Tissue-derived microparticles reduce inflammation and fibrosis in cornea wounds

Abstract

Biological materials derived from the extracellular matrix (ECM) of tissues serve as scaffolds for rebuilding tissues and for improved wound healing. Cornea trauma represents a wound healing challenge as the default repair pathway can result in fibrosis and scar formation that limit vision. Effective treatments are needed to reduce inflammation, promote tissue repair, and retain the tissue’s native transparency and vision capacity. Tissue microparticles derived from cornea, cartilage and lymph nodes were processed and screened in vitro for their ability to reduce inflammation in ocular surface cells isolated from the cornea stroma, conjunctiva, and lacrimal gland. Addition of ECM particles to the media reduced expression of inflammatory genes and restored certain tear film protein production in vitro. Particles derived from lymph nodes were then applied to a rabbit lamellar keratectomy corneal injury model. Application of the tissue particles in a fibrin glue carrier decreased expression of inflammatory and fibrotic genes and scar formation as measured through imaging, histology and immunohistochemistry. In sum, immunomodulatory tissue microparticles may provide a new therapeutic tool for reducing inflammation in the cornea and ocular surface and promoting tissue repair.

1. Introduction

The cornea serves as a window to the eye, and transparency is central to its function. The unique organization of the extracellular matrix (ECM) of the cornea comprising collagens and proteoglycans allows for tissue transparency and robust mechanical integrity [1,2]. When this highly organized matrix structure is damaged, fibrotic scar tissue develops that causes corneal opacity and obstructs vision. There is a significant need for efficacious therapeutic options to treat traumatic injuries from physical and chemical abrasions, as well as cornea damage associated with the rising number of elective cornea surgical procedures [3,4].

Cornea wound healing occurs through a complex cascade involving multiple cell types and cytokines. Corneal keratocytes in the wound environment undergo apoptosis following stromal injury that is mediated by cytokines including interleukin (IL)-1 and tumor necrosis factor alpha (TNFα) [4]. Transforming growth factor beta (TGF-β) mediates differentiation of previously quiescent keratocytes into fibroblasts and also myofibroblasts that migrate to the site of injury [5,6]. The rapid production of unorganized ECM from activated myofibroblasts significantly reduces corneal stromal transparency and leads to scar formation. The cornea however does not function in isolation. It is linked together with the conjunctiva and tear film as the ocular surface system [7], which is exposed to growth factors and cytokines during inflammatory processes relevant to trauma and chronic disease [8,9].

A number of topical ointments and eye drops designed to decrease inflammation and reduce scar formation following corneal injuries and keratoplasty are clinically available today [10]. However, these treatments have significant shortcomings [11–13]. Cellular therapies are emerging as potential therapies to prevent scarring and to reduce the need for expensive and invasive allograft transplantation. In clinical trials, autologous and allogeneic limbal stem cells (LSCs) were transplanted to restore corneal epithelium after ocular surface burns to prevent chronic inflammation and corneal scarring [14–17]. Mesenchymal stem cells (MSCs), purported to be immunomodulatory and secrete anti-inflammatory molecules, also demonstrated efficacy in reducing corneal scar formation in preclinical models [18,19]. Corneal stromal stem cells transplantation has also shown promising outcomes in reducing corneal fibrosis and scarring in animal studies [20,21]. While these studies are promising, cell therapies have numerous challenges, including high manufacturing cost and challenges in batch-to-batch reproducibility. Difficulties of application and time sensitivity also arise when considering cell therapies for reducing inflammation and promoting functional wound healing after surgical trauma [22]. Acellular materials such as amniotic membrane [16] and fibrin glue [23] are biological alternatives to cell therapies that demonstrate some efficacy. However, these options have practical challenges that limit efficacy and leads to questionable reproducibility [24–26].

Biological materials derived from tissue ECM are gaining recognition for their pro-regenerative and immunomodulatory capacity [27,28]. Tissue ECMs, from allogeneic and xenogeneic sources, are processed mechanically and chemically to remove cellular remnants [29]. Further, the processed ECM can be manipulated into various forms depending on the need of the application. Biological materials are used clinically today for reconstruction of numerous tissues and clinical indications such as hernia, rotator cuff, and wound healing applications [27,30,31]. Preclinical studies and clinical trials are further expanding into cardiac, lung, and soft tissue reconstruction. While these biological materials were thought to primarily impact stem cells and their differentiation, their immunomodulatory capacity is now the center of interest. Biological scaffolds exhibit immunoregulatory effects by supporting type 2 pro-regenerative immune responses that include reducing pro-inflammatory cytokines, driving macrophage response towards an anti-inflammatory lineage, promoting Th2 effector cell responses, and interleukin-4 (IL-4) production [32,33].

In the current study, we applied tissue-derived ECM particles from several tissue sources, including porcine lymph nodes (LN), cornea (CO) and cartilage (CA), to facilitate cornea wound healing. The ECM particles were processed to remove cell remnants and milled into micro-particles. We investigated the anti-inflammatory effects of ECM particles on multiple ocular surface cells in vitro. The microparticles reduced inflammatory gene expression in ocular surface cells cultured in an inflammatory cytokine environment. The microparticles were then applied to a rabbit lamellar keratectomy model where they improved corneal stromal reconstruction and reduced scar formation. Overall, the ECM particles successfully regulated the wound healing process by decreasing the expression of inflammatory mediators, reduced immune cell infiltration in the wound, and reduced corneal fibrosis.

