Inhibition by the Antimicrobial Peptide LL37 of Lipopolysaccharide-Induced Innate Immune Responses in Human Corneal Fibroblasts

Abstract

Purpose: The synthesis of cytokines and adhesion molecules by corneal fibroblasts contributes to the innate immune response to corneal infection. The effects of the antimicrobial peptide LL37 on cytokine and adhesion molecule expression induced by bacterial lipopolysaccharide (LPS) in human corneal fibroblasts were examined.

Methods: The release of the cytokines IL-6 and IL-8 into culture supernatants and the expression of intercellular adhesion molecule (ICAM)-1 at the cell surface were measured with ELISAs and by flow cytometry. The abundance of mRNAs was quantitated by RT and real-time PCR analysis, and the phosphorylation of signaling proteins was examined by immunoblot analysis. The subcellular localization of ICAM-1 and the transcription factor nuclear factor (NF)-κB was determined by immunofluorescence analysis. Neutrophil infiltration in a mouse model of LPS-induced keratitis was evaluated by immunohistofluorescence analysis.

Results: The antimicrobial peptide LL37 inhibited the up-regulation of IL-6, IL-8, and ICAM-1 both at protein and mRNA levels in corneal fibroblasts induced by LPS without affecting those elicited by TNF-α. Furthermore, LL37 attenuated the LPS-induced phosphorylation of the NF-κB inhibitor IκBα and the mitogen-activated protein kinases extracellular signal-regulated kinase, p38, and c-Jun NH₂-terminal kinase, as well as the translocation of NF-κB to the nucleus in corneal fibroblasts. Lipopolysaccharide-induced keratitis in mice was also suppressed by topical application of LL37.

Conclusions: The inhibition of LPS-induced cytokine and adhesion molecule expression in human corneal fibroblasts by LL37 suggests that this peptide might promote the resolution of corneal inflammation associated with bacterial infection.

The cornea serves as a barrier to protect the eye against external agents including microbial pathogens and antigens. Whereas corneal epithelial cells are not responsive to bacterial components, corneal fibroblasts (activated keratocytes) recognize and react to the presence of various bacterial constituents as part of the innate immune response to protect the host from infectious organisms.^1,2 Corneal fibroblasts thus express Toll-like receptors (TLRs) and synthesize various chemokines and adhesion molecules in response to stimulation with TLR ligands such as bacterial lipopolysaccharide (LPS),³ and they are thought to function as sentinels in defense of the cornea against various external pathogens. Corneal fibroblasts also contribute to maintenance of the normal structure of the corneal stroma through the synthesis and degradation of stromal collagen. Corneal fibroblasts mediate the degradation of collagen by releasing matrix metalloproteinases, and such collagen degradation can become excessive through interaction of corneal fibroblasts with infiltrated neutrophils and bacteria during infectious keratitis.^4,5 The uncontrolled and prolonged activation of corneal fibroblasts by pathogens or inflammatory cells may result in destruction of the corneal stroma and corneal scarring. The regulation of corneal fibroblast activation is thus important to minimize corneal scarring during treatment of infectious keratitis. Although corticosteroids are potent immunosuppressants and inhibit collagen degradation by corneal fibroblasts,⁶ they may also promote the proliferation of infecting pathogens. There are currently no drugs available that are able to restrain an excessive inflammatory response without adverse effects in the treatment of infectious keratitis. 

Among the soluble factors secreted by cells into tear fluid, antimicrobial peptides play a particularly important role in host defense and innate immunity. Corneal and conjunctival epithelial cells produce β-defensins and the cathelicidin LL37,^7,8 the latter of which is one of several mature cathelicidin peptides derived from the 18-kDa proprotein hCAP18 by kallikrein-mediated cleavage. The antimicrobial peptide LL37 possesses broad antimicrobial activity and is able to kill bacteria, viruses, and fungi. In addition to its antimicrobial activity, however, LL37 is able to modulate inflammation by inducing the expression of proinflammatory cytokines and chemokines—such as IL-6 and IL-8, respectively—in monocytes, mast cells, and epithelial cells.⁹ In the cornea, LL37 stimulates the expression of cytokines in, as well as the migration of, epithelial cells.^10,11 However, the effects of LL37 on corneal stromal fibroblasts have remained unknown. 

To examine the possible role of LL37 in inflammatory reactions mediated by the innate immune system in the cornea and to explore possible new treatment options for individuals with bacterial keratitis, we now investigated the effects of LL37 on LPS-induced inflammatory responses in human corneal fibroblasts, as well as in a mouse model of LPS-induced keratitis. 

