Tear Lipid Metabolites As Potential Diagnostic Biomarkers for Ocular Chronic Graft-Versus-Host Disease

Chronic graft-versus-host disease (cGVHD) remains a common threat after allogeneic hematopoietic cell transplantation (allo-HSCT), and ocular manifestations occur in up to 60% to 90% of cGVHD patients. We sought to reveal major metabolic dysregulation and to determine tear metabolites as potential biomarkers for ocular cGVHD. Twenty-three ocular cGVHD and 16 control tear samples were collected for this study. Differential metabolites were identified using a liquid chromatography-mass spectrometry system. Spearman's test was used to analyze the correlation between metabolites and ophthalmic indexes (National Institutes of Health [NIH] eye score, fluorescein tear film break-up time [T-BUT], corneal fluorescein staining [CFS], and Schirmer's test). Receiver operating characteristic (ROC) curve was analyzed to evaluate the prediction potential of identified metabolites for ocular cGVHD. Differential metabolites were mainly observed in lipid metabolites, and we highlighted the lipid dysregulation in glycerophospholipid metabolism, sphingolipid metabolism, and biosynthesis of unsaturated fatty acids. In glycerophospholipid metabolism, phosphatidylcholine (34:1) (PC [34:1]) exhibited the strongest correlation with NIH eye score (r = 0.80), T-BUT (r = 0.79), CFS (r = 0.77), and Schirmer's test (r = 0.69). In sphingolipid metabolism, sphingomyelin (SM) was the most consistent with T-BUT (r = 0.74) and CFS (r = 0.71), whereas lactosylceramide (LacCer) was the most consistent with NIH eye score (r = 0.76) and Schirmer's test (r = 0.64). In biosynthesis of unsaturated fatty acids, docosahexaenoic acid (DHA) had the highest correlation with NIH eye score (r = 0.73), T-BUT (r = 0.60), CFS (r = 0.67) and Schirmer's test (r = 0.67) (P < .0001 for all). ROC analysis revealed that area under the curve (AUC) values for PC (34:1) (AUC = 0.967), LacCer (AUC = 0.946), SM (AUC = 0.932), and DHA (AUC = 0.929) were significantly correlated with cGVHD (P < .0001 for all). Our study identified PC (34:1), SM, LacCer, and DHA as promising tear biomarkers to indicate metabolic dysregulation and ophthalmic manifestations in ocular cGVHD.