Lysyl Oxidase-Like 2 Level and Glaucoma Surgical Outcomes

Abstract

Purpose.: To evaluate the levels of TGF-β and lysyl oxidases in both the aqueous humor and in Tenon's tissue in patients with POAG, and to assess their association with bleb morphological features and outcomes of glaucoma surgery.

Methods.: Eighteen POAG patients scheduled to undergo glaucoma surgery and 17 healthy controls were enrolled. At the time of surgery, approximately 0.1 mL aqueous humor through an anterior chamber paracentesis and a 4 × 4-mm sample of Tenon's tissue were obtained. Concentrations of TGF-β and lysyl oxidases (types 1 and 2) were analyzed by ELISA. The bleb morphologic features based on the Moorfields Bleb Grading System were evaluated by bleb photographs.

Results.: The TGF-β in Tenon's tissue and lysyl oxidase-like 2 in aqueous humor and in Tenon's tissue were significantly elevated in POAG patients. When POAG patients were divided by outcome 1 year after surgery, lysyl oxidase-like 2 protein was significantly higher in both aqueous humor and Tenon's tissue of the failure group. The level of lysyl oxidase-like 2 protein was significantly correlated with that of TGF-β in the Tenon's tissue. Lysyl oxidase-like 2 level in the aqueous humor (r = 0.563, P < 0.001) and Tenon's tissue (r = 0.760, P < 0.001) had significantly negative correlation with bleb area.

Conclusions.: The TGF-β and lysyl oxidase-like 2 levels at the time of surgery were significantly correlated with 1-year outcomes of glaucoma surgery. The level of lysyl oxidase-like 2 was significantly correlated with that of TGF-β, suggesting induction by TGF-β. Modulation of lysyl oxidase-like 2 levels may have important applications in glaucoma surgery.

Introduction

The goal of glaucoma treatment is control of the patient's IOP, which is achieved by medication or by glaucoma surgery. To prevent blindness, the success of glaucoma filtration surgery is crucial. Glaucoma surgery is performed under the conjunctival and subconjunctival spaces, and fibrosis at the site of surgery affects the surgical outcome. Excessive fibrosis and extracellular matrix (ECM) remodeling at the site of surgery is the most common cause of failure after glaucoma surgery. ^1,2 Tenon's fibroblasts are the main effector cells in this process, and excessive deposition of collagen and other ECM components results in fibrotic scar formation at the surgical site. ³  

Lysyl oxidases are a family of enzymes that play a role in the maturation of collagen, the formation of elastin, and the cross-linking of collagen and elastin. ^4,5 One family member, lysyl oxidase-like 2, is reported to be involved in many fibrotic diseases, such as pulmonary fibrosis, cardiac fibrosis, and several rheumatologic diseases. ^6–8 Activated fibroblasts secrete lysyl oxidase-like 2, along with increased amounts of collagen. ⁹ Recently, one study reported that inhibiting lysyl oxidase-like 2 decreased pathologic angiogenesis, inflammation, and fibrosis at the site of glaucoma surgery. ¹⁰ However, the role of lysyl oxidase enzymes in glaucoma surgery remains to be elucidated. 

Many studies have investigated the role of TGF-β in the activation of Tenon's fibroblasts. ^11–13 The application of TGF-β to human ocular fibroblasts increases lysyl oxidase expression, as well as synthesis of ECM components, resulting in ECM remodeling. ¹⁴ Transforming growth factor-β has been reported to induce lysyl oxidases through several pathways in cardiac fibroblasts. ¹⁵ Thus, it is possible that fibrosis at the site of glaucoma surgery may be activated by TGF-β through the induction of lysyl oxidases. Therefore, this study examined the levels of TGF-β and lysyl oxidases in both the aqueous humor and in Tenon's tissue at the time of glaucoma surgery, in patients with POAG. Relationship between the levels of TGF-β, the induction of lysyl oxidases, and surgical outcomes after 1 year were analyzed. 

