Genipin increases extracellular matrix synthesis preventing corneal perforation

Abstract

Purpose:

Corneal melting and perforation are feared sight-threatening complications of infections, autoimmune disease, and severe burns. Assess the use of genipin in treating stromal melt.

Methods:

A model for corneal wound healing was created through epithelial debridement and mechanical burring to injure the corneal stromal matrix in adult mice. Murine corneas were then treated with varying concentrations of genipin, a natural occurring crosslinking agent, to investigate the effects that matrix crosslinking using genipin has in wound healing and scar formation. Genipin was used in patients with active corneal melting.

Results:

Corneas treated with higher concentrations of genipin were found to develop denser stromal scarring in a mouse model. In human corneas, genipin promoted stromal synthesis and prevention of continuous melt. Genipin mechanisms of action create a favorable environment for upregulation of matrix synthesis and corneal scarring.

Conclusion:

Our data suggest that genipin increases matrix synthesis and inhibits the activation of latent transforming growth factor-β. These findings are translated to patients with severe corneal melting.

Introduction

The stroma comprises 90% of the corneal tissue and consists of water, and extracellular matrix─ mainly various collagens and proteoglycans.[1–4] The structure and hierarchical organization of the corneal extracellular matrix are a major reason for its unique properties: clarity, avascularity, tissue strength and shape.[1–4] Loss of stromal integrity caused by infection, chemical injuries or autoimmune disease (e.g., melting) can lead to corneal perforation and permanent vision loss. Full thickness corneal transplantation is the standard treatment when corneal perforation is imminent or has occurred. The long-term prognosis of corneal transplantation in the setting of active inflammation is poor.[5]

Keratocytes, neural crest–derived cells that reside in the corneal stroma, are believed to play an important role in the acquisition and maintenance of the normal properties of the corneal stroma including transparency.[1, 6, 7] During corneal stromal development, keratocytes regulate the synthesis and deposition of extracellular matrix and organize collagen fibrils.[8–10] Keratocytes are mitotically quiescent and have a dendritic morphology with extensive intercellular contacts.[7] The expression of keratocan, a proteoglycan found almost exclusively in the corneal stroma, is regarded as a marker of corneal keratocyte phenotype.[11, 12]

During standard culture conditions on plastic dishes using fetal bovine serum, keratocytes (murine,[13] bovine,[14] rabbit,[15] primate,[16] and human [17–19]) lose their dendritic morphology and expression of keratocan, thereby transforming to fibroblasts without morphological or biosynthetic characters of keratocytes (e.g., shut down keratocan expression).[11, 12, 17] These corneal derived fibroblasts cultured at low densities or stimulated by transforming growth factor-β1 might further differentiate into myofibroblasts.[15, 20] After corneal stroma acute injuries, quiescent keratocytes are believed to activate into fibroblasts and become mitotically active increasing their collagen and extracellular matrix synthesis needed to repair injuries and close stromal defects. Such processes will preserve the integrity and function of the eyeball at the expense of corneal clarity. Eventually, fibroblasts can transform into α-smooth muscle actin-expressing myofibroblasts. Transformed fibroblasts and myofibroblasts do not express keratocan. [1, 21, 22]

Corneal wound healing studies are focused in promoting tissue regeneration and preventing scar formation by regulating cytokines, tissue repair and immune system infiltration and response. Immediately following injury that results in loss of tissue integrity, host responses attempt to promptly restore integrity to prevent infection and loss of function. Genipin is a crosslinking compound, isolated from Gardenis jasminoides and Genipa americana. It is also known as the gardenia and genipap fruits. We and others have shown that genipin solutions can be safely used to crosslink corneal stromal with minimal risk of corneal toxicity, [23, 24] although a recent report suggests retinal toxicity when in direct contact with the posterior segment of the eye.[25] Previous work has shown that genipin efficiently and significantly increases corneal tissue stiffness in rabbit, porcine and human stroma.[26, 27] Besides the effects on increasing tissue stiffness by cross-linking the extracellular matrix, we recently demonstrated that genipin increased tissue resistance to bacterial collagenase digestion and retards stromal melting.[28] In this project, we investigate the potential action of genipin in decreasing the risk of corneal perforation by activating stromal cells synthetic activity. We further examine a mechanism that underlies increased stromal cells extracellular matrix synthesis and describe the clinical application of genipin in a group of patients with threatening perforations.