2. Materials and methods

2.1. Preparation of particulate ECM

ECM particles were prepared as previously described [34,35]. Briefly, fresh porcine lymph nodes, corneas and cartilage were purchased from a US certified butcher (Wagner’s Meats, Mt. Airy, MD) and frozen at −20 °C. The thawed tissue was ground manually and decellularized in 3% peracetic acid (Sigma-Aldrich, St. Louis, MO) at 37 °C for 4 h, and then in 1% Triton-X100 (Sigma-Aldrich) with 2 mM Ethylenediaminetetraacetic acid (EDTA; Sigma-Aldrich) at room temperature for 48 h under continuous agitation. After cellular components were removed, 0.03% DNase I (Roche Diagnostics Corp, Indianapolis, IN) treatment was performed at 37 °C for 24 h. The decellularized ECM was lyophilized, pulverized using a cryomill (SPEX 6770, SPEX SamplePrep, Metuchen, NJ). Particles were sterilized with ultraviolet radiation at 253.7 nm wavelength for 1 h and stored at −20 °C until use.

2.2. Morphology and size distribution of ECM particle

ECM particle morphology was characterized by field-emission scanning electron microscopy (SEM; Carl Zeiss, Jena, Germany). Lyophilized ECM particles were taped on the SEM sample stage with carbon tape, and sputter coated with 20 mA, 5 cm working distance at 0.1 mbar, for 120 s, to achieve a 10-nm coating. Images were taken under 1.0 kV, 2 mm working distance.

Average particle size and particle size distribution of each type of ECM particle were determined by dynamic light scattering, using Zetasizer Nano ZS90 (Malvern Instruments, Malvern, UK) at 25 °C. Samples were diluted with distilled water before measurement.

2.3. Ocular surface cell isolation and culture

Cells on the ocular surface were isolated from young New Zealand white rabbit tissues (eyes and lacrimal glands), purchased from Pel-Freez Biologicals (Rogers, AR).

2.3.1. Rabbit lacrimal gland acinar cells (LGACs)

Inferior lacrimal glands were removed aseptically and finely minced. Cell isolation was carried out as previously described [36]. Briefly, the tissue was enzymatically digested by collagenase (350 U/ml; Worthington Biochemical Corp, Lakewood, NJ), DNase I (40 U/ml; Roche) and hyaluronidase (300 U/ml; Sigma-Aldrich) for 45 min at 37 °C, and the digests were filtered through a 100 μm cell strainer and then centrifuged at 200g for 5 min. Cells were resuspended in HepatoStim culture medium supplemented with epidermal growth factor (EGF; 5 ng/mL; Corning, Tewksbury, MA). The LGACs were seeded on Matrigel^® (growth factor reduced, Corning) coated wells at a density of 5 × 10⁵ cells/cm², and incubated at 37 °C with 5% CO₂.

2.3.2. Rabbit conjunctival epithelial cells (CECs)

Conjunctival tissue was removed from rabbit eyes, and digested with Dispase^®II (1.2 U/ml; Roche) at 4 °C overnight. The loosened epithelial aggregates were dispersed from the surface with a cell scraper, and then collected by centrifugation at 200g for 5 min. The CECs were separated into single cells by a secondary digestion with Accutase^® (Sigma-Aldrich) for 10 min at 37 °C and filtered through a 100 μm cell strainer. The final cell pellet was resuspended in supplemented bronchial epithelial cell growth medium (BEGM; Lonza Walkersville Inc, Walkersville, MD). Isolated cells were seeded at a density of 5 × 10⁴ cells/cm² on tissue culture plates at 37 °C with 5% CO₂.

The air-lifting culture of CECs was performed as described in a previous study [37]. CECs stratified into multiple layers in air-lift culture conditions, and were maintained for 7 days at 37 °C with 5% CO₂ for downstream experiments.

2.3.3. Rabbit corneal keratocytes

Rabbit corneal epithelium and endothelium were scraped and peeled off before processing. Keratocytes were isolated using a sequential collagenase (3.3 mg/mL; Worthington) digestion [38]. Briefly, corneas were diced into quarters and digested in collagenase solution on an orbital shaker at 37° C. The digests were collected and filtered through 70 μm cell strainer after 30, 60 and 180 min of digestion, and each time fresh collagenase solution was added. Keratocytes were pelleted by centrifugation at 200 g for 10 min, and seeded at a density of 2 × 10⁴ cells/cm² on tissue culture plates with EpiLife^® medium (Thermo Fisher Scientific, Waltham, MA) at 37 °C with 5% CO₂.

2.4. IL-1β stimulation and particulate ECM treatment in vitro

Ocular surface cells were stimulated with pro-inflammatory cytokine IL-1β (20 ng/ml, Thermo Fisher Scientific) after the cells reached confluence. To compare their anti-inflammatory and secretion stimulatory effects, various types of ECM particles (lymph nodes, cornea and cartilage) were added to the media at a concentration of 1 mg/mL (0.1% w/v) based on a previous study [33]. Cells were maintained for 48 h and then harvested for gene and protein expression analyses. Cells cultured under standard conditions without IL-1β stimulation or ECM particles were used as controls.

