Downregulation of IL-7 and IL-7R Reduces Membrane-Type Matrix Metalloproteinase 14 in Granular Corneal Dystrophy Type 2 Keratocyte

Abstract

Purpose: Granular corneal dystrophy type 2 (GCD2) is caused by a point mutation (R124H) in the TGF-β–induced gene (TGFBI). However, the mechanisms underlying the accumulation of TGF-β–induced protein (TGFBIp) are poorly understood. Therefore, we evaluated the signaling cascade affecting the expression of TGFBIp using patient-derived cells.

Methods: Keratocyte primary cultures were prepared from corneas from the eye bank or from heterozygous or homozygous patients with GCD2 after penetrating or lamellar keratoplasty. GCD2 diagnoses were based on the results of a DNA analysis for the R124H TGFβI mutation. Keratocytes were treated with various cytokines and then analyzed using quantitative PCR (qPCR) array, qPCR, flow cytometry, ELISA, and Western blotting.

Results: TGFBI expression was counterregulated by IL-7 in corneal fibroblasts. IL-7 expression was significantly reduced in corneal fibroblasts from patients with GCD2. TGF-β and TGFBI expression were reduced on IL-7 treatment in corneal fibroblasts. Interestingly, the interplay between TGF-β and IL-7 was regulated by the RANKL/RANK signaling cascade. Also, IL-7 regulates the expression of a membrane-type matrix metalloproteinase (MT-MMP), which plays a crucial role in migration and neovascularization in the cornea.

Conclusions: These studies demonstrate that impaired IL-7 expression in patients with GCD2 affects disease pathogenesis via a failure to control TGF-β expression. The RANKL/RANK axis regulates TGF-β and TGFBI expression via IL-7–mediated MT-MMP regulation in corneal fibroblasts. These findings improve our understanding of the pathogenesis of GCD2.

Granular corneal dystrophy type 2 (GCD2) is an autosomal-dominant type stromal dystrophy caused by a point mutation (R124H) in the TGF-β–induced gene (TGFBI) on chromosome 5q31.¹ Typically, small white irregularly shaped deposits attributed to the accumulation of the TGF-β–induced protein (TGFBIp) are found in the corneal stroma, resulting in the disruption of corneal transparency and eventually a loss of vision.^2,3 Studies of GCD2 have established the importance of a TGFBI mutation in the pathogenesis; however, the mechanisms underlying the accumulation of TGFBIp and its clearance are poorly understood. In the steady state, most TGFBIp, also known as kerato-epithelin, is produced by the corneal epithelium. During the process of wound healing in the normal human cornea, TGFBIp is found in both the epithelium and keratocytes near the site of the wound, suggesting that keratocytes produce TGFBIp.^4,5 The mechanism by which wild-type (WT) or mutant (MUT) TGFBIp is degraded is highly complicated and many unresolved issues remain, including the cell types that are functional, the types of proteases that play important roles, and how amyloidogenesis occurs in a single mutated peptide.^6,7 Another interesting point is that the cornea is the only tissue affected by this disease. Other organs (or tissues) show no clear functional or morphological abnormalities during the whole life span of humans, despite universal expression of the MUT form of TGFBIp. 

IL-7 was initially identified as a factor required for the growth of murine B-cell precursors⁸; however, subsequent studies have shown that IL-7 plays an important role in T cells, dendritic cells, and fibroblasts.^9–11 The role of IL-7 in rheumatoid arthritis (RA) has been studied owing to its elevated levels in the serum of patients with RA and its increased expression in the RA synovium, synovial fibroblasts, and chondrocytes.^11,12 IL-7 is now known to be produced by many types of fibroblasts, not only in immune tissues, such as the thymus, bone marrow, and lymph nodes, but also in nonimmune tissues, like the skin,^13,14 intestine,¹⁵ and liver.¹⁵ With respect to eye diseases, IL-7 has been found in the pseudophakic vitreous¹⁶ and tear fluids¹⁷ from thyroid ophthalmopathy.¹⁸ Interestingly, IL-7 inhibits fibroblast TGF-β production and tissue fibrosis.^19,20 Thus, it is possible that IL-7 contributes to the pathogenesis of GCD2 by affecting TGF-β and subsequent TGFBIp expression in corneal fibroblasts (i.e., keratocytes). Despite the relatively high level of TGF-β expression in ocular tissues, especially the cornea, which maintains immune privilege, to the best of our knowledge, the relationship between corneal fibroblasts and IL-7 expression has not been evaluated in any corneal pathologies. 

