Calreticulin accelerates corneal wound closure and mitigates fibrosis: Potential therapeutic applications

Abstract

The processes involved in regeneration of cutaneous compared to corneal tissues involve different intrinsic mechanisms. Importantly, cutaneous wounds involve healing by angiogenesis but vascularization of the cornea obscures vision. Previous studies showed that topically applied calreticulin (CALR) healed full-thickness excisional animal wounds by a tissue regenerative process markedly enhancing repair without evoking angiogenesis. In the current study, the application of CALR in a rabbit corneal injury model: (1) accelerated full wound closure by 3 days (2) accelerated delayed healing caused by corticosteroids, routinely used to prevent post-injury inflammation, by 6 days and (3) healed wounds without vascularization or fibrosis/hazing. In vitro, CALR stimulated proliferation of human corneal epithelial cells (CE) and corneal stromal cells (keratocytes) by 1.5-fold and 1.4-fold, respectively and induced migration of CE cells and keratocytes, by 72% and 85% compared to controls of 44% and 59%, respectively. As a marker of decreased fibrosis, CALR treated corneal wounds showed decreased immunostaining for α-smooth muscle actin (α-SMA) by keratocytes and following CALR treatment in vitro, decreased the levels of TGF-β2 in human CE cells and α-SMA in keratocytes. CALR has the potential to be a novel therapeutic both, to accelerate corneal healing from various injuries and in conjunction with corticosteroids.

Keywords: TGF‐β; calreticulin; corneal epithelial cells; corneal injury; fibrosis; keratocytes; migration; proliferation; tissue regeneration; α‐smooth muscle cell actin.

FIGURE 1

Calreticulin accelerates corneal wound closure in a rabbit partial keratectomy wound model. (A) Anatomy of eye; enlarged corneal region. Diagram was created using Biorender.com. The avascular cornea acts as a protective structural barrier and with the tear interface provides 2/3 refractive power of the eye., , , , The lens refracts light and images for focus onto the retina. The iris regulates pupil size for the amount of light accessing the eye. The limbus forms the border of the cornea and opaque sclera and harbours limbal stem cells that replenish both the corneal epithelium and corneal stromal cells; the limbus is involved in blood vessel sprouting due to injury. The ciliary body aids in changing the shape of the lens. The aqueous humour is a nourishing clear fluid between the cornea and the lens that maintains ocular pressure. The vitreous humour maintains an oxygen gradient and shape of the eye. The choroid is vascular and provides oxygen and nutrients to the retina. The retina provides the neural signal to the optic nerve and the optic nerve transmits visual information to the brain. The cornea is composed of 4–5 layers of non‐keratinized stratified squamous epithelial cells (40–50 μm thick). The basal cell layer contacts the EBM (Bowman's layer) containing laminins, perlecan, nidogens and collagen I, III, V and VI. The stromal layer, comprising 90% of the corneal thickness contains and keratocytes that produce crystallins and collagens I, V, VI. Below the stromal layer is Descemet's membrane and a single layer of endothelial cells. Experiment #1: CALR dose ranging and efficacy in accelerating corneal wound closure. CALR (0.035 mL) was applied to the right non‐surgical and left eye injured by vacuum to the central corneal surface, BID for 14 days. Treatment groups: (1) negative control, vehicle, TBS; (2) 0.1% dexamethasone (anti‐inflammatory [Dex]); (3) CALR at 2.5 mg/mL [87.5 μg/application]; (4) CALR at 0.25 mg/mL [8.75 μg/application]; (5) CALR at 0.025 mg/mL [0.875 μg/application]; n = 5/group. Approximately 0.05 mL of 1% sodium fluorescein solution was applied to the ocular surface, followed by rinsing with phosphate buffered saline (PBS). Corneal staining was photographed under cobalt blue light using a Nikon Digital SLR camera on a tripod. The area in pixels of fluorescein staining at each time point was determined by Image J Software (NIH) of masked images. The fluorescein solution was applied two times per day (8–12 h apart), until Day 6 and once daily thereafter, and the area of corneal wound remaining open was measured as pixels [wound survival] (B) Rate of healing was measured by dividing the wound size on Day 0 by the wound size at the time of complete wound closure for each injured cornea. (C) Kaplan–Meier plot: the time to wound healing from 100% open at wound initiation is shown as per cent survival of the wound (y axis) over time (h; x axis) with the last animal to show full closure bisecting the x axis. CALR at 2.5, 0.25 and 0.025 mg/mL healed wounds by 96 h; TBS at 168 h, and Dex at 240 h. (D) Matrix Scatter Dot Plot: The time to wound closure for each injured rabbit cornea from each treatment group is shown by the Matrix Scatter Dot Plot. The mean values show accelerated wound closure in the three calreticulin treatment groups compared to TBS vehicle (p ≤ 0.01) and Dex (p ≤ 0.01). (E) The table provides the average hours to wound closure derived from the Matrix Dot Scatter plot, for the CALR treatment groups as compared to the TBS control. Data is represented as Mean ± SD; n = 5. [p ≤ 0.05 (*); p ≤ 0.01 (**)].