Methods

A previous study ¹⁶ reported an SD ranging between 40 and 49 pg/mL and an approximately 78 pg/mL difference in TGF-β level between glaucomatous and normal eyes. The minimum sample size required with a significance level of 5% and a power of 80% was calculated to be 7 in each group. Thus, the minimal requirement of sample size was 7 for statistical power of 80%, which was achieved in our study. 

Participants

After approval by the institutional review board of our institution, 18 patients with POAG who were scheduled for trabeculectomy due to uncontrolled IOP, were enrolled. To be included, the patients had to have an IOP higher than 21 mm Hg, visual field or optic disc changes characteristic of glaucoma, and to be taking the maximum tolerated dose of medication. In addition, 17 healthy controls scheduled to undergo cataract surgery were enrolled during the same period. Written informed consent was obtained from each patient, and all procedures were performed in accordance with the Declaration of Helsinki. 

The study excluded patients younger than 18 years, those diagnosed with secondary glaucoma, patients with angle-closure glaucoma, and angle-closure suspects. Further reasons for exclusion included ocular trauma; previous laser therapy or intravitreal injection; ocular surgery other than cataract surgery; active or past inflammatory disease of the cornea, conjunctiva, episclera, or sclera; and active, chronic, or recurrent uveitis. 

Aqueous Humor and Tenon's Tissue Sampling and Quantification

Trabeculectomy was performed by one surgeon (CKP). Before surgery, approximately 0.1 mL aqueous humor was collected through an anterior chamber paracentesis. After a fornix-based conjunctival flap incision, the conjunctiva and Tenon's tissue were separated by blunt dissection. A 4 × 4-mm piece of Tenon's tissue was cut from the eye at the margin of the conjunctival flap incision, from areas with no apparent vessels. Minimal manipulation was done during the dissection and cutting. No significant bleeding occurred during the procedure. Each sample was stored immediately at −70°C. 

The concentrations of TGF-β, lysyl oxidase-like 1, and lysyl oxidase-like 2 in the aqueous humor and Tenon's tissue samples were determined using ELISA (Quantikine ELISA Kits; R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions. For Tenon's tissue, proteins were extracted before the ELISA using a lysing kit (hard tissue-homogenizing kit CK28; Precellys, Bertin, France) according to the manufacturer's instructions. The tissue extracts were incubated for 10 minutes on ice and clarified by centrifugation at 10,000g for 25 minutes at 4°C. Total protein from the tissue was determined using a standard bicinchoninic acid assay (Pierce, Rockford, IL, USA). Duplicates of tissue extract (each 30 μg total protein) and diluted standards of each substrate were loaded for the ELISA analysis. After color development was halted with a stop solution, the optical density was measured by spectrophotometry (DU-530; Beckman Instruments, Inc., Fullerton, CA, USA). From the average of the duplicate readings for each standard and sample, the average zero standard optical density was subtracted. Standard curve was created and protein concentration of each substrate was calculated. 

Clinical Data

The preoperative IOP (three recordings on separate days) by Goldmann applanation tonometry, the number of IOP-lowering medications taken before the glaucoma surgery, and the mean deviation (MD) of the last visual field before surgery using the 24-2 Swedish Interactive Threshold Algorithm standard program (Humphrey Visual Field Analyzer; Carl Zeiss Meditec, Dublin, CA, USA) were recorded. All IOP-lowering medications were discontinued 1 day before trabeculectomy and during postoperative periods until there was need to lower IOP. Topical antibiotics and steroids were used for 2 months postoperatively. At each postoperative visit, Goldmann applanation tonometry and slit-lamp biomicroscopy were performed. The number of IOP-lowering medications used over 12 months postoperatively was recorded. The IOP at 2 weeks and 1, 2, 3, and 6 months were recorded. The mean of two recordings of the IOP at 12 months postoperatively also was recorded, and the mean of two recordings of the IOP at 12 and 18 months postoperatively was considered the final IOP. The patients were divided into success or failure groups based on the final IOP. Failure was defined as an IOP of 21 mm Hg or higher, with or without IOP-lowering medications. 

