TSG-6 Protects Corneal Endothelium From Transcorneal Cryoinjury in Rabbits

Abstract

Purpose.: To investigate the effect of an anti-inflammatory protein, TNF-α stimulated gene/protein (TSG)-6 and an antiapoptotic protein, stanniocalcin (STC)-1 on corneal endothelium in rabbits with transcorneal cryoinjury.

Methods.: Transcorneal freezing (−80°C) was applied to rabbit corneas for 30 seconds. Immediately post injury, either TSG-6 (10 μg/100 μL), STC-1 (10 μg/100 μL), or the same volume of balanced salt solution (BSS) was injected into the anterior chamber. Each eye was examined for corneal opacity, corneal thickness, endothelial cell density, and endothelial hexagonality every 2 to 6 hours for 48 hours post injury. The concentrations of myeloperoxidase (MPO) and IL-1β were measured in the aqueous humor every 6 hours. At 48 hours post injury, each cornea was assayed for TNF-α, IL-1β, IL-6, and MPO, and histologically evaluated with alizarin red-trypan blue staining, hematoxylin-eosin staining, and immunostaining for neutrophils.

Results.: Tumor necrosis factor-α stimulated gene/protein-6 significantly decreased the development of corneal opacity and edema after cryoinjury compared with STC-1 or BSS. The corneal endothelial cell density and hexagonality were markedly preserved by TSG-6. The mRNA levels of TNF-α, IL-1β, and IL-6 in the cornea and the protein levels of MPO and IL-1β in the aqueous humor and cornea were significantly lower in TSG-6–treated eyes than BSS-treated controls. Similarly, the expression of fibroblast growth factor-2 was reduced by TSG-6 treatment. Histologic evaluation demonstrated that neutrophil infiltration of the cornea was decreased in TSG-6–treated eyes.

Conclusions.: Tumor necrosis factor-α stimulated gene/protein-6 protected corneal endothelial cells from transcorneal cryoinjury through suppression of inflammation

Introduction

Corneal transparency is maintained by the corneal endothelium through a barrier mechanism and through a Na⁺/K⁺-ATPase pump. Corneal endothelial dysfunction causes irreversible loss of corneal transparency, and is a major cause of severe visual impairment and corneal blindness. The principal therapy for corneal endothelial dysfunction is corneal transplantation, and over 40% of all corneal transplants performed in the Unites States involve endothelial keratoplasty to replace dysfunctional corneal endothelium. ^1–4 Ongoing efforts to develop pharmaceutical agents to treat corneal endothelial dysfunction demonstrate promising results ^5–10 ; however, no medical therapies are yet available. 

The mechanism of corneal endothelial dysfunction has been explained by endothelial-mesenchymal transformation (EMT). ^11–14 Studies demonstrate that after damage to the cornea, neutrophils infiltrate the corneal endothelium and produce IL-1β. Interleukin-1β subsequently activates phosphatidylinositol (PI) 3-kinase, which induces synthesis of FGF-2. Fibroblast growth factor-2 triggers EMT, and forms the retrocorneal fibrous membrane characteristic of end-stage corneal endothelial disease. Another mechanism of corneal endothelial dysfunction is apoptosis of corneal endothelial cells. Our group previously demonstrated that human corneal endothelium is more susceptible to transcorneal freezing injury than corneal stroma or corneal epithelium, and corneal endothelial cells underwent apoptotic death after injury. ¹⁵ These findings suggest that treatment, which either targets the upstream of EMT or prevents apoptosis may improve corneal endothelial dysfunction. 