2.5. F-actin staining of keratocytes

Alexa Fluor 488 Phalloidin (Thermo Fisher Scientific) was used to visualize the morphology of keratocytes. The cells were rinsed with phosphate buffered saline (PBS; Thermo Fisher Scientific) and fixed with 4% paraformaldehyde (PFA; Sigma-Aldrich) solution for 10 min, followed by permeabilization with 0.1% Triton X-100 for 20 min at room temperature. After thorough rinses with PBS, keratocytes were stained with Phalloidin for 20 min in the dark, followed by counterstaining for nuclei with DAPI (4′,6-diamidino-2-phenylindole dihydrochloride; Thermo Fisher Scientific). The cells were imaged using a fluorescent microscope (Axio Imager 2, Carl Zeiss, Jena, Germany).

2.6. β-hexosaminidase secretion assay

β-hexosaminidase, or N-Acetylglucosaminidase (NAG) is a lysosomal enzyme in the tear fluid secreted by lacrimal glands after carbachol stimulation and 4-Nitrophenyl N-acetyl-β-D-glucosaminide (NP-GlcNAc) was used as the substrate for NAG in this assay. Four replicates were carried out in each group. The culture medium was replaced with DMEM/F12 (Thermo Fisher Scientific) and incubated at 37 °C for 2 h. After removing a baseline sample, carbachol (Sigma-Aldrich) was added to the medium to a final concentration of 100 μM, and incubated at 37 °C for 30 min. The culture medium was collected and centrifuged at 700 rpm for 5 min. The resulting supernatants were analyzed for NAG catalytic activity with a NAG assay Kit (Sigma-Aldrich), per the protocol. Absorbance was measured at 405 nm using a microplate reader (BioTek, Winooski, VT).

2.7. Confocal microscopy measurement of secreted mucin in vitro

Following seven days of air-lift culture, CECs cultured on Transwell^® (Corning) were exposed to IL-1β (20 ng/ml) and each type of ECM particles for 48 h. Mucus membrane is typically labeled with conjugated dextran, which dissolves within the mucus and is relatively impermeable across the epithelium [39]. Texas Red-dextran (10 kDa, 2 mM in PBS; Thermo Fisher Scientific) solution was loaded onto the upper chamber of Transwell^® insert. Z-stack scanning was performed on the samples using a laser scanning confocal microscope (LSM 510, Carl Zeiss). The serial images were analyzed and stacked to generate 3-dimensional images for the measurement of mucin thin film thickness using ZEN imaging software (Carl Zeiss).

2.8. Animal surgeries and the application of ECM particles on corneal wound

New Zealand White rabbits, weighing between 2 and 3 kg, were used in the corneal wound healing study. The animals were housed and treated in accordance with the guidelines in the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Visual Research and also with the approval of the Animal Care and Use Committee at Johns Hopkins University. The rabbits were anesthetized by intramuscular injection of ketamine (15 mg/kg) and xylazine (5 mg/kg). Corneal stromal injuries were introduced with superficial lamellar keratectomy (SLK) using an automated microkeratome (Bausch & Lomb, Rochester, NY). The dimensions of the wound were 6 mm in diameter and 100 μm in average depth. The corneal flaps were surgically removed, leaving the corneal stroma exposed (untreated; Fig. 2A).

Fig. 2.

In vivo application of ECM particles on cornea after lamellar keratectomy. (A) Surgery procedures. The flap was removed, and in the untreated group, corneal stroma was left exposed; FG sealant with or without ECM was applied on the wound. ECM particles were visible within the FG gel. Rabbit corneas examined at different stages by slit lamp using direct diffuse (B) and focal/beam (C) illumination. Arrowheads indicated diffused beam corresponding to corneal opacity.

Fibrin glue (FG) sealant (Baxter, Deerfield, IL) was used as a positive control treatment for wound healing, as well as the delivery agent of ECM particles. After SLK, 10 μL fibrinogen solution was first pipetted on top of the corneal wound, followed by the addition of 10 μL thrombin solution. The mixture was allowed to gel at room temperature for 5 min to form a layer covering the wounded area. LN-ECM powder was mixed with fibrinogen solution before adding to the corneal wound, and the final ECM concentration in FG gel was 30 mg/mL (3% w/v). ECM particles could not be more concentrated due to the high viscosity of fibrinogen solutions and difficulty during corneal application. A topical antibiotic ointment (0.5% erythromycin; Wilson, Mustang, OK) was applied after the procedures and every day afterwards, while no systemic and local immunosuppressive agents were used.

2.9. Clinical observations of re-epithelialization and corneal haze analysis

The epithelialization of cornea was evaluated by fluorescein staining under blue light every day in the first week post-surgery until completed healing. The area of cornea stained with fluorescein was measured using ImageJ software (National Institute of Health, NIH, Bethesda, MD), and normalized to the whole cornea and this value was used for comparison between groups (Fig. S3A, C).

