Article Text

Human conjunctival epithelial cells express functional Toll-like receptor 5
  1. K Kojima,
  2. M Ueta,
  3. J Hamuro,
  4. Y Hozono,
  5. S Kawasaki,
  6. N Yokoi,
  7. S Kinoshita
  1. Department of Ophthalmology, Kyoto Prefectural University of Medicine, Kyoto, Japan
  1. M Ueta, Department of Ophthalmology, Kyoto Prefectural University of Medicine, Kajii-cho 465, Hirokoji-agaru, Kawaramachi-dori, Kamigyo-ku, Kyoto 602-0841, Japan; mueta{at}ophth.kpu-m.ac.jp

Abstract

Purpose: The expression and function of Toll-like receptor 5 (TLR5) was analysed in human conjunctival epithelial cells (HCjEC).

Methods: The expression of TLR5 in HCjEC was studied by reverse transcriptase (RT) PCR and flow cytometry. The amount of interleukin (IL) 6 and IL-8 proteins was determined by ELISA. Messenger RNA expression elicited by stimulation with flagellins derived from Pseudomonas aeruginosa, Serratia marcescens, Salmonella typhimurium, and Bacillus subtilis was assayed by quantitative RT–PCR. The localisation of TLR5 protein in human conjunctival epithelium was detected immunohistochemically.

Results: HCjEC expressed TLR5-specific mRNA and TLR5 protein. In HCjEC stimulated with flagellins derived from P aeruginosa and S marcescens, IL-6 and IL-8 production was increased and IL-6 and IL-8 mRNA was upregulated. Flagellins from S typhimurium and B subtilis did not induce the upregulation of these genes and proteins. TLR5 protein was detected on the basolateral but not the apical side of human conjunctival epithelium.

Conclusions: Human conjunctival epithelium harbours functional TLR5. Considering the spatially selective basolateral localisation of TLR5 protein, it was postulated that flagellins from ocular pathogenic bacteria induce inflammatory responses when disruption of the epithelial barrier permits their transmigration to the basolateral side but not under healthy physiological conditions on the ocular surface.

Statistics from Altmetric.com

The ocular surface consists of mucosal epithelium composed of corneal and conjunctival epithelial cells. As are other mucosal epithelia, the ocular mucosal epithelium is continuously challenged by pathogenic and non-pathogenic microbes. The mucosal epithelium is thought to represent a first line of defence against diverse microbes, not only by providing a physical barrier, but also by producing pro-inflammatory cytokines and antimicrobial peptides.1 Ocular surface epithelium such as corneal and conjunctival epithelium produce antimicrobial peptides similar to other mucosal epithelium.2 3

Toll-like receptors (TLR), pattern recognition receptors that sense conserved pathogen-associated molecular patterns, are the key receptors for the recognition of microbes.4 5 TLR were initially shown to be expressed and functional in immunocompetent cells such as macrophages and dendritic cells (DC); subsequently, their expression in a variety of cells including intestinal and airway epithelial cells was documented.6 7 Signalling through TLR induces the activation of transcription factors for antimicrobial genes and cytokines.8 TLR5 recognises bacterial flagellin,9 a component protein of bacterial flagella. Flagella are present in both Gram-positive and Gram-negative bacteria; they are essential for bacterial motility, invasion, and chemotaxis. Pseudomonas aeruginosa and Serratia marcescens are the bacteria associated with ocular surface with flagella.10 11

We previously reported that human corneal epithelial cells (HCEC) express TLR2 and TLR4 intracellularly but not on their cell surface, and that HCEC do not secrete interleukin (IL) 6 or IL-8 in response to peptidoglycan and lipopolysaccharide stimulation. This defective response may contribute to the immunosilent environment in the epithelium for commensal bacteria inhabiting the ocular surface.12 We documented that TLR3 on poly I : C-treated HCEC responded by producing pro-inflammatory cytokines and interferon β.13 Others1416 reported that functional TLR in HCEC play a pivotal role in innate immunity to pathogen-associated molecular patterns. Elucidation of the unique innate immune responses of HCEC that are distinct from those of immune-competent cells will lead to a better understanding of the host–commensal symbiosis on the ocular surface.

P aeruginosa flagellin elicits the inflammatory responses of HCEC in a TLR5-dependent manner.17 We reported that HCEC express TLR5 and secrete pro-inflammatory cytokines in response to flagellin derived from ocular bacteria that are pathogenic, but not to non-pathogenic ocular bacteria.18 We found that flagellin derived from S. typhimurium, a bacterium pathogenic in the intestine but not on the ocular surface, evoked a response by human intestinal epithelial cells but not HCEC.18

Whereas the critical homeostatic role of TLR in corneal epithelium has been discussed extensively in the literature, there are few reports on the role of TLR in conjunctival epithelium. Cook et al19 and Bonini et al20 documented the expression of TLR2, TLR4, and TLR9 in human conjunctival epithelial cells (HCjEC). The expression and function of TLR5 in conjunctival epithelium has not, however, been clarified yet.

