Using Computational Drug-Gene Analysis to Identify Novel Therapeutic Candidates for Retinal Neuroprotection

Our plain-English summary of Using Computational Drug-Gene Analysis to Identify Novel Therapeutic Candidates for Retinal Neuroprotection, by E. Xie, U. Nadeem, B. Xie, et al (International Journal of Molecular Sciences, 2022).

The study

• Authors: E. Xie, U. Nadeem, B. Xie, M. D'Souza, D. Sulakhe, D. Skondra
• Journal: International Journal of Molecular Sciences
• Published: 2022;23(20)
• Cited by: 5 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Retinal cell death is responsible for irreversible vision loss in many retinal disorders. No commercially approved treatments are currently available to attenuate retinal cell loss and preserve vision. We seek to identify chemicals/drugs with thoroughly-studied biological functions that possess neuroprotective effects in the retina using a computational bioinformatics approach. We queried the National Center for Biotechnology Information (NCBI) to identify genes associated with retinal neuroprotection. Enrichment analysis was performed using ToppGene to identify compounds related to the identified genes. This analysis constructs a Pharmacome from multiple drug-gene interaction databases to predict compounds with statistically significant associations to genes involved in retinal neuroprotection. Compounds with known deleterious effects (e.g., asbestos, ethanol) or with no clinical indications (e.g., paraquat, ozone) were manually filtered. We identified numerous drug/chemical classes associated to multiple genes implicated in retinal neuroprotection using a systematic computational approach. Anti-diabetics, lipid-lowering medicines, and antioxidants are among the treatments anticipated by this analysis, and many of these drugs could be readily repurposed for retinal neuroprotection. Our technique serves as an unbiased tool that can be utilized in the future to lead focused preclinical and clinical investigations for complex processes such as neuroprotection, as well as a wide range of other ocular pathologies.

In plain English

This was a computational study of genes and drug databases. It did not study people, animals or asbestos exposure. The authors searched the NCBI Gene database and identified 117 genes linked to protecting cells in the retina (the light-sensing layer of the eye). They used the ToppGene software to find chemicals statistically associated with these genes in drug–gene databases. Of 77,146 candidate compounds, 14,984 passed the significance cutoff, a count that included duplicates. The authors manually removed duplicates and compounds they considered harmful or without clinical use, including asbestos, particulate matter and ozone, and kept 25. Asbestos therefore appears only as a compound that was excluded. The 25 retained compounds included antioxidants such as N-acetylcysteine, plant compounds such as curcumin, the diabetes drug metformin, cholesterol-lowering statins and blood-pressure drugs. The authors stated that their results are meant only to guide future laboratory and clinical studies. They noted that the databases contain false positives and draw on data from animals and cell lines. The authors concluded that the approach could help identify existing drugs worth investigating for new uses.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Computational (bioinformatics) drug–gene enrichment analysis; no humans or animals studied
• Who was studied: 117 genes associated with retinal neuroprotection, compiled from the NCBI Gene database (accessed 24 August 2021), analyzed against 77,146 drug annotations in ToppGene
• Exposure: Asbestos was not studied. It was one of the compounds manually removed from the list of results as harmful to human health or lacking clinical use.
• Main results:
  • Of 77,146 candidate compounds, the enrichment analysis returned 14,984 chemical compounds at a false-discovery-rate-adjusted p-value cutoff of 0.05. This number included redundancies.
  • After removing redundancies and known harmful compounds (including asbestos), the authors manually selected the top 25 compounds.
  • In the text, N-acetylcysteine (p = 3.67 × 10^-56) was the compound most significantly associated with retinal neuroprotection, affecting 51 of the 117 genes. The table gives different figures for acetylcysteine (p 3.67 × 10^-55, 59 hits).
  • Curcumin was the most significant polyphenol nutrient identified, affecting 58 of the 117 genes.
  • Metformin, simvastatin and nifedipine were associated with 31, 53 and 24 of the 117 genes, respectively, according to the text. The table lists 32 hits for metformin.
  • Genes central to the drug–gene network included SOD2, TP53, NFE2L2, NOS2, BCL2, BAX, TNF, CASP3, TGFB, AKT1 and IL6.
• The authors' conclusion: The authors concluded that their approach can serve as a guide for identifying future preclinical and clinical targets and existing drugs that could be repurposed, such as metformin and statins.
• Limitations the authors state:
  • The source datasets (STITCH, CTD) contain inherent errors, including false positives and indirect chemical–gene associations.
  • Predictions depend on published literature, so heavily studied conditions and drugs may be overrepresented.
  • The analysis did not account for whether gene variants were protective or risk variants.
  • Some predicted interactions came from animal models and transformed cell lines and may not apply to humans.
  • As a computational analysis, the results are only meant to guide future preclinical and clinical studies.

Related studies

• Novel LRRK2 mutations and other rare, non-BAP1-related candidate tumor predisposition gene variants in high-risk cancer families with mesothelioma and other tumors — M. Cheung, Y. Kadariya, et al. (2021)
• Identification of Potential Hub Genes and Therapeutic Drugs in Malignant Pleural Mesothelioma by Integrated Bioinformatics Analysis — X. Zhang, L. Yang, et al. (2020)
• Promising investigational drug candidates in phase I and phase II clinical trials for mesothelioma — A. Guazzelli, E. Bakker, et al. (2017)
• Interleukin-6 as a therapeutic target in candidate inflammatory diseases — N. Nishimoto (2010)
• Identification of novel candidate oncogenes and tumor suppressors in malignant pleural mesothelioma using large-scale transcriptional profiling — G.J. Gordon, G.N. Rockwell, et al. (2005)
• Combined Analysis of SNP Array Data Identifies Novel CNV Candidates and Pathways in Ependymoma and Mesothelioma — G. Wajnberg, B.S. Carvalho, et al. (2015)

Browse more research in International Journal of Molecular Sciences, or return to medical research.