Tear matrix metalloproteinases-9 and ocular surface parameters in diabetics: a cross-sectional study in Shenyang, China

Abstract

Background

Overexpression of tear matrix metalloproteinases-9 (MMP-9) on the ocular surface tissues has been reported to result in ocular surface damage. MMP-9 levels in tears have been listed as one of many tools for confirming dry eye disease (DED).

Objective

This investigation aimed to compare MMP-9 levels and ocular surface parameters in diabetic patients with and without DED.

Design

A cross-sectional study.

Setting

He Eye Specialist Hospital, Shenyang, China.

Participants

This study recruited 144 right eyes of 144 diabetic patients between November and December of 2023, and 110 patients with similar propensity scores were included in the analyses.

Main outcome measures

Non-invasive breakup time (NITBUT), tear film lipid layer (TFLL), conjunctival hyperaemia (redness score (RS)), corneoconjunctival staining (CS), corneal sensitivity and Ocular Surface Disease Index (OSDI) questionnaire were evaluated. MMP-9 was measured using an immunochromatography assay.

Results

In total, 55 patients (55 eyes) were grouped as diabetic dry eye (DDE) and 55 patients (55 eyes) as diabetic non-dry eye (DNDE). The mean MMP-9 concentrations were higher in patients with DDE than DNDE (70.63±52.06 ng/mL vs 33.98±33.93 ng/mL; p<0.001). The optimal cut-off value of MMP-9 to predict DED in diabetic patients was>52.5 ng/mL, with 58.2% sensitivity and 78.2% specificity.

Conclusions

MMP-9 concentration was higher in patients with DDE than DNDE. The MMP-9 test is a potential diagnostic tool for DDE. It may help follow-up diabetic patients with DED and guide clinicians in deciding on anti-inflammatory treatments for these patients.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study assessed and compared the levels of tear matrix metalloproteinases-9 (MMP-9) and its correlation with other ocular parameters in individuals with diabetes, both with and without dry eye disease.

• A comprehensive evaluation of tear film, tear MMP-9 and dry eye symptoms was performed.

• The inclusion of a group with non-diabetic patients would have further bolstered the results of this study.

• The inclusion of multiple inflammatory factors could have better reflected the ocular surface status of the study group.

Introduction

Diabetes mellitus (DM) is a multifaceted chronic metabolic disease caused by decreased insulin production and varied degrees of peripheral insulin resistance resulting in hyperglycaemia.¹ Diabetes is a continuing global epidemic that affects approximately 422 million people globally, and the WHO stated that its prevalence is expected to climb to 592 million by 2035.² In China, it was reported to have reached 12.42% in 2018.³

While diabetic retinopathy (DR) is considered to be the significant ocular consequence of DM and the primary cause of blindness among adults of working age in the majority of industrialised nations,⁴ 47% of patients with DM suffer ocular experience ocular surface damage as a result of detrimental alterations to various components, including corneal endothelium, corneal nerve, central corneal thickness, corneal epithelium and tear film.^{5 6} Tear Film and Ocular Surface Society Dry Eye Workshop II (TFOS DEWS II) reports that dry eye (DE) is a multifactorial disease with DM identified as one of the risk factors.⁷ The consensus is that the harmful effect of DM on corneal nerve sensitivity and the microvasculature of the lacrimal gland, which impairs homeostasis of the tear film contributes to ocular surface insult that ultimately causes the symptoms and signs of DE.^{8 9} However, increasing biological evidence suggests various pathogenesis is at play regarding diabetic dry eye (DDE) versus non-diabetics DE.^{10 11}

Inflammation plays a definite role in the development of diabetes and the onset of dry eye disease (DED). Matrix metalloproteinase-9 (MMP-9) is a gelatinase produced by the lacrimal glands, corneal epithelium, infiltrating leucocytes and fibroblasts.^{12 13} MMP-9 partakes in the breakdown of extracellular matrix and epithelial cell loss.¹⁴ Overexpression of MMP-9 on the ocular surface tissues has been reported to result in ocular surface damage. Therefore, MMP-9 levels in tears have been listed as one of many tools for confirming DE.^{13 15 16} Furthermore, Iyengar et al¹⁷ documented that MMP-9 concentration increased in diabetic peripheral neuropathy; thereby, making the ocular surface characteristics of DDE more complex. Elevated levels of MMP-9 in tear samples are not exclusive to DDE alone. Further studies are needed to understand the difference between the mechanisms of elevated tear MMP-9 levels in diabetic patients with DED.

This study aimed to assess and compare the levels of tear MMP-9 and its correlation with other ocular parameters in individuals with diabetes, both with and without DED.