Methods

Animals

Wild-type (WT) C57BL/6 mice were used for studies evaluating the effects of genipin on stromal healing after injury. To obtain a pure population of keratocyte derived expanded fibroblasts, an inducible KeraRT/tetO-Cre mouse model obtained from Professor CY Liu laboratory [29] was bred with the Rosa26^mTmG mouse (Stock 008463, Jackson Labs, Bar Harbor, Maine). We generated a novel bitransgenic conditional KeraRT/tetO-Cre/mTmG mouse strain that we named I-KeramTmG. This mouse line expressed eGFP (Enhanced Green Fluorescein Protein) as a marker of keratocan induction (keratocyte phenotype marker) when fed with oral doxycycline supplementation. All mice were housed and treated in accordance with NIH’s Guide for the Care and Use of Laboratory Animals.

Stromal injury model

Only left corneas were subjected to injury. All experiments conformed to the use of Laboratory Animals and ARVO statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee of the University of South Florida College of Medicine. Animals were anesthetized with intraperitoneal injection of ketamine (100 mg/kg) and xylazine (10 mg/kg). Under microscope visualization, a stromal debridement injury technique was created on the corneal surface using a diamond burr. Approximately, eighty percent of the anterior stromal surface was debrided by the same surgeon for consistency (Christopher Donovan).[30]. At the time of injury, different concentrations of a genipin solution at different times of exposure were applied to the anterior stroma to evaluate their effect on scarring, including a control group where only vehicle was applied to allow genipin penetration into the anterior stroma after epithelial debridement. Genipin diluted in Hank’s balanced salt solution (HBSS, Sigma, St. Louis, MO, USA) vehicle was applied to the debrided surface on a filter paper (Whatman qualitative filter paper, Sigma).

Grading of corneal scarring was performed by two different corneal specialists in a masked fashion at 3 weeks post injury. Images were obtained at 3 weeks, a sufficient time for scar formation and maturation in murine corneas.[31] Eight mice per condition were used.

RNA isolation and quantification of mRNA

Whole corneas were dissected from adult, postnatal day 90. Epithelium was removed with dispase as previously published.[32] The stroma of cornea was cut into small pieces and total RNA was extracted using QIAzol Lysis Reagent (Qiagen, Venlo, Limburg) and RNeasy MinElute Cleanup Kit (Qiagen, Venlo, Limburg). Total RNA was extracted from expanded fibroblasts too. Reverse transcription and quantitative real-time PCR analysis was performed as described before. ( Sun et al., 2020b) The following primer sequences were used: Col1a1 FW: 5’-TTCTCCTGGCAAAGACGGACTCAA-3’, Col1a1 RV: 5’-AGGAAGCTGAAGTCATAACCGCCA-3’; Actb FW: 5’-AGATGACCCAGATCATGTTTGAGA-3’, Actb RV: 5’-CACAGCCTGGATGGCTACGT-3’. Acta2 FW: 5’-GTCCCAGACATCAGGGAGTAA-3’, Acta2 RV: 5’- TCGGATACTTCAGCGTCAGGA-3’. Serpine-1 FW: 5’ GTCTTTCCGACCAAGAGCAG-3’ , Serpine-1 RV: 5’- GACAAAGGCTGTGGAGGAAG-3’. Each sample was run in duplicate PCR reaction, statistical analysis was performed from 3–6 corneas of different mice at each time point.

