Highlights
- •
Ocular surface mucins were imaged using fluorescein-tagged wheat germ agglutinin dye and a custom imaging system.
- •
Mucin density appeared greatest on the bulbar conjunctiva and lowest on the cornea.
- •
Contact lens wear reduced F-WGA dye binding across the ocular surface, with the most marked effect at the cornea.
- •
F-WGA associated fluorescence appeared reduced in the lid wiper region for symptomatic contact lens wearers.
- •
F-WGA imaging is a key tool in furthering our understanding of contact lens discomfort and ocular surface disease.
Abstract
Purpose
Fluorescein-labelled wheat germ agglutinin (F-WGA) acts as a marker for ocular surface mucins. This clinical study sought to investigate whether the degree of F-WGA fluorescence observed at the ocular surface differed between symptomatic contact lens wearers, asymptomatic contact lens wearers and non-contact lens wearers, using a novel imaging system.
Methods
Twenty-five participants (10 symptomatic contact lens wearers, 10 asymptomatic contact lens wearers and 5 non-contact lens wearers) attended a single study visit. Photographs of the cornea, bulbar and tarsal conjunctiva were captured following application of F-WGA solution.
Results
The imaging system captured high-resolution images of F-WGA fluorescence at the ocular surface. The degree of fluorescence differed between the ocular surface regions ( p < 0.001). A significant difference in fluorescence was observed between participant groups for the cornea ( p = 0.01), with both the symptomatic and asymptomatic contact lens wearers showing lower fluorescence than the non-lens wearers. F-WGA associated fluorescence appeared diminished in the lid wiper region of the symptomatic lens wearers, compared to the asymptomatic group ( p = 0.025).
Conclusion
The use of F-WGA as a clinical marker for ocular surface mucins allows an improved understanding of their distribution across the ocular surface. Contact lens wear appears to negatively impact mucin density across the ocular surface, with the most marked effect on the cornea. F-WGA fluorescence appeared diminished in the lid wiper region for the symptomatic contact lens wearing group, indicating that mechanical interaction in this region may play a role in the aetiology of contact lens discomfort. Given the ability of F-WGA to disclose mucin distribution across the ocular surface it is likely to be a key clinical tool in furthering our understanding of (i) the aetiology of contact lens related discomfort, (ii) contact lens designs/materials to minimise interaction with the ocular surface and (iii) dry eye disease and other ocular surface diseases.
1
Introduction
At the interface between the tear film and the ocular surface, the stratified epithelium expresses a range of membrane-associated mucins, including MUC1, MUC4 and MUC16 . These mucins form a dense glycocalyx layer, which acts as a barrier to prevent the penetration of pathogens, lubricates the motion of the eyelids, prevents adhesion of facing cells (i.e. corneal epithelium and tarsal conjunctiva) during blinking/sleeping and provides a wettable surface to aid tear film adherence . This structure is however vulnerable to damage and is known to be altered by contact lens wear and a range of ocular surface diseases including allergy, non-autoimmune dry eye, autoimmune dry eye and infection .
It is well documented that wheat germ agglutinin (WGA) is able to bind to N-acetyl- d -glucosamine and sialic acid , which are found in the human cornea . By attachment of a fluorescent conjugate, the presence of WGA can be visualised using fluorescence microscopy . Recent work has shown that fluorescein-labeled WGA (F-WGA) can act as a clinical marker for the presence of ocular surface mucins and the glycocalyx . However, due to the low intensity of fluorescence observed with F-WGA, this approach has previously required the use of a fluorophotometer (an ophthalmic instrument which measures fluorescent intensity across a 2 mm diameter region) and thus has lacked the spatial resolution of a conventional imaging system. In addition, previous F-WGA research has focused on the cornea and bulbar conjunctiva, whereas recent work has highlighted the importance of the tarsal conjunctiva and in particular the lid wiper region to ocular comfort . The objectives of this work were (i) to develop an imaging system to characterise F-WGA staining at the ocular surface, and (ii) to use this instrument to investigate whether F-WGA fluorescence at the ocular surface differed between symptomatic contact lens wearers, asymptomatic contact lens wearers and non-contact lens wearers.
