Integrated Analysis of Distant Metastasis-Associated Genes and Potential Drugs in Colon Adenocarcinoma

Miaowei Wu, Weiyang Lou, Meng Lou, Peifen Fu, Xiao‐Fang YuFrontiers in Oncology 2020 · DOI

TL;DR

This study identifies key genes associated with distant metastasis in colon adenocarcinoma (COAD) and proposes three existing drugs as potential therapeutic agents to combat this deadly aspect of the disease. By analyzing transcriptomic data, researchers pinpointed four genes (LEP, DLX2, CLSTN2, and REG3A) that are significantly linked to metastasis and patient survival, offering new targets for drug development.

What Problem Does This Paper Address?

Distant metastasis is the primary cause of mortality in colon adenocarcinoma (COAD) patients, despite available treatments for metastatic disease. However, the specific genetic signatures that drive this metastasis and lead to poor prognosis remain largely unclear. This knowledge gap hinders the development of effective therapeutic strategies. This paper aims to bridge this gap by identifying genes uniquely expressed in metastatic COAD and exploring potential drug candidates that target these genes, ultimately seeking to improve patient outcomes.

Methods

The researchers employed a multi-pronged bioinformatics approach. They began by analyzing transcriptomic data from The Cancer Genome Atlas (TCGA) cohort to identify differentially expressed microRNAs (DEmiRNAs) and messenger RNAs (DEGs) between COAD patients with and without distant metastasis (M1 vs. M0 stages). This initial screening utilized the "edgeR" package in R.

Next, they predicted potential target genes for the identified DEmiRNAs using the mirDIP v4.1 database. The overlap between these predicted targets and the DEGs yielded a set of "candidate DEGs." To further refine these, they constructed protein-protein interaction (PPI) networks using the STRING database and identified "hub genes" with high connectivity using Cytoscape.

The prognostic significance of these candidate genes was then evaluated using the Gene Expression Profiling Interactive Analysis (GEPIA) and PROGgeneV2 databases. Survival analyses, including Kaplan-Meier curves, were performed to assess the association between gene expression levels and patient survival.

Finally, potential therapeutic drugs were predicted using the Connectivity Map (CMap) database, which links gene expression profiles to drug perturbations. These drug candidates were then further validated using the Comparative Toxicogenomics Database (CTD) for their known associations with colon cancer.

Datasets

  • The Cancer Genome Atlas (TCGA) cohort: Transcriptomic (RNA-Seq) and clinical data for COAD patients (n=514, with 334 non-distant metastatic and 62 distant metastatic). This served as the primary discovery dataset.
  • GEPIA database: Used for validating the prognostic significance of candidate genes.
  • PROGgeneV2 database: Also used for validating the prognostic relevance of key genes.
  • mirDIP v4.1 database: Used for predicting miRNA-target gene interactions.
  • STRING database: Used for constructing protein-protein interaction networks.
  • Cytoscape software: Used for analyzing PPI networks and constructing miRNA-gene regulatory networks.
  • DIANA-miRPath (v3.0): Used for KEGG pathway enrichment analysis of DEmiRNAs.
  • Metascape: Used for functional annotation and pathway enrichment analysis of candidate DEGs.
  • AmiGO database: Used for analyzing Gene Ontology (GO) terms of prognosis-related genes.
  • Comparative Toxicogenomics Database (CTD): Used for validating drug-cancer associations.
  • ONCOMINE database: Used for analyzing the expression levels of the identified genes in colorectal cancer versus normal tissues.
  • Tumor Immune Estimation Resource (TIMER): Used for analyzing the correlation between gene expression and immune cell infiltration.
  • Connectivity Map (CMap) database: Used for predicting potential therapeutic drugs.

Key Insights and Ideas

  • Metastasis has a distinct genetic signature: The study strongly suggests that distant metastasis in COAD is driven by a specific set of genes, differentiating it from non-metastatic disease. This moves beyond simply comparing tumor vs. normal tissue.
  • Four key prognostic genes identified: LEP, DLX2, CLSTN2, and REG3A emerged as significant players in COAD metastasis and patient survival, offering potential biomarkers and therapeutic targets.
  • REG3A as a potential tumor suppressor in metastasis: While some genes were associated with worse outcomes, REG3A showed a correlation with better survival and was downregulated in metastatic tumors, suggesting a potential tumor-suppressive role in the context of metastasis, despite potentially promoting proliferation.
  • Existing drugs as potential anti-metastasis agents: The identification of ajmaline, TTNPB, and dydrogesterone as drugs that may target these metastasis-associated genes is a significant finding, offering a faster path to clinical translation compared to developing entirely new drugs.
  • Combined gene signature outperforms individual genes: A risk score derived from the four identified genes demonstrated better predictive accuracy for patient survival than any single gene alone, highlighting the power of multi-gene panels.