In this study, we evaluated the expression and regulation of IL-7 and its corresponding receptors in the cornea as well as their functional roles in TGF-β–mediated TGFBIp accumulation and clearance in a GCD type 2 disease model using immortalized human corneal fibroblasts from healthy subjects and patients with GCD type 2. 

Materials and Methods

Antibodies and Other Reagents

An FITC-conjugated mouse anti-human CD127 (IL-7Ra) antibody was obtained from BioLegend (San Diego, CA, USA). The PE-conjugated mouse anti-human/mouse CD265 (RANK) antibody was obtained from Thermo Fisher (Rockford, IL, USA). Human TGF-β1 ELISA Duo Set and Human TGFBIp ELISA Duo Set were obtained from R&D Systems (Minneapolis, MN, USA). Human RANKL (receptor activator of nuclear factor kappa B ligand) was donated by Eun Ju Chang, PhD (Asan Bio Institute, Seoul, Korea). Human IL-7, human TGF-β1, and human TGFβI protein were obtained from R&D Systems. 

Isolation of Corneal Fibroblasts From Healthy and Granular Dystrophy Type 2 Patients

In accordance with the tenets of the Declaration of Helsinki and with the permission of the Institutional Review Board (No. 4-2013-0472), human corneal tissue samples were harvested from recipient corneal buttons following keratoplasty as described before.²¹ Briefly, three kinds of recipient corneal buttons were obtained. Five normal corneas were obtained from human donors younger than 20 years. A corneal button was also obtained from a 37-year-old man with a heterozygotic R124H mutation who underwent penetrating keratoplasty. Homozygotic corneal buttons were obtained from a 6-year-old boy and 14-year-old girl after lamellar keratoplasty. The obtained corneas were washed three times with Dulbecco's modified Eagle's medium (DMEM) containing 50 g/mL gentamicin and 1.25 μg/mL amphotericin B. Then, the corneal epithelium was removed by scraping the epithelial surface with a #15 Bard Parker blade. The endothelium was removed and the Descemet's membrane was detached with fine forceps and a blade. After the removal of the epithelium and Descemet's membrane, the washing procedure was repeated three times to remove contaminated epithelium and endothelium. Then, the corneal tissues were placed on a 30-mm culture dish coated with 100 μL of fetal bovine serum (FBS) for 0.5 hour in a 37°C incubator to induce tissue attachment on the dish. Then, 10% FBS-supplemented DMEM was added gently to the dish. After 48 hours, migratory fibroblasts from the corneal button could be found. The corneal fibroblasts (CFs) were maintained at 37°C under 95% humidity and 5% CO₂. The medium was changed every other day, and cell outgrowth was monitored daily for 3 weeks by inverted phase microscopy (IX70; Olympus, Tokyo, Japan). When the cultured corneal epithelium was approximately 80% to 90% confluent, cells were subcultured with 0.25% trypsin and 5.0 mM EDTA at a 1:3 ratio. 

Culture of Corneal Fibroblasts From Healthy Subjects and Patients With Granular Corneal Dystrophy Type 2

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Severance Hospital Institutional Review Board (IRB-2006-0139). Keratocyte primary cultures were prepared from corneas from the eye bank or from heterozygous or homozygous patients with GCD2 after penetrating or lamellar keratoplasty. GCD2 diagnoses were based on the results of a DNA analysis for the R124H TGFβI mutation. Stromal explants were prepared by first removing the epithelium and endothelium and then culturing the explants in 10% FBS/ DMEM at 37°C in 5% CO₂ in six-well tissue culture plates. Corneal fibroblasts migrated from the explants along the bottom of the plates. Cells were confluent within 15 to 21 days. Cells were harvested by enzymatic detachment using 0.05% trypsin at 37°C for 3 minutes, followed by centrifugation at 280g for 5 minutes and aspiration of the supernatant. The cell pellet was then resuspended in 20 mL of the medium and cultured in 75-mL flasks at 37°C in 5% CO₂ until confluent. Cells were then serially trypsinized and passaged three to five times before use in experiments. Cells were plated at 3 × 10³ to 5 × 10³ cells per well in 96-well tissue culture plates and incubated in 1 mL of 10% FBS/DMEM at 37°C in 5% CO₂ for 24 to 48 hours. 

Treatment of Human Keratocytes

Keratocytes were used between passages 3 and 5. Keratocyte suspensions (5 × 10⁵ cells/mL in DMEM) were transferred to six-well plates and incubated at 37°C with DMEM containing 10% FBS overnight. Keratocytes were separately treated with TGF-β1 (10 ng/mL), TGFβIp (20 ng/mL), RANKL (50 ng/mL), and increasing concentrations (0, 1, and 100 ng/mL) of IL-7 for 24 hours for quantitative RT-PCR and 2 days for ELISA and fluorescence-activated cell sorting (FACS). The same amount of vehicle was added to the control keratocytes. The culture media were collected for ELISA. 