Methods

Materials

Primary human keratocytes were obtained from ScienCell Research Laboratories (Carlsbad, CA, USA); Eagle's MEM, PBS, fetal bovine serum (FBS), and trypsin-EDTA were from Invitrogen-Gibco (Grand Island, NY, USA); 24- and 96-well culture plates, as well as 35-mm culture dishes, were from TPP Techno Plastic Products (Trasadingen, Switzerland); and 8-well culture slides were from BD Biosciences (Bedford, MA, USA). The LPS derived from Pseudomonas aeruginosa serotype 10 was obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA), LL37 was from Anaspec (Fremont, CA, USA), human serum was from MP Biomedicals (Santa Ana, CA, USA), and recombinant human TNF-α and paired antibodies for human IL-6, IL-8, IL-1α, and IL-1β ELISAs were from R&D Systems (Minneapolis, MN, USA). Rabbit polyclonal antibodies to human IκBα and to the p65 subunit of nuclear factor (NF)-κB were obtained from Santa Cruz Biotechnology (Dallas, TX, USA), and horseradish peroxidase (HRP)-conjugated secondary antibodies, mouse monoclonal antibodies to human phosphorylated IκBα, and rabbit polyclonal antibodies to human extracellular signal-regulated kinase, to phosphorylated ERK, to c-Jun NH₂-terminal kinase (JNK), to phosphorylated JNK, to p38 mitogen-activated protein kinase (MAPK), to phosphorylated p38 MAPK, and to β-actin were from Cell Signaling (Beverly, MA, USA). Mouse monoclonal antibodies to intercellular adhesion molecule–1 (ICAM-1) were from Novus Biologicals (Littleton, CO, USA), and the rat monoclonal antineutrophil antibody NIMP-R14 was from Abcam (Cambridge, MA, USA). Alexa Fluor 488–conjugated secondary antibodies to mouse, rabbit, or rat immunoglobulin G (IgG) were from Thermo Fisher Scientific (Waltham, MA, USA). All media and reagents used for cell culture were endotoxin minimized. 

Assay of IL-6, IL-8, IL-1α, and IL-1β Secretion

Corneal fibroblasts were cultured in 24-well plates until they achieved confluence in MEM supplemented with 10% FBS, after which the culture medium was replaced with serum-free MEM for 1 day. The cells were then incubated first for 2 hours with various concentrations of LL37 in serum-free MEM and then for 48 hours in the additional absence or presence of LPS (100 ng/mL) and 0.5% human serum or of TNF-α (10 ng/mL). The medium was then collected and centrifuged at 120g for 5 minutes, and the resultant supernatant was frozen at −80°C for subsequent assay of IL-6, IL-8, IL-1α, and IL-1β with the use of solid-phase ELISAs. 

RT and Quantitative PCR Analysis

Total RNA was isolated from cells with the use of a NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany) and was subjected to RT also with the use of a kit (Qiagen, Hilden, Germany). The resulting cDNA was then subjected to real-time PCR analysis with the use of a StepOne Plus Real-Time PCR System (Thermo Fisher Scientific). Transcripts of the constitutively expressed gene for hypoxanthine phosphoribosyl transferase (HPRT) served to normalize the amount of target mRNA in each sample. Primers for IL-6, IL-8, ICAM-1, and HPRT were obtained from Qiagen. Real-time PCR data were analyzed with StepOne software V.2.2 (Thermo Fisher Scientific).^12,13 

In Situ Whole-Cell ELISA for ICAM-1

Expression of ICAM-1 was measured with the use of an in situ whole-cell ELISA as previously described.^3,14 Corneal fibroblasts (1 × 10³) were cultured in 96-well flat-bottomed microtiter plates for 72 hours, after which the culture medium was changed to serum-free MEM, and the cells were incubated for an additional 24 hours. The cells were then incubated first for 2 hours with or without LL37 (1 μM) in serum-free MEM and then for 24 hours in the additional absence or presence of LPS (100 ng/mL) and 0.5% human serum or of TNF-α (10 ng/mL). They were washed twice with PBS, fixed for 15 minutes at room temperature with PBS containing 1% paraformaldehyde, washed with PBS containing 0.1% BSA, incubated for 1 hour at 37°C with antibodies to ICAM-1 (1:10,000 dilution; Novus Biologicals) in PBS containing 1% BSA (PBS-BSA), washed three times with PBS-BSA, and incubated for 1 hour at 37°C with HRP-conjugated goat antibodies to mouse IgG (Thermo Fisher Scientific) in PBS-BSA. The cells were washed again three times with PBS-BSA before incubation for 15 minutes in the dark with 100 μL 3,3′,5,5′-tetramethylbenzidine. The reaction was terminated by the addition of 50 μL of 1 M H₂SO₄, and the absorbance of each well was determined at a wavelength of 450 nm with a microplate reader. 

