Non-coding RNAs and related molecules associated with form-deprivation myopia in mice

Abstract

The role of miRNAs and its regulatory mechanism in myopia are indeterminate. Our study aimed to investigate potential myopia-associated non-coding RNAs and related molecules by performing a comprehensive bioinformatic analysis of miRNA expression profile of mice with form-deprivation myopia (FDM). Differentially expressed miRNAs in two raw microarray data sets (GSE58124 and GSE84220) from Gene Expression Omnibus (GEO) database were comprehensively analysed using GEO2R. Target genes were predicted using miRDB and enriched with Metascape online tool. Protein-protein interaction (PPI) networks were constructed utilizing STRING and Cytoscape. Significant differentially expressed miRNAs were validated by real-time polymerase chain reaction (qRT-PCR) using RNA extracted from monocular FDM ocular tissues. As result, we identified three upregulated miRNAs (mmu-miR-1936, mmu-miR-338-5p, and mmu-miR-673-3p) significantly associated with myopia in the two microarray data sets (p < 0.05 and |Log (Fold Change) |>1). GO functional analysis suggested these three miRNAs were targeted in genes mostly enriched in morphogenesis and developmental growth of retinal tissues. Enrichment analysis revealed top eight transcription factors, including PAX6 and Smad3, related to myopia. Ten hub genes, including Rbx1, Fbxl3, Fbxo27, Fbxl7, Fbxo4, Cul3, Cul2, Klhl5, Fbxl16 and Klhl42, associated with ubiquitin conjugation were identified. qRT-PCR confirmed the increased expression of mmu-miR-1936 and mmu-miR-338-5p (p < 0.05), but no statistical difference was observed in mmu-miR-673-3p expression in myopic retinas. Our findings indicated mmu-miR-1936, mmu-miR-338-5p and mmu-miR-673-3p upregulation may be associated with myopia development via post-transcriptional gene regulation, and identified potential molecules that could be further explored in future studies of the mechanism in myopia.

Keywords: bioinformatics analysis; myopia; non-coding RNAs.

FIGURE 1

Flow chart of the current study. DE‐miRs, differentially expressed miRNAs; GO, gene ontology; KEGG, Kyoto Encyclopedia of genes and genomes; TF, transcription factors; and PPI, protein‐protein interaction

FIGURE 2

Heat maps of differentially expressed miRNAs between a treated eye (myopia retina) and contralateral untreated eye (control retina) in two microarray profiles. (A) Heat map of GSE84220 data set; (B) heat map for GSE58124 data set (only 49 unique miRNAs are shown of a total of 115 microRNAs)

FIGURE 3

Volcano plots and Venn diagram of this study. (A, B) Volcano plots of the distribution of all differentially expressed miRNAs that were selected with a p value < 0.05 and |Log (Fold Change) | >1. Upregulated, downregulated and not significantly changed miRNAs are marked, respectively, as red, green and black dots. From the retina tissues in the data set, a total of 115 miRNAs were differentially expressed in the GSE58124 data sets, and all were upregulated. Sixteen differentially expressed miRNAs (5 upregulated and 11 downregulated) were screened out from the GSE84220 data sets. (C) Venn diagram of co‐expressed miRNAs within GSE58124 and GSE84220, with consensus defined as DE‐miRs. (D) mmu‐miR‐1936, mmu‐miR‐338‐5p and mmu‐miR‐673‐3p were screened out as DE‐miRs; their specific p and logFC values are listed in the two data sets

FIGURE 4

Enrichment analysis of genes predicted from DE‐miRs using Metascape including (A) molecular functions, (B) cellular components, (C) biological processes, (D) KEGG pathways analysis, (E) transcription factor enrichment and (F) tissue/cell expression analysis. DE‐miRs, differentially expressed miRNAs. The length of each bar is equal to −log10(P)

FIGURE 5

Protein network analysis of 1340 unique target genes. (A) Protein‐protein‐interaction (PPI) network constructed using STRING database and visualized by NetworkAnalyst. The degree of connectivity is shown by the size of the node, with larger circles indicating greater connectivity. The colour of the node represents the value of the betweenness from red (large) to purple (small). (B, C) Top two clusters of functional modules as determined by using the Molecular Complex Detection (MCODE) plug‐in of Cytoscape (version 3.8.0)

FIGURE 6

Network of the top 10 hub genes based on cytoHubba plug‐in of Cytoscape. (A) The expanded subnetwork is displayed. The warm colours represent the genes; the blue circles represent genes that directly interact with them. (B) The shortest path is displayed. The darker the node, the higher the ranking using the degree algorithm. (C) Network map of miRNA–hub gene interactions. Rbx1, Fbxl3, Fbxo27, Cul3, Cul2, Klhl42, Klhl5 and Fbxl16 are target genes of mmu‐miR‐338‐5p; Fbxo4 and Fbxl7 are target genes of mmu‐miR‐1936; mmu‐miR‐673‐3p has no target genes that is the core gene in the molecular network

FIGURE 7

Validation of the relative expressed level of mmu‐miR‐1936, mmu‐miR‐338‐5p and mmu‐miR‐673‐3p. *p < 0.05. The fold changes were normalized to small nuclear RNA U6 and calculated using the 2^−∆∆CT method. The validation results show that the expression levels from myopia of mmu‐miR‐1936 and mmu‐miR‐338‐5p in retinal tissues were statistically significantly higher than those of controls