The eyes were also examined and photographed throughout the study by a slit lamp to assess corneal scar and haze formation. The amount of corneal haze was quantified with Photoshop software (Adobe Systems Inc., San Jose, CA) as previously described [40]. Images of the corneas were converted to gray scale for better contrast. The “set black point” option in Photoshop was used for background correction, in which a point on the cornea above the pupil with no haze was set to be black. Hence the scarring area of each image was normalized to its respective unscarred regions. The area above the pupil was selected using the “elliptical marquee tool” and the measurement was recorded (Fig. S3B). The value of “integrated density” was used for analysis, as it is related to both brightness and area of the haze.

Scar tissue within corneal stroma was imaged at various time points by an in vivo confocal microscope (ConfoScan 3, Nidek Inc., Fremont, CA, USA) similar to a previous study [41]. Briefly, rabbits were anesthetized and placed on a customized platform with the heads immobilized and one of the eyes facing the objective lens. Corneal stromal tissues of different layers (anterior, mid and posterior) were imaged using the Confoscan 3 software. Lubricant gel (Genteal; Novartis, East Hanover, NJ) was used throughout the imaging as an immersion fluid.

2.10. Quantitative real-time polymerase chain reaction (RT-qPCR)

Cornea in vivo samples were collected at one- and four-week post-surgery. After euthanasia, corneas were dissected off the eye and frozen at −80 °C in TRIzol^® reagent (Sigma-Aldrich), and cryogenically homogenized with a mortar and pestle in liquid nitrogen. The homogenate was collected and used for total RNA isolation. Gene expression level was studied for each individual cornea and each group had 4 replicates.

Total RNA from cells growing in vitro and cornea samples in vivo was isolated using the RNeasy Mini Kit (Qiagen, Germantown, MD), according to manufacturer’s protocol. The RNA concentration was measured using a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE). cDNA was synthesized from 1 μg of total RNA with the high-capacity cDNA reverse transcript kit (Thermo Fisher Scientific) on a MyCycler (Bio-Rad, Hercules, CA) thermal cycler. qPCR was conducted on the StepOnePlus System (Thermo Fisher Scientific) using SYBR Green PCR Master Mix (Thermo Fisher Scientific) to quantify the expression level of target genes. Primer sequences are listed in Table 1. Relative quantification of the signals was carried out by the ΔΔC_T method with the expression level normalized to that of control samples. β-actin was used as internal reference.

Table 1.

Primer sequences for RT-qPCR.

Gene		Sequence
Conjunctiva specific	MUC1	forward	GAG TCA CAG TGC GTG ATG TT
		reverse	GGC CAG GGC TAT GAA ATA GAT G
	MUC5AC	forward	CGC CTT CTT CAA CAC CTT CA
		reverse	TGG GCA AAC TTC TCG TTC TC
Lacrimal gland specific	AQP5	forward	CAA CGC GCT CAA CAACAA
		reverse	GTG AGT CGG TGG AAG AGA AA
	LTF	forward	GAT GCC ATG ACC CTG GAT AG
		reverse	GTC TGT GGC TTC GCT TCT
Inflammatory	IL-6	forward	GAA TAA TGA GAC CTG CCT GCT
		reverse	TTC TTC GTC ACT CCT GAA CTT G
	IL-8	forward	TGG ACC TCA CTG TGC AAA T
		reverse	GCT CAG CCC TCT TCA AGA AT
	TNF	forward	GTA GTA GCA AAC CCG CAA GT
		reverse	GGT TGT CCG TGA GCT TCA T
	MMP9	forward	AGT ACC GAG AGA AAG CCT ACT T
		reverse	TGC AGG ATG TCA AAG CTC AC
	iNOS	forward	CCA TCC CTG CAT CCT CAT T
		reverse	CCG GAG CCC TTT GTA CTC
	Arg1	forward	ACT CCA CTG ACA ACC ACA AG
		reverse	CCT GGT ACA TCA GGG ATC TTT C
Fibrosis	TGF	forward	CCT GTA CAA CCA GCA CAA CC
		reverse	CGT AGT ACA CGA TGG GCA GT
	CTGF	forward	AGG AGT GGG TGT GTG ATG AG
		reverse	CCA AAT GTG TCT TCC AGT CG
	COL1	forward	TTC TGC AGG GCT CCA ATG AT
		reverse	TCG ACA AGA ACA GTG TAA GTG AAC CT
	αSMA	forward	AGA GCG CAA ATA CTC CGT CT
		reverse	CCT GTT TGC TGA TCC ACA TC
	IL-17a	forward	ATC TGT GTC ACT GCT GCT G
		reverse	GAG TCC AAG GTG AAG TAG ATC G
	IL-23a	forward	GAG GGA GAT GAA GAG ACT ACC A
		reverse	CAG GCA GAA CTG AGT GTT GT
Internal reference	β-actin	forward	GCT ATT TGG CGC TGG ACT T
		reverse	GCG GCT CGT AGC TCT TCT C

