FBXW4 Acts as a Protector of FOLFOX-Based Chemotherapy in Metastatic Colorectal Cancer Identified by Co-Expression Network Analysis

High-accuracy prediction of colorectal cancer chemotherapy efficacy using machine learning applied to gene expression data

Introduction: FOLFOX and FOLFIRI chemotherapy are considered standard first-line treatment options for colorectal cancer (CRC). However, the criteria for selecting the appropriate treatments have not been thoroughly analyzed. Methods: A newly developed machine learning model was applied on several gene expression data from the public repository GEO database to identify molecular signatures predictive of efficacy of 5-FU based combination chemotherapy (FOLFOX and FOLFIRI) in patients with CRC. The model was trained using 5-fold cross validation and multiple feature selection methods including LASSO and VarSelRF methods. Random Forest and support vector machine classifiers were applied to evaluate the performance of the models. Results and Discussion: For the CRC GEO dataset samples from patients who received either FOLFOX or FOLFIRI, validation and test sets were >90% correctly classified (accuracy), with specificity and sensitivity ranging between 85%-95%. In the datasets used from the GEO database, 28.6% of patients who failed the treatment therapy they received are predicted to benefit from the alternative treatment. Analysis of the gene signature suggests the mechanistic difference between colorectal cancers that respond and those that do not respond to FOLFOX and FOLFIRI. Application of this machine learning approach could lead to improvements in treatment outcomes for patients with CRC and other cancers after additional appropriate clinical validation.

SSLpheno: a self-supervised learning approach for gene-phenotype association prediction using protein-protein interactions and gene ontology data

MOTIVATION

Medical genomics faces significant challenges in interpreting disease phenotype and genetic heterogeneity. Despite the establishment of standardized disease phenotype databases, computational methods for predicting gene-phenotype associations still suffer from imbalanced category distribution and a lack of labeled data in small categories.

RESULTS

To address the problem of labeled-data scarcity, we propose a self-supervised learning strategy for gene-phenotype association prediction, called SSLpheno. Our approach utilizes an attributed network that integrates protein-protein interactions and gene ontology data. We apply a Laplacian-based filter to ensure feature smoothness and use self-supervised training to optimize node feature representation. Specifically, we calculate the cosine similarity of feature vectors and select positive and negative sample nodes for reconstruction training labels. We employ a deep neural network for multi-label classification of phenotypes in the downstream task. Our experimental results demonstrate that SSLpheno outperforms state-of-the-art methods, especially in categories with fewer annotations. Moreover, our case studies illustrate the potential of SSLpheno as an effective prescreening tool for gene-phenotype association identification.

AVAILABILITY AND IMPLEMENTATION

https://github.com/bixuehua/SSLpheno.

Identification of co-regulated genes associated with doxorubicin resistance in the MCF-7/ADR cancer cell line

Introduction

The molecular mechanism of chemotherapy resistance in breast cancer is not well understood. The identification of genes associated with chemoresistance is critical for a better understanding of the molecular processes driving resistance.

Methods

This study used a co-expression network analysis of Adriamycin (or doxorubicin)-resistant MCF-7 (MCF-7/ADR) and its parent MCF-7 cell lines to explore the mechanisms of drug resistance in breast cancer. Genes associated with doxorubicin resistance were extracted from two microarray datasets (GSE24460 and GSE76540) obtained from the Gene Expression Omnibus (GEO) database using the GEO2R web tool. The candidate differentially expressed genes (DEGs) with the highest degree and/or betweenness in the co-expression network were selected for further analysis. The expression of major DEGs was validated experimentally using qRT-PCR.

Results

We identified twelve DEGs in MCF-7/ADR compared with its parent MCF-7 cell line, including 10 upregulated and 2 downregulated DEGs. Functional enrichment suggests a key role for RNA binding by IGF2BPs and epithelial-to-mesenchymal transition pathways in drug resistance in breast cancer.

Discussion

Our findings suggested that MMP1, VIM, CNN3, LDHB, NEFH, PLS3, AKAP12, TCEAL2, and ABCB1 genes play an important role in doxorubicin resistance and could be targeted for developing novel therapies by chemical synthesis approaches.

Overexpressed XRCC2 as an independent risk factor for poor prognosis in glioma patients

Background

XRCC2, a homologous recombination-related gene, has been reported to be associated with a variety of cancers. However, its role in glioma has not been reported. This study aimed to find out the role of XRCC2 in glioma and reveal in which glioma-specific biological processes is XRCC2 involved based on thousands of glioma samples, thereby, providing a new perspective in the treatment and prognostic evaluation of glioma.

Methods

The expression characteristics of XRCC2 in thousands of glioma samples from CGGA and TCGA databases were comprehensively analyzed. Wilcox or Kruskal test was used to analyze the expression pattern of XRCC2 in gliomas with different clinical and molecular features. The effect of XRCC2 on the prognosis of glioma patients was explored by Kaplan–Meier and Cox regression. Gene set enrichment analysis (GSEA) revealed the possible cellular mechanisms involved in XRCC2 in glioma. Connectivity map (CMap) was used to screen small molecule drugs targeting XRCC2 and the expression levels of XRCC2 were verified in glioma cells and tissues by RT-qPCR and immunohistochemical staining.

Results

We found the overexpression of XRCC2 in glioma. Moreover, the overexpressed XRCC2 was associated with a variety of clinical features related to prognosis. Cox and meta-analyses showed that XRCC2 is an independent risk factor for the poor prognosis of glioma. Furthermore, the results of GSEA indicated that overexpressed XRCC2 could promote malignant progression through involved signaling pathways, such as in the cell cycle. Finally, doxazosin, quinostatin, canavanine, and chrysin were identified to exert anti-glioma effects by targeting XRCC2.

