Abstract
Purpose.:
Matrix metalloproteinases (MMPs) are degrading enzymes which maintain and remodel tissue architecture. Upregulation of MMP-9 has been associated with corneal erosions and ulceration. As these conditions are often exacerbated on waking, suggesting that degrading activity is upregulated overnight, this study set out to determine the diurnal variation of MMP-9, Tissue Inhibitor of Metalloproteinase (TIMP)-1, and Neutrophil Gelatinase-Associated Lipocalin (NGAL).
Methods.:
Flush tears were collected from 46 healthy, non-contact lens wearers at midday, before sleep, and immediately on waking. Total protein content (TPC) was measured using the bicinchoninic acid method, and MMP-9, TIMP-1, and NGAL concentrations were measured using sandwich enzyme-linked immunoassay. Statistical analysis was performed using repeated measures analysis of variance.
Results.:
TPC was 3.4 ± 1.5 mg/mL, 5.0 ± 3.7 mg/mL and 15.5 ± 8.4 mg/mL for midday, before sleep, and on waking respectively, the latter being significantly greater than the other two (P < 0.001). MMP-9 concentrations at the corresponding time points were 9.8 ± 14.3 ng/mL, 8.5 ± 11.7 ng/mL, and 2000.7 ± 1950.7 ng/mL. Again, the value on waking was significantly greater than the previous two visits (P < 0.001). TIMP-1 concentrations exceeded those of MMP-9 at midday but the ratio of the two reversed on awakening.
Conclusions.:
Concentrations of MMP-9 are negligible during the day and completely inhibited by TIMP-1. On awakening, MMP-9 increases 200-fold, an increase that is not completely inhibited by TIMP-1. This diurnal change, along with the presence of NGAL which protects MMP-9 from degradation, suggests that the closed eye is an environment conducive to extracellular matrix remodeling.
Matrix metalloproteinases (MMPs), also called matrixins, are a family of degrading enzymes whose function is to maintain and remodel the tissue architecture by their ability to degrade the structural proteins of the extracellular matrix (Blanco A, et al.
IOVS 2008;49:ARVO E-Abstract 3404).
1,2 They have been implicated in diverse physiological processes including embryonic development, tissue morphogenesis, and wound repair, as well as pathologic processes such as inflammatory diseases, epidermolysis bullosa,
2 rheumatoid arthritis,
2,3 and the progression of cancer due to the breakdown of the structural barriers.
2
There are at least 23 MMPs and these can be grouped into subfamilies: collagenases,
4,5 gelatinases,
4 stromelysins,
4 matrilysins,
4 and membrane-types.
1 In the cornea, MMP-9, also known as gelatinase B, is the primary matrix-degrading enzyme produced by basal corneal epithelial cells
6 and neutrophils
7 and is known to degrade the major components of the epithelial basement membrane such as collagen type VII
8 and impede re-epithelialization of the cornea.
9 As the fibrils responsible for anchoring the basement membrane to the stroma are composed of type VII collagen, an upregulation in MMP-9 may result in a weakening of these attachments, contributing to the cascade of events resulting in corneal erosion.
10,11 An upregulation of MMP-9 has been associated with the epithelial defect in corneal ulceration
6,12 as well as in corneas with ocular rosacea,
9 keratoconus,
13 and pterygia.
14
MMP-9 can be inhibited, and hence regulated, by binding to tissue Inhibitor of metalloproteinase (TIMP)-1.
15 The MMP/TIMP ratio is an indicator of the regulation of matrix degradation and an imbalance in this ratio toward excess MMP-9 may result in collagen degrading effects.
16 Neutrophil Gelatinase-Associated Lipocalin (NGAL) which has a number of possible roles in the tear film,
17 –19 can also bind to MMP-9 when released from neutrophils. Its role however is thought to be protective of MMP-9 degradation, thereby increasing its activity.
18
Conditions such as recurrent corneal erosions and corneal infections, as well as contact lens-related adverse events, are exacerbated on waking, suggesting that degrading activity is upregulated overnight. Thus it is essential to understand how MMP-9 and its associated factors, vary as a function of the diurnal cycle in healthy individuals and what levels are tolerated by the ocular surface. This study therefore set out to establish the diurnal variation of MMP-9, TIMP-1, and NGAL, in the healthy ocular surface.