2
Methods
2.1
Custom imaging system
In the development of a custom imaging system to characterise F-WGA fluorescence at the ocular surface, a number of technical challenges had to be overcome, including the low level of fluorescence produced by F-WGA, the curvature of the ocular surface, continuous microscopic eye movements and photobleaching of the fluorophore (reduction in fluorescence due to excessive light exposure). The imaging system used a high sensitivity back-illuminated CMOS (complementary metal-oxide-semiconductor) imaging sensor (8.4 μm pixel pitch), combined with small aperture macro optics and flash illumination to obtain the images captured in this study. A digital camera (Sony α 7S II, Sony Europe Ltd., Weybridge, UK) was mounted on an ophthalmic instrument base (Takagi Ltd., Manchester, UK) incorporating a chin rest, allowing the camera to be aligned with the participant’s eye during clinical imaging. To allow high resolution imaging across the curved ocular surface, a macro lens was selected (Sony FE 90 mm Macro lens and 26 mm extension tube, Sony Europe Ltd., Weybridge, UK) and a small aperture employed (f/20) to optimise depth of field. To minimise image blur whilst ensuring sufficient fluorescence from the F-WGA, a macro flash illumination source (Canon MT-24EX, Canon Europe Ltd., Uxbridge, UK) and a short exposure time (1/25 s) were used. Excitation optical bandpass filters (MF475-35 ø50 mm, Thorlabs Ltd., Ely, UK) were positioned in front of both flash heads and an emission optical bandpass filter (MF530-43 ø50 mm, Thorlabs Ltd., Ely, UK) positioned over the camera objective with custom mounts. These optical bandpass filters were selected to match the spectral characteristics of F-WGA (excitation of 494 nm and emission of 518 nm). The flash was used in a manual mode (1/1), with the flash heads positioned on each side of the objective and rotated 45 degrees inwards to ensure even illumination. An additional low intensity LED lamp (LIU470A, Thorlabs Ltd., Ely, UK) was mounted below the objective with an excitation optical bandpass filter (MF475-35 ø25 mm, Thorlabs Ltd., Ely, UK) to allow camera alignment and focusing, prior to image capture. A live video feed from the camera was displayed on a wall-mounted monitor to aid in this process. To ensure that the illumination was consistent, images of a fluorescent calibration slide were captured (FSK5 Green, Thorlabs Ltd., Ely, UK). The imaging system was calibrated with a concentric square target (R3L3S3P, Thorlabs Ltd., Ely, UK). Prior to the clinical study, testing was undertaken to investigate whether use of the custom imaging system resulted in any significant F-WGA photobleaching. In this work, F-WGA solution was dispensed between two glass microscopy slides with a 10 μm separation to model its distribution within the tear film. A series of 7 images of this F-WGA film were then captured (10 s between each captured image) and the intensity of fluorescence was then assessed for each image using custom image analysis software.
2.2
F-WGA preparation
Fluorescein labelled-wheat germ agglutinin (F-WGA) (ThermoFisher Scientific Inc., UK) was suspended in sterile ophthalmic saline (Ami-dose, Albatron Ltd., UK) to form a 5% suspension (as previously described by Mochizuki et al. ). This suspension was then passed through a sterile 0.22 μm Millex syringe filter (Merck Inc., Darmstadt, Germany) into a sterile 1.5 ml microcentrifuge tube (Sigma–Aldrich Ltd., UK). The microcentrifuge tube was stored for up to one week in a laboratory refrigerator (5 ± 3 °C). After one week the microcentrifuge tube and contents were discarded and a new suspension prepared. This F-WGA solution was used for both the laboratory imaging (described in Section 2.1 ) and the clinical study.
