation data). Sources for the databases include GEO, EGA, and TCGA. The correlation between the target gene mRNA expression levels and disease-free survival rate (DFS), the overall survival (OS) rate, distance metastasis free survival (DMFS) and post progression survival (PPS) in BC groups were calculated by the Kaplan-Meier curve and log-rank test. The results were shown in the Kaplan–Meier survival plots. Hazard ratio (HR) and 95% confidence were calculated automatically by website tool. The values of each group are shown as the mean ± SD. P-value < 0.05 was regarded as statistically significant by using Log-rank test.
University of California Santa Cruz (UCSC) cancer genomics browser
UCSC Xena functional genomic browser is a database maintained by the University of California, Santa Cruz (UCSC). It is a new generation of online data analysis and visualization platform integrating analysis, visualization and galaxy. This tool contains the common standardized the data from TCGA, ICGC, TARGET, GTEX and CCLE datasets [32]. We used the UCSC Xena browser (http://xena.ucsc.edu/) to explore the correlation between FBXO1 and co-expression genes expression in different BC subtypes. The result of the comparison was evaluated by Spearman's correlation and represented in heat-map form.
Catalogue of Somatic Mutations in Cancer (COSMIC) database
COSMIC is the world's largest and most comprehensive resource for exploring the impact of somatic mutations in human cancer(https://cancer.sanger.ac.uk/cosmic). It includes somatic mutation data from different research institutions and databases, and provides convenient browsing, retrieval and downloading functions. The main goal is to conduct in-depth study on cell samples commonly used in cancer research and analyze their mutation information [33]. We used the pie charts to depict the mutations in FBXO1 in BC and the distribution and substitutions on the coding strand.
Functional Enrichment Analysis
COXPRESdb is a comprehensive dataset that comparing coexpression-gene in seven model animals (https://coxpresdb.jp/) [34]. We used cBioPortal database and COXPRESdb dataset to screen out the top human 150 genes with the strongest correlation with FBXO1, and obtained the intersect genes from both of databases. The functions of FBXO1 and the genes significantly associated with FBXO1 were predicted by gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. GO enrichment includes biological process (BP), cell component (CC), molecular function (MF). By referring to STRING database (https://string-db.org/), we screened the items with corrected P value ≤ 0.05. A total of 313 biological processes, 36 molecular functions and 56 cell components are related. Using R 3.6.3 software, we installed clusterProfiler, enrichplot and ggplot2 package to draw the histogram and bubble chart of the most remarkable results of GO and KEGG enrichment analysis. Using the Database for Annotation, Visualization and Integrated Discovery (DAVID) database(http://david.abcc.ncifcrf.gov/), we annotated the key targets in hsa04110 via Fisher's exact test: Cell cycle pathway to reveal the possible pathogenesis mediated by critical genes in breast adenocarcinoma.
Protein-protein interactions (PPI) network analysis
The PPI of co-expressed genes was retrieved from STRING database with an interaction score > 0.4, and we reconstructed the data via Cytoscape software (version 3.6.1) [35]. Molecular Complex Detection (MCODE) plug-in was employed to locate the densest connected module to find hub genes of clusters based on topology. The parameter standard as follows: MCODE score > 5 points, degree cut-off = 2, node score cut-off = 0.2, Max depth = 100, and k-Score = 2. The top 10 hub genes were vertified according to the degree-rank by CytoHubba plug-in. Next, we analyzed the potential biological process of hub genes by using BINGO plug-in. We seleted hypergeometic text and Benjamini & Hochberg False Discoverary Correction (FDR) method. The significance p-value set to 0.05.