FIGURE 2

Calreticulin enhances wound closure of rabbit partial corneal keratectomies treated with Dexamethasone (Experiment #2). Rabbit left corneas were injured using the partial keratectomy vacuum model described in Methods and 0.035 mL CALR in PBS was applied to the left injured and right uninjured eyes one time per day for 4 days alone or with Dex applied twice per day for 7 days. Treatment groups: (1) negative control, vehicle PBS; (2) CALR at 0.0125 mg/mL (0.438 μg/application); (3) CALR at 0.025 mg/mL (0.875 μg/application); (4) 0.1% Dex; (5) Dex plus CALR at 0.0125 mg/mL (0.438 μg/application); (6) Dex plus CALR at 0.025 mg/mL (0.875 μg/application); (n = 5/treatment group). As described Figure 1, fluorescein solution was applied to the ocular surface before treatment, two times per day, until Day 6 and once daily thereafter, and the area of corneal wound remaining open was measured as mean pixel area (n = 6 animals) ± SD [wound survival]. (A) Rate of healing was measured by dividing the wound size on day 0 by the wound size at the time of complete wound closure for each injured cornea. (B) Kaplan–Meier Plot shows the time to wound healing from 100% open at wound initiation, shown as per cent survival of the wound (y axis) over time (h) with the last animal to show full closure bisecting the x axis: time of full closure: 0.025 mg/mL CALR at 120 h, PBS at 153 h, Dex at 288 h [range 152–288 h], 0.025 mg/mL CALR plus Dex at 168 h. (C) Matrix Scatter Dot plot represents the time of wound closure of each rabbit cornea from each treatment group. The mean values show time of wound closure (h) in the Dex plus CALR treated groups compared to Dex alone (not significant). (D) Calculated from the average healing value shown in Matrix Scatter Dot plot, the Table shows faster healing by CALR in combination with Dex compared to Dex‐treatment alone. One rabbit from each of three Dex‐treated groups was excluded due to lack of wound closure through 336 h post‐surgery. There was also an outlier in one DEX‐treated animal (healed at 288 h) skewing the average to be of slower healing. Data is represented as Mean ± SD; n = 5 [ns = not significant].

FIGURE 3

(A, B) Ocular Examination scores indicate that calreticulin treatment of rabbit injured corneas does not induce inflammation or fibrosis during healing. Corneal surface morphology and anterior inflammation of the cornea were evaluated using the Hackett McDonald grading system to obtain Ocular Examination (OE) scores for both the right uninjured and left surgical eyes as base line and on Days 1,3,7,14 and 17 (Exp#1) or Day 28 (Exp#2) using a Slit Lamp Biomicroscope and an indirect ophthalmoscope. Treatment groups are described in Figure 1. The highest score range for OEs is 34, which is weighted as follows, 10 attributable to the conjunctiva, 10 for the cornea, 13 for the iris, and 1 for the lens. (A) Exp #1: Total mean OE scores for the surgical eyes for each treatment group is shown over time (days). Mean OE scores for CALR 2.5 mg/mL treatment were higher from Days 3–17 than all other treatments whereas CALR 0.25 mg/mL and CALR 0.025 mg/mL treatments had mean OEs that that were similar to the vehicle control. (B) Exp #2: Treatment groups are described in Figure 2. Mean OE scores for the surgical eyes were highest on Day 1 post‐surgery for all treatment groups, which declined and showed comparable resolution of inflammation and surface irregularities for all six treatments from Day 7 to 28. (C, D) Anterior segment corneal Optical coherence tomography (OCT) indicates that calreticulin does increase corneal thickness or cause stromal fibrosis during healing. OCT was performed for all injured and corneas receiving treatments, as described in Exp 1 and Exp 2, to image the cornea for depth of lesion, overall corneal thickness and stromal fibrosis before and immediately after injury and then on Days 7 and 17 (Exp#1) and in addition, Day 28 (Experiment #2) post‐injury. Callipers, as part of the ophthalmoscope software, were used for measurements with scoring on a scale of 0 = Normal, 1 = Mild superficial area opacity, 2 = Moderate superficial corneal opacity, 3 = severe corneal opacity without full thickness involvement, 4 = Diffuse, sever, full thickness corneal opacity. (C) The bar graph shows OCT scores for fibrosis for Exp#1. Only CALR at 2.5 mg/mL gave a statistically significant score for fibrosis on Days 7 and 17 (p ≤ 0.0133, p ≤ 0.0009) post‐wounding compared to the Tris Vehicle control. (D) The bar graph shows that all treatment groups in Exp#2 gave similar fibrosis scores to the vehicle control and Dex scored on post‐surgical Days 7 and 28. (E) The bar graph shows the average for all the histology scores by staining with haematoxylin and eosin and Masson's Trichrome that are represented in Table 1 for each rabbit surgical eye in Exp#2. The compilation of the pathological scores for the 0.025 mg/mL CALR plus Dex treatment group were lower than the Dex alone group. PBS vehicle = 8.4 ± 3.4; 0.0125 mg/mL CALR = 8.2 ± 1.3; 0.025 mg/mL CALR = 5.2 ± 2.0; Dex = 4.0 ± 1.4; 0.0125 mg/mL CALR plus Dex = 4.4 ± 2.3; 0.025 mg/mL CALR plus Dex = 1.2 ± 0.8. (F) CALR 0.025 mg/mL plus Dex treatment of corneal injury shows similar histology to non‐surgical right eye after 28 days of healing from Exp#2. Images of corneal tissue stained for collagen density with Masson's Trichrome. Right non‐surgical eye normal eye, 0.1% Dex‐treated corneal injury, 0.025 mg/mL CALR plus 0.1% Dex treated injury. Data is represented as Mean ± SD; n = 5. [p ≤ 0.05 (*); p ≤ 0.01 (**)].