RNA Isolation, Quantitative RT-PCR, and Quantitative PCR Array

Total RNA was extracted from rat corneas using QIAzol reagent (QIAGEN, Germantown, MD, USA) according to the manufacturer's instructions and reverse transcribed to cDNA using a cDNA Synthesis Kit (Prime Script RT Master Mix; Takara, Shiga, Japan). Quantitative real-time PCR was performed using a PCR detection system (StepOnePlus Real-Time PCR; Applied Biosystems, Carlsbad, CA, USA) and a commercial detection kit (SYBR Premix Ex Taq Kit; Takara Bio Co., Dalian, China) according to the manufacturer's instructions. The amplification program included an initial denaturation step at 95°C for 30 seconds, followed by 40 cycles of denaturation at 95°C for 5 seconds and annealing and extension at 60°C for 30 seconds. Fluorescence signals were measured after each extension step and the specificity of the amplification was evaluated by a melting curve analysis. The primer sequences used in this study are summarized in the Table. The quantitative PCR (qPCR) results were analyzed by the comparative Ct method and normalized against the levels of GAPDH. RT²profiler PCR array human cytokine and chemokines were purchased from SA Biosciences (QIAGEN). The QIAGEN online Web analysis tool was used to calculate fold changes in gene expression. 

Table

 

qPCR Primers Used in the Study

 

Analysis of Surface Markers by Flow Cytometry

Cell surface antigens were analyzed by flow cytometry. The cells were incubated with anti-human CD127 (IL-7Rα)-APC (Catalog No. 351316, clone A019D5, isotype, mouse IgG1; BioLegend) or anti-human CD265 (RANK)-PE antibody (Catalog No. FAB683P, Clone #80704, isotype, mouse IgG1; R&D Systems) for 30 minutes at 4°C. The cells were then washed twice with 0.5% BSA-PBS and analyzed using the FACS CANTO II (Becton Dickinson, San Jose, CA, USA). 

ELISA of TGF-β1 and TGFβI in the Culture Supernatant

The expression levels of TGF-β1 and TGFβI in both WT and MUT keratocytes were assessed by ELISA (R&D Systems), following the manufacturer's instructions. Standard curves were obtained for a reference range of 31.30 to 2000 pg/mL for TGF-β1 and 31.25 to 4000 pg/mL for TGFβIp. 

Scratch Wound Healing Assay

A scratch wound healing assay was performed by making a scratch wound through the center of confluent cell monolayer in each well using a 1-mL pipette tip. Debris was removed from the wound and the edges were smoothed by rinsing with PBS. The cells were then incubated with 10% FBS/DMEM at 37°C. To assess wound healing, digital images were obtained at 0 and 18 hours. 

Knockdown of IL-7 or IL-7R in Keratocytes

Lentiviral particles containing small hairpin RNA (shRNA) targeting human IL-7 (sc39629-V), IL-7Rα (sc-35664-V), or control shRNA lentiviral particles (sc-108080) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Keratocytes were cultured in a 24-well plate (1 × 10⁵ cells/well) and infected with each lentiviral particle according to the manufacturer's protocol. Control keratocytes generated using control shRNA lentiviral particles, IL-7 knockdown (KD) keratocytes generated using IL-7 shRNA lentiviral particles (IL-7^KD keratocytes), and IL-7Rα KD keratocytes generated using IL-7Rα shRNA lentiviral particles (IL-7Rα^KD keratocytes) were cultured in puromycin-containing (4 to 6 μg/mL) medium for 2 weeks to select stable clones. Puromycin-resistant colonies were picked. The expression of IL-7 was examined with an anti-IL-7 antibody by Western blotting and the expression of IL-7Rα was examined with an anti-IL-7Rα-APC antibody by FACS. 

Preparation of Cell Lysates and Western Blot Analysis

Keratocytes were lysed in lysis buffer containing 150 mM NaCl, 20 mM Tris-HCl, pH 7.5, 10 mM EDTA, 1% Triton X-100, 1% deoxycholate, 1.5% aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Cellular debris was removed by centrifugation. Cell lysates were resolved by SDS-PAGE and electroblotted onto polyvinylidene difluoride membranes. After blocking with 5% skim milk in TBS-T (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 0.05% Tween 20), the membranes were probed with antibodies. For Western blotting, anti-TGFBI and anti-GAPDH antibodies were used. The proteins were detected using the Enhanced Chemiluminescence System (Thermo Scientific, Waltham, MA, USA). 