This study was approved by the Institutional Ethics Committee of the University of New South Wales (UNSW) for the first 37 participants (Approval numbers: 084055 and 084056) and the Vision CRC and Institute for Eye Research Ethics Committee (VIHEC) Human Ethics Committee with ratification from UNSW for the last nine participants (Approval number 10/09 and Human Research Ethics Committee [HREC] 10244). The tenets of the Declaration of Helsinki were followed. Informed consent was obtained from 46 participants (27 females, 19 males) aged between 18 and 46 years of age (mean 28 ± 6 years) after explanation of the nature and possible consequences of the study. Participants were recruited through the general university population and through approved email notices.
Participants were healthy and were not taking medications for any general health conditions. A requirement of the study was that the participants were either non-contact lens wearers or non-habitual lens wearers who had not worn contact lenses in the month preceding the study. Pregnant or lactating women were excluded from the study.
Diurnal Study.
Day-To-Day Repeatability Study.
Flush Tear Collection.
Total protein content (TPC) was determined using the bicinchoninic acid (BCA) method and using reagents (Pierce; Thermo Fisher Scientific) and flat-bottom 96-well microplates (Nunc-F Maxisorp; Thermo Fisher Scientific). Serial dilutions of bovine serum albumin (BSA) were used as standard. This was loaded in triplicate starting at 2 mg/mL down to 0.1 mg/mL in ultrapure laboratory grade water (MilliQ; Millipore, Billerica, MA) and 10 μL was added to each well. Tear samples were loaded at a 1:10 dilution. Tears were analyzed in duplicate and a 10 μL volume was added to each well. Solution A (BCA Protein Assay Reagent A; Pierce, Thermo Fisher Scientific) and solution B (BCA Protein Assay Reagent B; Pierce, Thermo Fisher Scientific) were combined in respective volumes of 20 mL and 0.4 mL. This mixture was then added at a volume of 200 μL per well. The optical density was read with a microplate reader (Spectra Fluor Plus, Tecan Multifunction Microplate reader using the X-Fluor 4 software; Tecan, Männedorf, Switzerland) at 595 nm after 30 minutes and a standard linear curve generated using the BSA as a reference.
Gelatinolytic activity of MMP-9 was determined using gelatin zymography. Ten micrograms of each tear sample was added to 5 μL of sodium dodecyl sulfate (SDS) sample buffer (2× Novex Tris-Glycine; Invitrogen, Carlsbad, CA) and the volume was made up to 12 μL with ultrapure laboratory grade water (MilliQ; Millipore). A concentration of 50 ng/mL of MMP-9 standard (R&D Systems) was activated by incubating with 1 mM 4-aminophenyl-mercuric acetate
26 (APMA; Sigma-Aldrich) at 37°C for 1 hour. A 1-μL volume of this was then added to a 5 μL volume of 2 × sample buffer and topped up to a volume of 12 μL with ultrapure laboratory grade water (MilliQ; Millipore). This was incubated for 30 minutes at room temperature (not shown).
Each sample was resolved using 10% zymogram gelatin gel (Novex; Invitrogen) under denaturing but nonreducing conditions. Gels were electrophoresed at a constant voltage of 120 V for 3.5 hours in running buffer (glycine, Tris base and 1% SDS) at 4°C. As proteolytic activity can be reversibly inhibited by SDS during electrophoresis and recovered by incubating the gel in aqueous Triton X-100,
27 the gels were washed with 2.5% (vol/vol) Triton X-100 (Astral Scientific, Amresco, Solon, OH) for 1 hour. This was then decanted and the gels were equilibrated with developing buffer (Novex zymogram developing buffer; Invitrogen) for 30 minutes. The gels were then transferred to fresh developing buffer and incubated overnight at 37°C for 16 to 20 hours. Proteins were detected by staining with Coomassie Blue R-250 for a minimum of 2 hours and destaining in 30% (vol/vol) ethanol, 10% (vol/vol) acetic acid.
Disodium ethylene-diaminetetra-acetic acid (EDTA), which is known to inhibit MMPs, was added to a separate renaturing and developing buffer. After electrophoresis, the final well containing an identical sample to that in the third well, was cut off and renatured in buffer containing 20 mM EDTA, followed by incubation in developing buffer containing EDTA. These bands should have no proteolytic activity on staining and were used as controls (results not shown). Cancer serum samples were used as a positive control, courtesy of Michele Madigan, SaveSight Foundation.
