Meibomian Gland Shortening Is Associated With Altered Meibum Composition

Abstract

Purpose: The purpose of this study was to investigate differences in the composition of meibum extracted from human meibomian glands displaying different morphological characteristics.

Methods: Adult participants with evidence of meibomian gland shortening were enrolled. Glands were observed using infrared meibography, and meibum was collected from one short and one long gland from the lower eyelid of the same eye. Total protein concentration was evaluated using the detergent compatible (DC) protein assay and lipid analysis was performed using liquid chromatography mass spectrometry.

Results: Fifteen participants (8 women and 7 men) completed the study (mean age = 34.7 ± 7.7 years). The average volume of meibum collected from the short and long glands was 0.02 ± 0.01 and 0.05 ± 0.03 µL, respectively (P = 0.01). Average protein concentration in the short glands was significantly higher compared to the long glands (0.10 ± 0.03 vs. 0.08 ± 0.02 mg/mL, P = 0.01). Among the non-polar lipids, the mole percent of cholesterol esters (CEs; P = 0.02), triacylglycerols (TAGs; P = 0.04), and ceramide (Cer; P = 0.03) was significantly lower in short glands compared with long glands. Among the polar lipids, the mole percent of (O-acyl)-ω-hydroxy fatty acid (OAHFA; P = 0.01) was significantly lower in short glands, whereas phosphatidylcholine (PC; P = 0.02) and sphingomyelin (SM; P = 0.01) were significantly higher in short glands than long glands.

Conclusions: Meibum composition differed between short and long glands. The lower mole percent of CE, TAG, OAHFA, and Cer and a higher mole percent of PC and SM among short glands may indicate disease activity. This information may clarify the natural history of meibomian gland dysfunction and future targets for therapy.

Meibomian gland dysfunction (MGD) is a leading cause of dry eye and is more prevalent in populations over the age of 40 years.¹ The condition is associated with ocular discomfort, visual disruption, alterations to tear film composition, and damage to the ocular surface, impacting functional ability and quality of life.² MGD is characterized by meibomian gland obstruction and changes in the quality or quantity of the gland secretions (meibum).³ 

Meibum is rich in lipids but also contains proteins that are produced by fully differentiated meibocytes in the tubulo-acinar structure of meibomian glands (MGs) in humans and most mammals.⁴ Meibum protects the ocular surface from environmental hazards, desiccation,⁵ and prevents contamination of the tear film by skin lipids.⁶ Any morphological changes to the MGs may alter their meibum production, affecting tear film stability, and, ultimately, disturbing ocular surface homeostasis, leading to epithelial damage.⁷ 

There are approximately 30 to 40 glands located in the upper eyelid with a length of approximately 5.5 mm and 20 to 30 glands in the lower eyelid with a length of approximately 2 mm to 4 mm. Significant gland shortening has been observed in individuals aged 30 years and above,^8,9 and gland shortening has also been associated with dry eye disease.^10,11 These glands are formed by multiple acini, comprised of meibocytes which terminally differentiate to produce meibum.¹² Meibum flows through ductules connecting the acini to a central gland duct which terminates in an opening on the eyelid margin.^13,14 Animal models suggest that when the composition of meibum secretions changes, the light absorption and transmission characteristics of these glands are also altered,¹⁵ which may be reflected in the morphological gland changes observed in vivo. MG shape has been assessed widely using infrared meibography and may vary from full-length glands in the normal eye to short, atrophied, tortuous, or completely absent glands.^16–18 However, whether visible structural changes are associated with changes to the gland of the meibum composition in humans is unknown. 

The aim of this exploratory study was to investigate whether there are differences in composition between meibum extracted from shortened meibomian glands (defined for the purposes of this study as glands that are half the length of the everted eyelid or less), and full-length glands from the same eye in humans. Understanding the morphology of the glands and the potential impact on secretions is important to understand the natural history of MGD and inform MGD management and possible prevention. 

Methods

Study Design

This prospective cross-sectional study was conducted at the School of Optometry and Vision Science, University of New South Wales (UNSW), Sydney, Australia. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Human Research Ethics Committee, UNSW Sydney, Australia (HC 210617). All subjects provided signed informed consent before enrollment. 