Statistical Analysis

The generation of graphs and statistical analyses were performed using GraphPad Prism (GraphPad, La Jolla, CA, USA) or Microsoft Excel 2010 (Microsoft, Redmond, WA, USA). Results are expressed as means ± SEM. Statistical comparisons between groups were performed using independent t-tests, and P values were adjusted by Bonferroni correction. P < 0.05 was considered statistically significant. 

Results

Basal Expression Levels of IL-7, IL-7R, TGF-β1, and TGFβI in WT and MUT Keratocytes

TGF-β1 and TGFβI are highly expressed in keratocytes during wound healing; accordingly, we hypothesized that immune factors, such as interleukins, interferons, growth factors, and chemokines, would be expressed in keratocytes. Furthermore, factors that differ in expression between keratocytes from patients with GCD2 and normal keratocytes may play an important role in the pathogenesis of GCD2. To identify differential expression, we used a PCR array with a variety of immune factors to evaluate WT and MUT (GCD2) keratocytes. Interestingly, many factors were differentially expressed in MUT keratocytes (Fig. 1A; Supplementary Table S1). In particular, IL-7 exhibited lower expression in MUT than in WT keratocytes (P < 0.001; Fig. 1B). In addition, IL-7R expression was reduced in MUT keratocytes (Fig. 1C). As expected, TGF-β1 and TGFβIp levels were higher in MUT than in WT keratocytes (P < 0.001; Figs. 1D, 1E). We confirmed that there were more IL-7R–positive cells for WT compared with MUT keratocytes using a FACS analysis (Fig. 1F; Supplementary Fig. S1). Because IL-7 expression may be regulated in an autocrine manner, we examined IL-7 and IL-7R expression in keratocytes over time. In particular, we treated WT and MUT keratocytes with 10 ng/mL recombinant IL-7. We found that IL-7 expression increased 18 hours after treatment (Fig. 1G). IL-7R expression increased slightly 3 hours after treatment. (Fig. 1H). Notably, IL-7 and IL-7R expression levels were very limited in MUT keratocytes, regardless of conditions, indicating that IL-7 regulation plays a crucial role in the pathogenesis of GCD2 (Figs. 1G, 1H). 

Figure 1

 

Expression of IL-7, IL-7R, TGF-β1, and TGFBIp in human keratocytes (WT, MUT, GCD type 2 keratocytes). (A) Gene expression profile of GCD2 corneal keratocytes. Comparison between MUT and WT keratocytes. Total RNAs were isolated from each sample and a PCR array was used as described in the Materials and Methods. (B–E) Total RNAs were isolated from each sample and a quantitative RT-PCR analysis was performed with specific primers targeting the IL-7, IL-7R, TGF-β, and TGFBI genes. Means ± SEM of mRNA levels of (B) IL-7, (C) IL-7R, (D) TGF-β, and (E) TGFβIp were compared by independent t-tests. Real-time PCR revealed significant differences in gene expression levels between WT and MUT keratocytes. Independent t-test: *P < 0.05, **P < 0.01, ***P < 0.001. Error bars indicate the SE. (F) Cell surface expression levels of IL-7R were detected by FACS analysis. (G, H) WT and MUT keratocytes were treated with or without rhIL-7 (10 ng/mL) for various time periods. (G) IL-7 and (H) IL-7R mRNA expression levels were determined by quantitative RT-PCR. **P < 0.01 after Bonferroni correction. Results are presented as mean (bars) ± SEM (error bars) of three independent experiments.

Reciprocal Inhibition of TGF-β1 and TGFβI Expression by IL-7

To investigate the relationships among IL-7, TGF-β1, and TGFBI, we treated keratocytes with TGF-β1 and measured IL-7 expression by qPCR, and vice versa. Interestingly, IL-7 expression was reduced by treatment with recombinant TGF-β1 (Fig. 2A). Moreover, TGF-β1 and TGFBI expression levels were also reduced by treatment with recombinant IL-7 (Figs. 2B, 2C). Next, we examined whether the reciprocal inhibition between IL-7 and TGF-β1/TGFBI also occurred at the protein level. Both ELISA (Figs. 2D–F) and immunoblotting (Fig. 2G) showed that IL-7 treatment decreased TGF-β1 expression. As expected, TGFBI was highly expressed on TGF-β1 treatment in MUT keratocytes (Fig. 2G). 