Immunofluorescence microscopy

Fresh eyes were harvested from C57BL/6 mice 3 weeks after stromal injury. They were immediately embedded in OCT medium frozen with isopentane (Sigma Aldrich, St. Louis, MO) on dry ice. Corneal sections 5–7 μm were blocked using 10% donkey serum (Sigma Aldrich, St. Louis, MO). Sections were incubated with rabbit anti-mouse Collagen IV antibody (Southern Biotech, Birmingham, Al, 1:100 dilution) and fibronectin (Abcam, Cambridge, MA, 1:100) for 1 hour at room temperature. The secondary antibody was Alexa Fluor 568 Equus asinus raised anti–rabbit IgG (Invitrogen, Carlsbad, CA) used at 1:200 dilution. DAPI Fluoromount-G^® clear mounting solution (SouthernBiotech) with DAPI was used as a nuclear marker. Images were captured using a fluorescence microscope (Leica DM5500B). Identical brightness and exposure settings during image acquisition and negative controls were used to facilitate comparisons between samples.

Second harmonic generation microscopy

Corneal tissue from mice 10 days after creating a full-thickness keratotomy wound was evaluated using standardized protocol.[30, 33, 34] Following euthanasia the injured left cornea from each mouse was embedded and frozen in OCT medium (Sakura Finetek, Torrance, CA). Cross sections of 15 μm were cut using an NX 50 cryostat. Each section was imaged using an Olympus MPE-RS microscope using a 25× (0.95 NA) water-immersion objective (Olympus). Propidium iodine (Thermo Fisher Scientific, Waltham, MA) in a 1:100 concentration was added. Two-photon second harmonic generation (SHG) signals were generated using a mode-locked titanium:sapphire laser at 960 nm. The SHG forward-scattered signals passing through the corneal sections were collected using a 0.8 NA condenser lens with a narrow band-pass filter (465–485 nm). All samples were scanned using a 2-μm z-axis step size from the back to the front of the section. The two-photon excited fluorescent signal from propidium iodine was captured with a band pass filter (575–630 nm).

Isolation and culture of corneal fibroblasts

After euthanasia, the eyes of P60 I-KeramTmG strain mice, induced with dietary oral Doxycycline for 1 week, were copiously washed with Betadine ophthalmic solution, and then incubated in DMEM containing 15 mg/ml dispase II (Roche Applied Science) at 4 °C for 18 h. The entire corneal epithelium sheet loosened by this treatment was removed by vigorous shaking. Under a dissecting microscope, the corneal stroma was separated from the sclera at the corneoscleral limbus by pressing down the limbus with a 27-gauge needle. Isolated corneal stromas were incubated overnight at 37 °C in DMEM containing 1.25 mg/ml collagenase A (Roche Applied Science) and 25 μg/ml gentamicin. A keratocyte-containing cell suspension was then seeded on T25 flasks (Thermo Fischer Scientific) in DMEM containing ITS (5 μg/ml insulin, 5 μg/ml transferrin, and 5 ng/ml sodium selenite), and 25 μg/ml gentamicin supplemented with 5% FBS. The suspension of keratocytes prepared from 12 mouse corneal buttons was seeded into each flask. When cells reached confluence, they were isolated and treated with trypsin for cell flow study. Cells were cultured on dishes of different stiffness (CytosSoft, Elastic modulus plates) that were obtained from Sigma.

Flow cytometry and cell sorting

Primary stromal cells (passage 1) from I-KeramTmG mice, fed with diet doxycycline for 1 week before sacrifice, were sorted via flow cytometry. Single cells suspension was obtained by trypsinization, washed with Dulbecco’s phosphate-buffered saline (DPBS) plus 25mM HEPES, suspended into same DPBS/HEPES buffer plus 1% dialyzed fetal bovine serum (FBS), then transferred into a 5 ml polypropyleane FACS tube with cell strainer filter cap. Sorting of the GFP positive cells was done on a Becton Dickinson (San Jose, CA) FacsMelody with a 100 mM nozzle, using FACSChorus software. GFP positive single cells were identified by using no-eGFP matching cells as negative control. The brightest 30% of the positive cells were sorted with NERL Diluent 2 (Thermo Scientific, Waltham MA) as sheet fluid. After sorting the cells were immediately pelleted and resuspended in growth media.

