The Effects of Different Histone Deacetylase Inhibitors on Hypoxia-induced Intestinal Epithelial Barrier Dysfunction and a Preliminary Exploration of the Mechanisms

Abstract

Abstract Aim We aimed to compare the effects and underlying mechanisms of 6 different representative HDAC inhibitors (MS-275, MC-1568, Tubastatin-A (Tub-A), EX-527, Vorinostat Suberoylanilide hydroxamic acid (SAHA) and Valproic Acid (VPA)) in hypoxia-induced intestinal epithelial barrier dysfunction based on in vitro Caco-2 cell experiments and network pharmacology combined with bioinformatics analysis. Methods In vitro study: Human intestinal epithelial cells (Caco-2) were divided into 3 groups: (1) sham (normoxia); (2) vehicle (anoxia, with DMSO treatment), and (3) treatment (anoxia, and treatment with 5 μmol/L MS-275, MC-1568, Tub-A, EX-527, SAHA and VPA). After 12 h in an anoxic chamber, the cells were examined for epithelial permeability, cellular viability, cytotoxicity and tight junction component Claudin-3 protein expression. Network pharmacology study: The structures of HDAC inhibitors were downloaded from NCBI-PubChem compound databases (https://www.ncbi.nlm.nih.gov/pccompound/). The possible targets of the core active ingredients of HDAC inhibitors were obtained from PharmMapper (http://www.lilab-ecust.cn/pharmmapper/index.html) and SuperPred (https://prediction.charite.de). The compound-target network was constructed by Cytoscape software version 3.8.2. Full gene names corresponding to these human-derived target proteins were obtained from the UniProt database (https://www.uniprot.org/). An UpSet diagram of all the human-derived target proteins obtained from PharmMapper and SuperPred was constructed using R software version 3.6.3 (R package: UpSetR (v1.4.0)). Functional annotation information was obtained from Gene Ontology (GO) (http://geneontology.org/) and KEGG (https://www.kegg.jp/) databases. To predict the underlying biological function and corresponding pathways of these significant genes, the DAVID database and Merascape were introduced to perform functional enrichment analysis, including three GO terms (BP: biological process; CC: cellular component; MF: molecular function) and KEGG pathways. ​Results In vitro study: In cultured Caco-2 cells, anoxia markedly increased the permeability of Caco-2 monolayer cells (P=0.000), while Tub-A, SAHA and VPA significantly attenuated the alteration (P=0.021, 0.000, 0.017). Anoxia significantly decreased cellular viability (P=0.000) and increased cytotoxicity (P=0.000) compared to the sham group, while MC-1568, Tub-A, EX-527 and VPA treatment offered significant protection. Moreover, the expression of Claudin-3 was markedly decreased in vitro compared to that in the sham group, whereas this decrease was significantly attenuated by Tub-A (P =0.002). Network pharmacology study: The “Herbs-Components-Targets” network of HDAC inhibitors from the PharmMapper database included 116 nodes and 287 edges. In the SuperPred database, the “Herbs-Components-Targets” network included 124 nodes and 158 edges. Six genes were selected by taking the intersection of 6 HDAC inhibitor key target gene sets from PharmMapper. Twelve genes were selected from SuperPred by taking the intersection of 4 HDAC inhibitor key target gene sets. GO and KEGG enrichment analyses were conducted to identify 6 target genes from PharmMapper and 12 target genes from SuperPred. ​Conclusion HDAC inhibitors can promote cellular viability and prevent the loss of intestinal TJ proteins during anoxia. Based on the existing recognized public databases and bioinformatics analysis of the structure, target proteins and functions of different HDAC inhibitors, specific genes and functional pathways could be involved in the underlying mechanism.