FIGURE 4

Calreticulin stimulates proliferation and migration of corneal epithelial cells and stromal keratocytes in vitro. (A, C, D) Corneal epithelial cells (CE) and (B, E, F) stromal keratocytes. Proliferation: CE cells and keratocytes were synchronized for 24 h in basal media in 96‐well plates in triplicate and then, treated with increasing concentrations of CALR (0–100 ng/mL) in keratinocyte basal medium (KBM) or for keratocytes in DMEM‐F12 containing 0.5% FBS respectively, for 24 h. Proliferation: (A) CE cells: Controls: KBM‐2 (negative), epidermal growth factor (EGF) at 5 and 10 ng/mL (positive). (B) keratocytes: Controls: DMEM‐F12 (negative) and DMEM‐F12 containing 5%FBS or FGF (positive). CCK‐8 solution was added and cells incubated for 2 h; absorbance was measured at 450 nm. Cellular proliferation was quantified as fold increase relative to the negative control set at 1.0. CALR stimulates proliferation with peak activities for CE cells at 50 pg/mL and keratocytes at 10 ng/mL. Migration: In vitro wound healing scratch plate assay: cell monolayers in 24‐well plates were wounded with a 200 μL pipette, detached cells removed by PBS wash, followed by treatment with increasing concentrations of CALR, as described. Mitomycin C (5 μg/mL; 1 h) was used in initial experiments to ensure that migration did not involve proliferation. (C, D) CE cells: controls: KBM‐2 basal media (negative) and KGM (positive). Cells were incubated for 8 h. (E, F) Keratocytes: controls: DMEM‐F12 with 5% FBS (positive). Cells incubated for 16 h. (C–F). For measuring migration, the cells were stained with Coomassie blue in 10% acetic acid, 45% methanol for 15 min, washed with PBS, and grey scale images captured using Olympus CK2 microscope. Wound area was determined using ImageJ macros wound healing tool. Per cent wound closure was calculated for each treatment group by measuring the area of the wound (pixels) at 0 time compared to the end of the experiment. Statistical analysis was performed using the unpaired students t‐test. Data are represented as Mean ± SEM of three independent experiments. CALR stimulates migration of CE cells and keratocytes at peak concentrations of 10 and 1.0 ng/mL, respectively. The difference in epithelial cuboidal and stromal spindle morphology between the CE cells in Figure 5C and keratocytes in Figure 5E is well‐displayed. Data is represented as Mean ± SEM; n = 3. [p ≤ 0.05 (*); p ≤ 0.01 (**); p ≤ 0.001 (***); p ≤ 0.0001 (****)].