TGF-β latency in crosslinked corneal tissue

Four human corneas from donors ranging in age from 59 to 65 years were maintained at 4°C in Optisol (Chiron Vision, Irvine, CA, USA) for less than seven days after death. They were obtained from the Tampa Lions Eye Bank (Tampa, FL, USA) and handled according to the tenets of the Declaration of Helsinki. The epithelium of each cornea was removed with a 15-degree blade. Twelve 2-mm diameter stromal discs were trephined from each cornea using a 2-mm disposable dermatology punch (Integra Miltex, York, PA, USA). Trephined stromal discs were tested in different conditions to evaluate if genipin crosslinking of the stromal matrix interfered with TGF-β activation. Discs were immersed in HBSS (vehicle) for 30 minutes or in 2 mg/ml genipin in HBSS for 30 minutes. To create a dose response curve, stromal discs were treated with different genipin concentrations for 30 minutes: 0.2 mg/ml, 1 mg/ml and 2 mg/ml genipin in HBSS. Each disc was then transferred to our TGF-β reporter cell system, as described previously.[33] Briefly, transformed mink lung cells transfected with luciferase cDNA driven by plasminogen activator inhibitor (PAI-1) promoter (A generous gift of Dr Daniel Rifkin, New York University) were seeded and allowed to attach to a 24-well plate for 4 hours. DMEM from each stromal disc was then instilled into individual wells within the 24-well plate. 10 μmol/L TGF-β type I receptor/activin receptor-like kinase 5 (ALK5) inhibitor, SB431542 (Tocris Bioscience, Minneapolis, MN) was used as a negative control. Luciferase assay was performed by using Promega’s Luciferase Assay System (Promega, Madison, WI), and luminance was measured with Synergy HT plate reader (BIO-TEK, Winooski, VT).

Topical genipin treatment in patients with progressive corneal thinning

A total of seven patients with progressive corneal melt unresponsive to conventional treatments were treated with topical application of 2 mg/ml genipin (Sigma). An expert committee was consulted according to the Declaration of Helsinki (article 37) and the patients met the criteria for the genipin crosslinking therapeutic modality accepted in the protocol. The protocol was approved by IRB 014–149, Universidad Nacional de Colombia, School of Medicine, Bogota, Colombia. Prior to genipin treatment, the informed consent was reviewed with each patient, who consented to treatment with a single application of topical genipin. Demographic data and prior treatments of patients is listed in Table 1. A solution of 2 mg/ml genipin solution suspended in sterile saline buffered was prepared. A sterile 5 mm sponge was soaked in this solution and applied to the cornea for 5 minutes. The sponge was then removed, and the eye was copiously irrigated with BSS. A bandage contact lens (Proclear 1 day, Cooper Vision) was applied for 4 days. Each patient was started on topical moxifloxacin and preservative free artificial tears four times daily after genipin treatment. Each patient was observed daily for at least 5 days after genipin application, then monitored based on clinical findings. Slit lamp photos were taken prior to genipin treatment, at day 10, day 30, and 10 months after treatment.

Table 1.

Demographics of patients that received genipin treatment.

Patient	Underlying Cause	Age	Gender	Treatment regimen at presentation
1	Mycobacterium	56	Male	Topical kanamycin, levofloxacin; PO cycloserine, ethionamide, and pyrazinamide
2	Fusarium	26	Male	Topical natamycin; PO fluconazole
3	Unknown	58	Male	Topical moxifloxacin
4	Fusarium	25	Male	Topical natamycin
5	Pseudomonas	85	Female	Topical moxifloxacin
6	Unknown	70	Female	Topical moxifloxacin, cefazolin
7	Pseudomonas	67	Female	Topical moxifloxacin, amikacin

Statistical analysis

GraphPad Prism version 9.1.2 (San Diego, CA, USA) was used for all statistical analyses and all data are shown as means ± standard deviation. Statistical analyses were done by the Research Methodology and Biostatistics Core, Morsani College of Medicine, University of South Florida. Statistical significance between two continuous dependent variables was evaluated by t-student analyses and by simple ANNOVA for more than 2 groups of dependent variables. Values with P < 0.05 were considered statistically significant, (*P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001).