In this study we confirmed the expression and function of TLR5 in HCjEC. We document that flagellins derived from P aeruginosa and S marcescens can trigger the innate immune response in HCjEC expressing TLR5, and TLR5 was selectively expressed on the basolateral but not the apical side of conjunctival epithelium.

MATERIALS AND METHODS

Human conjunctival epithelial cells

This study was approved by the institutional review board of Kyoto Prefectural University of Medicine, in Kyoto, Japan. All experimental procedures were conducted in accordance with the principles set forth in the Helsinki Declaration. The purpose of the research and the experimental protocol were explained to all patients and their informed consent was obtained.

For reverse transcriptase (RT) PCR, we obtained HCjEC from healthy volunteers by brush cytology. A tiny brush (Cytobrush S; Medscand AM, Malmo, Sweden) was used to scrape epithelial cells from the bulbar conjunctiva. For ELISA, real-time quantitative PCR, and flow cytometric analysis, we harvested primary HCjEC from conjunctival tissue obtained at conjunctivochalasis surgery.21 The cells were cultured using a modification of previously described methods.22 Briefly, conjunctival tissues were washed and immersed for 1 h at 37°C in 1.2 U ml−1 purified dispase (Roche Diagnostic Ltd, Basel, Switzerland). Epithelial cells were detached, collected, and cultured in low-calcium k-SFM medium supplemented with 0.2 ng ml−1 human recombinant epidermal growth factor (Invitrogen, Carlsbad, California, USA), 25 mg ml−1 bovine pituitary extract (Invitrogen), and 1% antibiotic-antimycotic solution. Cell colonies usually became obvious within three or four days. After reaching 80% confluence in seven to 10 days, the cells were seeded and after their subconfluence, they were used in subsequent procedures.

Bacterial flagellins used in this study

Flagellins derived from P aeruginosa and S marcescens were obtained from Inotek Pharmaceuticals (Beverly, Massachussetts, USA); flagellins from Salmonella typhimurium and Bacillus subtilis were from InvivoGen (San Diego, California, USA). S typhimurium flagellin is a strong pathogen that gains entrance into intestinal mucosa by penetrating enterocytes;23 it does not penetrate to the ocular surface mucosa. The pathogen S marcescens is implicated in approximately 5–10% of Gram-negative corneal ulcers related to contact lens wear.10 11 The pathogenic potential of B subtilis is reportedly weak or absent.24 P aeruginosa is an opportunistic ocular pathogen that can initiate a highly destructive corneal infection in humans10 11 and is involved in up to 70% of contact lens-related bacterial keratitis cases.25

Purification of mononuclear cells from peripheral blood

The procedures were described elsewhere.13 Briefly, we obtained peripheral venous blood samples from volunteers who had given their previous informed consent. Mononuclear cells isolated in lymphoprep tubes (Daiichi Pure Chemicals, Tokyo, Japan) were stimulated for 24 h with the different flagellins.

Reverse transcriptase PCR

We analysed HCjEC for TLR1–10 RNA expression as described in our previous study.13 Briefly, total RNA was isolated from HCjEC and human peripheral mononuclear cells using Trizol Reagent (Life Technologies, New York, New York, USA) according to the manufacturer’s instructions. For the RT reaction we used the SuperScript Preamplification Kit (Invitrogen). PCR amplification was with DNA polymerase (cTaq; Toyobo, Japan); the conditions were 38 cycles at 94°C for one minute, annealing for one minute, and 72°C for one minute on a commercial PCR machine (GeneAmp; Perkin-Elmer Applied Biosystems, Foster City, California, USA). The primers were as described in our previous study (table 1).13 RNA integrity was assessed electrophoretically in ethidium bromide-stained 1.5% agarose gels.

Table 1 The primer list of Toll-like receptor (TLR)

Enzyme-linked immunosorbent assay

Primary HCjEC were examined to quantify cytokine secretion as described in our previous report.18 Briefly, primary HCjEC were plated in 24-well plates and after reaching subconfluence were either left untreated or incubated for 24 h with each flagellin at a final concentration of 100 ng/ml or with 10 ng/ml human IL-1α (R&D Systems, Inc, Minneapolis, Minnesota, USA). The amount of IL-6 and IL-8 secreted in response to flagellin exposure increased with the flagellin concentration; we used 100 ng/ml. As early as 24 h, we found high levels of IL-6 and IL-8 in supernatants from flagellin-exposed cultures of primary HCjEC. IL-6 and IL-8 release into culture supernatants was quantitated using the Opteia IL-6 and IL-8 set (BD Pharmingen, San Diego, California, USA) according to the manufacturer’s instructions.