2.11. Histology and immunohistochemistry (IHC)

Cornea samples in vivo were processed for histology and IHC as previously described [42]. Briefly, samples were fixed in 4% PFA, dehydrated, embedded in paraffin (Tissue-Tek, Sakura Finetek, Torrance, CA) and sectioned at 5 lm thickness. Hematoxylin & Eosin (H&E; Sigma-Aldrich) staining was carried out per the protocol to examine corneal stromal and epithelial reconstruction. IHC (fluorescence) was performed to visualized the expression of wound healing related markers. Primary anti-CD11b antibody was purchased from BD Biosciences (BDB550282) and anti-alpha smooth muscle actin (αSMA) antibody was from Dako (Clone 1A4). Goat anti-mouse and anti-rat antibodies (conjugated with Alexa Fluor 594; Thermo Fisher Scientific) were used as secondary antibodies. The sections were either imaged with Axio Imager 2 or confocal laser scanning microscope (LSM 510; Carl Zeiss). Final images shown in Fig. 5 were combined from several microscopic images, showing the region with the most severe scar tissue and fibrosis within each cornea.

Fig. 5.

The anti-fibrotic feature of ECM on corneal wound healing in vivo. (A, B) The expression of genes related to fibrosis (TGFβ1, CTGF, COL1, αSMA, IL-17a and IL-23a) was quantified by RT-qPCR at both week one and week four. (C) H&E staining of reconstructed corneal stroma and epithelium after SLK. Arrowheads are showing stromal scars formed after the surgery. The arrow indicated possible epithelial hyperplasia (thickening). Scale bar: 100 μm. (D) αSMA immunostaining (red) showing fibrotic tissue in the reconstructed corneal stroma at week 4. Scale bar: 50 μm. *p < 0.05; **p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2.12. Statistical analysis

Data are expressed as mean values ± standard deviation, and quantitative experiments were performed with at least triplicates. Significance was analyzed by one-way analysis variance (ANOVA). Statistical comparisons were between IL and 1β stimulated group and each ECM treated group, if not stated otherwise. Conventionally, p value < 0.05 was considered significant.

3. Results

3.1. Morphology and size distribution of ECM particles

Three tissue sources (lymph nodes (LN), cartilage (CA), and cornea (CO)) were processed to remove cellular components and milled into particles (Fig. S1). Particles of LN-ECM were spherical in shape while the CA-ECM particles were a mixture of spherical particles, elongated fibers and porous particles (Fig. S1). The CO-ECM particles contained lamellar sheets with fibrillar structures, similar to the native tissue structure. The size distribution of ECM particles also varied with tissue source (Fig. S1). LN-ECM particles demonstrated a broad size distribution (68.1–6439.0 nm), with a mean value of 1502.3 nm. Particles from CA-ECM exhibited a size range of 825.0–6439.0 nm with the largest mean particle size of 2967.9 nm. The CO-ECM particles were smaller with a mean size of 231.5 nm. The gross morphological observations demonstrated a tissue-dependent particle ultrastructure and size.

3.2. Immunomodulation of ocular surface cells by ECM particles in vitro

Cells were isolated and cultured from rabbit eyes to study the potential immunomodulatory effects of the biological ECM particles on ocular surface cells (Fig. 1A). First, keratocytes isolated from the cornea stroma were cultured with IL-1β to mimic an inflammatory environment. Exposure to the cytokine significantly increased mRNA levels of the pro-inflammatory genes IL-6, tumor necrosis factor alpha (TNFα), and matrix metalloproteinase 9 (MMP9) (Fig. 1B). Addition of LN-ECM and CO-ECM particles to the culture medium significantly decreased TNFα expression and treatment with CA-ECM particles reduced IL-6 expression. All three ECM particles lowered MMP9 expression, although it did not return to control levels with microparticle treatments. Exposure to IL-1β and various ECM particle treatments also influenced the morphology of keratocytes in vitro, as revealed by F-actin staining (Fig. 1B). Control primary keratocytes demonstrated a characteristic stellate shape with multiple cellular processes (Fig. 1B arrowheads). In the presence of IL-1β, very few cellular processes were present and keratocytes became more fibroblast-like. Treatment with ECM particles resulted in further morphology change. Keratocytes were more elongated compared to control and IL-1β treated groups.

Fig. 1.

The anti-inflammatory effect of ECM on ocular surface epithelial cells in vitro. (A) Ocular surface cells (lacrimal gland acinar cells, corneal keratocytes and conjunctival epithelial cells) were isolated and subject to IL-1β stimulation and ECM treatment. (B) Changes in gene expression (RT-qPCR) and morphology of keratocytes (F-actin staining) after exposure to IL-1β and ECM particles. F-actin was stained by phalloidin (green); cell processes were indicated with arrowheads. Scale bar: 50 μm. (C) RT-qPCR of tear film related gene expression of conjunctival epithelial and lacrimal gland acinar cells. (D) Mucin layer secreted by conjunctival epithelial cells in air-lifting culture was imaged by confocal microscope stained with Texas Red conjugated dextran. Images shown here were constructed from the x-z plane of a series of z-stacking images. Scale bar: 100 μm. Thickness values were measured from corresponding mucin layer images by Photoshop software and “+” indicated the mean values. *p < 0.05; **p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The ocular surface cells responsible for tear secretion were also evaluated; conjunctival epithelial cells (CECs) and lacrimal gland acinar cells (LGACs). Stimulation of CECs and LGACs with IL-1β also significantly upregulated gene expression of multiple pro-inflammatory cytokines including TNFα, IL-6, IL-8 and MMP9 (Fig. S2A, B). Treatment with LN-ECM lowered expression of some of these genes; CEC expression of IL-8 and MMP9 and LGAC expression of MMP9. CA-ECM and CO-ECM particles had no significant influence on inflammatory gene expression.