Participants

Fifteen healthy participants aged 18 years and over, with or without symptoms of dry eye, and with evidence of at least one short and one long meibomian gland in the lower eyelid of one eye, were enrolled. Anyone who had undergone intensive MG treatment, including moist heat treatment, thermal pulsation, or intense pulsed light therapy in the last 12 months, was excluded. Other exclusion criteria included an active eye infection, inflammation or allergy, eye injury, or surgery within the past 6 months, any systemic disease that may affect ocular physiology, and use of any corticosteroids, immunosuppressant, or antihistamine medications 12 weeks prior to enrollment. Pregnant or lactating women were also excluded from the study. 

Study Procedures

The lower eyelid was everted, and the MGs were viewed using the Oculus Keratograph 5M (Oculus, Wetzlar, Germany). One shortened gland and one full-length gland were marked (Fig. 1A) on the lower lid using a surgical pen (Livingston surgical marker [LIVSKMRN: NSW, Australia]). The choice of either the left or right eye was made based on the presence of both a short and long gland in the same eye for each participant. Short and long glands that were adjacent to each other were not considered for meibum collection. Samples for all participants were collected at approximately the same time of day, by the same investigator to avoid diurnal variations.¹⁹ A micro-capillary tube (Drummond Microcap 100 µL, Drummond Scientific Company, Broomall, PA, USA) was placed vertically on the face of the orifice²⁰ and each marked gland was expressed using a Korb expressor (Tear Science, Australia; Fig. 1B) placed below the eyelash line of the lower eyelid for 10 seconds²¹ to ensure the expressed meibum was from the marked glands. In addition, all the sample collection was conducted by the same investigator. Meibum from the same marked gland was collected at two time points under a slit lamp biomicroscope (Carl Zeiss Pty Ltd, New South Wales, Australia), approximately 2.5 hours apart, as the recovery time of MGs after expression is approximately 2 hours for healthy subjects.²² The aim was to collect approximately 1 mm of meibum measured in the microcapillary tube, with an expected range of 0.25 mm to 1.5 mm.²⁰ 

Figure 1.

 

(A) Example of short (yellow arrow) and long (white arrow) glands (B) capillary tube placed at the orifice for collection of meibum from individual glands.

Lid margins were not cleaned before collection as cleaning the lid margin was not found to affect the composition of lipids expressed from the glands.²³ In all cases during meibum collection, the eyelid was gently turned away from the eye to reduce contamination by tears.²⁴ The meibum sample was dissolved in 300 µL chloroform in a glass vial and stored at −80°C until analysis.²¹ 

Quantification of Total Protein

The samples were dried under a stream of nitrogen gas at room temperature and resuspended in chilled (−20°C) acetone to precipitate proteins in meibum samples. Samples tubes were vortexed and incubated for 60 minutes at −20°C. Samples were centrifuged for 20 minutes at 10,000 g. The supernatant of each sample, containing lipids, was moved to a new glass vial (with caution to not dislodge the protein pellet) and used for lipid analysis. The vials containing protein pellets were left at room temperature with an open cap for 2 minutes to facilitate the evaporation of excess acetone. Proteins were dissolved in phosphate buffer saline (PBS) and quantified using a standard detergent compatible (DC) protein assay protocol kit (ThermoFisher Scientific, Australia) following the manufacturer's instructions.²⁵ Three to 5 dilutions of a protein standard containing from 0.2 mg/mL to about 1.5 mg/mL protein were prepared. Standards were prepared in the same buffer as the sample. Five microliters (5 µL) of standards and samples were pipetted into clean, dry microtiter plates. Twenty-five microliters (25 µL) of reagent A was added and a further 200 µL of reagent B was added into each well and gently agitated to mix the reagents. After 15 minutes, absorbance was read at 750 nm through a Multiskan SkyHigh UV/Vis microplate spectrophotometer (ThermoFisher Scientific, Australia) 

Quantification of Total Lipids

Chemicals

Most chemicals used were as previously reported.²⁶ Methanol, chloroform, and Methyl tert-butyl ether (MTBE) were obtained from Sigma-Aldrich (Castle Hill, New South Wales, Australia). Ammonium acetate (liquid-chromatography mass spectrometry [LC-MS] grade) and analytical grade butylated hydroxytoluene (BHT) were also purchased from Sigma-Aldrich. The 18:1/16:0 (O-acyl)-ω-hydroxy fatty acid (OAHFA standard) was synthesized by Life Science Private Limited (Kolkata, India); phospholipid standards were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Wax ester (WE), cholesterol ester (CE) standards were from Nu-Chek Prep (Elysian, MN, USA), and triacylglycerol standards from CDN isotopes (Point-Claire, Quebec, Canada) 

