Localization and Gene Expression of Human β-Defensin 9 at the Human Ocular Surface Epithelium

Abstract

Purpose.: Antimicrobial peptides (AMPs) are multifunctional host defense molecules. Human β-defensin 9 (HBD9) has previously been shown to be downregulated during ocular surface (OS) infection or inflammation. Here, the authors aimed to study localization of HBD9 protein in different OS regions and to determine the role of Toll-like receptors (TLRs), nucleotide oligomerization domain (NOD)-like receptors, and proinflammatory cytokines in HBD9 expression.

Methods.: Immunolocalization of HBD9 protein was carried out on the normal human OS regions (cornea, limbus, and conjunctiva). Quantitative PCR analysis of HBD9 mRNA was performed in SV40-transformed human corneal epithelial cells (hCECs) treated for different durations with synthetic pathogen-associated molecular patterns (PAMPs) and recombinant cytokines.

Results.: HBD9 protein was constitutively expressed on OS epithelia. Corneal and limbal epithelia and corneal stroma demonstrated modest levels of HBD9, whereas conjunctival epithelium demonstrated high levels of HBD9 protein. TLR02, TLR03, TLR04, and TLR05 were shown to modulate HBD9 mRNA in hCECs. Similarly, NOD2 and IL-1β were also shown to alter HBD9 in a time-dependent manner. In response to infection-related PAMPs and inflammatory cytokines, an initial increase in HBD9 mRNA levels was observed, followed by a significant downregulation.

Conclusions.: This is the first demonstration of HBD9 protein expression at different OS regions. The authors also determined the role of various innate immune receptors in HBD9 mRNA modulation. Further understanding of the signaling mechanisms involved in the initial response of HBD9 to infection or inflammation is likely to indicate future therapeutic directions with this AMP.

The interaction of pathogens and epithelial cells is recognized to be an active process in which epithelial cells participate in innate host defense by expressing a variety of proinflammatory cytokines, chemokines, and antimicrobial peptides (AMPs). ¹ Several studies have demonstrated the localization of different families of AMPs, including defensins and cathelicidins in ocular surface (OS) epithelium. ² Localization of human β-defensin (HBD)-1 to HBD-3 was reported in superficial layers of normal and inflamed corneal and conjunctival epithelium, ^3–5 whereas LL-37 distribution was shown only in normal human corneal tissue. ⁶ A newer member of the β-defensin family, DEFB109, also known as HBD9, has been shown to be constitutively present in gingival keratinocytes and to be downregulated in response to Candida albicans. ⁷ In our previous report, we demonstrated a reduced expression of HBD9 in bacterial-, viral-, acanthamoeba-, and dry eye-related OS disease compared with controls. ⁸ To further analyze the expression of HBD9, we characterized the distribution of this AMP in normal cornea, conjunctiva, and limbal tissue using an immunofluorescence technique. 

The ocular surface epithelium has been shown to provide immune surveillance against invading pathogens by means of a wide gamut of pathogen recognition receptors such as Toll-like receptors (TLRs) and nucleotide oligomerization domain (NOD)-like receptors (NLRs). TLRs are highly conserved and homologous receptors found in different species of plant and animal kingdoms. ¹ To date, 11 human and 13 mouse TLRs have been identified, and each is capable of recognizing pathogen-associated molecular patterns (PAMPs) derived from various infection-causing microbes. ⁹ TLR02, TLR04, TLR05, and TLR11 reside on the cell membrane and recognize lipopeptides of Gram-positive bacteria, lipopolysaccharide (LPS) of Gram-negative bacteria, flagellin and propellin, respectively. ⁹ TLR03, TLR07/08, and TLR09 are sequestered in endosomes and respond to double-stranded RNA from Herpes-simplex virus and single-stranded RNA from influenza virus and West Nile virus, respectively. ⁹ As yet the PAMP for TLR10 is unknown. Of all NOD-like receptors, NOD-1 and NOD-2 are well-characterized proteins, present endosomally, and recognize γ-D-glutamyl-meso-diaminopimelic acid (iE-DAP) and muramyl dipeptide (MDP), the synthetic components of peptidoglycan, respectively. ^9,10  

Here, we investigated the role of known human TLRs and of NOD-1 and NOD-2 at the corneal epithelium in relation to the expression of HBD9. In addition, we studied the effects of IL-1β and TNFα on HBD9 expression. 