3.3. In vitro tear protein production modulated by ECM particles

ECM particles normalized surface cell functional activity as measured by tear protein production such as soluble mucins and lysozyme. Mucin is produced by conjunctival goblet cells and serves as the base layer of the tear film on the cornea surface. Confocal microscopy with Z-scanning (thickness) of fluorescently labeled dextran markers revealed lower mucin production in conjunctival goblet cells following IL-1β exposure (Fig. 1D). The physical mucin film thickness of cultures exposed to IL-1β decreased to 20 μm compared to control levels of 30 μm. Subsequent ECM particle treatment increased conjunctival goblet cell mucin production despite the inflammatory environment. With all ECM particle exposure, the fluorescent dextran was thicker and brighter, suggesting an increased mucin layer. NAG (lysozyme) activity stimulated by carbachol is a measure of LGAC function. Lysozyme (β-hexosaminidase assay) secretion increased slightly with IL-1β exposure. Particles from CA- and CO-ECM did not alter the secretion, however, LN-ECM doubled the secretion of NAG compared to control cultures (Fig. S2C).

Gene expression changes in the presence of the inflammatory cytokine and therapeutic microparticles supported increased tear protein production. Stimulation of CECs with IL-1β did not influence the level of mucin 1 (MUC1, major membrane bound mucin on corneal epithelium) or mucin 5AC (MUC5AC, secreted gel-forming mucin from the conjunctiva) expression, however, the addition of LN-ECM particles increased MUC5AC expression more than 2-fold. LGACs had decreased aquaporin 5 (AQP5) mRNA levels when exposed to IL-1β, which reverted to normal values when treated with LN-ECM and CO-ECM. Lactotransferrin (LTF) mRNA significantly increased when LGACs were stimulated by IL-1β, and ECM treatment further upregulated LTF (Fig. 1C).

3.4. Scar reduction post keratectomy with tissue ECM particle treatment

Based on the in vitro studies on the ECM particle efficacy, the LN-ECM particles were selected for further in vivo testing. To create a clinically-relevant scar, an SLK wound 6.0 mm in diameter and 100 μm in depth was created on the rabbit cornea. The wounds were treated with ECM particles delivered in FG. Control wounds were treated with FG alone or left untreated and were compared to healthy (unwounded) controls (Fig. 2A). Corneal re-epithelialization, following SLK, was assessed by fluorescein staining. All wounds re-epithelialized within five days for all groups (Fig. S3). The two treatment groups (FG and FG + ECM) epithelialized marginally slower than untreated group, with larger epithelial defects at day 2 and 3 (Fig. S3A,C).

Corneal healing and scar formation varied with treatment. Slitlamp examination with diffused or focal illumination revealed severe scar formation in the untreated corneas (Fig. 2B, C). FG and FG + ECM groups developed some corneal haze one week after SLK, but the scar decreased over time and was no longer visible after 4 weeks. Corneal haze was quantified by converting images to gray scale and calculating integrated intensity within the wound area (Fig. S3B). Both FG and FG + ECM significantly reduced haze over the 4 weeks after injury. In contrast, the untreated group that had an increase in haze over the time. Furthermore, FG + ECM group demonstrated lower haze formation compared to FG alone, especially at 14 and 28 days.

The reduction in scar tissue and corneal haze with ECM particle treatment was further confirmed with in vivo confocal imaging (Fig. 3). Keratocytes in healthy (unwounded) corneas were visible across the stromal depth. Surgery without treatment resulted in severe scarring (bright and high intensity image) in the anterior to mid stromal region of the cornea 1 and 4 weeks after surgery. While FG treatment demonstrated reduced scarring after 4 weeks compared to the untreated control, residual scar remained in the anterior stroma. Corneas treated with LN-ECM particles demonstrated minimal scarring in the anterior stroma at 1 week after wounding, which was no longer visible by week 4.

Fig. 3.

Corneal stromal tissue imaged by in vivo confocal microscope. In control group, keratocytes were clearly imaged in all layers (anterior, mid and posterior) of stroma. High intensity (brighter) indicated the presence of scar tissue within the stroma. Images from both one-week (A) and four-week (B) time points were shown. Scar bar: 100 μm.

3.5. Immunomodulation and fibrosis in corneal wounds

Treatment of corneal wounds with FG and FG + ECM downregulated the pro-inflammatory cytokine TNFα gene expression at both day 7 and 28 (Fig. 4A). FG + ECM also significantly downregulated expression of MMP9 at both days 7 and 28 after surgery, while FG group decrease was only significant at day 28. Treatment with FG + ECM initially suppressed inducible nitric oxide synthase (iNOS) expression at day 7, however both FG and FG + ECM treated groups increased iNOS expression compared to healthy and untreated corneas at day 28. In contrast, arginase I (Arg1) levels were initially upregulated in all surgical groups at day 7 and subsequently decreased at day 28 in all groups except FG (Fig. 4B).