Extraction of Lipids From Meibum Samples

Meibum lipid extraction was performed using a previously established method.^27–29 In brief, lipid standards were dissolved in pure methanol to prepare a standard solution. The molar concentrations of the lipid's internal standard solution are presented in Table 1. Fifty microliters of the internal standard solution were added to each dry meibum sample. MTBE = methanol (2 mL; 10:3 vol/vol; containing 0.01% butylated hydroxytoluene) was added and samples were mixed in an orbital shaker for 10 minutes at room temperature. 

Table 1.

 

Molar Concentrations of Internal Standards in the Stock Solution

 

The internal standards used in these experiments did not share the same mass, but their chemical structures and relative concentrations were similar to the lipid species within the studied lipid classes. However, these standards were not found in human meibum. They served as valuable controls, accounting for both random and systematic uncertainties arising from sample preparation or instrument fluctuations. The resulting signals closely resembled those of the analytes in most aspects, yet they remained distinct enough for the instrument to differentiate between them. 

A biphasic lipid extraction was performed by adding 500 µL of 0.15 M ammonium acetate to the MTBE:methanol solution and tubes were vortexed and centrifuged for 20 minutes at 28°C at 1500 rcf. The supernatant was removed to a separate tube. The upper organic layer was dried under nitrogen gas at room temperature, resuspended in 100 µL methanol:chloroform (2:1 vol/vol), and stored at −80°C until analysis by mass spectrometry. 

Mass Spectrometry

Mass spectrometry was performed using a previously published method.²⁸ Twenty-two microliters of extracts were injected using a Q-Exactive Plus Mass Spectrometer (ThermoFisher Scientific, Waltham, MA, USA) coupled to a U3000 ultra pressure liquid chromatography (UPLC) system (ThermoFisher Scientific). Solvent A was 6:4 acetonitrile:water and solvent B was 1:9 acetonitrile:isopropanol, both with 10 mM ammonium formate and 0.1% formic acid. Lipids were chromatographed according to a previous method.³⁰ In brief, a 30-minute gradient running from 30% to 100% of solvent B was performed, eluting lipids in order of hydrophobicity. Chromatography was performed at 60°C on a Waters CSH C18 UHPLC column 2.1 × 100 mm, 1.8 uM with VanGuard (Waters Corporation, New South Wales, Australia) column. The column eluate was directed into the electrospray ionization source of the mass spectrometer where a heated electrospray ionization-I probe was used. Source parameters were optimized for a range of lipid standards prior to the analysis. 

Lipid Search software version 4.1 (Thermo Fischer Scientific, Waltham, MA, USA) was used for the identification of molecular species. The parent search mode, based on an accurate mass of precursor ions, and the product search mode, based on the mass of precursor ions and MS2 spectral pattern, were used. The precursor and product tolerance were set to a 5-ppm mass window. The relative intensity threshold of precursor ions was set to 1%, and the relative intensity threshold of product ions was set to 5%. The m-score threshold was set to 2.0.³¹ Data were then exported to an Excel spreadsheet for manual processing and statistical analysis. The raw abundances (peak areas) were normalized by dividing each peak area by the raw abundance of the corresponding internal standard for that lipid class. 

Data Analysis

Mean volume (µL) of collected samples were calculated¹⁶ and compared between short and long glands using the Student’s t-test. Sample protein concentrations were reported as microgram/microliter using a standard curve. 

For lipids, initially, individual lipid species in each sample were normalized with respect to total lipid in each sample and reported as median mole percentage (interquartile range) (Fig. 2). Species of the lipid classes having significant differences between a short and long gland were normalized to the sum of all lipids within the respective class (Figs. 3A–F) using IBM SPSS statistics version 27 (IBM Corp., Armonk, NY, USA). 

Each species of each lipid class was also presented as the median mole percentage (interquartile range) for continuous variables. The normality of continuous variables was checked using the Shapiro-Wilk normality test and histogram. Second, data were log transformed to reduce the skewness and multivariate analysis of variance (MANOVA) followed by descriptive discriminant analysis, which was performed using GraphPad Prism version 10.0 to compare each lipid class and species between short and long glands. An adjusted P < 0.05 was considered to be statistically significant. 