Key Results

  • 127 significantly differentially expressed genes (DEGs) were identified between metastatic and non-metastatic COAD patients.
  • Four prognosis-related genes were selected: LEP, DLX2, CLSTN2, and REG3A.
  • LEP and DLX2 were identified as "risky genes" associated with worse overall survival (OS).
  • CLSTN2 was significantly associated with worse OS and metastasis-free survival (MFS).
  • REG3A was significantly associated with better OS and MFS.
  • A combined gene signature panel (LEP, REG3A, DLX2) showed superior predictive accuracy for OS (5-year AUC of 0.672) compared to individual genes.
  • Three potential drugs (ajmaline, TTNPB, and dydrogesterone) were predicted to target these genes and potentially suppress COAD metastasis.
  • The expression of the four prognosis-related genes was significantly higher in colorectal cancer tissues compared to normal tissues.
  • The expression of these genes showed correlations with various immune cell infiltrates, suggesting a link between the tumor microenvironment and metastasis.

Limitations and Future Work

  • Limited sample size for M1 stage: The study notes that the number of patients in the M1 stage (n=62) was relatively small, which could impact the robustness of some findings.
  • In vitro and in vivo validation needed: While the study identified potential drug candidates, further experimental validation through in vitro and in vivo studies is crucial to confirm their efficacy and safety in treating COAD metastasis.
  • Understanding REG3A's dual role: The paper acknowledges that while REG3A is downregulated in metastatic tumors and associated with better survival, it might also promote tumor cell proliferation. Further research is needed to fully elucidate its complex role in metastasis.
  • Exploring larger cohorts: The authors suggest that analyzing larger cohorts and performing in vitro and in vivo experiments would improve the understanding of the roles of these four genes in colorectal cancer metastasis.
  • Investigating drug mechanisms: Future studies should focus on exploring the specific molecular targets and therapeutic effects of the identified drugs (ajmaline, TTNPB, and dydrogesterone) in the context of COAD.

Summary updated

Abstract

<b>Background:</b> Most colon adenocarcinoma (COAD) patients die of distant metastasis, though there are some therapies for metastatic COAD. However, the genes exclusively expressed in metastatic COAD remain unclear. This study aims to identify prognosis-related genes associated with distant metastasis and develop therapeutic strategies for COAD patients. <b>Methods:</b> Transcriptomic data from The Cancer Genome Atlas (TCGA; <i>n</i> = 514) cohort were analyzed as a discovery dataset. Next, the data from the GEPIA database and PROGgeneV2 database were used to validate our analysis. Key genes were identified based on the differential miRNA and mRNA expression with respect to M stage. The potential drugs targeting candidate differentially expressed genes (DEGs) were also investigated. <b>Results:</b> A total of 127 significantly DEGs in patients with distant metastasis compared with patients without distant metastasis were identified. Then, four prognosis-related genes (LEP, DLX2, CLSTN2, and REG3A) were selected based on clustering analysis and survival analysis. Finally, three compounds targeting the candidate DEGs, including ajmaline, TTNPB, and dydrogesterone, were predicted to be potential drugs for COAD. <b>Conclusions:</b> This study revealed that distant metastasis in COAD is associated with a specific group of genes, and three existing drugs may suppress the distant metastasis of COAD.
BibTeX
@inproceedings{wu2020integrated,
  author = {Miaowei Wu and Weiyang Lou and Meng Lou and Peifen Fu and Xiao‐Fang Yu},
  title = {{Integrated Analysis of Distant Metastasis-Associated Genes and Potential Drugs in Colon Adenocarcinoma}},
  booktitle = {Frontiers in Oncology},
  year = {2020},
  doi = {10.3389/fonc.2020.576615},
}
Figure