Analysis of Bleb Morphology

The bleb morphologic features based on the Moorfields Bleb Grading System (MBGS) were evaluated. ¹⁷ Photographs needed to complete bleb morphologic features analysis routinely are obtained in our glaucoma clinic of all patients who have undergone glaucoma filtration surgery. Scoring was carried out according to standard MBGS protocol (http://www.blebs.net/html/Protocol.html) by using reference photographs of the various aspects of the bleb (http://www.blebs.net/html/Images.html). Six criteria are assessed that pertain to the three main aspects of the bleb: area, height, and vascularity. Bleb area is assessed based on maximal bleb area and central demarcation area. A score of 1 to 5 is given based on extension, from 0% to 100% extension. Bleb height represents the highest point of the bleb and is graded from 1 to 4 against reference photographs, with 1 being the least height and 4 being the greatest height. Bleb vascularity is scored separately in the central part of the bleb, the demarcated part of the bleb, peripheral part of the bleb, and the peripheral nonbleb conjunctiva. Each of these parameters is scored from 1 to 5, with 1 corresponding to avascular and 5 corresponding to severe vessel inflammation. 

Statistical Analysis

Statistical analyses were performed using SPSS software (ver. 11.1; SPSS, Inc., Chicago, IL, USA). One-way analysis of variance and Scheffe's method for multiple comparisons were used to compare the continuous data among the three groups. Independent Student's t-tests were applied to compare data between two groups. The χ² test was used to compare categorical data. Spearman's correlation test was used to assess the relationship between the TGF-β and lysyl oxidase-like 2 levels in the aqueous humor and Tenon's tissue, the relationship between TGF-β and lysyl oxidase-like 2 levels with bleb morphological features at 18 months postoperatively. Univariate and multivariate linear regressions were used to assess the relationship between the clinical data and lysyl oxidase-like 2 levels in the aqueous humor and Tenon's tissue. For all analyses, the level of statistical significance was set at P less than 0.05. 

Results

The preoperative IOP (P < 0.001) and MD of the perimetry (P < 0.001) differed significantly between the healthy controls and patients with POAG. Of the POAG eyes that underwent glaucoma surgery, 11 eyes were classified into the success group, and 7 into the failure group 1 year after surgery. No significant differences existed in age (P = 0.884) or sex (P = 0.484) between the success and failure groups. Preoperative IOP (P = 0.912), number of preoperative medications (P = 0.720), and MD (P = 0.610) were similar between the groups. Postoperative IOP did not show difference at postoperative 2 weeks and 1 month between the success and failure groups. However, postoperative IOP at 2, 3, 6, 12, and final IOP (17.27 ± 3.71 vs 22.96 ± 2.73 mm Hg, respectively; P < 0.001) differed significantly between the groups (Table). 

Table

 

Demographics of the Study Participants

	Healthy Control, n = 17	POAG, n = 18	P Value	Success Group, n = 11	Failure Group, n = 7	P Value
Age, y, mean ± SD	60.94 ± 18.36	67.76 ± 10.21	0.176*	57.33 ± 24.54	62.90 ± 15.03	0.884*
Sex, M:F	10:7	7:11	0.247†	4:7	3:4	0.484†
IOP, mm Hg
Preoperative	13.26 ± 2.27	23.24 ± 4.09	<0.001*	23.11 ± 2.03	24.74 ± 7.83	0.912*
Postoperative 2 weeks		10.23 ± 3.88		10.16 ± 2.57	10.38 ± 3.26	0.572*
Postoperative 1 month		11.95 ± 3.24		11.24 ± 2.97	12.36 ± 4.26	0.139*
Postoperative 2 months		15.93 ± 3.59		15.71 ± 2.56	17.26 ± 4.71	<0.001*
Postoperative 3 months		16.94 ± 3.16		15.54 ± 3.01	18.31 ± 3.91	<0.001*
Postoperative 6 months		17.34 ± 4.72		16.32 ± 2.64	17.54 ± 4.88	<0.001*
Postoperative 12 months		18.51 ± 4.46		16.57 ± 3.22	19.24 ± 5.61	<0.001*
Final		20.24 ± 3.55		17.27 ± 3.71	22.96 ± 2.73	<0.001*
MD, mean ± SD	−0.43 ± 0.96	−15.16 ± 8.38	<0.001*	−15.63 ± 6.95	−14.52 ± 8.54	0.610*
Preoperative medications, n		2.84 ± 0.34		2.90 ± 0.74	2.73 ± 0.83	0.720*
α-agonists, n (%)		14 (77.8)		9 (81.8)	5 (71.4)	0.728†
β-blockers, n (%)		14 (77.8)		9 (81.8)	5 (71.4)	0.728†
Prostaglandins, n (%)		16 (88.9)		10 (91.0)	6 (85.7)	0.647†
Carbonic anhydrase inhibitors, n (%)		14 (77.8)		8 (72.7)	6 (85.7)	0.243†