Results

Genipin promotes scarring after injuries that disturb the stromal extracellular matrix homeostasis.

To evaluate the effects of genipin on stromal healing, we performed a single application of topical genipin using different dosages in mice models of stromal debridement.[30] At 3 weeks post injury (known time to induce stromal scarring in mouse models of corneal injury [31]), we found increased stromal scarring when higher doses of genipin were applied to the injured anterior stromal surface. Longer exposure to genipin resulted in denser haze and scarring. Control groups including stromal debridement without topical genipin and topical genipin without stromal debridement created no significant scarring (Figs. 1A–F). There was statistically significant difference in the scar grading with increasing genipin dosage, one way ANOVA (p < 0.0001), (Fig. 1, Bottom).

Figure 1. Genipin exposure increases corneal scarring at 3 weeks only if stromal injury is performed.

(A,B) No significant haze is noticeable after mild stromal injury without topical genipin application or epithelial debridement and exposure to genipin without stromal injury. (C, D, E and F) Increasing corneal haze and scarring noted with higher genipin concentrations 0.1%(C) , 0.125 % (D) or 0.25% (E) for 2 mins or 0.5% for 5 mins (F). Increasing dose or exposure time resulted in increased haze and scar as graded using a modified Fantes scale. **P < 0.01. NS: not significant.

Those findings noted during clinical corneal examination correlated with immunohistology. Evaluation of injured tissue exposed to genipin vs vehicle showed increased fibronectin and collagen IV expression (matrix components of provisional corneal matrix and wound healing[35, 36]) in the injured anterior stroma and demonstrated disorganization of the anterior stromal extracellular matrix, and hypercellularity in corneas that underwent stromal debridement followed by topical genipin in contrast to corneas that had topical vehicle only (Fig. 2A1 vs 2A2). These findings were confirmed by second harmonic generation imaging that demonstrated disorganization of the anterior stromal collagen network and increased forward scattered signaling suggesting increased collagen fibril density and/or disorganization when genipin was applied following stromal debridement compared to corneas treated with vehicle only (Fig. 2B1 vs 2B2). Similarly, quantitative PCR from tissue obtained from injured corneas demonstrated increased expression of Serpine1, Acta2, and Col1a1 in genipin treated group vs vehicle only (Fig. 2C). These results suggest that application of increasing doses of genipin to the cornea during the wound healing process upregulate stromal synthesis of extracellular matrix. Clinically, significant increase in scarring, haze and even neovascularization was noted.

Figure 2. Increased cellular synthetic activity and deposition of a disorganized matrix following genipin treatment.

Topical application of 2.5 mg/ml genipin for 5 minutes in vehicle (HBSS) following injury vs vehicle only for 5 mins. (A1) Denser scarring and haze are observed in eyes treated with genipin following mechanical burr injury. Collagen IV and fibronectin expression was upregulated. In contrast, in eyes treated with vehicle only following injury, Collagen IV and fibronectin expression was minimal (A2). Second harmonic generation imaging shows increased cellularity, higher forward scattering signaling suggesting a more disorganized deposition of collagen fibrils following injury when genipin was used (B1) compared to vehicle (B2). Increased expression of PAI-1, collagen I and α-smooth muscle actin in stromal matrix after genipin use. Negative controls showed no reactivity. Nuclei were stained with DAPI (blue). At least 3 samples were used. *P < 0.05; **P < 0.01. Bar 50μm.

Isolation and expansion of a keratocyte derived cell lineage expressing eGFP.

To elucidate the mechanism(s) explaining scar formation after genipin application in our corneal injury model, we used an in vitro system, in which cultured cells were obtained from our I-KeramTmG mouse strain. We created this mouse strain to isolate corneal stromal fibroblasts derived from a keratocyte lineage (Kera eGFP expressing) and to be able to track keratocyte derived cell lineage from other cell types (not eGFP expressing). We were concerned that contamination during tissue culture occurs with cells isolated and expanded from sclera, iris or other structures around the cornea or from other cell populations in the stroma that are not keratocytes.[18] A finding highly suggestive of a heterogeneous population of cells in the stroma is shown in histology sections of our I-KeramTmG mouse that demonstrate that some cells in vivo, in the stroma continue to express membrane tomato (red color) after 1 week of oral induction with doxycycline (Fig. 3A, see asterisks).