Figure 2

 

IL-7 counterregulates TGF-β1 in TGFBIp induction. (A–C) WT keratocytes were treated with (A) recombinant human (rh)TGF-β1 (10 ng/mL) or (B, C) rhIL-7 (10 ng/mL) for 18 hours. Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (A) IL-7, (B) TGF-β, and (C) TGFBI. The results are expressed as means ± SEM for five independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001. (D–F) WT keratocytes were treated with (D) rhTGF-β1 (10 ng/mL) or (E, F) rhIL-7 (10 ng/mL) for 24 hours. Culture supernatants were collected and then levels of (D) IL-7, (E) TGF-β1, and (F) TGFBIp were measured, as described in the Materials and Methods. The results are expressed as means ± SEM for five independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001. (G) WT and MUT keratocytes were treated with rhTGF-β1 (10 ng/mL) for 24 hours. Western blotting of cell lysates from each sample probed with an anti-TGFBI antibody. (H) WT keratocytes were treated with rhIL-7 (10 ng/mL) or rhTGF-β1 (10 ng/mL) for 24 hours. Cell lysates were collected for immunoblotting using an anti-TGFBI antibody.

TGF-β1 and TGFβI Expression in IL-7 and IL-7R KD Keratocytes

To determine whether the reciprocal inhibition between IL-7 and TGF-β1/TGFBI is regulated in an intrinsic manner, we generated IL-7 and IL-7R KD keratocytes (Supplementary Fig. S2). IL-7 expression was abolished in IL-7–deficient keratocytes, even on TGF-β1 treatment (Fig. 3A). Interestingly, TGF-β1 expression was slightly, but significantly higher in IL-7–deficient cells than in control cells, even in the steady state (Fig. 3B). TGFBI expression increased more highly in IL-7–deficient cells on TGF-β1 treatment (Fig. 3C). Similarly, IL-7, TGF-β1, and TGFBI expression levels were increased in IL-7R–deficient cells (Figs. 3D–F). The protein level of TGFBI was also increased in IL-7 and IL-7R–deficient cells, both in the cell lysate and the supernatant (Fig. 3G). 

Figure 3

 

IL-7 counterregulates TGF-β1 in TGFBIp induction. Upregulation of TGFBIp by TGF-β in IL-7^KD and IL-7R^KD keratocytes. Control keratocytes were generated using control shRNA lentiviral particles, IL-7 KD keratocytes were generated using IL-7 shRNA lentiviral particles (IL-7^KD keratocytes), and IL-7Rα KD keratocytes were generated using IL-7Rα shRNA lentiviral particles (IL-7Rα^KD keratocytes) and were cultured in puromycin-containing (4 to 6 μg/mL) medium for 2 weeks to select stable clones. (A–C) Control and IL-7^KD keratocytes were treated with or without rhTGF-β1 (10 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting IL-7, TGF-β, and TGFBI. The results are expressed as means ± SEM for three independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001. (D–F) Control and IL-7R^KD keratocytes were treated with or without rhTGF-β1 (10 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (D) IL-7, (E) TGF-β, and (F) TGFBI. The results are expressed as means ± SEM for three independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001. (G) Control, IL-7^KD, and IL-7R^KD keratocytes were treated with or without rhTGF-β1 (10 ng/mL). Whole cell lysates and cell culture supernatants were collected for immunoblotting using an anti-TGFBI antibody.

Interplay Between IL-7 and the RANK/RANKL Axis

Next, we determined the signaling cascade that regulates IL-7 in keratocytes. We focused on the fact that the expression of osteoprotegerin (OPG, also known as TNF receptor superfamily member 11B) was higher in MUT than in WT keratocytes (Fig. 4A; Supplementary Table S1). Because OPG is a decoy receptor for RANKL, OPG prevents RANK-mediated nuclear factor kappaB signaling, which is one of the central signaling cascades leading to the activation of many immune-related genes. Thus, we evaluated the expression of RANKL and RANK. As expected, RANKL expression was higher in MUT keratocytes than in WT keratocytes (Fig. 4B). Conversely, RANK expression was lower in MUT keratocytes, as confirmed by qPCR (Fig. 4C) and FACS analyses (Fig. 4D; Supplementary Fig. S3). To evaluate the relationships among IL-7, RANK, RANKL, and OPG, we measured the expression of RANK, RANKL, and OPG on TGF-β1 treatment in IL-7 and IL-7R–deficient cells. OPG expression increased in response to TGF-β1 treatment in IL-7 and IL-7R–deficient cells (Fig. 4E). RANK expression was reduced in response to TGF-β1 treatment in IL-7 and IL-7R–deficient cells (Fig. 4F). Finally, RANKL expression was reduced in response to TGF-β1 treatment in IL-7–deficient cells, but not in IL-7R–deficient cells (Fig. 4G), indicating that RANK signaling, along with IL-7, plays a crucial role in the pathogenesis of GCD2. 