Tumor necrosis factor-α stimulated gene/protein (TSG)-6 is a multifunctional endogenous protein expressed by a variety of cells including mesenchymal stem/stromal cells (MSCs) in response to stimulation by pro-inflammatory cytokines. ^16,17 Previous experiments show that TSG-6 modulates inflammation through the reduction of neutrophil migration, ^18–21 the stabilization of extracellular matrix via inhibition of proteases, ^22–25 or by the inhibition of angiogenesis through an interaction with FGF-2. ²⁶ Therapeutic potential of TSG-6 in ocular disease has been previously demonstrated in animal models for corneal chemical injury or corneal transplantation. ^27–29 Stanniocalcin-1 (STC-1) is an antiapoptotic protein that is secreted by MSCs in response to signals from apoptotic cells. ³⁰ Previous studies demonstrate that STC-1 rescues photoreceptors or retinal ganglion cells from apoptotic cell death through suppression of reactive oxygen species generation in models for retinal degeneration or optic nerve transection. ^31,32  

In the present study, we investigated the effects of an anti-inflammatory protein, TSG-6 on rabbit corneal endothelium that was injured by transcorneal freezing, and compared with those of an antiapoptotic protein, STC-1 on corneal endothelium after cryoinjury. Our results demonstrated that TSG-6, not STC-1, suppressed neutrophil infiltration, reduced the levels of pro-inflammatory proteins including IL-1β and FGF-2, and subsequently protected corneal endothelial cells from damage. 

Materials and Methods

Animals and Animal Model

Animal experiments were performed in accordance with the ARVO Statement for Use of Animals in Ophthalmic Vision and Research, and the protocols were approved by the Institutional Animal Care and Use Committee of Seoul National University Biomedical Research Institute (IACUC No. 12-0247-C1A0; Jongno-gu, Seoul, Korea). 

Eight-week-old New Zealand white rabbits weighing 2.0 to 2.5 kg (Orient Bio, Inc., Seongnam, Korea) were used in the study. Under anesthesia with zolazepam-tiletamine (Zoletil, Virbac, Carros, France), a drop of proparacaine solution (0.5%) was instilled to the eye, and a lid speculum was inserted. Then, one cycle of transcorneal freezing was performed by applying a 7.5-mm diameter round-shaped dry ice (−80°C) to the corneal surface involving the temporal limbus and the center for 30 seconds. The injury was made on the right eye of each animal, and the left eye was used as control without injury. Immediately after injury, either recombinant human (rh) TSG-6 (10 μg/100 μl; R&D Systems, Minneapolis, MN, USA) or rhSTC-1 (10 μg/100 μl; BioVender, Brno, Czech Republic) was injected into the anterior chamber of rabbit eyes using a 30-G needle. The same volume of balanced salt solution (BSS; BioWhittaker, Walkersville, MD, USA) was injected in control eyes. Levofloxacin drops (Cravit; Santen Pharmaceutical Co., Ltd., Osaka, Japan) were instilled to prevent infection at the end of the procedure. 

Corneal Opacity Grading and Thickness Measurement

At 2, 4, 6, 12, 24, 36, and 48 hours after injury and treatment, the animals were killed by an intravenous injection of potassium chloride (1 mg/kg) under deep anesthesia, and each cornea was examined for opacity under a microscope, and photographed. Corneal opacity was independently graded by two ophthalmologists (JK, JYO) in a blinded manner as previously described ³³ : grade 0, completely transparent cornea; grade 1, minimal corneal opacity, but iris clearly visible; grade 2, moderate corneal opacity, iris vessels still visible; grade 3, moderate corneal opacity, pupil margin but not iris vessels visible; and grade 4, complete corneal opacity, pupil not visible. Corneal thickness was serially measured at the center of the cornea with an ultrasonic pachymeter (Pachymeter echograph; Quantel Medical, Clermont-Ferrand, France) at 2, 4, 6, 12, 24, 36, and 48 hours post injury. The average value of three repeated measurements at each time point was obtained. 

In Vivo Corneal Endothelial Cell Analysis

The central corneal endothelium of each live rabbit was observed at 2, 4, 6, 12, 24, 36, and 48 hours after injury with a noncontact specular biomicroscope (Konan specular microscope SP-8800; Konan Medical, Inc., Nishinomiya, Japan). Cell density was recorded as the number of endothelial cells per square millimeter, and the percentage of hexagonal cells (hexagonality) was used as an index of variation in cell shape (polymorphism). The average value of three consecutive measurements at each time point was obtained. 