Fig. 4.

The anti-inflammatory effect of ECM on corneal wound healing in vivo. (A, B) Inflammatory gene expression (TNFα, MMP9, iNOS and Arg1) of reconstructed cornea at day 7 and 28 in different groups, measured by RT-qPCR. (C) CD11b immunostaining (red) showing myeloid cell infiltration in the wounded area at week one. Scale bar: 200 μm. (D) Ki67 staining showing the presence of proliferating cells within the healing corneas at both week one and week four. Scale bar: 100 μm. *p < 0.05; **p < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Immunohistochemistry of CD11b, a marker of leukocytes in the innate immune system, including monocytes and macrophages [43], revealed differences in immune cell migration into the wounds with treatment (Fig. 4C). Without treatment, significant numbers of immune cell infiltrated the wounded stroma. FG treated wounds had lower numbers of CD11b positive cells and FG + ECM had even fewer. Staining with the cell proliferation marker Ki67 (Fig. 4D) demonstrated epithelial and stromal cell proliferation at 1-week post-SLK. In contrast, most corneal cells were quiescent at the 4-week time point in all groups, except a small number of epithelial cells which expressed Ki67 in the FG treated group.

Treatment with FG + ECM and FG also downregulated expression of genes associated with fibrosis; TGFβ1, connective tissue growth factor (CTGF), Type I collagen (COL1) and alpha smooth muscle actin (αSMA). FG + ECM treatment downregulated the expression of these genes to a greater degree compared to FG alone (Fig. 5A). FG and FG + ECM treatments also downregulated IL-17a expression 1 week post-surgery, and reached control levels by 4 weeks. Treatment with FG or FG + ECM initially upregulated IL-23a expression at week 1. However, after 4 weeks, IL-23a expression was lower in the FG + ECM treatment group, but continued to be upregulated in corneas treated with FG alone (Fig. 5B).

Fibrosis of the cornea wounds was also visualized by histology and immunohistochemistry. Dense, fibrotic stromal tissue below the epithelium was visible in control wounds with H&E staining (Fig. 5C; arrowheads). Both FG and FG + ECM treatments reduced severity of the fibrosis, with FG + ECM treatment demonstrating minimal fibrotic tissue formation. The greatest difference in FG and FG + ECM treatment was observed with the αSMA immunostaining. αSMA, a marker for myofibroblasts, was present in the untreated and FG cornea wounds (Fig. 5D). FG + ECM treatment displayed the lowest αSMA positive staining. Epithelial hyperplasia (thickened epithelium) observed in untreated and FG-treated corneas, was notably absent in the FG + ECM treated cornea wounds (Fig. 5C; arrow).

4. Discussion

The cornea is the outermost part of the eye and is thus susceptible to traumatic injury. Surgical procedures also create trauma in the cornea. Cornea refractive surgeries continue to increase with nearly 20 million patients undergoing laser-assisted in situ keratomileusis (LASIK) surgery in the U.S. alone [3,4]. When the ocular surface epithelium is injured, cytokines such as IL-1 are released from damaged epithelial cells, causing inflammatory cell infiltration into the ocular surface [4,8]. Excessive inflammation can lead to acute or chronic dry eye disease, as well as neovascularization and scarring following corneal stromal injuries [4]. Patients with severe corneal and/or ocular surface injuries often develop dry eye symptoms as well. Reduced levels of anti-inflammatory components in the tear film can also induce inflammation and further lead to abnormal wound healing and fibrosis [44,45].

Tissue-based materials provide a biological scaffold for wound healing. Multiple tissues provide the base material for manufacturing (i.e., bone, bladder, intestine, cardiac, and skin) and application for multiple clinical indications. These biological scaffolds are a complex mixture making the exact composition and thus mechanism of action unclear. However, they likely serve both a structural role for cell migration and biologically impact multiple processes including inflammation, differentiation, and vascular development [46–48]. Recent studies highlight the impact of biological scaffolds on innate and adaptive immune responses and the resulting pro-regenerative immune environment [33]. In vitro results frequently do not correlate with in vivo performance but do provide insight into efficacy and potential mechanisms. Therefore, we studied multiple cell types from the ocular surface to evaluate the potential efficacy of tissue particles in corneal wound healing. This recognizes the interactions of the tissues that comprise the ocular surface and the tear film. In an artificial inflammatory environment (Figs. 1 and S2), tissue particles reduced pro-inflammatory cytokines (TNFα, IL-6 and IL-8) and proteinase (MMP9, a major indicator for dry eye disease [8]) expression and normalized mucin and lysozyme secretions.