Figure 3. Keratocyte derived cell lineage responds to matrix stiffness by regulating cell morphology and function.

Our I-KeramTmG mouse strain facilitates isolation of keratocyte derived cell lineages expressing eGFP. (A) In this corneal histology section, epithelium and endothelium express dTm while most but not all stromal cells express keratocan, see asterisks (eGFP). (B) Increasing matrix stiffness regulates keratocyte derived cells morphology. Broader cells noted with increasing matrix stiffness. (C) Increasing matrix stiffness activates TGF-β signaling as suggested by increased PAI-1 mRNA in keratocyte derived fibroblasts. Collagen I mRNA synthesis is upregulated by increased matrix stiffness in keratocyte derived fibroblasts. White bar represents mRNA obtained from expanded cells on 2kPa and black bars from cells expanded on 64 kPa. Nuclei were stained with DAPI (blue). At least 3 samples were used. *P < 0.05; **P < 0.01. Bar 50μm.

A keratocyte derived cell lineage responds to increased matrix stiffness by upregulating extracellular matrix synthesis.

Once we obtained a keratocyte derived cell lineage after inducing eGFP expression in vivo and selecting eGFP expressing cells by flow cytometry, we were able to study the effects of matrix stiffness in vitro in keratocyte derived cells. We were interested in studying the effects of increased matrix stiffness in keratocyte lineage cells because it is well established that genipin increases corneal stroma stiffness. We found that like other cells expanded on matrices of different stiffness, our I-KeramTmG derived cells showed significant changes in morphology with increased matrix stiffness (Fig. 3B). By inducing eGFP expression in adult I-KeramTmG mice in vivo, by feeding them with oral doxycycline for 1 week, we were able to expand corneal fibroblasts derived from keratocyte lineage in vitro (passage 1) and isolate only those cells expressing eGFP by flow cytometry. Cells were barely attached and had a round shape when expanded on a matrix of 0.2 or 2 KPa. Cells did not expand well in the presence of fetal bovine serum at these surfaces. In contrast cells expanded and proliferated well on stiffer surfaces, 32 kPa and 64 kPa, and had a broader cell body. Quantification of Serpine1 and Col1a1 mRNA suggested that keratocyte derived fibroblasts increase collagen I synthesis significantly with increasing stiffness. Increased Serpine1 mRNA synthesis is suggestive of upregulated TGF- β signaling with higher stiffness (Fig. 3C). Together, these data demonstrate that increasing matrix stiffness is itself a regulator of cell function and synthesis in a keratocyte derive cell lineage.

Crosslinking of stromal matrix by genipin retards TGF- β activation and maintains latency in the extracellular matrix in vitro.

To simulate matrix crosslinking and stiffening, human stromal discs were chemically crosslinked using 2 mg/ml of a genipin solution.[28] Genipin crosslinked tissue and tissue exposed to vehicle solution as control for 30 minutes were heated at 80°C for 10 minutes to activate stored latent TGF-β in the stroma. The amount of activated TGF-β was quantified by stimulating transformed mink lung cells. We found a statistically significant reduction in the activation of total TGF-β in the genipin crosslinked tissue compared to vehicle-treated control tissue (Fig. 4A). To further explore the hypothesis that tissue crosslinking decreases the activation and/or release of latent TGF-β stored in matrix, we immersed human stromal discs in different concentrations of genipin creating a dose response experiment. We used three different genipin concentrations: G2 (2 mg/ml), G1 (1 mg/ml) and G0.2 (0.2 mg/ml), (Fig. 4B). Exposure to higher doses of genipin showed decreased concentration of TGF-β released into solution. These results suggest that stromal crosslinking by genipin downregulates latent TGF-β activation.