Methods

Reagents and Antibody

Synthetic PAMPs—namely Pam3CSK4, LPS, Poly I-C, Flagellin, R848, CpG ODN, MDP, and iE-DAP—were purchased from Invivogen (San Diego, CA). Recombinant human IL-1β and human TNFα were procured from Calbiochem (Nottingham, UK). Purified polyclonal rabbit anti–human HBD9 was custom synthesized (Eurogentec, Southampton, UK) for immunofluorescence analysis. 

Tissue Processing

Human cadaver eyes were collected after obtaining local ethics committee approval and prior consent from the donors or their relatives in adherence with the tenets of the Declaration of Helsinki. Cadaver ocular tissue was processed within 48 hours of death under aseptic conditions. A 15-mm trephine was used to dissect the corneoscleral button, maintaining a 3-mm frill of conjunctiva around the limbus. The button was divided into eight segments and then snap-frozen in prechilled isopentane. Frozen tissue blocks were stored at −80°C until further use. 

Immunofluorescence

Six-micrometer sections of normal ocular surface tissue on 2% 3-aminopropyltriethoxysilane–coated glass slides were acetone fixed and blocked with normal goat serum (Invitrogen, Paisley, UK). Washing step was performed with 1% BSA + 0.1% Triton-X premixed solution. The sections were then incubated with primary antibody against HBD9 (1:50; polyclonal rabbit anti–human antibody; Eurogentec, Liege, Belgium). The sections were washed and incubated with secondary antibody Alexa Fluor 555 (anti–rabbit; Invitrogen) for 30 minutes, followed by counterstaining with DAPI (4′, 6-diamidino-2-phenylindole) for 4 minutes. The slides were mounted with fluorescent mounting medium (Dako, Ely, UK) and examined under a fluorescent microscope (B51X; Olympus, Tokyo, Japan) using imaging software (CELL-F; Olympus, Southend-on-Sea, UK). The images were edited in image editing software (Photoshop CS4; Adobe Systems, San Jose, CA). 

SV40-Transformed Human Corneal Epithelial Cell Line

SV40-transformed human corneal epithelial cells (hCECs) were a kind gift from Felicity Rose (School of Pharmacy, The University of Nottingham, Nottingham, UK). Cells were maintained in medium (Epilife; Cascade Biologics, Paisley, UK) containing human keratinocyte growth supplement (HKGS; Cascade Biologics), antibiotic mixture (gentamicin and amphotericin B; Cascade Biologics), and anti–mycoplasma agent (Plasmocin; Invivogen). The hCECs (8 × 10⁴) were plated onto 12-well culture plates and grown until they reached 70% confluence in humidified conditions (5% CO₂, 37°C). Before treatment, hCECs were starved overnight in HKGS-free media. 

Total RNA Extraction

Cells were collected in lysis buffer (Buffer RLT; Qiagen, Crawley, UK), and total RNA was extracted according to the manufacturer's instructions (RNeasy mini kit; Qiagen). Briefly, the cell lysate was homogenized with a spin column (QIAshredder; Qiagen) and mixed with 70% ethanol in equal volume (1:1). The mixture was then applied onto spin columns and centrifuged at 10,000 rpm for 15 seconds. The filtrate was discarded, and the spin column was washed with buffer RW1 and then with buffer RPE by centrifugation. Total RNA was eluted in RNase-free water and quantified (Nanodrop Spectrophotometer; Thermo Fisher Scientific, Loughborough, UK). 