*  Independent Student's t-tests.

†  χ² test.

The TGF-β level was elevated in the aqueous humor and Tenon's tissue of patients with POAG compared with healthy controls (Fig. 1). When the POAG patients were subdivided into success and failure groups, the level of TGF-β of failure group (157.98 ± 40.17 pg/mL) was statistically different from that of both the normal controls (89.15 ± 22.32 pg/mL; P = 0.031) and the success group (93.00 ± 27.25 pg/mL; P = 0.047) in Tenon's tissue. However, the level in aqueous humor was 43.15 ± 12.96 pg/mL in the healthy controls, and was elevated to 53.516 ± 18.78 pg/mL in the success group and 68.41 ± 20.94 pg/mL in the failure group, differences that were not statistically significant (P = 0.240). Lysyl oxidase-like 1 levels did not differ among groups in either the aqueous humor or the Tenon's tissue. 

Figure 1

 

The TGF-β, lysyl oxidase-like 1, and lysyl oxidase-like 2 in aqueous humor and Tenon's tissue in each group. Error bars indicate the upper SD of the mean value in each group. *P < 0.05, Mann-Whitney U test.

Lysyl oxidase-like 2 levels were elevated in the aqueous humor and Tenon's tissue of patients with POAG compared with the healthy controls (Fig. 1). When POAG patients were subdivided into success and failure groups, the lysyl oxidase-like 2 level in the Tenon's tissue of the failure group (121.20 ± 26.57 pg/μL) was statistically different from that of both the healthy controls (53.37 ± 20.62 pg/μL; P < 0.001) and the success group (67.16 ± 32.37 pg/μL; P < 0.001). Similarly, the level in the aqueous humor was 23.94 ± 22.71 pg/μL in the healthy controls and was elevated to 36.36 ± 26.21 pg/μL in the success group and to 67.16 ± 32.37 pg/μL in the failure group, significantly different from both control and success groups (P < 0.001). 

The relationship between TGF-β and lysyl oxidase-like 2 levels is shown in Figure 2. The level of TGF-β in the aqueous humor was not significantly related to the level of lysyl oxidase-like 2 in aqueous humor or in the Tenon's tissue. However, TGF-β in Tenon's tissue was significantly correlated with lysyl oxidase-like 2 level in aqueous humor (r = 0.592, P < 0.001) and in Tenon's tissue (r = 0.711, P < 0.001). 

Figure 2

 

The association between TGF-β and lysyl oxidase-like 2 in aqueous humor and in Tenon's tissue. R, correlation coefficient; P values are from Spearman's correlation.