Conclusions

This study analyzed the expression pattern of XRCC2 in gliomas and its relationship with prognosis using multiple datasets. This is the first study to show that XRCC2, a novel oncogene, is significantly overexpressed in glioma and can lead to poor prognosis in glioma patients. XRCC2 could serve as a new biomarker for glioma diagnosis, treatment, and prognosis evaluation, thus bringing new insight into the management of glioma.

Supplementary Information

The online version contains supplementary material available at 10.1186/s10020-021-00316-0.

Identification of Hub Genes and Construction of a Transcriptional Regulatory Network Associated With Tumor Recurrence in Colorectal Cancer by Weighted Gene Co-expression Network Analysis

Tumor recurrence is one of the most important risk factors that can negatively affect the survival rate of colorectal cancer (CRC) patients. However, the key regulators dictating this process and their exact mechanisms are understudied. This study aimed to construct a gene co-expression network to predict the hub genes affecting CRC recurrence and to inspect the regulatory network of hub genes and transcription factors (TFs). A total of 177 cases from the GSE17536 dataset were analyzed via weighted gene co-expression network analysis to explore the modules related to CRC recurrence. Functional annotation of the key module genes was assessed through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses. The protein and protein interaction network was then built to screen hub genes. Samples from the Cancer Genome Atlas (TCGA) were further used to validate the hub genes. Construction of a TFs-miRNAs–hub genes network was also conducted using StarBase and Cytoscape approaches. After identification and validation, a total of five genes (TIMP1, SPARCL1, MYL9, TPM2, and CNN1) were selected as hub genes. A regulatory network of TFs-miRNAs-targets with 29 TFs, 58 miRNAs, and five hub genes was instituted, including model GATA6-MIR106A-CNN1, SP4-MIR424-TPM2, SP4-MIR326-MYL9, ETS1-MIR22-TIMP1, and ETS1-MIR22-SPARCL1. In conclusion, the identification of these hub genes and the prediction of the Regulatory relationship of TFs-miRNAs-hub genes may provide a novel insight for understanding the underlying mechanism for CRC recurrence.

Identification of LncRNAs Associated With FOLFOX Chemoresistance in mCRC and Construction of a Predictive Model

Oxaliplatin, fluorouracil plus leucovorin (FOLFOX) regimen is the first-line chemotherapy of patients with metastatic colorectal cancer (mCRC). However, studies are limited regarding long non-coding RNAs (lncRNAs) associated with FOLFOX chemotherapy response and prognosis. This study aimed to identify lncRNAs associated with FOLFOX chemotherapy response and prognosis in mCRC patients and to construct a predictive model. We analyzed lncRNA expression in 11 mCRC patients treated with FOLFOX chemotherapy before surgery (four sensitive, seven resistant) by Gene Array Chip. The top eight lncRNAs (AC007193.8, CTD-2008N3.1, FLJ36777, RP11-509J21.4, RP3-508I15.20, LOC100130950, RP5-1042K10.13, and LINC00476) for chemotherapy response were identified according to weighted correlation network analysis (WGCNA). A competitive endogenous RNA (ceRNA) network was then constructed. The crucial functions of the eight lncRNAs enriched in chemotherapy resistance were mitogen-activated protein kinase (MAPK) and proteoglycans signaling pathway. Receiver operating characteristic (ROC) analysis demonstrated that the eight lncRNAs were potent predictors for chemotherapy resistance of mCRC patients. To further identify a signature model lncRNA chemotherapy response and prognosis, the validation set consisted of 196 CRC patients from our center was used to validate lncRNAs expression and prognosis by quantitative PCR (qPCR). The expression of the eight lncRNAs expression between CRC cancerous and adjacent non-cancerous tissues was also verified in the validation data set to determine the prognostic value. A generalized linear model was established to predict the probability of chemotherapy resistance and survival. Our findings showed that the eight-lncRNA signature may be a novel biomarker for the prediction of FOLFOX chemotherapy response and prognosis of mCRC patients.

MUC1 plays an essential role in tumor immunity of colorectal cancer stem cell vaccine

Increasing knowledge of colorectal cancer stem cells (CCSCs) and tumor microenvironment improves our understanding of cellular mechanisms involved in the immunity against colorectal cancer (CRC). Tumor associated antigens were evaluated via RNA-seq and bioinformatics analysis, evoking promising targets for tumor immunotherapy. MUC1 has been demonstrated to participate in the maintenance, tumorigenicity, glycosylation and metastasis of CCSCs, which may provide a new priority for CSC vaccination. In the present study, the vaccination with CCSCs with high expression of MUC1 was evaluated in a murine model for the vaccine's immunogenicity and protective efficacy against CRC. CD133⁺ CCSCs were isolated from SW620 cell line using a magnetic-activated cell sorting system, and shMUC1 was further used to knock down the expression of MUC1 in CD133⁺ CCSCs. Mice were subcutaneously immunized with the cell lysates of CCSCs and shMUC1 CCSCs, followed by a challenge with SW620 cells at ten days after final vaccination. The results indicated CCSC vaccine significantly reduced the tumor growth via a target killing of CCSCs as evidenced by a decrease of CD133⁺ cells and ALDH⁺ cells in tumors. Moreover, CCSC vaccine resulted in the elevated NK cytotoxicity, production of perforin, granzyme B, IFN-γ, memory B cells, and anti-MUC1 antibodies. Of note, MUC1 knockdown partly impaired the anti-tumor efficacy of CCSC vaccine. Importantly, the CCSC vaccine has no toxic damage to organs. Overall, CCSC vaccine could serve as a potent and safe vaccine for CRC treatment, and MUC1 might play an essential role in CCSC vaccine.