Complementary DNA Synthesis

Total RNA (2000 ng) was reverse transcribed into complementary DNA (cDNA) with a reverse transcription kit (Quantitect; Qiagen). Briefly, total RNA was mixed with genomic DNA wipeout buffer (Qiagen) on ice, and the volume was adjusted with water. The mixture was incubated at 42°C for 3 minutes and then placed on ice. The reverse transcription enzyme mix (Quantitect reverse transcriptase, Quantitect RT buffer, and RT primer mix; Qiagen) was prepared on ice and mixed with total RNA mixture. The final mixture was then incubated at 42°C for 30 minutes, followed by reverse transcriptase deactivation at 95°C for 3 minutes. Samples were stored at −20°C until further analysis. 

Quantitative Polymerase Chain Reaction

Quantitative polymerase chain reaction (qPCR) analysis of HBD9 mRNA was performed using pre-optimized assay (TaqMan, Hs02760065_g1; Applied Biosystems, Warrington, UK) on the Mx3005p real-time PCR instrument (Stratagene, La Jolla, CA). For relative quantification, HBD9 levels were normalized against 18S rRNA endogenous control (4319413E; Applied Biosystems) levels. The qPCR experimental setup was carried out as detailed in the manufacturer's protocol (Applied Biosystems). Initially, template cDNA was diluted to 1:2 using nuclease-free water to perform the qPCR in triplicate. Each reaction was prepared to 20 μL final reaction volume with 10 μL of 2× gene expression master mix, 1 μL of 20× assay (primers and probe mix; TaqMan; Applied Biosystems), 5 μL diluted cDNA, and 4 μL nuclease-free water. Appropriate negative (nontemplate control and reverse transcriptase control) and positive (human reference RNA; Stratagene) controls were also run in each experiment. Raw data were acquired (MxPro software, version 4.01; Stratagene) on the computer linked to the qPCR machine. The data were then analyzed on a spreadsheet (Excel; Microsoft, Redmond, WA) using the ΔΔ-Ct method. 

Statistical Analysis

qPCR data were statistically analyzed on statistical software (SPSS, 16.0 version; IBM, Chicago, IL) with significance set at P < 0.05. Using Student's t-test, we statistically compared HBD9 mRNA expression in treated samples with those obtained in untreated control. All data were represented as mean ± SE of three independent experiments performed in triplicate. 

Results

Immunolocalization of DEFB109/HBD9 in Normal OS Tissue Sections

As shown in Figure 1, DEFB109/HBD9 was constitutively expressed in all regions of normal OS tissue sections. It was predominantly present in the uppermost layers of the corneal epithelium, with a decline in intensity in basal layers. A few keratocytes in corneal stroma also stained for HBD9. HBD9 was also localized in all layers of limbal epithelium, with no staining demonstrated in limbal keratocytes. Notably, the conjunctival epithelium demonstrated strong expression of HBD9 protein, with decreased staining in basal layers. Human tonsil tissue section, which served as a control, was strongly positive for HBD9. 

Figure 1.

 

Localization of HBD9 protein. Six-micrometer frozen section of full-thickness normal OS tissue and human tonsil tissue was stained with polyclonal antibody against HBD9 (rabbit anti–human; 1:50 dilution). Secondary antibody (goat anti–rabbit IgG; 1:300 dilution) was used to detect primary antibody. Top: merged images of both HBD9 (yellow) and nuclei (DAPI; blue) staining at 200× magnification. Bottom: HBD9 and DAPI merged images at 400× magnification. Negative control staining with secondary antibody alone showed no immunoreactive staining.