Immunofluorescence Analysis of ICAM-1 Expression

Corneal fibroblasts were transferred to eight-well culture slides and then incubated for measurement of ICAM-1 expression by whole-cell ELISA. The cells were then washed twice with PBS, fixed for 15 minutes at room temperature with PBS containing 1% paraformaldehyde, washed three times with PBS-BSA, and incubated for 1 hour at room temperature with antibodies to ICAM-1 (1:1000 dilution in PBS-BSA; Novus Biologicals). They were then washed again three times with PBS-BSA, incubated for 1 hour at room temperature with Alexa Fluor 488–conjugated goat antibodies to mouse IgG (1:500 dilution in PBS-BSA; Thermo Fisher Scientific), washed with PBS, and mounted in mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA) for observation with a BIOREVO BZ-9000 fluorescence microscope (Keyence, Osaka, Japan). 

Flow Cytometric Analysis of ICAM-1 Expression

Flow cytometric analysis of ICAM-1 expression was performed as previously described.¹ Corneal fibroblasts (1 × 10⁴) were incubated in 35-mm culture dishes for measurement of ICAM-1 expression by whole-cell ELISA. They were then washed twice with PBS, harvested by incubation with cell dissociation buffer (Thermo Fisher Scientific) for 4 minutes at room temperature, washed twice with MEM supplemented with 10% FBS, incubated for 30 minutes on ice with PBS-BSA, washed with fluorescence activated cell sorting (FACS) staining buffer (PBS containing 1% FBS and 0.1% sodium azide), and incubated for 30 minutes on ice with antibodies to ICAM-1 (1:100 dilution; Novus Biologicals) or control mouse IgG1 (Dako Japan, Tokyo, Japan) in FACS staining buffer. The cells were washed with FACS staining buffer, incubated for 30 minutes on ice with Alexa Fluor 488–conjugated goat antibodies to mouse IgG (1:100 dilution in FACS staining buffer; Thermo Fisher Scientific), washed with FACS staining buffer, fixed for 20 minutes on ice with PBS containing 1% paraformaldehyde, and washed with FACS staining buffer. They were finally analyzed for ICAM-1 expression with an Attune flow cytometer (Thermo Fisher Scientific). 

Immunoblot Analysis

Immunoblot analysis of total and phosphorylated forms of ERK, p38 MAPK, JNK, and IκBα was performed as described previously.^15,16 In brief, cell lysates were subjected to SDS-PAGE on a 10% gel under reducing conditions, and the separated proteins were transferred electrophoretically to a nitrocellulose membrane with the use of a Trans-Blot Turbo device (Bio-Rad, Hercules, CA, USA). Nonspecific sites of the membrane were blocked, and it was then incubated with primary antibodies before detection of immune complexes with HRP-conjugated secondary antibodies and enhanced chemiluminescence reagents (GE Healthcare, Piscataway, NJ, USA). Protein band intensity was measured with the use of Multi Gauge software (Fujifilm, Tokyo, Japan) and was normalized by that for β-actin. 

Immunofluorescence Staining for NF-κB

Immunostaining for NF-κB was performed as described previously.¹⁴ In brief, cell monolayers in eight-well chamber slides were incubated at 37°C for 24 hours in serum-free MEM. They were then incubated first for 2 hours with or without LL37 (1 μM) in serum-free MEM and then for 1 hour in the additional absence or presence of LPS (100 ng/mL) and 0.5% human serum or of TNF-α (10 ng/mL). The cells were then washed twice with PBS, fixed with 4% paraformaldehyde in PBS, and washed an additional three times with PBS before permeabilization with 100% methanol for 6 minutes at −20°C. Nonspecific adsorption of antibodies was blocked by incubation for 30 minutes with PBS containing 3% BSA, and the cells were then incubated for 1 hour at room temperature with antibodies to the p65 subunit of NF-κB (1:100 dilution in PBS-BSA; Santa Cruz Biotechnology), washed, and incubated for 30 minutes at room temperature with Alexa Fluor 488–conjugated secondary antibodies (1:500 dilution in PBS-BSA; Thermo Fisher Scientific). The cells were finally washed, mounted in mounting medium (Vectashield; Vector Laboratories), and examined with a BIOREVO BZ-9000 fluorescence microscope. 