Normality was tested using the Shapiro-Wilk test and only the data for MMP-9 at midday and before sleep and TIMP-1 on awakening were not normally distributed when eyes were averaged. Repeated measures analysis of variance (ANOVA) was used to analyze the visits. Mauchly's test was used to verify the sphericity assumption. Where main effects were significant, Student's t-tests with Bonferroni correction were used post hoc. Software (PASW version 18.0 GP; IBM, Chicago, IL) was used for the analysis. In a small number of cases inadequate sample volume was obtained resulting in missing data. To avoid the need to delete all the data from affected subjects during repeated measures ANOVA, replacement was made using expectation maximization (EM), with any negative values generated by the process replaced with zeros. In cases where MMP-9 levels were not detectable, samples were allocated a value of zero. Results are reported as mean ± SD.
Repeatability between both days and eyes was evaluated by coefficients of repeatability defined as the SD of the mean difference between replicates multiplied by 1.96.
28
Table 3 lists the concentrations of the proteins measured at each time point. TPC was significantly greater on awakening than at midday and before sleep (
P < 0.001) and midday and before sleep were also significantly different from each other (
P = 0.03).
Table 3. The Diurnal Variation of Each of the Proteins Measured in the Tear Film
Table 3. The Diurnal Variation of Each of the Proteins Measured in the Tear Film
| Visit | TPC (mg/mL) | MMP-9 (ng/mL) | TIMP-1 (ng/mL) | MMP9/TIMP-1 Ratio | NGAL (ng/mL) | MMP-9:NGAL Ratio |
| Midday | 3.4 ± 1.5 | 9.8 ± 14.3 | 74.5 ± 39.7 | 0.3 ± 0.4 | 680.8 ± 523.3 | 0.01 ± 0.04 |
| Before sleep | 5.0 ± 3.7 | 8.5 ± 11.7 | — | — | — | — |
| Upon awakening | 15.5 ± 8.4 | 2000.7 ± 1950.7 | 277.8 ± 282.2 | 13.9 ± 15.8 | 3620.3 ± 1832.1 | 0.16 ± 0.14 |
The concentration of MMP-9 was significantly greater on awakening compared with both midday and on awakening (P < 0.001) and midday and before sleep were not significantly different (P = 0.6).
TIMP-1 was measured as it is the main inhibitor of MMP-9. Concentrations were significantly greater on awakening compared with midday (P < 0.001). The MMP-9/TIMP-1 ratio provides an indication of the potential for MMP-9 to cause tissue damage. The MMP-9/TIMP-1 ratio was significantly greater on awakening than at midday (P < 0.001). These ratios indicate that at midday, the concentration of TIMP-1 exceeded that of MMP-9, hence completely inhibiting MMP-9. In contrast, on awakening the ratio indicates that this balance is reversed, with an excess of MMP-9 in the tear film, indicating a shift from TIMP-1 dominance during the day to MMP-9 dominance on awakening.
NGAL was significantly elevated on awakening compared with midday (P = 0.002). When considered as an MMP-9/NGAL ratio, the ratio on awakening was significantly greater than that at midday (P = 0.006). At both the midday and on awakening visits, NGAL exceeded the concentration of MMP-9 present in the tear film.
Sex was not significantly different for TPC, MMP-9, NGAL or TIMP-1 and its ratio with MMP-9 (P = 0.86, 0.13, 0.39, and 0.23 respectively). The ratio of MMP-9 to NGAL showed a significant difference between the sexes, with males having a significantly greater ratio on awakening than the females and the reverse at midday (P = 0.048).
MMP-9 is a vital but potentially destructive enzyme by virtue of its ability to degrade corneal collagen. Control of this activity is thus critical to maintaining corneal health. When the eyes are open, there is little cause for concern as, in agreement with previous studies,
6,8 –10,13 these data suggest that MMP-9 concentrations are negligible and significantly exceeded by those of the inhibitor molecule TIMP-1. The potential for proteolytic activity is thus inhibited and hence the corneal epithelium and basement membrane are protected from degradation.