TLRs Play an Essential Role in DEFB109/HBD9 Expression in hCECs

To determine whether HBD9 expression is modulated in response to the activation of TLRs, we assessed mRNA expression of HBD9 using the qPCR method. As shown in Figure 2, treatment of hCECs for different durations (1–24 hours) with PAMPs of TLRs resulted in the immediate upregulation of HBD9 mRNA. Stimulation of hCECs with Pam3CSK4 resulted in significant upregulation of HBD9 at 1 hour (5.4-fold increase; P = 0.0077), with a sudden decline at the end of 3 hours that remained low until 9 hours. As indicated, after 24 hours of treatment, HBD9 expression reached 3.9-fold (P = 0.0237; Fig. 2A). In response to poly I-C, a time-dependent increase in HBD9 level (up to 3 hours; P = 0.0435) was observed. However, after a surge, HBD9 transcript levels reached baseline expression at 6 hours and remained at the same level until 24 hours (Fig. 2B). Similar to the TLR02-mediated response, treatment of hCECs with LPS, a TLR04 PAMP, resulted in increased expression of HBD9 mRNA at 1 hour (1.83-fold), with a rapid decline at 3 hours. However, after 6 hours, HBD9 transcript levels increased again (reaching the level seen at 1 hour) and remained unchanged until 24 hours (Fig. 2C). HBD9 expression pattern in response to Flagellin, a TLR05 stimulant, matched the response mediated by the Pam3CSK4/TLR02 axis except for a rise at 24 hours. A surge of HBD9 mRNA at 1 hour (1.9-fold; P = 0.0191) was followed by a plunge in levels below the baseline expression until 24 hours (Fig. 2D). Stimulation of TLR07/08 and TLR09 with R-848 (Fig. 2E) and CpG ODN (Fig. 2F), respectively, did not affect HBD9 mRNA expression in hCECs. 

Figure 2.

 

TLRs are the key modulators of HBD9 mRNA in hCECs. Cells were incubated with (A) Pam3CSK4, (B) poly I-C, (C) LPS, (D) flagellin, (E) R848, and (F) CpG ODN for the indicated time points. HBD9 mRNA levels were measured by qPCR. All data represent mean ± SEM of three samples repeated three times. *P < 0.05; **P < 0.001. Expression levels are normalized against the control untreated sample, which has been maintained at 1 in each of the graphs (note different scales in y-axes).

NOD2 Modulates DEFB109/HBD9 Expression in hCECs

To investigate whether NOD-like receptors regulate HBD9 mRNA expression, we treated hCECs with synthetic PAMPs of NOD1 and NOD2 in a time-dependent manner. As shown in Figure 3A, treatment of cells with iE-DAP demonstrated an insignificant upregulation of HBD-9 at all time points. However, a 70% decrease in baseline expression of HBD9 was noted at 24 hours. In contrast to iE-DAP, we observed an immediate response of HBD9 mRNA to NOD2 stimulant MDP at 1 hour (1.58-fold; P = 0.0228) that was followed by a time-dependent decrease in transcript levels, reaching 50% below the baseline expression at 6 hours (P = 0.0019). This was followed by a gradual increase in mRNA, with maximal levels noted at 24 hours (Fig. 3B). 

Figure 3.

 

NOD1 and NOD2 modulate HBD9 mRNA expression in hCECs. Cells were incubated with (A) iE-DAP or (B) MDP for the indicated time points. HBD9 mRNA levels were measured by qPCR. All data represent mean ± SEM of three samples repeated three times. *P < 0.05; **P < 0.001.

IL-1β Is an Important Regulator of DEFB109/HBD9 Expression in hCECs

To determine the effect of proinflammatory cytokines on HBD9 expression, we treated the cells without or with human recombinant IL-1β and TNFα, alone or in combination for 24 hours. As shown in Figure 4A, a 4.4-fold increase in HBD9 mRNA expression was noted in response to IL-1β (P = 0.0139), and a modest and insignificant upregulation of HBD9 transcript was observed after incubation with TNFα in combination with IL-1β or by itself. To further understand the role of IL-1β, we conducted a time-course treatment study. As indicated in Figure 4B, a significant downregulation of HBD9 was evident after 3 hours (0.04-fold; P = 0.0004) of stimulation, which remained at the same level until the end of 9 hours (0.09-fold; P = 0.0119). This was followed by a sharp increase in HBD9 mRNA levels at 24 hours (4.43-fold; P = 0.0419). 