2.3
Clinical study
This was a prospective, controlled, non-randomised, open-label, parallel group clinical study which used the custom imaging system to quantify fluorescence after the application of F-WGA. This technique was used to compare mucin distribution at the ocular surface between a group of symptomatic and asymptomatic contact lens wearers, in addition to a control group of non-contact lens wearing participants. In total 25 participants were recruited, with 10 symptomatic contact lens wearing participants (defined as a CLDEQ-8 questionnaire score ⩾14 and a difference between contact lens wear time and comfortable wear time ⩾2 h), 10 asymptomatic contact lens participants (defined as a CLDEQ-8 questionnaire score ⩽8 and a difference between contact lens wear time and comfortable wear time ⩽1 h) and 5 non-contact lens wearing participants (defined as no more than one day of contact lens wear experience in the last 2 years and a DEQ-5 questionnaire score <8). As this work represented the first time the custom imaging system had been deployed for the assessment of F-WGA no a priori power analysis was possible. These subject numbers were considered to be reasonable to (a) determine if the imaging system could be used, (b) to provide some initial evaluation of the utility of this system to compare symptomatic and asymptomatic contact lens wearers with non-wearers and (c) to provide a dataset for power analysis for future work in this area.
All participants provided written informed consent before inclusion in the study. The study was conducted in accordance with the principles of the Declaration of Helsinki and The University of Manchester Research Ethics Committee provided ethical approval. Individuals with a history of ocular/systemic disorders that would normally contraindicate contact lens wear, history of ocular surgery, use of topical ophthalmic medication, corneal distortion, pregnant or breastfeeding, a history of anaphylaxis or severe allergic reaction or any infectious or immunosupressive disease that would pose a risk to study personnel, were excluded.
Each participant attended for a single study visit. Here, the contact lens-wearing participants attended wearing their habitual contact lenses, having worn them for at least 4 h. The non-contact lens wearing participants attended clinic a minimum of 4 h after waking. Subjective ocular comfort was assessed using a 0–100 visual analogue scale (0 = very uncomfortable/100 = excellent comfort). If the participant was a contact lens wearer, habitual lens fit was assessed (horizontal and vertical centration, corneal coverage and movement), details of the habitual lenses/care system were recorded and their contact lenses were then removed. Corrected high contrast distance logMAR visual acuity was then recorded. Slit lamp biomicroscopy was performed with clinical signs graded to the nearest 0.1 unit using the Efron Grading Scales (without sodium fluorescein assessment to avoid adversely influencing F-WGA imaging). Baseline photographs were then captured (in a randomised order) of seven ocular surface regions (i) the cornea (primary gaze with lids retracted using an Eyegenie external lid retractor (Bernell Inc., Mishawaka, IN)), (ii to v) the bulbar conjunctiva (in superior, inferior, nasal and temporal gaze) and (vi and vii) the tarsal conjunctiva (following evertion of the upper and lower eyelids), using the custom imaging setup described in Section 2.1 . A 5 μl droplet of F-WGA solution was applied to the temporal conjunctiva of one eye, followed by the same process in the other eye 3 min later (approximately 15 min after contact lens removal or after slit lamp biomicroscopy in the case of a non-contact lens wearer). Approximately 5 min and 30 min after F-WGA application, photographs were again captured of the cornea, bulbar conjunctiva and tarsal conjunctiva in a randomised order (a single image captured for each of the seven regions previously imaged). Subjective comfort was also assessed at these two time points using a 0–100 visual analogue scale. Following imaging at the 30 min time point, slit lamp biomicroscopy was undertaken including assessment with sodium fluorescein. Corrected high contrast distance logMAR visual acuity was recorded and the participant was exited from the clinical study.