Methods

Participant selection

Between November and December of 2023, this single-centre, cross-sectional study recruited 144 right eyes of 144 diabetic patients at the Department of Clinical Research, He Eye Specialist Hospital, Shenyang, China. The study followed the Declaration of Helsinki’s guidelines. Every subject previously diagnosed with diabetes was informed and enrolled in this study.

The 2020 ‘Standards of Medical Care in Diabetes’ established by the American Diabetes Association¹⁸ were used to diagnose T2DM in the participants, while the TFOS DEWS II diagnostic criteria were used to identify DE.¹⁹ Participants were divided into the DDE group and diabetic non-DE (DNDE) group based on the TFOS DEWS II diagnostic criteria: (1) non-invasive tear breakup time (NITBUT) ＜10 s, (2) corneoconjunctival staining (CS) score >5 corneal spots, >9 conjunctival spots and (3) Ocular Surface Disease Index (OSDI) questionnaire≥13. To determine a positive DE diagnosis, two or more criteria had to be present.²⁰

The exclusion criteria were: (1) Patients with other ocular disorders that could affect ocular surface parameters and MMP-9, such as pterygium, glaucoma, ocular infection, conjunctivochalasis and allergic conjunctivitis. (2) Ocular surgical history in the last 6 months. (3) Contact lens wearers and (4) patients using topical eye drops. (5) Any other autoimmune disorders that impact tear production, such as Sjögren’s syndrome, rheumatoid arthritis, scleroderma, and polymyositis.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Sample size

PASS 15.0 software was used to calculate the sample size. The sample size calculation is based on the primary outcome measures, namely MMP-9 concentration. For the current study, a sample size of 44 achieves 90% power to detect a mean of paired differences of 10 with an estimated SD of paired differences of 20 and with a significance level (alpha) of 0.05 using a two-sided paired t-test.

Data collected

Age, sex, best corrected visual acuity, intraocular pressure, fasting blood glucose (FBG) and diabetes duration of participants were obtained from their medical records, and no data were missing. The glycosylated hemoglobin (HbA1C) was not included due to missing HbA1C data in half of the patients.

Assessment of symptoms

A quantitative evaluation of DE symptoms was offered using a validated Chinese version of the OSDI.²¹ To get an individual score ranging from 0 to 100 (no symptoms to severe symptoms) points, the questionnaire’s 12 items can be calculated. Patients self-reported based on symptoms in each eye, which may be the same or different in both eyes. In any event, only data from the right eye were included in the analyses.

Clinical evaluations

NITBUT, tear meniscus height (TMH) and conjunctival hyperaemia [redness score (RS)] were performed by Keratography 5M (Oculus, Germany). The RS can range from a low of 0.0 to a high of 4.0. The median value of the final reading was calculated after three consecutive measurements.

Conjunctivocorneal epithelial staining uses methodologies described by Arita et al²² to quantify damage to the cornea and conjunctival epithelium.

To evaluate the quality of the tear film lipid layer (TFLL), DR-1 interferometry (Kowa, Nagoya, Japan) was used. According to Hosaka et al,²³ the grading method used for this procedure categorised the layer’s quality, and one score indicates the highest TFLL quality.

The Cochet-Bonnet esthesiometer (Luneau Technology Operations, France) was used to evaluate the central corneal sensitivity, which uses a nylon monofilament to stimulate the cornea. Using a slider on the pen’s side, users can adjust the filament’s length (0–6 cm) to change the stiffness of the material.²⁴

To determine the concentration of MMP-9 in the tear samples, a commercial reagent card (S05B, Seinda Biomedical Corporation, Guangdong, China) based on colloidal gold and immunochromatographic analysis was used. A 1 ul tear sample was placed in the sampling hole on the reagent card, followed by three drops of diluent. After inserting the reagent card into the proprietary analyser (S03A, Seinda Biomedical Corporation, Guangdong, China), the MMP-9 concentration was determined after 15 min.²⁵

Statistical analyses

Statistical analyses were conducted using SPSS statistics software (V 25.0; SPSS, USA). Given the differences in the demographic characteristics between eligible patients in the two groups, propensity-score matching (PSM) was used to identify patients with similar demographic characteristics. Patients in the DDE and DNDE groups were matched for a 1:1 PSM with a calliper value of 0.1. Matching factors include gender, age, diabetes duration and FBG. For continuous variables, descriptive statistics were expressed using mean and SD, while for binary parameters, statistics were expressed using numbers (%). Categorical data were analysed using the χ² test. The Kolmogorov-Smirnov test was used to determine the variables' normality. The means of all ocular parameters were compared from the DDE and DNDE groups. The student’s t-test was used to compare parameters with a normal distribution before PSM, and the paired t-test was used to compare parameters with a normal distribution after PSM. The Mann-Whitney U-test was used for values that were not normally distributed. Categorical data were analysed using the χ² test. The receiver operating characteristics (ROC) curve and Youden index were used to calculate the optimal cut-off value predicted by DE, as well as the sensitivity and specificity of MMP-9. Pearson and Spearman’s rho correlation were used to establish the relationships between MMP-9 and variables. A p value<0.05 was judged statistically significant.