Figure 4.

 

IL-1β is an important regulator of HBD9 mRNA expression in hCECs. (A) Effect of proinflammatory cytokines on HBD9. Cells were incubated with IL-1β, TNFα or combination of both for 24 hours and HBD9 mRNA expression was evaluated. (B) Effect of IL-1β on HBD9. Cells treated with IL-1β for different durations were processed for measurement of HBD9 mRNA levels by qPCR. All data represents means ± SEM of three samples repeated three times. *P < 0.05; **P < 0.001; ***P < 0.0001.

Discussion

In this study, we have demonstrated the constitutive presence of the HBD9 protein in the cornea, conjunctiva, and limbus. HBD9 was shown to be highly concentrated in superficial layers of healthy corneal epithelium, suggesting an immediate role of these AMPs in OS defense against invading pathogens. Notably, we also observed the presence of HBD9 in the corneal stroma, the layer beneath the epithelium. In the normal state, the corneal stroma provides structural support to the cornea and a second barrier against microbial intrusion. The latter response is provided by the presence of Langerhans cells (dendritic cells), immunoglobulins (IgG and IgA), and AMPs. ^11–15 On infection or breach in the corneal epithelium, the local stromal immune response is activated, recruiting polymorphonuclear cells (PMNs) and lymphocytes (B- and T-cells) to the site. ^14,16,17 Thus, it could be proposed that HBD9, similar to its family members HBD2 ¹⁸ and HBD3, ¹⁹ might also play a role in adaptive immunity. We noted an increased expression of HBD9 in the conjunctiva compared with the cornea. The normal conjunctiva contains resident lymphocytes and neutrophils and has a well-defined vasculature. Inflammatory mediators and growth factors produced locally or arriving through the blood vessels could influence the differential expression of HBD9 in the conjunctiva compared with the avascular cornea. ^20,21  

β-Defensins are highly expressed at the OS. Ikeda et al. ²² have shown the absence of mRNA expression of the human β-defensin-2 ortholog, mBD-2, in the limbal region of the experimentally developed allergic conjunctivitis (EAC) mouse model compared with control. Notably, conjunctival and corneal epithelia showed positive expression of mBD2 in the EAC model. Terai et al. ⁵ have demonstrated stronger expression of HBD1 in the normal limbus and the conjunctiva than in the cornea. The expression of HBD9 is high at the limbus and the cornea, and, like its family members, it could play a significant role in allergic and infectious conjunctivitis. 

Downregulation of HBD9 mRNA in response to infectious stimuli has previously been demonstrated. ^7,8 Innate immune receptors such as TLRs and NLRs are widely studied because of their high affinity for pathogen recognition and modulation of cytokine and AMP expression. ¹ In this study we have demonstrated the role of TLRs, NLRs, and IL-1β in HBD9 mRNA expression. Activation of TLR02 and TLR04 has been implicated in HBD2 expression in various epithelia, such as lung, ²³ tracheobronchus, ²⁴ intestine, ²⁵ and cornea. ²⁶ Consistent with previous studies on other defensins, we have demonstrated an important role of TLR02 in the modulation of HBD9 expression in hCECs. Kumar et al. ²⁶ demonstrated an increased expression of HBD2 mRNA in hCECs after 6 hours of treatment with 10 μg/mL Pam3CSK4. However, using a similar cell line model and a much lower concentration of 1 μg/mL Pam3CSK4, we demonstrated TLR02-induced HBD9 mRNA expression at an earlier time point, 1 hour. Notably, after 6 hours, we found HBD9 mRNA rapidly declined below baseline expression, supporting our ex vivo study results. ⁸ Using intestinal cell lines, Vora et al. ²⁵ showed TLR04 induced HBD2 expression in response to treatment with LPS after 5 hours. In our study, we found that the LPS induced HBD9 after 1 hour and remained at the same level until 24 hours, suggesting that the family members of β-defensins work in tandem against infectious stimuli. 