Flow cytometric analysis

HCjEC were analysed for TLR5 expression by flow cytometry as previously described.11 18 We purchased mouse anti-human TLR5 monoclonal antibody from Abcam (Cambridge, UK); it recognises an intracellular epitope in the cytoplasmic domain of TLR5.

Immunohistochemical study of TLR5 in human conjunctival sections

Serial sections of human conjunctiva were prepared from samples obtained at conjunctivochalasis surgery. These were fixed for 30 minures with methanol, incubated overnight in a moist chamber at 4°C with mouse anti-human TLR5 monoclonal antibody (Abcam) or isotype control mouse IgG2a (DakoCytomation, Kyoto, Japan), and washed in phosphate buffered saline. Alexa Fluor 488 goat anti-mouse IgG (H+L) (Molecular Probes, Eugene, Oregon, USA) was applied for 1 h at room temperature, the slides were washed and then antifade mounting medium with propidium iodide was applied (Vectashield; Vector Laboratories, Burlingame, California, USA).

Real-time quantitative PCR

Real-time quantitative PCR was performed on an ABI-prism 7700 (Applied Biosystems) according to previously described procedures.13 18 The initial amount of RNA used for reverse transcribing to complimentary DNA was approximately 1 μg and the cDNA was used at the original concentration for quantitative PCR. The stimulation time used in this study (1 h) was optimal for the maximum induction of IL-6 and IL-8 mRNA expressions (see supplemental fig 2 published online only).

Figure 2 Human conjunctival epithelial cells (HCjEC) express Toll-like receptor (TLR) 5 protein. Intracellular fluorescence-activated cell sorter analysis showed that TLR5 is expressed in HCjEC. As the TLR5 antibody recognises the intracellular domain of the TLR5 complex, we performed intracellular staining. Lymphocytes were the positive control. Histogram data are representative of three separate experiments (solid line, TLR5 antibody; dotted line, isotype control).

The primers and probes for human IL-6 (Hs00174131), IL-8 (Hs00174103) and human GAPDH (Hs 4326317E) were from Perkin-Elmer Applied Biosystems. Quantitative PCR was used to measure the expression of IL-6 and IL-8 mRNA in HCjEC treated for 1 h with flagellin derived from four different microbial bacteria (100 ng/ml) or human IL-α. The quantification data were normalised to the expression of the housekeeping gene GAPDH.

Data analysis

Data were expressed as the mean ± SE and evaluated by Student’s t-test using the Excel program.

RESULTS

TLR-specific mRNA and TLR5 protein expression in HCjEC

TLR1–10-specific mRNA expression was present in HCjEC harvested by impression cytology from healthy volunteers; TLR5-specific mRNA was expressed at levels comparable to mononuclear cells (fig 1). This finding demonstrates that the TLR5 gene is constitutively expressed in HCjEC. The TLR5 protein was also expressed by primary HCjEC harvested from conjunctival tissue at levels comparable to lymphocytes (fig 2).

Figure 1 Human conjunctival epithelial cells (HCjEC) express Toll-like receptor (TLR) 1–10-specific mRNA. Total RNA was isolated from human conjunctival epithelial cells. For the reverse transcriptase (RT) reaction, we used the SuperScript preamplification system. PCR amplification was with DNA polymerase; primers are listed in table 1. Human mononuclear cells were the positive control (P).

Primary HCjEC respond only to flagellin derived from ocular pathogenic bacteria

To assess whether the flagellins used were able to induce an inflammatory response in immunocompetent cells, we stimulated fractionated human peripheral mononuclear cells for 24 h with P aeruginosa, S typhimurium, B subtilis or S marcescens-derived flagellin. All flagellins induced the production of significant levels of the inflammatory cytokines IL-6, and IL-8 (fig 3A).