Results

Fifteen participants (8 women and 7 men) with an average age of 34.7 ± 7.7 years completed the study. Average meibum volumes collected from short and long glands were 0.02 ± 0.01 and 0.05 ± 0.03 µL (P = 0.01), respectively. All participants were non-contact lens wearers. 

Proteins and lipids were detectable in all meibum samples from short and long glands. Average protein concentration was significantly higher in short glands compared to long glands (0.10 ± 0.03 vs. 0.08 ± 0.02 mg/mL, P = 0.01, respectively). 

Lipids in Short and Long Glands

The mole percentage of polar and non-polar lipids are described in Table 2 and the overall lipid class profiles in a short and long gland are shown in Figure 2. The WEs were the most abundant non-polar and overall lipid class in the meibum collected from short and long glands, but the mole percentages were not significantly different (46.7 ± 3.6 vs. 47.4 ± 3.4%, respectively, P = 0.06). There was a significantly lower mole percentage of CEs in the short glands compared to the long glands (29.9 ± 3.3 vs. 34.5 ± 2.0%, P = 0.02). Similarly, the mole percentage of OAHFAs (2.2 ± 0.3 vs. 3.2 ± 0.2%, P = 0.02) and TAG (4.8 ± 1 vs. 5.3 ± 1.7%, P = 0.03) were significantly lower in short glands compared to long glands. 

Table 2.

 

Mole Percentage of Polar and Non-Polar Lipids to Total Lipids

 

Figure 2.

 

Mole percentage of total lipids in short and long glands. CE, cholesterol esters; WE, wax esters; TAG, triacyl glycerides; DAG, diheptadecanoin; Cer, ceramide; OAHFA, (O-acyl)-ω-hydroxyl fatty acids; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine; LPC, Lys phosphatidylcholine; SM, sphingomyelin. Quantitative comparison of lipid profiles in short and long glands are shown as medians mole percent and interquartile range (N = 15, normalized to the total lipids). Asterisks denote significant differences achieved with the log transformed data: *P < 0.05.

Among the phospholipids, mole percentage of Cer was significantly reduced in short glands versus long glands (2.9 ± 0.5 vs. 3.5 ± 0.5%, P = 0.02), whereas mole percentage of phosphatidylcholine (PC; 1.4 ± 0.2 vs. 0.8 ± 0.2%, P = 0.02) and sphingomyelin (SM; 0.3 ± 0.04 vs. 0.1 ± 0.02%, P = 0.01) were significantly higher in short versus long glands. The mole percentage of other less abundant phospholipids including Lys phosphatidylcholine (LPC; 0.05 ± 0.02 vs. 0.06 ± 0.01%), phosphatidylethanolamine (PE; 2.5 ± 1.1 vs. 0.5 ± 0.1%), and phosphatidylserine (PS; 2.2 ± 0.5 vs. 1.2 ± 0.2%) were not significantly different between groups (P > 0.05). 

Error bars for certain lipids, for instance TAG, are asymmetrical due to the data being transformed (see Fig. 2). 

Speciation of Lipids Displaying Significant Differences Between Short and Long Glands

All species were identified in all meibum samples collected. Molecular lipid speciation of lipid classes that were significant between short and long glands were normalized to the total of lipids in the respective class (see Fig. 3). 

Figure 3.

 

Species of each lipid class having a significant difference between short and long glands. Molecular lipid speciation of (A) CEs, (B) TAG, (C) Cer, (D) OAHFA, (E) PC and (F) SM normalized to the total of each respective class. Values are shown as the median and interquartile range (N = 15). Asterisks denote significant differences achieved with the log transformed data: *P < 0.05.

At least 12 different species of CEs were observed, mainly those with long chains and very long chains ranging from C18 to C32. The most abundant compounds of the CE family were CE 18:1 and 24:2 and CE 28:1, but these were not significantly different between short and long glands (see Fig. 3A). The mole percentage of TAG was not significantly different, but it is interesting to note that all 15 species of TAG were numerically reduced in short glands compared to long glands (see Fig. 3B). There was a similar trend for Cer and PC species. Only six Cer species and five PC species were observed in all short and long glands (see Fig. 3C), but there was no significant difference in mole percentage of any Cer species between glands. Similarly, PC species with fatty acid chain lengths of 19:0, 20:0, and 26:0 exhibited a decrease in long glands when compared with short glands, although the reduction did not attain statistical significance (see Fig. 3E). 

