Association of Extracellular Signal-Regulated Kinase Genes With Myopia: A Longitudinal Study of Chinese Children

Abstract

Objective: The present study was designed to investigate whether the extracellular signal-regulated kinase (ERK) signaling pathway, a downstream component of dopamine signaling, is involved in myopia among Chinese children.

Methods: During a 3.5-year follow-up, 488 primary school students were enrolled in this study. Non-cycloplegic spherical equivalent refraction (SE) and other ocular parameters were assessed. Four variants of four genes in the ERK signaling pathway were selected: RASGRF1 rs6495367, PTPN5 rs1550870, PTPRR rs11178469, and PDGFRA rs6554163. SNPscan was used to genotype single-nucleotide polymorphisms (SNPs). PLINK software was used to assess the associations of the genetic variants with the occurrence or development of myopia, SE, and other ocular parameters. We created a protein-protein interaction (PPI) network and microRNA (miRNA)-gene network using String and Cytoscape and conducted enrichment analyses on the genes in these networks.

Results: In total, 426 children (baseline age: 7.28 ± 0.26 years; 236 (55.4%) boys and 190 girls) wereenrolled. After adjusting for confounding factors with 10,000 permutations, children with the CT or TT genotype of PTPN5 rs1550870 were more susceptible to myopia than those with the CC genotype (adjusted p = 0.011). Additionally, PTPN5 rs1550870 was correlated with significant myopic shift and increasing axial length (AL) and lens thickness (LT) but had a negative effect on central corneal thickness (CCT). RASGRF1 rs6495367 was negatively associated with myopic shift (additive: adjusted p = 0.034; dominant: adjusted p = 0.020), myopic SE and AL. PDGFRA rs6554163 TA or AA was negatively associated with increasing LT (adjusted p = 0.033). Evaluation of the effects of SNP-SNP combinations on incident myopia revealed a statistically significant one-locus model: PTPN5 rs1550870 [cross-validation consistency (CVC) = 10/10, adjusted p = 0.0107]. The genes in the PPI and miRNA-gene interaction networks were subjected to enrichment analyses, which suggested that these genes are involved mainly in eye development and dopaminergic synapse-related processes.

Conclusion: We identified genetic variants of crucial ERK signaling pathway genes that were significantly correlated with myopia and ocular parameter alterations in Chinese children. A combination of gene and miRNA functional analyses with enrichment analyses highlights the regulatory effects associated with ocular development and dopamine biological functions. This study offers novel clues to understand the role of dopamine in the molecular mechanisms of myopia.

Keywords: association analysis; biological networks; ocular parameters; schoolchildren myopia; single nucleotide polymorphism.

FIGURE 1

The three best models for predicting incident myopia given by GMDR analysis. (A) One-locus model of PTPN5 rs1550870. (B) Two-loci model of PTPN5 rs1550870-PTPRR rs11178469. (C) Three-loci model of RASGRF1 rs6459367-PTPRR rs11178469-PDGFRA rs6554163. A grid represents the specific combinations of SNP-SNP interactions. High-risk genotypes are shown in dark gray, while low-risk genotypes are shown in light gray. All the bars on the left of each grid represent children who were not myopic at baseline but developed myopia in the follow-up period, while the bars on the right represent children who did not develop myopia during the complete follow-up period.

FIGURE 2

The three best models for predicting significant myopic shift given by GMDR analysis. (A) One-locus model of RASGRF1 rs6495367. (B) Two-loci model of PTPRR rs11178469-PDGFRA rs6554163. (C) Three-loci model of RASGRF1 rs6459367-PTPN5 rs1550870-PTPRR rs11178469. A grid represents the specific combinations of SNP-SNP interactions. High-risk genotypes are shown in dark gray, while low-risk genotypes are shown in light gray. All the bars on the left of each grid represent cases, while the bars on the right represent controls.

FIGURE 3

The protein-protein interaction network. The genes with red squares (including PTPN5, RASGRF1, and PDGFRA) were input into String.

FIGURE 4

Venn diagrams of target miRNAs in different databases including miRWalk, miRDB, and mirDIP. (A) The miRNA distribution of PDGFRA in each database. (B) The miRNA distribution of PTPN5 in each database. (C) The miRNA distribution of RASGRF1 in each database.

FIGURE 5

The miRNA-gene regulatory network. The yellow nodes represent 28 miRNAs, and the blue nodes represent the 51 genes that have direct or indirect interactions with PDGFRA, RASGRF1, and PTPN5. The edges represent the interaction relationships between miRNAs and genes.