Intraocular Pressure Measurement

At 2, 4, 6, 12, 24, 36, and 48 hours after injury, IOP was measured with a handheld rebound tonometer (ICare; Tiolat Oy, Helsinki, Finland) to exclude the possibility of effects of IOP on corneal endothelium. 

Histologic Examination

Forty-eight hours after injury, each rabbit was killed, and each cornea was extracted. For live/dead cell assay, each cornea was placed endothelial side up and 0.25% trypan blue was applied drop-wise to the endothelium for 90 seconds. Each cornea was rinsed with BSS to remove excess stain. Alizarin red S (0.20%; pH 4.2) was applied to the endothelium for 90 seconds and rinsed with BSS. After mounting on a glass slide, the corneal endothelium was observed and photographed with an Olympus BX 50 light microscope (Olympus Optical Co. Ltd., Tokyo, Japan). 

Each cornea was cut into 4-μm sections and subjected to hematoxylin-eosin staining or immunohistochemical staining for neutrophils. The formalin-fixed corneal sections were deparaffinized with ethanol, and the antigen was retrieved using a steamer in epitope retrieval solution (IHC WORLD, Woodstock, MD, USA). A mouse monoclonal antibody to rabbit neutrophil defensin 5 (NP-5; GenWay Biotech, Inc., San Diego, CA, USA) was used as primary antibody, and DAPI solution was used for counterstaining (IHC WORLD). 

Real-Time RT-PCR

For RNA extraction, the whole corneal tissue including the epithelium, stroma, and endothelium was cut into small pieces with microscissors, lysed in RNA isolation reagent (RNA Bee, Tel-Test, Inc., Friendswood, TX, USA), and sonicated with a probe sonicator (Ultrasonic Processor, Cole Parmer Instruments, Vernon Hills, IL, USA). Total RNA was extracted with a RNeasy Mini kit (Qiagen, Valencia, CA, USA), and double-stranded cDNA was synthesized by reverse transcription (High Capacity RNA-to-cDNA Kit; Applied Biosystems, Carlsbad, CA, USA). Real-time amplification was performed with a TaqMan Universal PCR Master Mix (Applied Biosystems). A rabbit glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used for normalization of gene expression. The PCR probe sets of rabbit TNF-α, IL-1β, and IL-6 were purchased from Applied Biosystems (Taqman Gene Expression Assay kits). 

Enzyme-Linked Immunosorbent Assay

For protein assays, the whole cornea including epithelium, stroma, and endothelium, was minced into small pieces and lysed in PRO-PREPTM Protein Extraction Solution (Intron Biotechnology, Seongnam, Korea). The samples were sonicated on ice with an ultrasound sonicator (Ultrasonic Processor, Cole Parmer Instruments), and the supernatant was collected after centrifugation at 13,523g for 20 minutes. To evaluate proteins in the aqueous humor, 100 μL of aqueous humor was serially collected from the anterior chamber of each eye with a 30-G needle at 2, 4, 6, 12, 24, 36, and 48 hours after injury, and cell-free supernatant was obtained after centrifugation. The supernatants were assayed for the concentrations of rabbit myeloperoxidase (MPO; Cusabio Biotech Co. Ltd., HangZhou, China) and IL-1β (Uscn Life Science, Inc., Wuhan, China) by ELISA according to the manufacturer's protocol. 

Western Blot

Clear lysates of protein from each cornea were prepared as described above and measured for concentration. A total of 30 μg protein was fractionated by SDS-PAGE on 10% bis-tris gel (Invitrogen, Carlsbad, CA, USA), transferred to nitrocellulose membrane (Invitrogen), and blotted with antibodies against rabbit FGF-2 (Abnova, Taipei City, Taiwan) or β–actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). 

