Alterations in Faecal and Serum Metabolic Profiles in Patients with Neovascular Age-Related Macular Degeneration

Abstract

Neovascular age-related macular degeneration (nAMD) is a common and multifactorial disease in the elderly that may lead to irreversible vision loss; yet the pathogenesis of AMD remains unclear. In this study, nontargeted metabolomics profiling using ultra-performance liquid chromatography coupled with Q-Exactive Orbitrap mass spectrometry was applied to discover the metabolic feature differences in both faeces and serum samples between Chinese nonobese subjects with and without nAMD. In faecal samples, a total of 18 metabolites were significantly altered in nAMD patients, and metabolic dysregulations were prominently involved in glycerolipid metabolism and nicotinate and nicotinamide metabolism. In serum samples, a total of 29 differential metabolites were founded, involved in caffeine metabolism, biosynthesis of unsaturated fatty acids, and purine metabolism. Two faecal metabolites (palmitoyl ethanolamide and uridine) and three serum metabolites (4-hydroxybenzoic acid, adrenic acid, and palmitic acid) were selected as potential biomarkers for nAMD. Additionally, the significant correlations among dysregulated neuroprotective, antineuroinflammatory, or fatty acid metabolites in faecal and serum and IM dysbiosis were found. This comprehensive metabolomics study of faeces and serum samples showed that alterations in IM-mediated neuroprotective metabolites may be involved in the pathophysiology of AMD, offering IM-based nutritional therapeutic targets for nAMD.

Keywords: age-related macular degeneration; faecal; metabolomics; serum.

Figure 1

Experimental design of this study for investigating faecal and serum metabolic alterations in nAMD patients. Abbreviations: nAMD, neovascular age-related macular degeneration.

Figure 2

Multivariate analysis of faecal metabolomics of the nAMD and non-AMD groups detected in positive and negative ion modes. (A,B) PCA score plots were obtained from UPLC-MS data in positive and negative ion modes, respectively. (C,D) Supervised PLS-DA score plots in positive (C) and negative ion (D) modes, respectively. Abbreviations: nAMD, neovascular age-related macular degeneration; PCA, principal component analysis; PLS-DA, partial least squares discriminant analysis; UPLC-MS, ultra-performance liquid chromatography coupled with Q-Exactive Orbitrap mass spectrometry.

Figure 3

Differential faecal metabolic features of nAMD patients. (A,B) Hierarchical clustering heatmaps of faecal metabolites with significant differences (p-value < 0.05 and VIP score >1) between nAMD patients and non-AMD controls detected in positive and negative ion modes, respectively. The red or blue colour in each cell represents whether the levels of each metabolite were high or low, respectively. (C) Changes in faecal metabolomic pathways, mainly including glycerolipid metabolism, nicotinate and nicotinamide metabolism, and pentose and glucuronate interconversions. The bubbles represent the enriched pathways of metabolites and the color represents metabolites intensity from low (light yellow) to high (deep red). (D,E) The relative concentrations of palmitoyl ethanolamide (positive ion mode) and uridine (negative ion mode) in the nAMD group and non-AMD group. The color of dots represents different group: red, AMD; green, non-AMD. *, p-value < 0.05; **, p-value < 0.01. (F,G) ROC curve of two selected faecal biomarkers with the best diagnostic capacity. Abbreviations: nAMD, neovascular age-related macular degeneration; VIP, variable importance in projection; ROC, receiver operating characteristic; AUC, area under the curve; CI, confidence interval.

Figure 4

Multivariate analysis of serum metabolomics of the nAMD and non-AMD groups detected in positive and negative ion modes. (A,B) PCA score plots were obtained from UPLC-MS data in positive and negative ion modes, respectively. (C,D) Supervised PLS-DA score plots in positive and negative ion modes, respectively. Abbreviations: nAMD, neovascular age-related macular degeneration; PCA, principal component analysis; PLS-DA, partial least squares discriminant analysis; UPLC-MS, ultra-performance liquid chromatography coupled with Q-Exactive Orbitrap mass spectrometry.

Figure 5

Differential serum metabolic features of nAMD patients. (A,B) Hierarchical clustering heatmaps of serum metabolites with significant differences (p-value < 0.05 and VIP score >1) between nAMD patients and non-AMD controls detected in positive and negative ion modes, respectively. The red or blue colour in each cell represents whether the level of each metabolite was high or low, respectively. (C) Changed serum metabolomic pathways, mainly including glycerolipid metabolism, nicotinate and nicotinamide metabolism, and pentose and glucuronate interconversions. The bubbles represent the enriched pathways of metabolites and the color represents metabolites intensity from low (light yellow) to high (deep red). (D–F) The relative levels of 4-hydroxybenzoic acid (positive ion mode) and adrenic acid and palmitic acid (negative ion mode) in the nAMD group and non-AMD group. The color of dots represents different group: red, AMD; green, non-AMD. **, p-value < 0.01; ***, p-value < 0.001. (G,H) ROC curve of selected serum biomarkers with the best diagnostic capacity. Abbreviations: nAMD, neovascular age-related macular degeneration; VIP, variable importance in projection; ROC, receiver operating characteristic; AUC, area under the curve; CI, confidence interval.

Figure 6

Correlations between altered metabolites (AUC > 0.7) and the intestinal microbiota of nAMD patients and non-AMD controls. Spearman’s rank correlation between faecal and serum metabolites and the intestinal microbiota. The line width corresponds to the Spearman correlation, and line colour denotes the statistical significance based on 999 permutations. Pairwise comparisons of metabolites are also shown, with a colour gradient denoting Spearman’s correlation coefficient. The colour and thickness of the lines represent the strength of the correlation between the metabolites and the differential microbiome. The thicker the line is, the stronger the correlation. The orange colour represents strong significance, p-value < 0.01. The correlation heatmap in the upper right corner indicates the correlation relationship between different important metabolites. Numbers indicate the correlation r value and asterisks indicate statistical significance (*** p-value < 0.001; ** p-value < 0.01; and * p-value < 0.05).