Results

Demographic and ocular surface parameters

Among 144 eligible patients, 110 patients with similar propensity scores were included in the analyses. Before PSM was performed, there were differences between the two groups in the duration of diabetes and FBG. (table 1) Using PSM, 55 patients in the DNDE group were matched with 55 patients in the DDE group. The duration of diabetes and FBG showed no significant difference in the two groups. (table 2).

Table 1. Demographic characteristics before propensity-score matching (PSM).

	DNDE	DDE	P value
No. of eyes	70	74
Gender (male)	31.43%	43.24%	0.143
Age (years)	58.50±12.39	62.26±11.78	0.065
Diabetes duration (years)	6.97±3.88	9.86±6.36	0.001*
FBG (mg/dL)	7.12±1.35	8.40±2.47	<0.001*
BCVA (LogMAR)	0.12±0.19	0.14±0.16	0.314
IOP (mm Hg)	16.49±2.29	16.39±2.62	0.825

^*P value<0.05.

BCVAbest corrected visual acuityDDEdiabetic dry eyeDNDEdiabetic non-dry eyeFBGfasting blood glucoseIOPintraocular pressure

Table 2. Demographic characteristics after propensity-score matching (PSM).

	DNDE	DDE	P value
No. of eyes	55	55
Gender (male)	34.55%	49.09%	0.122
Age (years)	60.8±11.28	61.82±11.98	0.67
Diabetes duration (years)	7.89±3.61	7.38±3.14	0.338
FBG (mg/dL)	7.32±1.40	7.48±1.44	0.266
BCVA (LogMAR)	0.13±0.20	0.14±0.17	0.786
IOP (mm Hg)	16.30±2.27	16.46±2.72	0.747

BCVAbest corrected visual acuityDDEdiabetic dry eyeDNDEdiabetic non-dry eyeFBGfasting blood glucoseIOPintraocular pressure

Table 3 presents the comparison of ocular surface parameters between the groups after PSM. DE signs and symptoms quantified by NITBUT and OSDI, were significantly higher in the DDE group compared with the DNDE group (p<0.001), respectively. MMP-9 concentration was higher in the DDE group than in the DNDE group (70.63±52.06 vs 33.98±33.93, p<0.001). The groups found no differences in CS, TFLL, corneal sensitivity, TMH and RS (p=0.066, p=0.067, p=0.655, p=0.343 and p=0.836, respectively).

Table 3. Comparison of ocular surface parameters.

DNDE	DDE	P value
NITBUT (s)	11.91±6.85	4.82±2.54	<0.001*
CS score (0–9)	0.96±1.02	1.42±1.47	0.066
TFLL (1–5)	2.64±0.49	2.82±0.48	0.067
Corneal sensitivity (1–6)	5.62±0.81	5.55±0.86	0.655
TMH (mm)	0.21±0.08	0.19±0.09	0.343
RS score (0–4)	1.60±0.45	1.62±0.49	0.836
MMP-9 (ng/mL)	33.98±33.93	70.63±52.06	<0.001*
OSDI (0–100)	22.61±20.12	38.11±16.34	<0.001*

^*P value<0.05, NITBUT, non-invasive tear breakup time; CS, ; TFLL, ; TMH, ; RS, redness score; MMP-9, matrix metalloproteinase 9; OSDI, .

CScorneoconjunctival stainingMMP-9matrix metalloproteinase 9NITBUTnon-invasive tear breakup timeOSDIocular surface disease indexRSredness scoreTFLLtear film lipid layerTMHtear meniscus height

The correlation between MMP-9 concentration and ocular surface parameters in patients with DM is shown in table 4. NITBUT and corneal sensitivity negatively correlated with MMP-9. CS and OSDI have a positive correlation with MMP-9. However, the correlation analysis between the MMP-9 and other ocular surface parameters, including TFLL, TMH score and RS, showed no statistical significance.

Table 4. Correlation between matrix metalloproteinases-9 (MMP-9) concentration and ocular surface parameters.

	NITBUT	CS	TFLL	Corneal sensitivity	TMH	RS	OSDI
MMP-9	−0.246*	0.326*	NS	−0.257*	NS	NS	0.233*

NS: correlation was detected during correlation analysis.

