Ocular Surface Microbiome Alterations Are Found in Both Eyes of Individuals With Unilateral Infectious Keratitis

Abstract

Purpose: To analyze the ocular surface microbiome (OSM) profile in both eyes of individuals with unilateral keratitis.

Methods: In this prospective, cross-sectional study, the conjunctival OSM of adults (>18 years old) presenting to an ophthalmic emergency department with acute unilateral keratitis and controls without an acute infectious process was sampled. Samples underwent DNA amplification and 16S sequencing using Illumina MiSeq 250 and were analyzed using Qiime. Statistical analysis was performed using a two-sided Student t-test, diversity indices, and principal coordinate analysis. The main outcome measures included relative abundance and α and β diversity.

Results: Bacterial DNA was recovered from all 34 eyes of 17 individuals with keratitis (mean age, 49.3 ± 17.5 years) and 16 eyes of controls (mean age, 56.6 ± 17.0 years). In the two culture-positive eyes, 16S aligned with culture results. Significant differences in α diversities were noted when comparing both eyes of individuals with keratitis to control eyes (all P < 0.05), but no significant differences between the eyes of an individual with keratitis. Principal coordinate analysis plots confirmed this finding, demonstrating separation between either eye of patients with keratitis and controls (both P < 0.01), however not between eyes in patients with unilateral keratitis. Both eyes of individuals with keratitis had greater abundance of Pseudomonas compared with controls both on compositional analysis and linear discriminant analysis.

Conclusions: Alterations in the OSM profile are detected in both eyes of individuals with unilateral keratitis compared with controls. Beyond the causative organism, a greater abundance of potential pathogens and lesser abundance of commensal organisms were found.

Translational relevance: The OSM profile is altered in both eyes of individuals with unilateral keratitis, which may lend insight into the role of the microbiome in the pathophysiology of disease.

Figure 1.

Diversity matrices comparing the Shannon, Chao1, and observed α diversities between the control and keratitis eyes (A), control and noninfected eyes of patients with keratitis (B), and keratitis to contralateral eyes (C). Test of significance was Mann–Whitney U test and P values were absolute.

Figure 2.

Principal coordinate analysis plots (unweighted UniFrac) demonstrating separation keratitis and control eyes (A) and nonkeratitis and control eyes (B). Test of significance was Monte-Carlo permutation with Bonferroni's correction (FDR).

Figure 3.

Genus-level heat map with hierarchical clustering showing that there were compositional differences when comparing (A) control eyes to the eyes with keratitis and (B) control eyes to the noninfected eye of patients with unilateral keratitis.

Figure 4.

Heat map tree of entire hierarchy from phylum to genus displaying increases in abundance for certain phyla when comparing controls to the keratitis (A) and nonkeratitis (B) eyes.

Figure 5.

Bacterial classes which significantly differed between the control and keratitis eyes. Test of significance was Mann–Whitney U test and P values were absolute.

Figure 6.

Linear discriminant analysis comparing the control to keratitis (A) and control to nonkeratitis eyes (B), finding a high effect size for Pseudomonas in both eyes of infected individuals when each eye was compared with controls.