Mouse Model of LPS-Induced Keratitis

A mouse model of LPS-induced keratitis was essentially as previously described¹⁷ and was studied in strict accordance with the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. In brief, 8-week-old male C57BL/6 mice (Japan SLC, Shizuoka, Japan) maintained under standard (specific pathogen-free) conditions were anesthetized, and the central corneal epithelium was abraded by three contiguous scratches made with a 26-gauge needle. Two microliters of PBS alone or containing 500 ng LL37, 200 ng LPS, or both LL37 and LPS was then applied to the corneal surface. Eyes were enucleated at 24 hours after treatment, embedded in Optimal Cutting Temperature compound (Sakura Finetek Japan, Tokyo, Japan), and frozen in liquid nitrogen. Sections were cut at a thickness of 8 μm, fixed in 4% formaldehyde for 5 minutes, washed with PBS, and incubated first for 30 minutes on ice with PBS-BSA and then for 1 hour at room temperature with the antineutrophil antibody NIMP-R14 (1:100 dilution in PBS-BSA; Abcam). The sections were then washed with PBS, incubated for 45 minutes at room temperature with Alexa Fluor 488–conjugated donkey antibodies to rat IgG (1:200 dilution in PBS-BSA; Thermo Fisher Scientific), washed with PBS, and mounted in mounting medium containing DAPI for observation with a BIOREVO BZ-9000 fluorescence microscope. Infiltrating neutrophils in the corneal stroma of each section were counted independently by two observers in a blinded manner. 

Statistical Analysis

Data are presented as means ± SEM and were analyzed with Dunnett's test or the Tukey-Kramer test. All data for human corneal fibroblasts are representative of at least three independent experiments. P < 0.05 was considered statistically significant. 

Results

Inhibition of LPS-Induced Expression of IL-6 and IL-8 in Human Corneal Fibroblasts by LL37

We first examined the effects of LL37 on the release of the chemokine IL-8 and the proinflammatory cytokines IL-6, IL-1α, and IL-1β by corneal fibroblasts stimulated with LPS. We have previously shown that, like serum, tear fluid contains soluble CD14 and LPS binding protein (LBP)¹⁸ and that these factors promote the responsiveness of corneal fibroblasts to LPS.^3,14 We therefore included a low concentration (0.5%) of human serum as a source of soluble CD14 and LBP in the culture medium of corneal fibroblasts exposed to LPS in order to facilitate its detection by the cells. Corneal fibroblasts released IL-8 in response to stimulation with LPS derived from P. aeruginosa (100 ng/mL), consistent with our previous observations.³ This effect of LPS was inhibited by LL37 in a concentration-dependent manner, with this inhibition being statistically significant at LL37 concentrations of ≥0.5 μM (Fig. 1A). In contrast, LL37 had no effect on IL-8 release from corneal fibroblasts stimulated with the proinflammatory cytokine TNF-α (Fig. 1A). Reverse transcriptase and real-time PCR analysis revealed that LPS and TNF-α each increased the amount of IL-8 mRNA in corneal fibroblasts. Whereas LL37 blocked this action of LPS, it had no effect on that of TNF-α (Figs. 1B, 1C). 

Figure 1

 

Effects of LL37 on LPS- or TNF-α–induced IL-8 expression in human corneal fibroblasts. (A) Cells were incubated first for 2 hours with the indicated concentrations of LL37 and then for 48 hours in the additional presence of LPS (100 ng/mL) plus 0.5% human serum, 0.5% human serum alone, or TNF-α (10 ng/mL), after which the amount of IL-8 in the culture supernatants was determined. Data are means ± SEM of quadruplicates. **P < 0.01 (Dunnett's test) versus the corresponding value for cells incubated without LL37. (B, C) Cells were incubated first for 2 hours in the absence or presence of LL37 (1 μM) and then for 18 hours either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) (B) or in the additional absence or presence of TNF-α (10 ng/mL) (C). The amount of IL-8 mRNA in the cells was then determined by RT and real-time PCR analysis. Data were normalized by the abundance of HPRT mRNA, are expressed in arbitrary units, and represent means ± SEM of quadruplicates. **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37.