Poly (I-C), the TLR03 agonist, has been shown to augment HBD2 expression in gingival, ²⁷ uterine, ²⁸ and fallopian tube ²⁹ epithelial cells. In hCECs, we demonstrated a threefold increase in HBD9 mRNA expression in response to poly I-C in 3 hours, suggesting an immediate role of HBD9 against viral infections similar to HBD2. 

Flagellin, a bacterial flagella filament protein, has been reported to induce HBD2 expression by TLR05 in human carcinoma (Caco-2) and murine enteroendocrine (STC-1) cell line models. ^30,31 Here, using synthetic flagellin protein, we demonstrated a rapid induction of HBD9 mRNA expression in hCECs. After a surge, we noted a sudden decrease in mRNA levels, with the lowest concentration attained at the end of 9 hours. Thus, further investigation is needed to fully understand the mechanism behind the downregulation of HBD9 mRNA in response to bacterial infection. 

Uehara et al. ³² have demonstrated the involvement of TLR07/08 in the induction of HBD2 expression in several human epithelial cell line models. Similarly, bacterial DNA or synthetic oligonucleotides containing CpG motif (CpG ODN)–activated TLR09-dependent HBD2 expression has been reported in sinonasal ³³ and respiratory epithelial cells. ³⁴ In contrast to these studies, we noted an insignificant effect of TLR07/08 and TLR09 activation on HBD9 gene expression in hCECs. The role of these TLRs in the expression of other OS AMPs remains to be investigated. 

NOD-like receptors such as NOD1 and NOD2 are involved in the recognition of bacterial stimuli. ¹⁰ Similar to TLRs, NOD1 and NOD2 have been shown to trigger HBD2 expression. ^35,36 Mutation in NOD2 has been implicated in Crohn's disease and in the subsequent reduced expression of α-defensins in the Paneth cells of affected persons. ^37,38 In this study, we noted a significant role of NOD2 in HBD9 induction in hCECs. Recent studies ^39,40 have demonstrated that the TLRs and NLRs synergistically enhance the immune response of innate and adaptive immune cell types. Therefore, further studies are needed to fully understand the synergistic effect of NOD1 and NOD2 agonists on TLR-induced HBD9 expression in hCECs. 

Proinflammatory cytokines such as IL-1β and TNFα have been widely studied because of their pronounced effects on defensin and cathelicidin expression in a variety of surfaces. In addition to TLR02, ²⁶ IL-1β has been shown to induce AMPs at the human OS. ^41–43 In this study, we demonstrated a significant effect of IL-1β, but not of TNFα, in HBD9 expression in hCECs. Unlike TLRs, which induced expression at earlier time points, IL-1β induced HBD9 mRNA expression at 24 hours of treatment. Notably, IL-1β downregulated the expression of HBD9 from 3 to 9 hours. 

HBD9 is unique in that its expression is both downregulated and upregulated in a time-dependent manner in response to the activation of different TLRs and NLRs with specific PAMPs and IL-1β. The variable regulation of HBD9 expression may suggest that it may have other functions, such as an intermediary between innate and adaptive immune system and immunomodulation, in addition to the attributed antimicrobial effect. Furthermore, the finding of HBD9 in various layers of the ocular surface expands the body of knowledge related to this AMP and highlights the role that AMPs may play in ocular surface defense. 

Footnotes

 Supported by a British Eye Research Foundation–Fight for Sight PhD Studentship (IM) and by the Royal Blind Asylum and School/Scottish National Institution for the War Blinded and Royal College of Surgeons of Edinburgh.

Footnotes

 Disclosure: I. Mohammed, None; H. Suleman, None; A.M. Otri, None; B.B. Kulkarni, None; P. Chen, None; A. Hopkinson, None; H.S. Dua, None