Figure 4

 

IL-7 induces the expression of RANKL in keratocytes. (A–C) Total RNAs were isolated from WT and MUT keratocytes. Quantitative RT-PCR was performed with specific primers targeting (A) OPG, (B) RANKL, and (C) RANK. (D) Cell surface expression levels of RANK were detected by a FACS analysis. (E–G) Control, IL-7^KD, and IL-7R^KD keratocytes were treated with or without rhTGF-β1 (10 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (E) OPG, (F) RANKL, and (G) RANK. Means ± SEM of the mRNA levels of (A, E) OPG, (B, F) RANKL, and (C, G) RANK were compared by five independent experiments. t-tests. *P < 0.05, **P < 0.01, ***P < 0.001.

Correlation Between IL-7 and MMP Expression

After corneal injury, activated keratocytes are involved in extracellular matrix production and the regulation of inflammation, but also promote the functional properties of vascular endothelial cells by paracrine mechanisms via metalloproteinases and other angiogenic factors in corneal neovascularization. Thus, we determined the expression levels of MMP family members on IL-7 treatment in WT and MUT keratocytes. The levels of MMP3, MMP9, MMP14, and MMP15 were reduced in MUT keratocytes (Figs. 5A–E). The reduction in MMP14 in response to IL-7 and TGF-β1 was further confirmed by FACS (Fig. 5F; Supplementary Fig. S4). To confirm the functional relationship between MMP and IL-7, first evaluated cell migration using WT and MUT keratocytes. MUT keratocytes showed a defect in migration compared with that of WT cells (Fig. 5G; top) and, importantly, IL-7 and IL-7R–deficient cells recapitulated the migration defect in a comparison with control cells (Fig. 5G; bottom). 

Figure 5

 

IL-7 regulates cell migration and TM-MMP expression in keratocytes. (A–E) WT and MUT keratocytes were treated with or without rhIL-7 (10 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (A) MMP3, (B) MMP9, (C) MMP14, (D) MMP15, and (E) MMP16. The results are expressed as means ± SEM for five independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001. (F) WT and MUT keratocytes were treated with rhIL-7 (10 ng/mL) or rhTGF-β1 (10 ng/mL). Flow cytometric analysis of each single cell suspension showing the expression of MMP14. (G) A scratch wound healing assay was performed. When cells reached 70% confluency, a scratch wound was created in each well using a sterilized plastic tip. After the incubation of cells, wound healing was recorded by obtaining digital images at various time points.

Correlation Between RANKL and MMP Expression

To confirm the relationship between RANKL and MMP expression, we measured the expression of MMP family members on RANKL treatment in WT and MUT keratocytes. All tested MMP genes (MMP3, MMP9, MMP14, MMP15, and MMP16) were induced on RANKL treatment in WT keratocytes. Importantly, the increases in MMP levels were very limited in MUT keratocytes (Figs. 6A–E). We further used OPG and OPG/RANKL to treat WT and MUT keratocytes and evaluated the expression of MMP14, MMP15, and MMP16. Although MMP14 expression did not change substantially in response to OPG and OPG/RANKL treatment (Fig. 6F), MMP15 and MMP16 expression levels were clearly reduced in all conditions (RANKL, OPG, and RANKL/OPG) in MUT keratocytes (Figs. 6G, 6H). These results were recapitulated in RANKL-treated IL-7 and IL-7R–deficient keratocytes (Figs. 6I–K). 

Figure 6

 

RANKL-mediated TM-MMP induction is regulated by IL-7. (A–E) WT and MUT keratocytes were treated with or without rhRANKL (50 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (A) MMP3, (B) MMP9, (C) MMP14, (D) MMP15, and (E) MMP16. (F–H) WT and MUT keratocytes were treated with rhRANKL (50 ng/mL) and/or OPG. Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (F) MMP14, (G) MMP15, and (H) MMP16. (I–K) Control, IL-7^KD, and IL-7R^KD keratocytes were treated with or without RANKL (50 ng/mL). Total RNAs were isolated from each sample and quantitative RT-PCR was performed with specific primers targeting (I) MMP14, (J) MMP15, and (K) MMP16. The results are expressed as means ± SEM for three independent experiments. t-test: *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