The LPS also stimulated the release of IL-6 from corneal fibroblasts, and this effect was inhibited by LL37 in a concentration-dependent manner (Fig. 2A). This inhibitory effect of LL37 was also significant at concentrations of ≥0.5 μM. Again, TNF-α also stimulated the release of IL-6 by these cells, and this effect of TNF-α was not influenced by the additional presence of LL37 (Fig. 2A). Lipopolysaccharide and TNF-α each increased the abundance of IL-6 mRNA in corneal fibroblasts, with this action of LPS being abolished by LL37 but that of TNF-α being unaffected (Figs. 2B, 2C). 

Figure 2

 

Effects of LL37 on LPS- or TNF-α–induced IL-6 expression in human corneal fibroblasts. (A) Cells were incubated first for 2 hours with the indicated concentrations of LL37 and then for 48 hours in the additional presence of LPS (100 ng/mL) plus 0.5% human serum, 0.5% human serum alone, or TNF-α (10 ng/mL), after which the amount of IL-6 in the culture supernatants was determined. Data are means ± SEM of quadruplicates. **P < 0.01 (Dunnett's test) versus the corresponding value for cells incubated without LL37. (B, C) Cells were incubated first for 2 hours in the absence or presence of LL37 (1 μM) and then for 18 hours either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) (B) or in the additional absence or presence of TNF-α (10 ng/mL) (C). The amount of IL-6 mRNA in the cells was then determined by RT and real-time PCR analysis. Data were normalized by the abundance of HPRT mRNA, are expressed in arbitrary units, and represent means ± SEM of quadruplicates. **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37.

The release of IL-1α or IL-1β from corneal fibroblasts was not detected in the absence or presence of LPS or LL37 (data not shown). At the concentrations examined in the present study, LL37 did not manifest a cytotoxic effect on corneal fibroblasts as revealed by measurement of the release of lactate dehydrogenase (data not shown). 

Inhibition of LPS-Induced Expression of ICAM-1 in Human Corneal Fibroblasts by LL37

We next examined the effect of LL37 on expression of the adhesion molecule ICAM-1 in corneal fibroblasts with the use of a whole-cell ELISA. As we showed previously,^3,14 exposure of corneal fibroblasts to LPS or TNF-α resulted in a marked increase in ICAM-1 expression at the cell surface. This up-regulation of ICAM-1 by LPS was abolished by LL37, whereas that induced by TNF-α was unaffected (Fig. 3A). We also examined the expression of ICAM-1 in corneal fibroblasts by immunofluorescence staining. Although no specific fluorescence was detected in cells incubated without stimulant, cells exposed to LPS or TNF-α exhibited a marked increase in ICAM-1–specific immunofluorescence (Fig. 3B). Again, this effect of LPS, but not that of TNF-α, was blocked by LL37. Flow cytometric analysis further confirmed that both LPS and TNF-α increased the expression of ICAM-1 at the cell surface and that this effect of LPS, but not that of TNF-α, was inhibited by LL37 (Fig. 4). 

Figure 3

 

Effects of LL37 on LPS- or TNF-α–induced ICAM-1 expression in human corneal fibroblasts. Cells were incubated first for 2 hours in the absence or presence of LL37 (1 μM) and then for 24 hours either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) or in the additional absence or presence of TNF-α (10 ng/mL). The surface expression of ICAM-1 was then determined by whole-cell ELISA (A) or by immunofluorescence analysis (B) with antibodies specific for this protein. Nuclei were stained blue with DAPI in (B). Quantitative data are means ± SEM of quadruplicates. **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37. Scale bar: 50 μm.

Figure 4

 

Flow cytometric analysis of the effects of LL37 on LPS- or TNF-α–induced ICAM-1 expression in human corneal fibroblasts. Cells were incubated first for 2 hours in the absence or presence of LL37 (1 μM) and then for 24 hours either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) (A, B) or in the additional absence or presence of TNF-α (10 ng/mL) (C, D). The surface expression of ICAM-1 was then determined by flow cytometry. Representative histograms (A, C) and the geometric mean fluorescence intensity (B, D) are shown. Quantitative data are means ± SEM of quadruplicates. **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37.

Reverse transcriptase and real-time PCR analysis also revealed that LPS and TNF-α each increased the amount of ICAM-1 mRNA in corneal fibroblasts. The antimicrobial peptide LL37 significantly inhibited this effect of LPS without altering that of TNF-α (Fig. 5). 