FIGURE 5

Exogenous Calreticulin decreases critical markers for fibrosis, vimentin and TGF‐β2 in corneal epithelial cells and α‐smooth muscle actin in keratocytes in vitro, by immunoblot analysis. (A, B) Corneal epithelial cells and (C, D) corneal stromal keratocytes cultured in KBM and DMEM‐F12 containing 0.5% FBS, respectively, at 70%–80% confluency in 6‐well plates, were treated with increasing concentrations of CALR (0–100 ng/mL) for 24 h, and cell lysates prepared on ice with 1X RIPA buffer (#20188, SigmaMillipore) containing undiluted protease inhibitor cocktail (Sigma, #P8340). (A, C) Protein concentrations were determined by the Micro‐BCA protein assay kit (Pierce) and 15 μg/well in Laemmli buffer containing 5% β‐mercaptoethanol were electrophoresed by SDS‐PAGE (10% acrylamide), and then, transferred to a polyvinylidene fluoride (PVDF) membrane for immunoblotting. Membranes were blocked with 5% nonfat dry milk in tris‐buffered saline (TBS) with 0.1% Tween‐20 (blocking buffer; TBST) for 1 h followed by overnight incubation in primary antibodies at 4°C. The following primary antibodies diluted in 5% milk were used: Mouse vimentin (#sc6260, Santa Cruz Biotechnology) at 1:1000 or mouse anti‐α‐smooth muscle actin (α‐SMA; #A5228, Sigma‐Aldrich). For rabbit anti‐human TGF‐β2 peptide antibody, the membrane was blocked overnight in 5% nonfat dry milk at 4°C followed by incubation with the antibody at 5 μg/mL in TBST containing 3% of nonfat milk overnight at 4°C, as described. As loading controls for all immunoblots, anti‐β‐actin at 1:10,000 or anti‐GAPDH (#sc‐32,233, Santa Cruz Biotechnology) at 1:2000 in 5% nonfat dry milk/TBST were used blots for CE cell and keratocytes, respectively. After incubation with primary antibodies, the membranes were probed with the appropriate secondary antibody for 1.5 h; either goat anti‐mouse IgG (Invitrogen), 1:2000 in TBST in 5% nonfat dry milk or goat anti‐rabbit IgG (Invitrogen), 1:2000 in TBST in 5% nonfat dry milk. Protein detection was performed using chemiluminescence‐based SuperSignal West Femto Maximum Sensitivity Substrate (#34095, Invitrogen ThermoScientific) (B) CE cells (TGF‐β2, vimentin) (D) Keratocytes (α‐SMA): Images of blots were captured using a ChemiDoc MP Imaging System (Bio‐Rad) and protein levels for each sample in each well on the blots were determined by densitometry using ImageJ. Target protein band intensity was normalized compared to the intensity of β‐actin or GAPDH. Data is expressed as fold (y axis) change for each target protein with the untreated control assigned as one. Data is represented as Mean ± SEM; n = 3.

FIGURE 6

Calreticulin prevents the induction of fibrotic protein 𝜶‐SMA in vivo. (A) Paraffin embedded tissue sections 5.0 μm thick were placed on slides (NYU Experimental Pathology Core), sections baked overnight at 55°C, processed from xylene through graded alcohol and rehydrated. Endogenous peroxidase activity was quenched with 0.6% hydrogen peroxide for 30 min at RT followed by blocking non‐specific antigens with 3% normal goat serum in TBS containing 0.5% BSA (blocking buffer) for 20 min at RT. Primary antibodies were diluted in blocking buffer as follows: anti‐α‐SMA, 1:150, (#A5228, Sigma) and anti‐vimentin, 1:150 (#sc6260; Santa Cruz Biotechnology). The slides were incubated overnight with primary antibodies at 4°C, followed by goat‐anti‐mouse secondary antibody at 1:200 dilution for 1 h at RT, and protein expression levels detected using the VECTASTAIN® Elite® ABC‐HRP Kit; (Vector Laboratories) and DAB substrate (ab64238, Invitrogen, Abcam). Tissue sections were counterstained with haematoxylin (Modified Mayers; #ab220365, Abcam). Following graded dehydration to xylene, the slides were mounted with Permount Mounting Media (#SP15‐500, Fisher Chemical™ Co), imaged using Zeiss Axiophoton II microscope. Images of cornea at 20× and 50× digital magnification are shown for expression of 𝜶‐SMA in the rabbit eyes treated with PBS (Panel a); CALR (2.5 mg/mL; Panel b) CALR (0.025 mg/mL; Panel c); Dexamethasone (0.1%); (Panel d); Dexamethasone (0.1%) plus CALR (0.025 mg/mL); (Panel e). (B) Quantification of 𝜶SMA and (C) vimentin expression levels in rabbit corneal tissue sections. The intensity of protein expression was quantified by measuring intensity of DAB staining in each tissue section for each treatment group shown, by ImageJ. Data represents the intensity of region(s) of interest within each tissue section denoted by solid circles within each bar on the graph.