Statistical Analysis

The data are presented as the mean ± SEM. Differences between two values were analyzed by the two-tailed Student's t-test (GraphPad Prism; GraphPad Software, Inc., La Jolla, CA, USA), and considered significant at P less than 0.05. 

Results

TSG-6 Decreased Corneal Edema and Opacity

To evaluate clinical effects of TSG-6 or STC-1 on the cornea, transcorneal freezing up to −80°C was applied for 30 seconds to the corneal surface of each eye. Immediately after injury, either rhTSG-6 (10 μg/100 μL) or rhSTC-1 (10 μg/100 μL) was injected into the anterior chamber. Injections of the same volume of BSS were used in control eyes. At 2, 4, 6, 12, 24, 36, and 48 hours after injury, each cornea was examined for opacity under a microscope, photographed, and graded (Fig. 1A). Corneal thickness was measured with an ultrasonic pachymeter. We found that corneal opacity and thickness were markedly increased by cryoinjury, and reached a peak at 24 to 36 hours after injury (Figs. 1B–D). Treatment with TSG-6 significantly reduced the development of corneal opacity and edema at all examined time points (Figs. 1B–D); however, STC-1 was not effective in reducing corneal opacity or edema (Figs. 1C, D). We measured IOP to exclude the possibility that elevated IOP after injury or treatment might directly affect the corneal endothelium. However, serial measurement of IOP indicated that cryoinjury or treatment with TSG-6 did not alter IOP (Fig. 1E). 

Figure 1

 

Effect of TSG-6 or STC-1 on corneal opacity and edema after transcorneal cryoinjury. (A) After transcorneal cryoinjury, either rhTSG-6 or BSS was injected into the anterior chamber of rabbit eyes. At 2, 4, 6, 12, 24, 36, and 48 hours after injury, the eyes were examined for corneal opacity, thickness, and IOP. (B, C) Representative corneal photographs demonstrated that corneal opacity gradually increased after cryoinjury, and reached a peak at 24 hours. Tumor necrosis factor-α stimulated gene/protein 6 treatment, not STC-1, significantly inhibited the development of corneal opacity. (D) Similar to corneal opacity, corneal thickness also increased after cryoinjury as measured by ultrasonic pachymetry, and reduced by TSG-6 treatment. Stanniocalcin-1 treatment had no effects on corneal thickness. (E) Intraocular pressure was not affected by cryoinjury or treatment. n = 5 in each group. Data were presented as mean ± SEM. *P < 0.05, **P < 0.01.

TSG-6 Rescued Corneal Endothelial Cells From Damage

We evaluated the effects of TSG-6 on corneal endothelial cells with serial measurements of corneal endothelial cell density and hexagonality. Specular biomicroscopy in vivo demonstrated that the density and hexagonality of corneal endothelial cells markedly decreased after cryoinjury (Figs. 2A–C). The morphology of corneal endothelial cells was also shown to be severely altered (Fig. 2A). In contrast, the density and hexagonality of corneal endothelial cells was significantly higher in eyes treated with TSG-6 than in BSS-treated controls (Figs. 2A–C), and the morphology of corneal endothelium was better preserved (Fig. 2A). Similarly, live/dead cell staining of the corneal endothelium at 48 hours post injury demonstrated a higher number of trypan blue-stained dead cells in the corneal endothelium of BSS-treated controls, as compared with TSG-6–treated eyes (Fig. 2D). 

Figure 2

 

Effect of TSG-6 on corneal endothelial cells after transcorneal cryoinjury. (A) Noncontact specular biomicroscopy demonstrated that the morphology of corneal endothelial cells was severely altered at 36 hours after cryoinjury. In TSG-6–treated eyes, the shape of corneal endothelium was markedly preserved. (B, C) Serial measurement of corneal endothelial density and hexagonality demonstrated that corneal endothelial cell density and cell hexagonality decreased with time after injury. However, TSG-6 treatment significantly increased corneal endothelial density and hexagonality. (D) Live-dead cell staining of the corneal endothelium demonstrated many trypan blue-stained cells in the corneal endothelium at 48 hours post injury, and few trypan blue-stained cells in TSG-6-treated eyes. n = 5 in each group. Data were presented as mean ± SEM. *P < 0.05, **P < 0.01.