Our study makes several novel contributions. First, we showed that IL-7 is highly expressed in keratocytes without stimulation in an autocrine manner and is a functionally important cytokine for maintaining the naïve status of keratocytes. Second, based on comparative analyses of GCD2 keratocytes, we found that IL-7 counterregulated TGF-β–mediated TGFBIp induction and eventually upregulated the MT-MMP MMP14, but not secretory-type MMPs. Last, the upstream RANKL/RANK axis maintained the IL-7 expression in WT keratocytes. However, in GCD2 keratocyte, RANK, but not RANKL, was significantly downregulated, and this was related to the reduction in MMP14. In addition, MUT keratocytes showed low OPG levels. It is quite interesting that RANKL/RANK/OPG was highly expressed in osteoclasts, osteoblasts, T cells, B cells, and monocyte-macrophages.²² In addition, IL-7 is primarily produced by fibroblastic reticular cells found in lymph nodes and by nonhematopoietic stromal cells present in other tissues, including the skin, intestine, and liver. It is an essential survival factor for lymphocytes, playing a key antiapoptotic role in T-cell development and mediating peripheral T-cell maintenance and proliferation.^9,23,24 Taken together, corneal keratocytes are not likely to be simply involved in the maintenance of structural integrity, but are likely be immune-relevant stromal cells that regulate corneal immune functions as well as matrix protein homeostasis. 

IL-7 System and Keratocyte Functions

We initially performed a microarray analysis to identify expression differences between WT and MUT keratocytes from patients with GCD type 2 and found that IL-7 and IL-7R were significantly downregulated. Then, we examined the effect of IL-7 reductions on GCD2 keratocyte functioning and the mechanism underlying the IL-7 reduction. The reductions in IL-7 and IL-7 receptor were more severe in homozygous than in heterozygous patients with mutations in TGFBIp (data not shown). These results suggest that the loss of TGFBIp function affects the expression and activation of the IL-7 system. 

It is well known that the IL-7 receptor is a heterodimeric complex of the common cytokine-receptor gamma chain (IL-2RG, CD132, or γc) and the IL-7R α chain (IL7R, IL7RA, and CD127). Both chains are members of the type 1 cytokine family. In a comparison with WT keratocytes, we confirmed the complete lack of IL-7R expression in homozygotic MUT keratocytes by immunoblotting and FACS. The IL-7 system has mainly been investigated with respect to its roles in immunity and infectious diseases; IL-7R is downregulated by viral infectious disease,^25,26 IL-15,²⁴ and glucocorticoid use in lymphocytes.²⁷ Our results provide the first evidence for the downregulation of IL-7R expression in nonimmune and noninflammatory diseases. 

Recently, Ouyang et al.²⁸ found that IL-7R expression is upregulated by enhanced TGF-β signaling in CD4 T cells. They found that TGF-β signaling is required to develop high IL-7Rα expression for developing CD4+ T cells with low T-cell receptor levels in the thymus. These findings are not consistent with our results showing that IL-7R is downregulated in GCD2 keratocytes with high levels of TGF-β. However, the functional role of TGF-β differs among cell types, especially between immune cells and tissue fibroblasts. In fibroblasts, TGF-β usually enhances cell proliferation, migration, and matrix protein production. However, in immune cells, TGF-β typically reduces activity levels and proliferation. With respect to IL-7R in fibroblasts, Huang et al.¹⁹ found that IL-7 inhibits TGF-β production in pulmonary fibroblasts, and interferon-γ had no effect. Zhang et al.²⁰ reported that TGF-β–mediated collagen synthesis is effectively regulated by IL-7 in fibroblasts. These two previous studies support our results demonstrating the inhibition of TGF-β and accumulation of matrix protein, like TGFBIp, in corneal fibroblasts treated with IL-7. However, in our study, another IL-7 receptor, γc (IL-2RG, CD132) did not differ between WT and MUT cells (data not shown). 

MMP Levels in Normal and GCD2 Keratocytes

Despite extensive studies, the exact proteolytic enzymes, such as MMPs, for TGFBIp remain unidentified. However, our previous results suggest that proteolytic enzymes associated with extracellular matrix (ECM) turnover may be involved in TGFBIp deposition in GCD2.²⁹ In addition, Akhtar et al.³⁰ reported that altered ECM proteolytic enzyme activities affect TGFBIp deposition by degrading ECM molecules, either by the scission of covalent bonds or the cleavage of MUT TGFBIp. Korvatska et al.³ suggested that abnormal proteolysis is involved in the deposition of TGFBIp in the cornea in cases of corneal dystrophy. Accordingly, we previously investigated the expression patterns of collagen metabolism-related MMP-1 and -2, the prominent MMPs in corneal tissues. However, we found higher levels of MMP-1 and -2 in heterozygous and homozygous GCD2 than in WT keratocytes, suggesting that MMP-1 and -2 are not critical for the degradation on TGFBIp. Surprisingly, we found that MMP14, a membrane-type MMP, rather than secretory MMPs, was significantly reduced in GCD2 keratocytes. 