Figure 5

 

Effects of LL37 on LPS- or TNF-α–induced up-regulation of ICAM-1 mRNA abundance in human corneal fibroblasts. Cells were incubated first for 2 hours in the absence or presence of LL37 (1 μM) and then for 18 hours either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) (A) or in the additional absence or presence of TNF-α (10 ng/mL) (B). The amount of ICAM-1 mRNA in the cells was then determined by RT and real-time PCR analysis. Data were normalized by the abundance of HPRT mRNA, are expressed in arbitrary units, and represent means ± SEM of quadruplicates. *P < 0.05, **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37.

Effects of LL37 on LPS-Induced Signaling in Human Corneal Fibroblasts

Regulation of the expression of cytokines or adhesion molecules by LPS is mediated by NF-κB and MAPK signaling pathways in various cell types including corneal fibroblasts.¹⁴ To investigate the mechanism by which LL37 blocks the stimulatory action of LPS on IL-6, IL-8, and ICAM-1 expression in corneal fibroblasts, we therefore examined its effects on these signaling pathways. Immunoblot analysis with antibodies to total or phosphorylated forms of signaling molecules revealed that the total abundance of MAPKs (ERK, p38, or JNK) was not affected by stimulation with LPS and that the phosphorylated forms of these proteins and that of the NF-κB inhibitor IκBα were virtually undetectable under basal conditions. Stimulation with LPS for 30 minutes induced phosphorylation of the MAPKs ERK, p38, and JNK, as well as the phosphorylation and degradation of IκBα in corneal fibroblasts (Fig. 6A). Whereas LL37 did not affect the abundance or phosphorylation level of these proteins in the absence of LPS, it inhibited the phosphorylation of ERK, p38, and JNK, as well as the phosphorylation and degradation of IκBα induced by LPS, indicating that the LPS-induced activation of MAPK and NF-κB signaling pathways was suppressed by LL37. We further confirmed the effects of LPS and LL37 on the phosphorylation and degradation of IκBα by densitometric analysis of immunoblots in three independent experiments. The phospho-IκBα/β-actin band intensity ratio was thus significantly higher in cells incubated with LPS than in those incubated alone or with both LPS and LL37. Conversely, the total IκBα/β-actin band intensity ratio was lower in cells incubated with LPS than in those incubated alone or with both LPS and LL37 (Fig. 6B). 

Figure 6

 

Effects of LL37 on the LPS-induced phosphorylation of IκBα and MAPKs in human corneal fibroblasts. Cells were incubated first for 2 hours with or without LL37 (1 μM) and then for 30 minutes in the additional presence of 0.5% human serum with or without LPS (100 ng/mL), after which cell lysates were prepared and subjected to immunoblot analysis with antibodies to total or phosphorylated (p-) forms of ERK, p38 MAPK, JNK, or IκBα or with those to β-actin (loading control) (A). The densitometric analysis of immunoblots on the phosphorylation and degradation of IκBα (B). The phospho-IκBα/β-actin band intensity ratio (upper graph) and the total IκBα/β-actin band intensity ratio (lower graph). The data are means ± SEM of triplicates. *P < 0.05, **P < 0.01 (Tukey-Kramer test) versus the corresponding value for cells incubated without LPS or TNF-α; †P < 0.05 (Tukey-Kramer test) versus the corresponding value for cells incubated without LL37.

We then examined the effect of LL37 on the subcellular localization of the p65 subunit of NF-κB in LPS-stimulated corneal fibroblasts by immunofluorescence analysis (Fig. 7). Under basal conditions, NF-κB immunofluorescence was localized predominantly to the cytoplasm of corneal fibroblasts. No immunofluorescence was apparent in cells stained with normal rabbit IgG as a negative control (data not shown). Stimulation with LPS or TNF-α for 60 minutes resulted in translocation of NF-κB from the cytoplasm to the nucleus. Whereas LL37 did not affect the localization of NF-κB in the absence of stimulant, it attenuated the translocation of NF-κB to the nucleus induced by LPS without affecting that induced by TNF-α. 

Figure 7

 

Effects of LL37 on the subcellular localization of NF-κB in LPS- or TNF-α–stimulated human corneal fibroblasts. Cells were incubated first for 2 hours with or without LL37 (1 μM) and then for 60 minutes either in the additional presence of 0.5% human serum with or without LPS (100 ng/mL) or in the additional absence or presence of TNF-α (10 ng/mL). The cells were then subjected to immunofluorescence analysis with antibodies to the p65 subunit of NF-κB. Scale bar: 50 μm.

