Role of Polyunsaturated Fatty Acids (PUFAs) and Eicosanoids on Dry Eye Symptoms and Signs

Polyunsaturated fatty acids (PUFAs) generate pro- and anti-inflammatory eicosanoids via three different metabolic pathways. This study profiled tear PUFAs and their metabolites and examined the relationships with dry eye (DE) and meibomian gland dysfunction (MGD) symptoms and signs. A total of 40 individuals with normal eyelids and corneal anatomies were prospectively recruited. The symptoms and signs of DE and MGD were assessed, and tear samples (from the right eye) were analyzed by mass spectrometry. Mann–Whitney U tests assessed differences between medians; Spearman tests assessed correlations between continuous variables; and linear regression models assessed the impact of potential confounders. The median age was 63 years; 95% were male; 30% were White; and 85% were non-Hispanic. The symptoms of DE/MGD were not correlated with tear PUFAs and eicosanoids. DE signs (i.e., tear break-up time (TBUT) and Schirmer’s) negatively correlated with anti-inflammatory eicosanoids (11,12-dihydroxyeicosatrienoic acid (11,12 DHET) and 14,15-dihydroxyicosatrienoic acid (14,15, DHET)). Corneal staining positively correlated with the anti-inflammatory PUFA, docosahexaenoic acid (DHA). MGD signs significantly associated with the pro-inflammatory eicosanoid 15-hydroxyeicosatetranoic acid (15-HETE) and DHA. Several relationships remained significant when potential confounders were considered. DE/MGD signs relate more to tear PUFAs and eicosanoids than symptoms. Understanding the impact of PUFA-related metabolic pathways in DE/MGD may provide targets for new therapeutic interventions.


Introduction
Dry eye (DE) is a prevalent, multifactorial disease that consists of a wide range of clinical manifestations that include symptoms of ocular surface pain (characterized as "dryness", "burning", and "discomfort", to name a few) and visual disturbances [1].Signs of DE can include tear instability, insufficient tear production, and/or ocular surface disruption [2].Closely related to DE, meibomian gland dysfunction (MGD) is defined as "a chronic, diffuse abnormality of the meibomian glands, commonly characterized by terminal duct obstruction and/or qualitative/quantitative changes in the glandular secretion", commonly caused by epithelial gland hyperkeratinization [3].Symptoms, such as dryness and foreign body sensations, and signs, such as tear instability, can overlap between DE and MGD [3].Together, DE/MGD symptoms and signs impact the quality of life, negatively affect mood, and limit activities of daily living [4].Various external and internal factors can impact DE/MGD manifestations, including weather, air pollution, diet, and systemic comorbidities [5].Inflammation has been identified as an important intermediary between such external and internal factors and DE/MGD, and thus, many studies have examined the contributions of cellular (i.e., macrophages, regulatory T cells) [6] and soluble (i.e., tumor necrosis factor-alpha (TNF-α), interleukin-1B (IL-1B), and interleukin-6 (IL-6)) [7] mediators on disease pathophysiology.
While several studies have examined the relationships between eicosanoids produced from the LOX and COX metabolic pathways and DE/MGD, what is missing from the literature is an examination of eicosanoids produced from the AA cytochrome P450 epoxygenase pathway.Animal [15] and human studies [16] have demonstrated that cytochrome-derived eicosanoids can have potent pro-or anti-inflammatory properties, and as such, these metabolites may play a role in DE/MGD.To bridge this knowledge gap, we aimed to profile the levels of tear PUFAs, LOX, COX, and cytochrome pathway-derived eicosanoids and examine their relation to clinical symptoms and signs of DE and MGD.

Study Design and Population
This prospective, single-site, cross-sectional study was conducted in accordance with the tenets of the Declaration of Helsinki, complied with the requirements of the United States Health Insurance Portability and Accountability Act (HIPAA), and was approved by the Miami Veterans Affairs (VA) Institutional Review Board.Participants were recruited from 2016 to 2017 at the Miami VA Eye Clinic, and informed consent was obtained from the subjects after an explanation of the nature and possible consequences of the study.Individuals were excluded if they had concomitant ocular or systemic conditions that could confound DE, such as anatomic abnormalities of their eyelids (i.e., ectropion), conjunctiva (i.e., pterygium), and/or cornea (i.e., Salzman's nodular degeneration and edema); history of glaucoma, refractive, or retinal surgery; cataract surgery within the last 6 months; use of contact lenses; topical medications besides artificial tears; HIV; sarcoidosis; graft-versus host disease or a collagen vascular disease.

Data Collection
Demographic information, including age, sex, race, ethnicity, smoking, medical history, oral medications, and supplements, was collected for each patient.