OAHFAs constituted the bulk of the polar lipids and there were 28 distinct species with long carbon chains from 16:0 to 18:2. All detected OAHFAs were unsaturated and the mole percent of 16:1/32:1 (P = 0.002), 18:2/25:2 (P < 0.0001), 18:2/26:1 (P < 0.0001), 18:2/27:1 (P < 0.0001), 18:2/30:1 (P = 0.002), and 18:2/34:1 (P < 0.0001) were significantly lower in short compared with long glands (see Fig. 3D). Nine SM species were observed with long and very long carbon chains ranging from 32 to 42. Mole percent of SM 34:1, 40:2, and 42:1 was significantly higher whereas 34:2 and 42:2 was significant in short glands compared to long glands (see Fig. 3F). 

Discussion

Morphological changes in glands are associated with worsening MGD and altered meibum quality.^32,33 Gland length has been identified as the key functional metric of the lower lids.³⁴ This study suggests that these morphological alterations may be reflected in the characteristics of the meibum within the glands in the same eye of an individual. Non-polar lipids play a significant role in stabilizing the tear film³⁵ and these were reduced in the shorter glands. A significant decrease in the relative fraction of non-polar lipids has previously been observed in MGD and aging³⁶ and it may be possible that this change in meibum composition is associated with gland shortening. As MGD is characterized by alterations in gland shape or function, the reduction of non-polar lipids in the short glands could potentially indicate a dysfunctional gland.^37,38 It remains uncertain whether changes in meibum composition led to morphological variations or vice versa. Investigation of MG natural history through a longitudinal study could provide further insights. 

This study also confirmed CEs as the dominant lipid class in long glands, representing 34.5% of the mole percentage, compared to 29.9% in short glands. In humans, cellular levels of free cholesterol are regulated by CEs.³⁹ Functionally, CEs increase the lipid phase transition temperature and lipid hydrocarbon order in the tear film, which contributes to the stability of the tear film lipid layer.⁴⁰ The observation of significantly reduced CEs in short glands may thus imply that gland shortening is associated with reduced tear film stability. This would be consistent with a previous report which found 40% fewer CEs in eyes with MGD.⁴¹ It is unclear at this stage where the critical threshold for gland loss to cause altered tear stability may lie. Further work is required to evaluate this effect with different severity levels of gland loss. 

Among the polar lipids, (O-acyl)-ω-hydroxy fatty acid was the most abundant species in both short and long glands. These moieties are responsible for tear film stabilisation^42,43 as demonstrated in OAHFA-deficient mice displaying tear film instability, as well as meibomian gland obstruction.⁴⁴ Humans with dry eye disease also show reduced OAHFA levels in tears, accompanied by increased phosphatidylcholine,⁴⁵ and both these phenomena were observed in association with gland shortening in the present work. The same study also observed significantly lower levels of TAG in patients with moderate versus mild dry eye symptoms, changes which once again were mirrored in meibum from the short glands sampled in the current work. This finding is consistent with an earlier study in humans, given gland shortening has been associated with worsening of symptoms.⁴⁶ 

Sphingolipids (SPLs) play roles in cell signaling, inflammation, and apoptosis⁴⁷ and Cer is likely to assist in maintaining MG morphology and preventing inflammation.⁴⁸ Cers are also the central intermediates of sphingolipid metabolism.⁴⁹ Therefore, these two lipid classes have very important physiological functions. There is an association between poor meibum quality and lower Cer levels, and with higher SM levels.^50,51 This study reports less Cer and more SM in short glands compared to long glands, which suggests the presence of inflammation and cell necrosis in short glands. The ratio of Cer/SM relative to gland morphology and function may be important as potential biomarkers in MGD. 

The effects of SM at the ocular surface have received limited attention to date.⁵² Sphingomyelins, a key component of the plasma membrane, is critical in maintaining membrane homeostasis. However, excessive accumulation of SM at the neuronal plasma membrane can lead to a higher intracellular calcium level, resulting in oxidative stress and cell death.^53,54 Notably, our study showed a significant increase in the amount of SM in short glands, which may warrant further exploration relative to the role of inflammation in gland changes. 