Suppression of LPS-Induced Keratitis by LL37 In Vivo

Finally, we adopted a mouse model to assess the ability of LL37 to suppress LPS-induced keratitis in vivo. Immunohistofluorescence analysis revealed that the application of LL37 alone to the abraded cornea for 24 hours had no significant effect on the extent of neutrophil infiltration into the cornea compared with that observed after the application of PBS vehicle. In contrast, neutrophil infiltration was markedly increased by application of LPS, consistent with previous observations.¹⁷ This effect of LPS was significantly inhibited by coadministration of LL37 (Fig. 8). 

Figure 8

 

Effect of LL37 in a mouse model of LPS-induced keratitis. The cornea was scratched, and either LL37, LPS, both LPS and LL37, or PBS vehicle was applied. After 24 hours, the eye was enucleated for immunohistofluorescence staining of neutrophils in the cornea with the NIMP-R14 antibody (A). The corneal epithelium (ep) and stroma (st) are indicated. Scale bar: 200 μm. The number of infiltrating neutrophils in the corneal stroma was counted (B). Data are means ± SEM for one section examined for each of four eyes. **P < 0.01 (Tukey-Kramer test) versus the corresponding value for PBS; ††P < 0.01 (Tukey-Kramer test) versus the corresponding value for LPS alone.

Discussion

We showed that LL37 inhibited the LPS-induced expression of IL-6, IL-8, and ICAM-1 in human corneal fibroblasts without affecting the expression of these molecules induced by TNF-α. Given that LL37 has previously been shown to stimulate corneal epithelial wound healing and cytokine synthesis by corneal epithelial cells,^10,11 this antimicrobial peptide appears likely to have differential effects on corneal epithelial cells and stromal fibroblasts during the innate immune response associated with infectious keratitis. 

In general, epithelial cells play a key role in the first line of defense against invading pathogens by inducing an innate immune response. In contrast to epithelial cells in other tissues such as skin, however, corneal epithelial cells do not respond to bacterial components including peptidoglycan and LPS because TLR-2 and TLR-4 are expressed intracellularly.² The corneal epithelium thus appears to serve primarily as a physical barrier to bacteria, with the unresponsiveness of corneal epithelial cells to these microorganisms contributing to an immunosilent environment at the ocular surface. LL37 is expressed in corneal and conjunctival epithelial cells, however, and a combination of TLR agonists or heat-killed Candida albicans has been found to stimulate its production by corneal epithelial cells.^19–21 The antimicrobial peptide LL37 is also up-regulated in the regenerating human corneal epithelium and stimulates the migration of and cytokine production by corneal epithelial cells.¹¹ Together, these observations suggest that, in addition to its direct antimicrobial effects, LL37 may function as an “alarmin,” a danger signal triggered by infection or injury to stimulate wound healing or elicit an immune response to infectious pathogens. 

In contrast to corneal epithelial cells, stromal fibroblasts recognize bacterial components and elicit an innate immune response in the cornea.^3,14 Corneal fibroblasts also produce LL37 under basal conditions, and this production is up-regulated in response to mycobacterial infection.²² The effects of LL37 on the function of corneal stromal fibroblasts have remained unknown, however. We showed that LL37 acts directly on corneal fibroblasts to suppress the innate immune responses of these cells to LPS, in contrast to its effects on corneal epithelial cells. We also found that exogenous LL37 did not induce corneal inflammation but significantly suppressed LPS-induced keratitis in mice. These pleiotropic effects of LL37 may contribute to ocular defense against pathogens, to wound healing in the corneal epithelium, and to the resolution of inflammation during infectious keratitis. 

Serum LL37 levels in pediatric patients with postinfectious bronchiolitis obliterans and in healthy children were previously found to be ∼70 and 50 ng/mL, respectively.²³ Serum of patients with antineutrophil cytoplasmic antibody–associated vasculitis also contained higher levels of LL37 (∼100 ng/mL) than did that of healthy individuals (∼30 ng/mL).²⁴ The plasma LL37 level was found to be ∼70 ng/mL in subjects with type 2 diabetes mellitus.²⁵ The concentration of LL37 in tracheal aspirates was ∼5 μg/mL in newborn infants but was ∼25 μg/mL in infants with pulmonary or systemic infection.²⁶ Analysis of a human bronchial xenograft model revealed that the concentration of epithelial cell–derived LL37 in airway surface fluid was 1.7 to 1.8 μg/mL.²⁷ These various studies suggest that the physiologic concentration of LL37 is relatively high in lung and that both circulating and local LL37 levels are increased during inflammation. Most of the experiments with corneal fibroblasts in the present study examined the effects of LL37 at a concentration of 1 μM, which corresponds to 4.5 μg/mL. The concentrations of LL37 at the ocular surface, including that in tear fluid, remain to be determined. 