Ocular Surface Assessment
All participants underwent a comprehensive ocular surface examination of both eyes, which included the following, in the order performed: (1) Measurement of tear film osmolarity (TearLAB, San Diego, CA, USA).

Statistical Analysis
Statistical analyses were performed using SPSS Statistics Software version 25.0 (IBM Corp. Armonk, NY).Descriptive statistics were used to summarize participant demographics, comorbidities, medication use, DE/MGD symptoms, and signs.The normality of the distributions of variables of interest was assessed using the Shapiro-Wilk test.Given that measures were not normally distributed, Mann-Whitney U tests were run to assess the differences between medians, and Spearman correlation coefficients (ρ) were calculated to evaluate the relationship between demographics, DE signs/symptoms, pro, and anti-inflammatory markers.After inspecting residuals, linear regression models with the forward method were performed to predict the contribution of patient characteristics, comorbidities, tear PUFAs, and eicosanoids on DE/MGD symptoms and signs.p < 0.05 was considered statistically significant.

Study Population
The median age of the racially diverse, predominantly male population was 63 years (interquartile range (IQR): 14) (Table 1).DE symptoms ranged from none to severe, with 90% of individuals reporting mild or greater DE symptoms as determined by a DEQ-5 ≥ 6 and 83% as determined by an OSDI ≥ 13.The majority reported some degree of ocular pain (85%, NRS ≥ 1), with 48% reporting moderate or greater pain (NRS ≥ 4).Ocular surface signs varied, with 13% displaying tear instability in the right eye (OD) (as determined by TBUT < 5 s) and 8% showing aqueous tear deficiency (as determined by Schirmer's < 5 mm).All individuals had at least one sign of MGD, which included eyelid telangiectasias, MG plugging, or MG dropout.

Relationships between Tear Eicosanoids and Clinical Metrics
Mann-Whitney U tests were performed to compare differences in medians between demographics, co-morbidities, tear PUFAs, and eicosanoids.Only the significant differences are summarized in Table 2. Males had higher levels of tear pro-inflammatory eicosanoids (i.e., 12 HETE and 15-HETE) compared to females.A similar pattern was noted in Hispanic individuals (compared to non-Hispanics) and in those with diabetes (compared to nondiabetics).On the other hand, smokers had increased levels of an anti-inflammatory marker (14,15 DHET) compared to non-smokers.Also, subjects who reported taking fish oil supplements had a less inflammatory profile, with higher levels of anti-inflammatory mediators and lower inflammatory ratios compared to those not on supplements (EPA: 49.83 (IQR: 66.85) vs. 7.61 (IQR: 16.3), p = 0.03; AA:EPA: 35.42 (IQR: 44.18) vs. 74.11(IQR: 75.56), p = 0.04).Individuals who reported taking multivitamin supplements had higher levels of both pro-(i.e., 5-HETE, AA) and anti-(i.e., 14,15 DHET, 11,12 DHET, EPA, and DHA) inflammatory eicosanoids compared to those not on supplements.

Linear Regression Models
After inspecting residuals, forward linear regression models were built with significant DE/MGD signs (right eye only) from univariable analysis as dependent variables (Table 3) and all PUFAs and eicosanoids as independent variables.Other variables included in the models were demographics (i.e., gender and ethnicity), history of smoking, hypercholesterolemia, diabetes, sleep apnea, use of betablockers, anxiolytics, fish oil, and multivitamin supplements (Table 4).In all models, tear lipids remained when confounders were considered.Specifically, those with worse ocular surface parameters (lower TBUT and Schirmer, higher corneal staining, and more severe eyelid telangiectasias) had higher levels of eicosanoids with antiinflammatory properties (11,12-DHET, 14,15-DHET, and DHA) (Table 4, Figure 1).

