Potential Role of Ocular Surface Microbiome in Dry Eye: Microbial Interactions and Symptom Alleviation

Study design and sample collection This study is randomized, double-blind, and prospective trial. The inclusion criteria were as follows: patients aged 19 years or older with symptoms of dry eye syndrome. A total of 50 participants were randomly assigned to either the CsA group (n = 25, cyclosporine A twice a day and hyaluronic acid twice daily) or the control group (n = 25, hyaluronic acid four times daily). 0.05% cyclosporine A (Cyporin N Eye Drops, Taejoon Pharm, Republic of Korea) and 0.15% hyaluronic acid (New Hyaluni Eye Drops, Taejoon Pharm) were used in this study. Blinding was maintained by separating the examiner and the sample collector.

Exclusion criteria included the use of topical ocular drops or systemic medication (such as systemic steroids, immunomodulators, and tetracyclines) that could influence study outcomes, as well as the presence of conjunctivitis, anterior blepharitis, Demodex infestation, parasitic eye infections, unresolved ocular trauma, healing disorders of the ocular surface (OS), a history of penetrating keratoplasty, contact lens use, or any ophthalmic surgical procedure within three months before baseline. Additionally, subjects with a single functional eye, pregnant or breastfeeding women, and those at risk of pregnancy without contraception were excluded.

Clinical data and OS samples were collected at baseline before eye drop administration (V1) and at 4 weeks (V2) and 12 weeks (V3) after treatment initiation (Fig. 1a). Of the 50 participants at V1, 21 participants from each group completed follow-up at V2, while 14 participants in the CsA group and 18 participants in the Newhyaluni group remained at V3. OS samples for microbiome analysis were obtained from the inferior conjunctival sac using sterile mixed cellulose ester (MCF) membrane (Millipore, Merck). OS samples were collected from 13 participants in the CsA group and 13 participants in the Newhyaluni group. All OS samples were stored at -80 °C until DNA extraction for microbiome analysis.

Clinical data collection Symptoms were assessed using the Ocular Surface Disease Index (OSDI), Visual Analogue Scale (VAS) for pain, and the modified Standard Patient Evaluation of Eye Dryness (SPEED) questionnaire. Tear film stability was evaluated by measuring the tear break-up time (TBUT). Fluorescein staining of the cornea was performed by applying a drop of sterile saline to a sterile fluorescein strip. After the subject blinked naturally three times, the time between a normal blink and the first appearance of a dry spot in the tear film was recorded. The average of three repeated measurements was used for analysis.

OS inflammation was assessed using the fluorescein staining score (FSS) graded according to the Oxford scoring system. The severity of meibomian glands and eyelid inflammation, as well as lid margin hyperemia, was graded as none, mild, moderate, or severe. Tear secretion was measured using the Schirmer strip without anesthesia. A Schirmer strip was placed at the temporal third of the lower eyelid, between the lower palpebral conjunctiva and the lower bulbar conjunctiva. After 5 min, the length of tear fluid absorbed by the strip was measured in millimeters.

DNA extraction and whole metagenome sequencing Total DNA was extracted from the collected OS samples using the RNeasy PowerMicrobiome Kit (Qiagen, Inc., Valencia, CA, USA). DNA concentration was measured with a BioPhotometer D30 using a μCuvette G1.0 (Eppendorf, Hamburg, Germany). To eliminate potential contaminants during the experimental process, 12 negative controls were included, comprising sampling membranes, DNA-free water added to the DNA extraction kit, and DNA-free water added to the sequencing library preparation kit. These negative controls were sequenced alongside the 61 OS samples.

For whole metagenome sequencing, extracted DNA was fragmented using NEBNext dsDNA Fragmentase (New England Biolabs, Inc., Ipswich, MA, USA). Library preparation was performed with a Swift 2S Turbo DNA Library Kit (Swift Biosciences, Inc., USA) following the manufacturer's protocol. Purification and size selection were conducted using HiAccuBead (AccuGene, San Diego, CA, USA). Index PCR was carried out with the Swift 2S Turbo Combinatorial Dual Indexing Primer Kit (Swift Biosciences, Inc.), followed by additional purification and size selection using HiAccuBead. The library concentration for each sample was quantified using the Qubit™ dsDNA HS Assay Kit (Thermo Fisher Scientific, USA). Equimolar concentrations of each library (4nM) were pooled and sequenced on an Illumina NovaSeq 6000 system (250-bp paired end).

Raw sequence data obtained from the NovaSeq system were processed as previously described. Adapter sequences were trimmed, and quality filtering was performed using Trimmomatic. Paired-end sequences were merged using PEAR v.0.9.11. Human-derived sequences in the metagenome data were identified and removed using BBMap with a reference human genome. Taxon