Figure 4. Stromal crosslinking and stiffening impair human corneal stromal TGF-β activation.

(A) Luciferase assay was used to functionally quantify TGF-β availability in human corneal stromas crosslinked with genipin. Stromal discs, 2mm in diameter after epithelial removal, were crosslinked in a solution of 2 mg/ml of genipin for 30 minutes (G) or placed in HBSS (vehicle control, V). Tissue was then washed out 3 times and heated at 80°C for 10 mins to activate latent TGF-β. D denotes basal levels of TGF-β by culturing reporter cells only in DMEM, T denotes reporter cells stimulated by 10 ng/ml human recombinant TGF-β1. A TGF-β specific blocker, SB431542 (S) demonstrate that reporter cells were responding to TGF-β signaling and no other cytokines. ** p=0.005; (B) Dose response experiment using three different genipin concentrations: G2 (2 mg/ml). G1 (1 mg/ml) and G0.2 (0.2 mg/ml). RLU: relative light unit. **p=0.0019. Experiments were repeated 4 times to ensure reproducibility.

Topical genipin accelerates scar formation and slows progressive corneal melting in human patients.

Of seven patients treated with topical genipin, there was prompt reversal of progressive ulceration in six. In several patients a slight blue hue was observed in the treated cornea which gradually faded over the subsequent days. Each of these six patients were treated with different antimicrobials based on corneal cultures or empiric therapy based on clinical course at the time of presentation without success in stopping progression of corneal thinning (See Table 1). Each of these six patients were observed until re-epithelialization and a stable scar was achieved; none received additional treatments with genipin. Fig. 5 illustrates a case of progressive peripheral melting and thinning treated with genipin that seemed to halt progression to perforation while being treated. This case represents patient number three in Table 1. Supplemental Fig. 1 shows progressive corneal thinning and significant possibility of perforation. Two weeks after genipin application in the thin area and continuation of medical therapy, stromal thinning seems improved. This case represents patient number one in Table 1. One patient continued to display progressive ulceration and corneal thinning in spite of genipin use, ultimately leading to small perforation which was managed with a conjunctival flap. Following surgery, patient’s ulceration improved and re-epithelialized, eventually forming a stable scar. See Table 2 for a summary of clinical results.

Figure 5. Clinical case demonstrates genipin effects in case of active ulceration.

Illustrative case showing a red eye with progressive stromal thinning mostly located in the corneal periphery and midperiphery despite continuous broad spectrum antibiotic treatment (A). Eye appears quiet and stromal thinning improved 1 week after genipin application to affected area. Note light stromal blue hue (B). Quiet eye with peripheral vascularization and haze at no risk of perforation. Slight blue hue still present. This case represents patient number three in Table 1.

Table 2.

Detailed clinical description of clinical condition of each patient that received genipin treatment.

Patient	Underlying Cause	Ulcer size	Infiltrate depth	Descemetocele	Scleral involvement	VA pretreatment	VA posttreatment	Ocular complications	Subsequent Treatments
1	Mycobacterium	5mm	80%	Yes	No	LP	HM	None	None
2	Fusarium	8mm	80%	No	Yes	LP	20/400	None	None
3	Unknown	10mm	70%	No	No	HM	20/100	None	None
4	Fusarium	5mm	90%	Yes	No	LP	20/100	Corneal perforation	Conjunctival flap, topical amphotericin, PO posaconazole
5	Pseudomonas	6mm	70%	No	No	LP	HM	None	None
6	Unknown	7mm	70%	No	No	NLP	NLP	None	None
7	Pseudomonas	5mm	80%	No	No	NLP	NLP	None	None

HM: hand motion, LP: light perception, NLP: no light perception.