Patients were subdivided into two groups by the lysyl oxidase-like 2 level in the Tenon's tissue. By MBGS scoring, central and maximal bleb area, bleb height, central and peripheral vascularity, and nonbleb vascularity were graded (Fig. 3). Eyes in the upper half of lysyl oxidase-like 2 level had significantly reduced central and maximal bleb area compared with eyes in the lower half of lysyl oxidase-like 2 level at 6, 12, and 18 months after surgery. Bleb height was also significantly lower at 6 months after surgery in eyes with greater amounts of lysyl oxidase-like 2. There was no difference in the scores representing vascularity of the bleb between two groups. The relationship between levels of lysyl oxidase-like 2 or TGF-β with MBGS scores of central bleb area was evaluated (Fig. 4). Lysyl oxidase-like 2 level in the aqueous humor (r = 0.563, P < 0.001) and Tenon's tissue (r = 0.760, P < 0.001) had significantly negative correlation with central bleb area. The TGF-β level in the Tenon's tissue (r = 0.573, P = 0.004), not the level in the aqueous humor (r = 0.121, P = 0.497), showed significantly negative correlation with central bleb area. 

Figure 3

 

Average changes in MBGS scores from 18 eyes with trabeculectomy over an 18-month period. Patients were subdivided into two groups by the lysyl oxidase-like 2 level in the Tenon's tissue. Eyes in the upper half of lysyl oxidase-like 2 level had significantly reduced central and maximal bleb area compared with eyes in the lower half of lysyl oxidase-like 2 level at 6, 12, and 18 months after surgery. Bleb height was also significantly different between two groups at 6 months after surgery. There was no difference in the scores representing vascularity of the bleb between two groups at the central, peripheral, and nonbleb area. Error bars indicate the upper and lower SD of the mean value in each group. *P < 0.05, Mann-Whitney U test.

Figure 4

 

The relationship between the level of TGF-β and lysyl oxidase-like 2 in aqueous humor and Tenon's tissue with central bleb area at 18 months postoperatively by MBGS scores are shown. Lysyl oxidase-like 2 level in the aqueous humor (r = 0.563, P < 0.001) and Tenon's tissue (r = 0.760, P < 0.001) had significantly negative correlation with central bleb area. The TGF-β level in the Tenon's tissue (r = 0.573, P = 0.004), not the level in the aqueous humor (r = 0.121, P = 0.497), showed significantly negative correlation with central bleb area. P values are from Spearman's correlation.

To identify clinical factors related to lysyl oxidase-like 2 levels in the aqueous humor and Tenon's tissue, univariate and multivariate linear regression analyses were performed. However, no significant relationship with clinical factors were identified. 

Discussion

Transforming growth factor-β in the Tenon's tissue, and lysyl oxidase-like 2 in both aqueous humor and Tenon's tissue at the time of surgery were significantly related to the bleb area, final IOP, and 1-year glaucoma surgery outcomes in patients with POAG. The level of lysyl oxidase-like 2 in Tenon's tissue was significantly correlated with the level of TGF-β. These findings suggest that enhanced wound fibrosis due to lysyl oxidase-like 2 may affect the outcome of glaucoma surgery, and that lysyl oxidase-like 2 may be induced by TGF-β. The level of lysyl oxidase-like 1 did not show any relationship with surgical outcome. 