MT-MMPs, including MMP14 (MT1-MMP), are able to directly degrade various ECM components, including collagens, gelatin, fibronectin, vitronectin, and laminin, and directly activate other MMPs.^31–34 Accordingly, many studies have shown that the loss of MT-MMPs leads to a significant disturbance of connective tissue metabolism.^31,33 We also found that MT-MMP mRNAs are more highly expressed in the cornea, the target location for TGFBIp accumulation in disease development, than in other tissues (e.g., the liver or muscle) (data not shown). These results also imply that MT-MMPs have more important roles in the homeostasis of corneal matrix proteins than in other organs or tissues. Therefore, the elevated TGF-β activity with decreased MMP14 in GCD2 may facilitate the accumulation of ECM proteins in the cornea, including TGFBIp, and this might be the core mechanism underlying TGFBIp accumulation in GCD2 corneas. In addition to MMP14, we found that MMP15 and MMP16 were also reduced in CF and important for the degradation of TGFBIp. All three MT-MMPs, MMP14, 15, and 16, were reduced in CF from patients with GCD2. In addition, IL-7 stimulates MMP13 expression via the activation of the receptor for advanced glycation end products (RAGE) in chondrocytes.¹² Therefore, IL-7 may play an important role in the expression of TM-MMPs, not only in corneal tissues, but also in tissues where ECM accumulates. 

Another interesting finding of this study is that the RANKL/RANK axis is an upstream pathway for the activation of IL-7R signaling, inhibition of TGF-β, and TM-MMP expression to maintain ECM homeostasis in the cornea. RANKL/RANK is an important signaling pathway in osteoclasts, which functions in the absorption of the bone tissue. In addition to osteoclastogenic activity, extensive investigations have revealed that the RANKL/RANK axis has key regulatory functions in bone homeostasis, organogenesis, immune tolerance, and cancer metastasis.²² In the eye, RANKL, OPG, and RANK were detected, and these factors were upregulated in corneal stromal cells by epithelial injury³⁵; however, no studies have evaluated RANKL expression levels and ocular pathology. We found that RANKL is a key regulator of TM-MMP expression and ECM remodeling by reducing TGFBIp via IL-7 activation in keratocytes using a GCD2 model. Interestingly, GCD2 keratocytes expressed significantly reduced levels of RANK, thereby reducing IL-7 activation as well as MMP14 expression. Interestingly, as described previously, the RANKL/RANK axis is key element in osteoclasts and could directly induce osteoclastogenesis, which requires high MMP levels to break down bone tissues. Consistent with these findings and the results of Wilson et al.,³⁵ RANKL/RANK-activated keratocytes share similar characteristics with those of osteoclasts that produce MMPs and play an essential role in the breakdown of ECM proteins, including TGFBIp. Therefore, it is quite reasonable that GCD2 cells, which exhibit reduced RANK and IL-7 activity, show TGFBIp accumulation; eventually, the cornea turns opaque and patients lose their vision. 

This study had several limitations. First, we were unable to perform in vivo or human studies. Recently, Yamazoe et al.³⁶ reported a transgenic mouse model with the R124H mutation of human TGFBIp and found corneal opacities. Despite differences between mice and humans, this mouse model may be useful for verifying the results of our study. Second, we could not provide indirect evidence for MMP14-mediated TGFBIp digestion. As MMP14 is an MT-MMP, we could not measure proteolytic activity for TGFBIp in vitro. To maintain the biological activity of MMP14, we used an MMP14 inhibitor to measure the breakdown of TGFBIp using in vitro keratocyte culture. The full-length MMP14 peptide, which shows sufficient biological activity, may resolve the issue of MMP14-mediated TGFBIp digestion. 

In conclusion, keratocytes expressed RANKL/RANK as well as IL-7R, similar to immune cells. These findings suggest that keratocytes are not simple fibroblasts, but a type of immune cell with multifunctional roles in the maintenance of homeostasis in the cornea. In addition, we identified the MMP subtypes involved in the degradation of TGFBIp (i.e., TM-MMP, and not secretory-type MMPs). These results provide a basis for studies of the breakdown of other amyloid or hyaline-form proteins that accumulate in various diseases affecting the brain, kidney, and other major organs. 

Acknowledgments

Supported by Science Research Program Grant No. NRF-2018R1A2B3001110 and Grant No. NRF-2017M3A7B4041798 through the National Research Foundation of Korea. 

Disclosure: S.Y. Kim, None; A. Yeo, None; H. Noh, None; Y.W. Ji, None; J.S. Song, None; H.C. Kim, None; L.K. Kim, None; H.K. Lee, None