TSG-6 Suppressed Neutrophil Infiltration in the Cornea

From the previous results, we found that TSG-6 treatment increased the viability of corneal endothelial cells, and thus prevented the development of corneal edema after transcorneal cryoinjury. Since transcorneal freezing induces inflammation, and TSG-6 has an anti-inflammatory effect, we tested whether TSG-6 might have an effect on intraocular inflammation after transcorneal cryoinjury. We histologically assayed the cornea that was extracted at 48 hours post injury. Hematoxylin-eosin staining demonstrated that the cornea was highly edematous after injury, which indicated corneal endothelial dysfunction, and TSG-6 treatment markedly suppressed the development of corneal edema (Fig. 3A). Immunohistochemistry demonstrated extensive neutrophil infiltration of the cornea after injury (Fig. 3B). In contrast, neutrophil infiltration of the cornea was markedly lower in TSG-6–treated eyes, when compared with BSS-treated controls (Fig. 3B). 

Figure 3

 

Histologic analysis for corneal edema and neutrophil infiltration. (A) Hematoxylin-eosin staining of the cornea demonstrated highly edematous cornea 48 hours after cryoinjury, and dense inflammatory cell infiltration of the corneal stroma. Corneal edema and inflammatory cell infiltration were markedly reduced in TSG-6–treated eyes. (B) Immunohistochemistry revealed dense neutrophilic infiltration of the cornea after injury, and neutrophil infiltration was markedly reduced by TSG-6 treatment. Original magnification ×100.

TSG-6 Reduced Pro-Inflammatory Cytokines and FGF-2

We assayed for levels of inflammation-related molecules in the cornea and aqueous humor of each eye. Real-time RT-PCR revealed that mRNA levels of TNF-α, IL-1β, and IL-6 in the cornea were highly increased by cryoinjury, and significantly reduced by TSG-6 treatment (Fig. 4A). We assayed for the protein levels of MPO, which is most abundantly expressed in neutrophils, ³⁴ as a quantitative measure of infiltrating neutrophils in the cornea and of IL-1β. We found that the levels of MPO and IL-1β were significantly lower in TSG-6–treated eyes than in BSS-treated controls (Fig. 4B). Similarly, serial measurements of MPO and IL-1β in the aqueous humor indicated that the protein levels of MPO and IL-1β gradually increased with a peak at 24 hours after injury, and were significantly reduced by TSG-6 treatment (Figs. 4C, 4D). Based on the previous report that IL-1β induced FGF-2 production in the cornea after transcorneal freezing, we also analyzed the expression of FGF-2 in the cornea at 48 hours post injury by immunoblotting, and found that the level of FGF-2 was markedly decreased in TSG-6-treated eyes compared with controls (Fig. 4E). 

Figure 4

 

Effect of TSG-6 on pro-inflammatory cytokines and FGF-2 in the cornea and aqueous humor after transcorneal cryoinjury. (A) Real-time RT-PCR revealed that mRNA levels of TNF-α, IL-1β, and IL-6 were highly increased in the cornea 48 hours after cryoinjury, and significantly reduced by TSG-6 treatment. (B) Similarly, ELISA showed that the protein levels of IL-1β and MPO were also decreased by TSG-6 treatment. n = 5 in each group. Data were presented as mean + SEM. (C, D) Serial measurement of IL-1β and MPO levels in the aqueous humor indicated that the protein levels of MPO and IL-1β gradually increased in the aqueous humor with a peak at 24 hours after injury, and were significantly reduced by TSG-6 treatment. n = 7 in each group. Data were presented as mean ± SEM. *P < 0.05, **P < 0.01. (E) Western blotting demonstrated that FGF-2 production in the cornea was increased by injury, and was markedly reduced by TSG-6 treatment.