2.4
Image analysis
During the clinical study, seven ocular surface regions were imaged for each of the participant’s eyes: cornea, bulbar conjunctiva (in superior, inferior, nasal and temporal gaze) and tarsal conjunctiva (upper and lower everted lids). Captured images were exported and coded by the clinical investigator. Image analysis was undertaken by a different masked investigator who analysed all clinical images in a randomised order. For each image, custom MATLAB image analysis software was used to isolate the region of interest and calculate the degree of fluorescence (assessed as the mean pixel value of the RGB image). For the image captured of the cornea, the region of interest was the entire visible corneal region. For the four images captured of the bulbar conjunctiva (with the subject in superior, inferior, nasal and temporal gaze), the region of interest was the entire visible bulbar conjunctiva in each image (excluding the tear film meniscus and areas obscured by the lashes). For the images captured of the everted tarsal conjunctiva, the region of interest was the entire tarsal conjunctiva visible in each image. To calculate the degree of fluorescence attributable to the F-WGA, the baseline fluorescence was subtracted from the 5-min and 30-min post-FWGA application images. This value was termed the Background Subtracted Intensity (BSI). An image analysis algorithm was also developed to characterise fluorescence in the lid wiper region of the upper and lower eyelid. The masked image analysis investigator was instructed to draw a line along the line of Marx and the custom image analysis algorithm (MATLAB, The MathWorks, Natick, MA), then calculated the average BSI over the lid wiper region (extending 0.6 mm from the line of Marx as detailed by Efron et al. ). The BSI of a reference region of visible tarsal conjunctiva (all of the visible tarsal conjunctiva >1 mm from the line of Marx) was then subtracted from the BSI of the lid wiper region to allow an assessment of relative fluorescence in the lid wiper region. For this metric of relative lid wiper fluorescence, a value above 0 indicated elevated F-WGA fluorescence in the lid wiper region, whereas a value below 0 indicated a depressed fluorescence in the lid wiper region. A further image analysis algorithm was developed to characterise fluorescence in the corneal limbal region. The investigator was instructed to mark the corneal limbus on the primary gaze image. The algorithm then performed a series of radial pixel intensity scans centred on the cornea. The algorithm then averaged these scans to produce an averaged radial scan intensity profile for each primary gaze image. As with the regional analysis, the baseline fluorescence profile was subtracted from the F-WGA fluorescence profile, to generate the background subtracted profile for both the lid wiper region and the corneal limbus region.
2.5
Statistical approach
Habitual contact lens wearing time, comfortable hours of lens wear and CLDEQ-8 data were analysed using a Student’s t -test. Contact lens fitting data were analysed using a chi-square test. Biomicroscopy data were assessed using a linear mixed model, with participant group, participant ID (random effect), eye (nested within participant ID) and assessment period (pre or post-F-WGA imaging) as factors of interest. Subjective comfort scores (0–100 VAS) were assessed using a linear mixed model, with participant group, participant ID (random effect), age and assessment period (pre or post-F-WGA imaging) as factors of interest. Background-subtracted ocular surface fluorescence intensity data were assessed using a linear mixed model, with participant group, participant ID (random effect), age, eye (nested within participant ID), imaging time point and ocular surface region as factors of interest. Visual acuity data were assessed using a linear mixed model with assessment period, participant group and participant ID (random effect) as factors of interest. The difference in fluorescence between the lid wiper region and the reference region of the tarsal conjunctiva were assessed using a linear mixed model, with participant group, participant ID (random effect), age, eye (nested within participant ID), imaging time point and eyelid as factors of interest. For all linear mixed models, interaction factors were initially included in the model, but were removed if the factors were not significant at p > 0.2. Tukey post-hoc analysis was performed where appropriate. A p -value of less than 0.05 was considered statistically significant. All data were analysed using JMP 14, Version 14.3 (SAS Institute Inc. Cary, NC, USA).