Given that LL37 blocked the expression of IL-6, IL-8, and ICAM-1, as well as the activation of NF-κB induced by LPS but not those elicited by TNF-α in corneal fibroblasts, these effects of LL37 were not attributable to nonspecific inhibition of NF-κB. The stimulatory effects of LL37 on corneal and skin epithelial cell migration were previously found to be induced via heparin-binding epidermal growth factor (HB-EGF)–mediated transactivation of the EGF receptor,^10,28 whereas the stimulation of chemokine synthesis in epidermal keratinocytes by LL37 is mediated through induction of IL-36γ.²⁹ LL37 has been found to possess LPS binding activity and to suppress the interaction of LPS with LBP that mediates the transport of LPS to CD14 and thereby facilitates the activation of a CD14⁺ murine macrophage cell line by LPS. In addition, LL37 was shown to bind to CD14 at the surface of these cells, and prior incubation of the cells with LL37 inhibited the binding of LPS to the cell surface.³⁰ We have previously shown that human corneal fibroblasts express CD14 on the cell surface and that the detection of LPS by corneal fibroblasts is facilitated by the addition of human serum or of both LBP and soluble CD14.^3,14 We and others have also shown that tear fluid and serum of healthy subjects contain levels of these proteins that are sufficient to facilitate an innate immune response by corneal fibroblasts.^18,31 These observations suggest that corneal fibroblasts in vivo are poised to respond to a complex of LPS with soluble CD14 and LBP present in tear fluid. In the present study, we added a low concentration (0.5%) of human serum as a source of LBP and soluble CD14 to the culture medium of corneal fibroblasts exposed to LPS in order to facilitate the detection of LPS by the cells. The precise mechanisms of the inhibitory effects of LL37 on corneal fibroblasts revealed in our study remain to be clarified, but they might be mediated by the binding of LL37 to LPS or to CD14 at the cell surface. 

The antimicrobial peptide LL37 has previously been found to inhibit collagen synthesis by dermal fibroblasts,³² whereas LL37 secreted by epithelial cells was shown to promote collagen production by lung fibroblasts.³³ We have previously shown that fibroblasts from different tissues have different properties, with lung fibroblasts differing from fibroblasts from skin and the cornea with regard to the production of the chemokine CCL17.¹² LL37 might therefore also have different effects on fibroblasts from different tissues. Anti-inflammatory effects of LL37 similar to those observed in the present study with corneal fibroblasts have been demonstrated with LPS-stimulated gingival fibroblasts.^34,35 

We found that LPS and TNF-α had similar effects on the levels of IL-6 or IL-8 mRNAs in corneal fibroblasts, whereas the extent of IL-6 or IL-8 release induced by TNF-α was substantially lower than that induced by LPS. The reason for this difference is unclear, but it may be attributable to differences in posttranscriptional regulation of the production and secretion of these cytokines by LPS and TNF-α^36,37 and might contribute to differences in pathogenesis between infectious and noninfectious keratitis. 

In summary, we showed that LL37 inhibited the expression of the proinflammatory cytokine IL-6, the chemokine IL-8, and the adhesion molecule ICAM-1 induced by LPS in human corneal fibroblasts. The uncontrolled overproduction of such inflammatory mediators by corneal fibroblasts may prevent resolution of infectious inflammation and result in inappropriate tissue remodeling or destruction, eventually leading to corneal scarring. We also demonstrated that LL37 suppressed corneal inflammation in a mouse model of LPS-induced keratitis. Exogenous LL37 might be potentially beneficial to restrain an excessive inflammatory response without adverse effects in the treatment of infectious keratitis as a result of its inhibitory effects on activated corneal fibroblasts, as well as of its antimicrobial and pro–wound healing activities. In the current experiment, LL37 was given at the same time that LPS was given. Further studies testing the effects of exogenous LL37 at varying time points after infection in an animal model of infectious bacterial keratitis are necessary to examine the therapeutic potential of exogenous LL37 in the clinical setting. In addition, the concentrations of LL37 at the ocular surface in human subjects with or without inflammation or infection should be examined to know the role of LL37 at the ocular surface in humans. 

Acknowledgments

Disclosure: W. Ishida, None; Y. Harada, None; K. Fukuda, None; A. Fukushima, None