^*P<0.01.

CScorneoconjunctival staining MMP-9matrix metalloproteinase 9NITBUTnon-invasive tear breakup timeOSDIocular surface disease indexRSredness scoreTFLLtear film lipid layerTMHtear meniscus height

ROC curve of MMP-9

The area under the curve of MMP-9 in differentiating DDE and controls was 0.737. The optimal cut-off value of MMP-9 to predict DE in diabetic patients was >52.5 ng/mL, with 58.2% sensitivity and 78.2% specificity (figure 1).

Figure 1. Matrix metalloproteinases-9 (MMP-9) receiver operating characteristic curves of diabetic dry eye (DDE) predictors.

Discussion

While not as consequential as DR, an increasing number of studies are exploring the impact of DM on ocular surface disorders in diabetic patients, including DE, peripheral neuropathy and corneal ulcers.^{10 17 26} The prevalence of DE among individuals with diabetes is 54%, above the rate in the general population.²⁷ At the same time, the pursuit of stratifying anterior eye segment disease patients using tear proteinomics is a worthy cause that would greatly benefit clinicians and patients alike. Previous studies have documented that MMP-9 levels above 40 ng/mL produce positive results of DE in the general population.^{28 29} Still, the current result from our study suggests that the optimal cut-off value of MMP-9 above 52.5 ng/mL predicted DE in diabetic patients, which is higher in the general population. Similarly, Liu et al¹⁰ while assessing the concentrations of common tear inflammatory factors, and the results showed that the epidermal growth factor (EGF) in tears had significantly increased in DDE versus non-diabetes with DE. The low sensitivity can be explained by the inflammation that precedes the clinical signs and symptoms of DE. Elevated MMP-9 suggests the development of DE but cannot be used as a single indicator for diagnosing DE and needs to be combined with other tests. Clinically, although patients with DM temporarily do not show signs of a significant reduction in tear secretion, corneal epithelial staining or shortening of tear film breakup time, they already have markedly elevated levels of inflammatory factors and may be treated with neurotrophic and anti-inflammatory therapies as appropriate.

Furthermore, an animal model study suggests the overall increase in vascular MMP-9 activity in DM³⁰ and hyperglycaemia’s redox-sensitive MMP-9 expression could give a basis for antioxidant therapy to modify diabetic vascular consequences. Iyengar et al¹⁷ have reported that MMP-9 tear levels in diabetics are lower compared with diabetes in the neuropathy group. The findings of the current study and past studies imply that elevated MMP-9 levels in tears are a confirmatory tool for DE but can also be elevated due to systemic diseases such as diabetes.^{13 16 31 32} In our study, no statistically significant differences in corneal staining and corneal sensitivity were observed between the two groups; this may be due to corneal epithelial and nerve damage caused by hyperglycaemia, even in patients without DED.

Our results found statistically significant but weak positive correlations of MMP-9 levels with CS score and OSDI score, suggesting the role of MMP-9 in ocular surface endothelial cellular damage, which also contributed to more DE discomfort. In addition, the higher ocular surface inflammation leads to the destruction of tear film stability.⁸ Previous studies have reported elevated ocular surface inflammation associated with peripheral nerve damage in diabetic patients,^{17 33} which explained the negative correlation between MMP-9 and corneal sensitivity in our findings. Detection of ocular surface inflammation is beneficial as an essential indicator for managing patients with anti-inflammatory treatments, including topical glucocorticoids, cyclosporine A and tetracycline and its analogues.³⁴

The limitations of the current study are as follows: While strict environmental and testing protocol measures were taken to ensure the accuracy of the MMP-9 test, some tear molecule levels have interday and intraday variability, suggesting that repeated measurements of MMP-9 are also necessary, which may lead to varying results.³⁵ The inclusion of multiple inflammatory factors could have better reflected the ocular surface status of the study group. Furthermore, HbA1C should be included in further study as it is important for diabetes monitoring. However, the inclusion of a group with non-diabetic patients would have further bolstered the results of this study. Finally, the Schirmer test was chosen not to be performed in our study due to the overall dryness of the ocular surface and burdening the participants with tests, but TMH, a surrogate, was assessed.

In summary, MMP-9 concentration was higher in patients with DDE than DNDE. Our study revealed MMP-9 test is a potential diagnostic tool for DDE. It may help follow-up diabetic patients with DE and guide clinicians in deciding on anti-inflammatory treatments for these patients. Additional multi-centre study is required to understand the distinct mechanisms underlying the heightened levels of tear MMP-9 in patients with DED and diabetes.

Data availability statement

Data are available upon reasonable request.