Toward the Characterization of the Human Core Ocular Surface Microbiome

Abstract

Purpose: The field of ocular surface microbiome (OSM) research suggests its involvement in ocular surface (OS) health and disease. However, existing OSM data are heterogeneous. This study aims to provide a whole-metagenome shotgun sequencing-based description of the healthy core ocular surface microbiome (COSM), encompassing all taxonomic kingdoms at species-level resolution.

Methods: Swabs from the conjunctiva and lower lid margin, and tear fluid of 27 individuals without OS disease aged 40 to 60 years were collected at 3 time points. The OSM was sequenced and taxonomically and functionally profiled using Kraken2 and HUMAnN3, respectively. To validate sequencing results, human and microbial proteins of the tear fluid, termed the tear proteome (TP), were characterized by nano liquid chromatography-tandem mass spectrometry (nLC-MS/MS) and profiled by gene ontology. The COSM was defined as the microbiome present in most of the study population over time. Therefore, species present in > 50% of all samples across the three time points were considered to form the COSM.

Results: At species level, Cutibacterium acnes, Malassezia restricta, and Staphylococcus epidermidis formed the COSM, with Corynebacterium segmentosum additionally being part of the core lid microbiome (LM). No significant differences in the OSM and human TP were observed between the left and right eyes on individual levels. However, the variance between subjects mostly exceeded that between eyes within subjects, suggesting an individual-specific COSM and TP.

Conclusions: The description of the COSM provides the basis for future OSM research and potential targets for preventive and therapeutic interventions of the OS and associated diseases.

Figure 1.

Taxonomic characterization of the OSM at species level (n = 155). Relative abundances (%) of the top 31 species in the left and right eyes of subjects (n = 27) in all samples (n = 155) averaged across the 3 different time points. The left and right eyes of each subject are superimposed to allow for direct comparison (a). Mean abundance of all species (n = 155) with a mean relative abundance > 3% (b).

Figure 2.

Most prevalent phyla, genera, and species of the OSM and their core (n = 155). Heatmap illustrating the mean relative abundance (%) of the 10 most prevalent phyla (a), genera (b), and species (c) in OS samples across all three time points. The color scale represents the positive rate, ranging from 0% (blue) to 100% (red), indicating the prevalence of each species among the subjects. The bar plots show the phyla, genera, and species present in > 50% of the conjunctival samples (n = 155) constituting the core microbiome (d, e, and f, respectively).

Figure 3.

Shared genera and species across the three time points (n = 155). The Venn diagrams show the number of genera per time point (time point 1 = 49, time point 2 = 36, and time point 3 = 37) and the number of species per time point (time point 1 = 63, time point 2 = 59, and time point 3 = 54) and the shared number of genera and species at the 3 time points (a and b) with a list of the 15 genera and 16 species that are overlapping at all 3 time points (c and d). Tp = time point.

Figure 4.

Shared and unique pathways between Cutibacterium acnes and Staphylococcus epidermidis. Venn diagram showing the number of shared (red-violet) and unique metabolic pathways of Staphylococcus epidermidis (purple) and Cutibacterium acnes (red).

Figure 5.

Most prevalent LM at species level and its core (n = 74). Heatmap illustrating the mean relative abundance (%) of the 10 most prevalent species in the lower lid margin samples across all three time points. The color scale represents the positive rate, ranging from 0% (blue) to 100% (red), indicating the prevalence of each species among the subjects (a). Species present in > 50% of the lid samples that constitute the core LM (b).

Figure 6.

Variance of the OSM (n = 132). Mean sum of square pictures with the confidence intervals of the 21 species present in ≥ 10 of the subjects. The mean of the difference of mean sum of square of factor side – residual is 0.0070, with a 95% confidence interval between −0.0023 and 0.0162. Thus, the variance explained by the difference between left and right eyes was mostly smaller than the variance that is explained by other factors (a). The mean of the difference of mean sum of square of factor subjects (subject) – residual is −0.0027, with a 95% confidence interval between −0.0071 and 0.0018. Thus, the variance explained by the difference between subjects was mostly greater than the variance that is explained by other factors (b).

Figure 7.

Variance of the human TP (n = 156). Mean sum of square pictures with the confidence intervals of 144 human proteins of the TP. The mean of the difference of mean sum of square of factor side – residual is 0.3367, with a 95% confidence interval between 0.1731 and 0.5004. Thus, the variance explained by the difference between left and right eyes was mostly and significantly smaller than the variance that is explained by other factors (a). The mean of the difference of mean sum of square of factor subjects (subject) – residual is −3.1491, with a 95% confidence interval between −3.5144 and −2.7838. Thus, the variance explained by the difference between subjects was mostly and significantly greater than the variance that is explained by other factors (b).