Conversely, PE is also an important phospholipid in the cell membranes of mammals and is required for the stability of lipoproteins.⁵⁵ Dyslipidemia is one of the significant risk factors for MGD.⁵⁶ The proportions of PE (4.6 ± 3.2 vs. 16.7 ± 4%, P < 0.05) and SM (7.2 ± 4.8 vs. 14.0 ± 8.1%, P < 0.05) in patients with chronic blepharitis-keratoconjunctivitis sicca (CB-KCS) were significantly reduced compared with normal controls, respectively.⁵⁷ In the current study, PE was numerically but not statistically reduced in short glands, but SM was significantly higher in short glands compared with long glands. Potential reasons for differences in findings include: (i) the sample population as participants in the current study did not have CB-KCS; (ii) this study used LC-MS, whereas Shine and McCulley used gas chromatography MS; and (iii) the current study identified lipids using intensity and area compared to identification by retention time. 

Animal models have demonstrated that gland shortening is associated with changes in the lipid composition of the meibum.⁵⁸ It is assumed that with shortening of the length of a gland, there would be a corresponding decrease in the number of sebaceous lobules that produce the lipid, compared to a long gland. However, with current meibography techniques, it is difficult to visualize and count the number of lobules. Therefore, this was not explored in this study. Shortened or atrophied glands frequently exhibit structural changes in the ducts, including ductal dilation and obstruction, which may impede the flow of meibum and contribute to the altered lipid profile observed in these glands.⁵⁹ However, the underlying mechanism driving changes in lipid composition between glands of different length is not clear and requires further investigation. 

Along with lipids, proteins also play a major role in maintaining ocular surface homeostasis.⁵³ This study found the average protein concentration in short glands was significantly higher in short glands compared with long glands. A limitation of the study was that further protein analysis could not be conducted due to a low sample volume available. Previous studies on meibum proteomics revealed that meibomian glands secrete a number of proteins into the tear film and more than 90 proteins have been identified in human meibum involving keratins (K1, 5, 6, 7, 9, 10, 13, and 16), surfactant proteins (SP-B and SP-C), lipocalins, lactoferrin, phospholipid transfer proteins, lipophillins, cytochrome c, farnesoid X laminin α-3 chain, lysozyme c, and proteoglycans.^60–62 It is intriguing here that, overall, the lipids were reduced in short glands, but the protein concentration was higher. How proteins interact with lipids and their association with gland morphology needs to be explored further. 

A strength of the study is the comparison between glands from the same eye in the same individual. Limitations of the study include the lack of assessment of other gland specific parameters, such as meibum quality and expressibility, telangiectasia or plugging of orifices, or assessment of symptoms. It is unclear whether a short gland is representative of an abnormality, although this notion seems to be supported in the literature, given these areas are usually associated with missing glands and often described as atrophic.⁶³ The study inclusion criterion required morphological gland changes rather than a diagnosis of MGD, which may limit the generalizability of the findings. The consequences of a short and long gland within an eye with respect to dry eye disease requires further investigation. Population studies on meibomian gland length distribution could also help to establish a clear definition of what constitutes “normal” gland length. 

The results for the different lipid classes are presented as mole percentages relative to total lipids, which enables comparison with previous work, rather than absolute quantification. Conceivably, future studies may perform targeted analysis for those classes showing significant differences with gland length. Compositional repeatability of short and long glands has also not been assessed due to the complexity of the methods. However, future work could assess the intersubject variability of the two measurements from each gland. Future studies could adopt imaging metrics to quantify other morphological characteristics of the gland, and evaluation of the upper eyelids would also be of interest. Finally, whether the change in morphology leads to a change in composition or the variation in meibum secretions leads to the altered structure is still to be determined. 

Conclusions

This study indicates a relationship between meibomian gland length and the composition of expressed meibum within the same eye of an individual. There are differences in several lipid classes between short and long glands, specifically the mole percentages of CE, TAG, OAHFA, and Cer are lower, whereas PC and SM are higher among shorter glands. These differences may contribute to disease activity. 

Acknowledgments

Supported by the ARVO's publication financial assistance program. 

Author Contributions: F.I. was responsible for designing and writing the protocol, conducting the clinical and laboratory tests, analyzing, and interpreting results and writing the manuscript. S.M. was responsible for laboratory analysis and interpretation of the data. F.S., E.P., and J.T. contributed to the design of the study and critically reviewing the manuscript. 

Disclosure: F. Iqbal, None; F. Stapleton, None; S. Masoudi, None; E.B. Papas, None; J. Tan, None