Although lysyl oxidase is the key enzyme for collagen cross-linking and maturation of collagen fibrils, few studies have investigated its role in wound healing after glaucoma surgery. Lysyl oxidase-like 1 gene (LOXL1) expression has been investigated, particularly in the pathogenesis of pseudoexfoliation syndrome and glaucoma. ^18,19 However, lysyl oxidase-like 1 levels were not elevated in either aqueous humor or Tenon's tissue in POAG patients compared with healthy controls in the present study. Four types of lysyl oxidases have been identified, and differential expression of each type has been reported depending on cell and tissue type. ^20–22 In the eye, the lysyl oxidase enzymes have been detected in the trabecular meshwork, ciliary body, lens, retina, and the lamina cribrosa. ^23–25 Decreased activity of lysyl oxidase in the cornea has been suggested to be related to keratoconus, through impairment of collagen cross-linking. ²⁶ Recently, the role of lysyl oxidase in organ fibrosis, tumor progression, and several other diseases has been actively investigated. ^6,7,27 In particular, lysyl oxidase-like 2 is thought to be directly involved in disease-related fibrogenesis. It promotes cross-linking of fibrillar collagen I, a major component of desmoplastic stroma and fibrosis, and plays a crucial role in matrix remodeling, fibrogenesis, and formation of pathologic stroma. ²⁸ Lysyl oxidase-like 2 protein expression has been observed in fibroblastic foci and collagenous regions in several diseases. ²⁹ Because lysyl oxidase-like 2 is associated with areas of active fibrogenesis in diseased tissue, we hypothesized that the level of lysyl oxidase-like 2 might be related to surgical outcome after glaucoma surgery. In our study, lysyl oxidase-like 2 was elevated in the aqueous humor and Tenon's tissue of POAG patients with failed glaucoma surgery, suggesting a role in excessive fibrosis and ECM remodeling. Interestingly, lysyl oxidase and lysyl oxidase-like 2 are upregulated in the Tenon's capsule and conjunctiva after trabeculectomy in rabbits. ¹⁰ In that study, using lysyl oxidase-like 2 antibody to inhibit lysyl oxidase-like 2 in glaucoma surgery improved bleb survival after trabeculectomy. The effect of lysyl oxidase-like 2 level on the vascularity of the bleb was not significant, showing that the mechanism of ECM remodeling by lysyl oxidase-like 2 may be a direct effect, not through enhanced vasculogenesis. Also, the gradually greater difference between IOP in the failure and success groups from 1 to 12 months after surgery suggest that lysyl oxidase-like 2 may influence bleb maturation. 

The fibroblasts of Tenon's tissue, which produce collagen and elastin, are the most important mediators of ocular scar formation after glaucoma surgery. ¹ There is evidence that Tenon's fibroblasts are activated and converted into myofibroblasts by TGF-β ³⁰ ; TGF-β is a fibrogenic cytokine known to activate the synthesis of collagen and elastin. In several tissues, TGF-β strongly promotes lysyl oxidase expression in fibroblasts. ^8,15 We found that the level of lysyl oxidase-like 2 in Tenon's tissue was significantly correlated with that of TGF-β. Our results indirectly suggest that lysyl oxidase-like 2 may be upregulated in turn by the activation of TGF-β in Tenon's fibroblasts; TGF-β is abundant in the aqueous humor and aqueous humor has shown to stimulate proliferation of fibroblasts. ^31,32 Modulation of the composition of aqueous humor or suppressing TGF-β in aqueous humor and Tenon's tissue may in turn reduce lysyl oxidase-like 2 and improve surgical outcome. Methods that could be applicable in clinical settings are needed to be elucidated in glaucoma eyes with high lysyl oxidase-like 2 undergoing surgery. 

Several factors also may influence surgical outcome in trabeculectomy. Especially inflammation due to previous glaucoma medication or benzalkonium chloride is identified. ^33–35 Glaucoma patients in our study were scheduled to undergo trabeculectomy because of uncontrolled IOP even with maximum tolerated dose of medication. We checked preoperative glaucoma medication between groups and there was no difference in the number and classification of drug use. However, the duration of medication use was not compared and the effect of medication on inflammatory markers and cytokines was not evaluated in the present study. Both preoperative medication and inflammatory cytokines could be associated with the induction of TGF-β and lysyl oxidase. Other than lysyl oxidase-like 1 and 2, other types may be related to surgical outcomes. These should be further evaluated. 

In summary, we evaluated the level of TGF-β and of lysyl oxidase in aqueous humor and Tenon's tissue in patients with POAG. The TGF-β and lysyl oxidase-like 2 level in aqueous humor and Tenon's tissue was significantly related to the bleb morphological features and outcome of glaucoma surgery. Lysyl oxidase-like 1 showed no relationship to surgical outcome. The level of lysyl oxidase-like 2 was significantly correlated with TGF-β in the Tenon's tissue, suggesting induction by TGF-β. Modulation of lysyl oxidase-like 2 levels may have important applications in glaucoma surgery. 

Acknowledgments

Supported by Seoul St. Mary's Hospital Clinical Medicine Research Program year of 2013 through the Catholic University of Korea. The authors alone are responsible for the content and writing of the paper. 

Disclosure: H.-Y.L. Park, None; J.H. Kim, None; C.K. Park, None