Discussion

Our results demonstrate that transcorneal cryoinjury induces neutrophil infiltration in the cornea, and increases the level of pro-inflammatory cytokines and FGF-2. These findings are consistent with a well-established paradigm for the mechanism of corneal endothelial dysfunction: EMT mediated by inflammation and IL-1β–induced FGF-2. ^11–14 Tumor necrosis factor-α stimulated gene/protein-6, an anti-inflammatory protein, protected corneal endothelial cells from cryoinjury by the inhibition of neutrophil infiltration and the suppression of IL-1β and FGF-2 production. In contrast, parallel experiments with the antiapoptotic protein STC-1 did not improve corneal endothelial cell dysfunction. These findings collectively suggest that inflammation has a significant role in corneal endothelial dysfunction after cryoinjury, and blocking the upstream of the injury-mediated inflammation is effective in preventing development of corneal endothelial dysfunction. 

Tumor necrosis factor-α stimulated gene/protein-6 is expressed by many different cells in response to pro-inflammatory cytokines, and has multiple actions that are linked to modulation of inflammation and stabilization of the extracellular matrix. ^16–25 Of note, TSG-6 was identified as being responsible for the beneficial effects of mesenchymal stem/stromal cells in the treatment of diseases of the heart, ³⁵ cornea, ^27,28 brain, ^36,37 lung, ³⁸ and peritoneum. ^39,40 One anti-inflammatory mechanism of TSG-6 is the inhibition of neutrophil migration. ^18–21 A recent study ²¹ revealed that TSG-6 inhibits neutrophil migration by directly binding to the chemokine CXCL8 and suppressing CXCL8-mediated chemotaxis of neutrophils, and identified TSG-6 as the first soluble mammalian chemokine-binding protein described to date. Our study demonstrates that neutrophil inhibition by TSG-6 is also effective in preserving corneal endothelium from injury-mediated inflammation. Hence, these findings suggest that TSG-6 might be valuable for therapeutic or preventive measures to protect the corneal endothelium in situations when inflammation might possibly damage the cornea, such as accidental trauma, phacoemulsification, or acute angle closure glaucoma. 

Another possibility is that TSG-6 might have beneficial effects on the corneal endothelium by directly enhancing the proliferation of corneal endothelial cells and wound healing. We observed that the cryoinjury used in this study did not create the defect in the corneal endothelium in the rabbit eye (Supplementary Fig. S1); however, we cannot rule out the possibility that TSG-6 might have other effects than inflammation modulation, such as promotion of corneal endothelial cell proliferation. In this context, several reports recently demonstrated that topically-administered Rho kinase (ROCK)-inhibitor enhanced corneal endothelial cell density and functional recovery through the promotion of corneal endothelial cell proliferation in similar animal models and in humans. ^8–10,41 Hence, it is possible that combined therapy with TSG-6 and ROCK inhibitor may have synergistic effects through the protection of corneal endothelial cells from inflammation and the stimulation of cell proliferation, and provide more effective strategy for the treatment of corneal endothelial disease. Further studies to validate this hypothesis would help to guide the development of novel pharmaceutical treatments for corneal endothelial dysfunction, and shed light on an emerging field of corneal endothelial regenerative medicine. Since rabbit corneal endothelium is able to proliferate in response to injury in contrast to human corneal endothelium, further evaluation of TSG-6 in a primate model would be needed prior to clinical testing in humans. 

In conclusion, our results demonstrate that TSG-6 protected corneal endothelial cells from cryoinjury by inhibiting neutrophil infiltration and reducing the levels of pro-inflammatory proteins and FGF-2. 

Supplementary Materials

pdf S1 

Acknowledgments

Supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare (A112023; Sejong City, Seoul, Republic of Korea). 

Disclosure: J.-A. Kim, None; J.H. Ko, None; A.Y. Ko, None; H.J. Lee, None; M.K. Kim, None; W.R. Wee, None; R.H. Lee, None; S.F. Fulcher, None; J.Y. Oh, None