Generally, tissue-based materials, a complex mixture, can be processed into various physical forms based on the clinical need including sheets, gels, and particles. In this study, we fabricated particles from cornea, cartilage and lymph node tissues. While forms of human cornea and cartilage tissue are used clinically [49,50], they are not typically processed to remove cellular components. We selected cornea tissue since cornea wounds were the target application. Cartilage was selected as another tissue source for particles as it also has a dense ECM with fibrillar collagens and proteoglycans [51]. Finally, lymph nodes were also processed into particles. Lymph nodes are filled with reticular networks that contain fibrillar collagens and basement membranes [52]. The different outcomes observed from the particles are possibly due to the variation in ECM composition of the native tissues. As basement membrane components are beneficial for the growth and differentiation of epithelial cells, this could explain the effectiveness of lymph nodes particles in regulating the inflammatory environment in ocular surface epithelia. Cartilage and cornea are load bearing tissues, and their ECMs are composed of structural collagens in general [51]. Therefore, despite the beneficial effects they offer to keratocytes, CA- and CO-ECM did not provide significant advantages as supplements for epithelial cells (CECs and LGACs) under inflammatory condition in vitro.

The experiments in vitro proved that LN-ECM was the most efficient in regulating ocular surface homeostasis, especially in reducing inflammatory gene expression and increasing tear production (Fig. 1). To localize the LN-ECM particles in the corneal wound we used FG sealant as a carrier. FG itself is used frequently as a sealant or wound dressing during corneal surgeries [24]. In our study, ECM particles were applied directly on the corneal wound. Compared to subconjunctival injection, the soluble and insoluble factors within ECM particles could have a direct influence on the wound healing process. However, FG suffers from weak physical properties and fast degradation. FG alone did improve cornea repair, however, FG containing ECM particles even further reduced haze and scarring (Figs. 2 and 3). The LN-ECM particles also decreased immune cell infiltration and expression of inflammatory and fibrotic cytokines TNFα and TGFβ respectively, comparing to the group only FG was applied (Figs. 4 and 5). The cytokine IL-17 and IL-23 are involved in tissue repair and fibrosis [53,54]. In a cardiac fibrosis model reported by Wu et al., reducing IL-17 signaling by using IL-17a knockout model inhibited fibrosis and scar formation [55]. In the corneal wound healing process, IL-17a was downregulated by ECM particles in the early time point and maintained stable level in the later healing stage.

Limbal stem cells residing in the corneoscleral junction play a critical role in repairing damaged epithelium [15]. Cornea stroma damage activates TGF-β1 that induces the transition of keratocytes into fibroblasts and myofibroblasts, and consequently scar formation and wound contracture. Gene therapy approaches that decrease TGF-β1 transcription, and TGF-β inhibitors such as rosiglitazone and ROCK inhibitors have shown some promise in managing corneal fibrosis [4,40]. In our study, TGF-β1 was highly upregulated due to corneal epithelial and stromal injuries. LN-ECM particles reduced the gene expression of TGF-β1, as well as downstream fibrotic markers, including CTGF, COL1, Arg1 and α-SMA (Fig. 5A). Therefore, the anti-fibrotic effect of LN-ECM in reducing scar formation is possibly acted upon the TGF-β system as well, and further studies to elucidate the role of these cytokines in cornea fibrosis are needed. In a previous study it was reported that soluble fibrinogen induced inflammation in macrophages [56]. We also observed high level of MMP9, Arg1 and IL-23a in FG-treated group (Figs. 4 and 5), indicating a possible pro-inflammatory effect due to degraded fibrin fragments. However, this effect was counteracted in FG + ECM treated group.

Subconjunctival-injections of mesenchymal stem cells (MSCs) reduced inflammation during the acute phase of corneal chemical burns [19], such as TNFα downregulation and reduced inflammatory cell infiltration. The paracrine effects of MSCs may be responsible for the observed lower MMP levels [57], reduced immune cell infiltration, and improved tissue repair [4]. In this study, the benefits offered by LN-ECM resemble the paracrine effects observed from MSC therapy, and the mechanism is possibly the unique components originates from the lymph node reticular meshwork. Further proteomic analysis would be necessary for exact mechanism and for identifying the key signaling molecules. Besides ophthalmic application, LN-ECM particles could be utilized in other systems as well where both inflammation control and proper ECM remodeling are needed. ECM particles might be chemically crosslinked to increase the retention time in wounded tissue for optimal functions.

In summary, biological materials have unique immunomodulatory and reparative properties that can be applied in the ocular environment. Tissue derived ECM scaffolds are effective and accessible therapeutics for the management of corneal reconstruction after traumatic injuries to the ocular surface and refractive surgeries. They provide a convenient, off-the-shelf alternative to cell therapies. Particle processing, loading and delivery and can be further optimized for application and maintenance in the eye.

Statement of Significance.

Damaged cornea will result in scar tissue formation that impedes vision, and new therapies are needed to enhance wound healing in the cornea and to prevent fibrosis. We evaluated the effects of biological scaffolds derived extracellular matrix (ECM) during corneal wound healing. These ECM particles reduced inflammatory gene expression and restored tear film production in vitro, and reduced scar formation and fibrosis genes in the wounded cornea, when applied to in vivo lamellar keratectomy injury model. The immunomodulatory tissue microparticles may provide a new therapeutic tool for reducing inflammation in the cornea and ocular surface and promoting proper tissue repair.

Data availability statement

The raw and analyzed data required to reproduce the findings during this study are available from the corresponding author on reasonable request.