Discussion

In the setting of corneal ulceration causing progressive thinning or impeding ocular perforation, the goal of treatment is preservation of tissue integrity by removal of inciting agent of injury and through the stimulation of extracellular matrix synthesis. In this manuscript, we show that genipin treatment has beneficial effects in the context of corneal melting with stromal thinning. In areas of corneal thinning or active melting, crosslinking of stromal tissue by genipin upregulates extracellular matrix synthesis probably by at least two different mechanisms: 1- activation of fibroblasts collagen and matrix synthesis by stiffening of the corneal extracellular matrix, and 2- regulation of latent TGF- β activation in the stromal matrix. Through these potential mechanisms of action, topical genipin treatment reduces or slows the rate of corneal melt, increases matrix synthesis, promotes scar formation, and diminishes the risk of corneal perforation.

We and others have demonstrated the effects of genipin in crosslinking tissue and stiffening the corneal matrix.[24, 26–28, 37] It is well established that a stiffer matrix favors cell synthesis of extracellular matrix components including fibronectin and Latent Transforming Growth Factor Beta Binding Protein 1 (LTBP-1). A stiffer matrix also regulates mechanical or enzymatic activation of latent TGF- β. In vitro and in vivo experiments have demonstrated the upregulation of matrix synthesis by increased activation of latent TGF- β in a stiffer matrix.[38–41] Our findings show significant morphology changes in fibroblast derived from a keratocyte lineage when exposed to a stiffer matrix. When these fibroblasts were expanded on matrices of increasing stiffness, their cell size and shape increased with increasing matrix stiffness. These changes in morphology came with upregulation of new matrix formation. We hypothesized that increased matrix and stromal stiffness created by genipin upregulates matrix synthesis and deposition.

We explored the possible role of TGF- β signaling after genipin exposure in our models. It is well known that TGF- β is a major regulator of corneal wound healing and upregulates matrix synthesis in activated fibroblasts.[42, 43] Our data demonstrated increased matrix synthesis and increased PAI-1 signaling, suggesting increased TGF- β activity with increasing stiffness. We used an ex vivo assay using human stromal tissue to evaluate how genipin crosslinking regulates TGF- β activation. In our ex-vivo models, genipin treated corneas did not activate latent TGF- β as efficiently as those not treated. How can decreased latent TGF- β activation increase haze formation? Our functional experiments using reporter cells show that activation of latent TGF- β is decreased in genipin crosslinked stromal tissue. This inverse correlation suggests that activation of latent TGF- β is impaired. We hypothesize that the crosslinking effects of genipin might preserve available latent TGF- β deposits in the matrix that upregulates matrix synthesis and deposition during later stages of wound repair.

Based on the two mechanisms discussed above, increased matrix stiffness and regulation of latent TGF- β stores, genipin increases extracellular matrix synthesis. Stromal haze and scar formation are upregulated by application of topical genipin when stromal tissue is mechanically debrided, and the process of wound healing is activated. Our In vivo experiments in mouse models of injury, indicate a dose-dependent relationship with increased haze in those groups that underwent higher doses of genipin when the wound healing process is active. The degree of haze and scar formation increased with higher concentrations of genipin treatment and with longer exposure to genipin. Both experiments demonstrate the role that genipin plays in upregulating matrix synthesis. Clearly, activation of the wound healing process is needed for genipin to regulate matrix synthesis. Stromal debridement only or topical genipin application without stromal injury did not significantly upregulate scarring.

There are several limitations of this study. Although genipin treatment has shown to have a favorable safety profile in prior studies, the long-term effects of topical treatments to the cornea are unknown. Nonetheless topical treatment with genipin shows significant promise in treating progressive corneal ulcers that are unresponsive to traditional therapies. Given the high risk of vision loss and ocular compromise coupled with a lack of proven treatment modalities, we believe that topical genipin could be a valuable tool in managing corneal melting and in preventing corneal perforations.

Supplementary Material

1

Supplemental Figure 1 Genipin treatment decreases perforation risk. Corneal thinning is evident with significant possibility of perforation noted in this patient. Two weeks after genipin application in the thin area and continuation of medical therapy, stromal thinning seems improved. Note blue hue in area of genipin treatment and presence of neovascularization in the regenerated stromal matrix. This case represents patient number one in Table 1.

Figure 6.

Graphical summary of mechanism of action for genipin.