During sleep however, this situation changes dramatically as MMP-9 concentration increases roughly 200-fold, while TIMP-1 barely triples. The shift in the balance from TIMP-1 to MMP-9 dominance, which may be a downstream effect of increased neutrophil activity,
7,24 suggests that there may be an increased propensity for basement membrane degradation overnight. This is further supported by the excess NGAL in the tear film compared with MMP-9, both at midday and on awakening, a finding that is in agreement with previous studies.
7 While this diurnal variation may be necessary to facilitate increased cell desquamation during sleep, or it may be that this increase in MMP-9 is a byproduct of epithelial cell breakdown during eye closure, it also apparently sets the scene for tissue destruction by virtue of the excess MMP-9 present. In reality however, such tissue damage was not evident among the participants in this study, nor does it typically occur overnight in the general population. Clearly then, our understanding of the regulatory mechanisms of MMP-9 in the tear film is incomplete.
Looking at the zymography results indicates that the MMP-9 in the closed-eye tear film is in the zymogen form rather than the active form (
Fig. 1), suggesting that this may be one contributory factor. Additionally, some MMP-9 was detected at molecular weights in excess of 220 kDa indicating it to be in the dimeric state which does not differ in enzymatic activity from the inactive form at 92 kDa.
30 It may also be that MMP-9 is being regulated by nonspecific inhibitors such as α
2-macroglobulin which traps activated MMP so that the complex can be removed by the receptor.
34
Many conditions such as recurrent corneal erosions and corneal ulcers are known to be exacerbated by eye closure.
35 Due to its action on collagen and hence the anchoring complexes of the corneal epithelium, MMP-9 could contribute to the increased incidence of recurrent erosions by virtue of its diurnal variation, particularly when the regulatory mechanisms are not controlled.
11,36 The diurnal profile of MMP-9 and its associated factors could also indicate why the epithelium is particularly prone to erosion and infection with overnight contact lens wear,
37,38 as well as corneal ulceration
10,12,13,39 and corneal desiccation on awakening.
MMP-9 concentration in the tears varies considerably within the population at any given time as demonstrated by the coefficients of repeatability presented in
Table 2. This may be in part attributable to the variability in the flush tear collection rate as shown in
Table 1. The lack of correlation with clinical signs of redness suggests that this ubiquitous sign of ocular distress is not particularly helpful as a marker for raised MMP-9 levels however. It may also be that this presence of elevated levels of the inactive form of MMP-9 does not produce redness because it does not constitute an insult to the ocular surface.
Finally, it may be that these results also have implications for certain topical therapies, in particular the use of prostaglandin analogues. These compounds are used for the treatment of glaucoma and work by increasing the concentration of MMP-9
40 as a means of degrading ciliary muscle extracellular matrix and increasing uveoscleral outflow.
41 Concurrent increased MMP-9 concentrations have also been found in the tear film,
42 increasing the potential for corneal damage as has been found in both animal
43 and human studies.
44 It may then be advisable to adjust the instillation timing for prostaglandin drops to avoid the periods before sleep and immediately on awakening, particularly for those already prone to corneal erosions.
In conclusion, this study has shown that there is a substantial diurnal variation of MMP-9 and its associated factors. Concentrations of MMP-9 are negligible during the day and completely inhibited by TIMP-1. On awakening MMP-9 increases 200-fold, an increase that is not completely inhibited by TIMP-1. This diurnal change, along with the presence of NGAL, suggests that the closed eye is an environment conducive to extracellular matrix remodeling. The fact that we typically do not see corneal damage on awakening suggests that other regulatory mechanisms are active to prevent excess extracellular matrix degradation.
Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2010; and the British Contact Lens Association Conference in Birmingham, United Kingdom, May 2010.
Supported by the Australian Government through the Australian Postgraduate Award, the American Optometric Foundation, the George and Jill Mertz Foundation, and CIBA VISION through the William C. Ezell Fellowship, the Cornea and Contact Lens Society of Australia through a Postgraduate Award, the OVRF-Maki Shiobara Scholarship, and the Brien Holden Vision Institute through a postgraduate grant to cover facilities and supervision.
Disclosure:
M. Markoulli, None;
E. Papas, None;
N. Cole, None;
B.A. Holden, None
The authors thank Michele Madigan for her assistance with zymography.