Discussion
In this study, we found some relationships between DE/MGD signs, PUFAs, and their derivatives, even when considering potential confounders.Specifically, the multivariable analysis found negative correlations between tear film parameters (stability and production) and anti-inflammatory eicosanoids (11,12-DHET and 14,15-DHET), suggesting compensatory mechanisms to a pro-inflammatory state driven by tear abnormalities.Similar findings were noted when considering corneal staining and eyelid telangiectasias, as they were related to higher levels of the anti-inflammatory eicosanoid DHA.
A unique finding in our study was the detection of tear eicosanoids metabolized via the CYP pathway (i.e., 11,12 DHET, 14,15-DHET).Studies outside the eye have related DHET molecules to pro-inflammatory conditions such as coronary heart disease (CHD), with individuals with CHD having higher levels of 14,15-DHET compared to controls (2.53 ± 1.60 ng/L vs. 1.65 ± 1.54 ng/L, p = 0.036) [16].We similarly found significant correlations between low tear stability and production and DHET molecules, indicating that higher levels of anti-inflammatory eicosanoids may be involved in disease modulation and restoration of homeostasis across disease states and organs [16].
Beyond tear and ocular surface parameters, several patient-related factors, including gender, ethnicity, the presence of comorbidities, and smoking, are related to levels of various tear eicosanoids, with ethnicity and diabetes remaining in multivariable models.Specifically, males, Hispanics, and individuals with diabetes had higher levels of proinflammatory eicosanoids (i.e., 12-HETE and 15-HETE) compared to their counterparts, while current or previous smokers had higher levels of an anti-inflammatory eicosanoid (i.e., 14,15-DHET) compared to non-smokers.These findings share similarities and differences with prior literature.While previous studies have not found differences in tear lipids by gender [10,13], mixed findings were noted by ethnicity, with higher [31], lower [31], and similar [32,33] levels of plasma anti-inflammatory PUFAs reported.Similar to our study, increased pro-inflammatory and angiogenic eicosanoids (i.e., LTB4 and 12-HETE) were noted in individuals with chronic diseases such as diabetes [34] and CHD [35].Interestingly, prior studies have also noted increased plasma pro-(i.e., 5-HETE) and anti-(i.e., 11,12-DHET and 14,15-DHET) inflammatory eicosanoids in smokers, suggesting that compensatory mechanisms may occur both in blood and tears [36].Although the aim of our study was not to describe the impact of oral supplements such as fish oil and multivitamins on tear composition, we noted that individuals who reported taking supplements had a less inflammatory tear film (lower AA: EPA, AA: DHA and ω6: ω3) compared to their counterparts.Similar findings were noted in our prior study with respect to individuals taking ω3 (DHA and EPA) supplements [13].While the results of studies on the benefits of ω3 and ω6 supplementation in DE have been mixed [37][38][39][40][41][42], our findings suggest the need for further research on this topic.
As with all studies, our findings need to be considered in light of the limitations of this study, which included its cross-sectional nature, predominately male population, limited sample size, and sensitivity of the tear eicosanoid assay.Furthermore, the origin of lipid mediators in tears is not known, with possible sources, including the MG, lacrimal gland, and/or ocular surface epithelium.Finally, our assay did not include some proresolving eicosanoids, such as resolvins, maresins, and neuroprotectin D1.Despite these limitations, in this study, we found both pro-and anti-inflammatory markers in subjects with tear abnormalities and MGD, which is similar to prior studies.We also detected new findings, such as CYP-derived metabolites, which have been reported to possess anti-inflammatory properties.

Conclusions
Overall, our findings highlight the complexity of studying tear eicosanoids in DE/MGD, as their presence may contribute to or be a compensatory mechanism for an abnormal ocular surface environment.It is plausible that the ocular surface constantly autoregulates itself to re-establish homeostasis, and thus, longitudinal studies are needed to evaluate our findings more robustly.A better understanding of the role of eicosanoids in DE/MGD is needed, as this knowledge may improve treatment algorithms by suggesting which medications (i.e., corticosteroids that block phospholipase A 2 enzyme preventing production of AA and all downstream products vs. nonsteroidal anti-inflammatory drugs (NSAIDs) that only block the COX pathway) would be optimal in an individual patient [43].Moreover, beyond COX and LOX inhibition, selective cytochrome P-450 inhibitors may have a beneficial role, which needs to be further defined in the DE/MGD field.
(AG).This work was also supported by a center grant to the UVA Comprehensive Cancer Center from the National Cancer Institute (2P30CA044579-26), National Eye Institute NIH Center Core Grant P30EY014801 (institutional) and Research to Prevent Blindness Unrestricted Grant (institutional).The contents of this manuscript do not represent the views of the Department of Veterans Affairs or the United States Government.The funding sources were not involved in the study design, data collection, interpretation, analysis, writing, or decision to submit this article for publication.

Institutional Review Board Statement:
The study was conducted in accordance with the tenets of the Declaration of Helsinki, complied with the requirements of the United States Health Insurance Portability and Accountability Act (HIPAA), and was approved by the Miami Veterans Affairs (VA) Institutional Review Board (protocol code 3011.08 and date of approval 8 June 2020).
Informed Consent Statement: Informed consent was obtained from all subjects after explanation of the nature and possible consequences of the study.

Table 1 .
Demographics and clinical information of the study population.

Table 2 .
Significant differences in medians between demographics, medical history, and eicosanoids.

Table 3 .
Spearman correlations demonstrating relationships between DE/MGD symptoms and signs and pro-and anti-inflammatory markers.

Table 4 .
Linear regression models examining impact of tear PUFAs, eicosanoids, demographics, and comorbidities on DE/MGD symptoms and signs.