Bioinformatics Analysis of Lactylation-related Biomarkers and Potential Pathogenesis Mechanisms in Age-related Macular Degeneration

Abstract

Background: Lactylation is increasingly recognized to play a crucial role in human health and diseases. However, its involvement in age-related macular degeneration (AMD) remains largely unclear.

Objectives: The aim of this study was to identify and characterize the pivotal lactylation-related genes and explore their underlying mechanism in AMD.

Methods: Gene expression profiles of AMD patients and control individuals were obtained and integrated from the GSE29801 and GSE50195 datasets. Differentially expressed genes (DEGs) were screened and intersected with lactylation-related genes for lactylation-related DEGs. Machine learning algorithms were used to identify hub genes associated with AMD. Subsequently, the selected hub genes were subject to correlation analysis, and reverse transcription quantitative real-time PCR (RT-qPCR) was used to detect the expression of hub genes in AMD patients and healthy control individuals.

Results: A total of 68 lactylation-related DEGs in AMD were identified, and seven genes, including HMGN2, TOP2B, HNRNPH1, SF3A1, SRRM2, HIST1H1C, and HIST1H2BD were selected as key genes. RT-qPCR analysis validated that all 7 key genes were down-regulated in AMD patients.

Conclusion: We identified seven lactylation-related key genes potentially associated with the progression of AMD, which might deepen our understanding of the underlying mechanisms involved in AMD and provide clues for the targeted therapy.

Keywords: Age-related macular degeneration; bioinformatics; hub genes; immune infiltration; lactylation; machine learning.

Fig. (1)

Data processing and selection of DEGs. (A-B) PCA plot of GSE50195 and GSE29801 before (A) and after (B) merging. Boxplots exhibited the batch effects of expression profiles (C) before and (D) after homogenization. (E) Volcano plots showed the gene expression pattern in the merged gene expression matrix. (F) Heatmap exhibited the top 20 differently expressed genes.

Fig. (2)

GO and KEGG analyses for DEGs. (A-C) DEGs were subject to GO enrichment analysis and the enriched terms in (A) biological process (BP), (B) cellular component (CC), and (C) molecular function (MF) were presented. (D) The KEGG enrichment results of DEGs.

Fig. (3)

Screening of lactylation-associated DEGs. (A-B) A total of 17 lactylation-associated up-regulated genes and 51 down-regulated genes were presented in Venn diagrams. (C-F) GO and KEGG analyses of lactylation-associated DEGs.

Fig. (4)

Expression pattern of lactylation-associated DEGs between AMD and control samples. (A) Volcano plot and (B) Heatmap exhibited the expression pattern of lactylation-associated DEGs. (C) Expression comparison of lactylation-associated DEGs between AMD and control samples.

Fig. (5)

Hub gene selection. (A) LASSO algorithm, (B) random forest algorithm and (C) SVM-RFE algorithm were employed for selecting the feature genes. (D) A Venn diagram showed the common feature genes selected by three algorithms. (E) Expression correlation of selected hub genes was analyzed and visualized using a chord plot. The red lines indicate positive correlations, while the green lines signify negative correlations. The higher the color depth, the stronger the correlation.

Fig. (6)

Evaluation of immune cell enrichment. (A) The enrichment correlation between immune cells. (B) Abundance of immune cells was compared between AMD and control groups. (C) The correlation between hub gene expression and immune cell enrichment. Only results with statistical significance (p < 0.05) were shown.

Fig. (7)

Consensus molecular clustering based on hub genes. (A) Consensus matrix heatmap illustrates the identification of several clusters and their correlation area. (B) Differential expression of hub genes between cluster (A) and (B). (C) Heatmap of hub genes.

Fig. (8)

Difference between clusters A and B. (A) PCA plot of AMD patients in two clusters. (B) Volcano plots of DEGs between clusters A and B. (C-D) GO and KEGG enrichment analyses of DEGs between clusters (A) and (B). The top 20 were displayed.

Fig. (9)

The co-expressed genes with hub genes in AMD samples. The heatmap showed top 50 co-expressed genes with seven hub genes in AMD samples.

Fig. (10)

GSEA of key genes. The top 20 GSEA terms of seven hub genes were presented.

Fig. (11)

Construction of the miRNAs-TFs- hub genes networks. The miRNAs and TFs potentially regulating the 7 hub genes were searched on the RegNetwork database. The red color indicates the hub genes, and the blue color indicates the miRNAs or TFs. The lines represent predicted interactions.

Fig. (12)

Expression validation of seven key genes. RT-qPCR validated the down-regulation of (A) TOP2B, (B) HMGN2, (C) SRRM2, (D) HNRNPH1, (E) SF3A1, (F) HIST1H1C and (G) HIST1H2BD in the serum of AMD patients compared with control individuals.