3
Results
3.1
Clinical assessment
Table 1 details the participant demographics, habitual lens information, typical lens wear time and typical subjective ocular comfort for the three participant groups. A statistically significant difference was observed between the two contact lens wearing groups, for the comfortable wear time ( F = 12.4, p = 0.002) and uncomfortable wear time (i.e. total hours per day – comfortable hours per day) ( F = 44.5, p < 0.0001). No significant difference was observed between the subject groups for subject age ( F = 3.2, p = 0.06), days per week of contact lens wear ( F = 3.2, p = 0.09) or hours per day of contact lens wear ( F = 3.0, p = 0.10). A statistically significant difference was seen between the two contact lens wearing groups for the CLDEQ-8 score ( F = 193.7, p < 0.0001). All habitual contact lenses were seen to fit acceptably, with no difference observed between the symptomatic and asymptomatic contact lens wearing groups for the percentage of optimum lens fits ( χ 2 = 1.9, p = 0.17).
| Parameter | Symptomatic | Asymptomatic | No lens wear |
|---|---|---|---|
| Gender | 9 female/1 male | 7 female/3 male | 3 female/2 male |
| Age | 39.2 ± 13.6 | 28.4 ± 4.7 | 40.6 ± 13.8 |
| Habitual lens material | 8 hydrogel/2 SiH | 5 hydrogel/5 SiH | – |
| Habitual lens modality | 8 DD/2 FR | 5 DD/5 FR | – |
| Days of CL wear per week | 4.3 ± 2.2 | 5.9 ± 1.9 | – |
| Hours of CL wear per day | 10.2 ± 1.8 | 13 ± 4.8 | – |
| Comfortable hours of lens wear per day | 6.3 ± 2.4 | 12.6 ± 5.1 | – |
| Uncomfortable wear time (hours per day – comfortable hours per day) | 3.9 ± 1.6 | 0.4 ± 0.5 | – |
| CLDEQ-8 (DEQ-5 for non-lens wearers) | 20.7 ± 2.5 | 5.3 ± 2.5 | 2.0 ± 1.6 |
High contrast visual acuity and subjective ocular comfort scores prior to and following the F-WGA imaging procedure are shown in Table 2 . A statistically significant improvement in visual acuity was observed following the F-WGA imaging procedure ( F = 6.1, p = 0.02), although this did not reach the level of clinical significance. Subjective comfort scores (0–100 VAS) prior to F-WGA application showed a significant difference between the participant groups ( F = 9.8, p = 0.0009), with a Tukey post-hoc analysis demonstrating a significantly lower comfort for the symptomatic group, than both the asymptomatic and non-lens wearing group. The linear mixed model analysis for the entire 0–100 VAS subjective comfort data set highlighted that the effect of participant group ( F = 5.1, p = 0.02), assessment period ( F = 6.04, p = 0.02) and the interaction term group * assessment period ( F = 6.6, p = 0.006) reaching statistical significance; with the effect of age not reaching significance ( F = 3.4, p = 0.08). A Tukey post-hoc analysis demonstrated a statistically significant reduction in comfort following F-WGA application. A Tukey post-hoc analysis on the interaction term (group * assessment period), demonstrated that the change in comfort between the entrance and exit measures was not consistent between the participant groups, with the comfort of the lens wearing groups not changing significantly prior to and following F-WGA application, whereas the non-lens wearing group showed a significant reduction in subjective comfort.
| Parameter | Symptomatic group | Asymptomatic group | No lens wear group |
|---|---|---|---|
| Entrance VA | −0.07 ± 0.09 | −0.03 ± 0.16 | 0.01 ± 0.19 |
| Exit VA | −0.12 ± 0.07 | −0.06 ± 0.15 | 0.00 ± 0.17 |
| Entrance 0–100 VAS comfort | 63.8 ± 20.0 | 87.5 ± 10.9 | 96.0 ± 8.9 |
| Post F-WGA 0–100 VAS comfort | 72.1 ± 17.4 | 78.2 ± 18.0 | 68.4 ± 19.5 |
Slit lamp biomicroscopy grading scores were statistically similar when assessed prior to and following the F-WGA imaging procedure ( Table 3 ). Due to the sensitivity of the imaging system to ocular surface fluorescence, sodium fluorescein was applied only on completion of the F-WGA imaging procedure, with the degree of staining observed typical of that reported in the literature following contact lens wear and on the bare eye for the non-contact lens wearing participants .