Figure 3 Production and mRNA expression of IL-6 and IL-8 in human conjunctival epithelial cells exposed to various flagellins. (A) To determine whether the flagellins induce an inflammatory response, fractionated human periphera mononuclear cells were stimulated for 24 h with P aeruginosa, S typhimurium, B subtilis, or S marcescens-derived flagellin at a concentration of 100 ng/ml. The supernatants were harvested and the concentration of IL-6 and IL-8 was measured. (B) To quantify inflammatory cytokine secretion, human conjunctival epithelial cells (HCjEC) were plated in 24-well plates. Upon reaching subconfluence they were left untreated or exposed for 24 h to flagellins derived from the four different microbial bacteria (100 ng/ml) or 10 ng/ml human IL-1a. The supernatants were harvested to measure IL-6 and IL-8 concentrations. (C) Quantitative reverse transcriptase PCR was used to measure the expression of IL-6 and IL-8 mRNA in HCjEC treated for 1 h with flagellin derived from four different microbial bacteria (100 ng/ml) or 10 ng/ml human IL-α. The quantification data were normalised to the expression of the housekeeping gene GAPDH. The Y axis shows the increase of specific mRNA over unstimulated samples. Data show the representative results from two independent experiments (A) or three independent experiments (B, C). Data represent mean ± SEM from the representative experiments with triplicated dishes (*<0.05, **<0.005).

Our results indicated that the flagellins derived from the four distinct microbes were able to induce pro-inflammatory cytokine production by human mononuclear cells. As we previously documented that HCEC produced IL-6 and IL-8 in response to flagellins derived from ocular pathogenic, but not non-pathogenic, bacteria,18 we investigated whether HCjEC exhibited a similar response. We found that in primary HCjEC, flagellins derived from P aeruginosa and S marcescens, but not those derived from S typhimurium and B subtilis elicited a significant increase in the secretion of IL-6 and IL-8 protein (fig 3B). The mRNA expression levels specific for IL-6 and IL-8 were considerably elevated in primary HCjEC stimulated with P aeruginosa and S marcescens-derived flagellin. In contrast, neither IL-6 nor IL-8-specific mRNA levels were significantly elevated in primary HCjEC stimulated with S typhimurium and B subtilis-derived flagellin (fig 3C). These results suggest that, like HCEC, primary HCjEC respond to ocular pathogenic but not to ocular non-pathogenic flagellins.

We decided on the dose for ELISA and real-time quantitative PCR according to our dose analysis and the recommended concentration of flagellin written in the manufacturer’s instructions. As per the manufacturer’s instructions, the recommended concentration to achieve TLR5 stimulation is 0.1–100 ng/ml, and the strongest response was achieved with 100 ng/ml P aeruginosa and S marcescens-derived flagellin (see supplemental fig 1 published online only).

TLR5 is expressed basolaterally in conjunctival epithelium

We subjected conjunctival tissues obtained at conjunctivochalasis surgery to immunohistochemical study to determine the presence and localisation of TLR5 expression in stratified conjunctival epithelium. TLR5 protein was consistently and abundantly expressed in human conjunctival epithelium. It was detected only at basal and wing sites, indicating its spatially selective presence on the basolateral but not the apical side (fig 4).

Figure 4 Localisation of Toll-like receptor (TLR) 5 in a human conjuncitval epithelium. TLR5 was detected by immunofluorescence staining. Frozen cryostat sections of a human conjunctiva were incubated with anti-TLR5 antibody (A, C). Isotype control incubation was the negative control (B, D). Bound antibodies were visualised by Alexa Fluor 488 goat anti-mouse IgG, nuclei by propidium iodide (PI) staining. Merged staining of TLR5 and PI shows no TLR5 staining associated with the apical layer of the epithelium.

DISCUSSION

This is the report of the expression and function of TLR5 on HCjEC. HCjEC stimulated with flagellins derived from P aeruginosa and S marcescens but not from S typhimurium and B subtilis induced the elevated production and mRNA upregulation of IL-6 and IL-8. TLR5 was selectively expressed on the basolateral but not the apical side.

The ability to detect microbes is an important task of the immune system. Exaggerated host defence reactions of the epithelium to endogenous bacterial flora may start and perpetuate inflammatory mucosal responses.26 Among ocular surface-related bacteria, only a few common causative ocular pathogens, ie P aeruginosa and S marscesens, have flagella; Staphylococcus epidermidis and Pseudomonas acnes, commensal ocular bacteria,27 28 do not. Most, if not all, of the responses to bacterial flagellin are thought to be mediated by TLR5.29 S typhimurium flagellin, strongly pathogenic and pro-inflammatory to intestinal epithelial cells,23 had little effect on HCjEC. Although B subtilis is one of the related species of Bacillus cereus, which is a major cause of severe keratitis, endophthalmitis, and panophthalmitis,30 B subtilis flagellin elicited no immune response in HCjEC. This suggests that TLR5 in HCjEC may discriminate specifically between bacteria that are ocular pathogenic and non-pathogenic.

Whereas P aeruginosa and S marcescens are common pathogens in keratitis,10 11 they do not usually cause conjunctivitis. Conjunctivitis attributable to these bacteria tends to be seen primarily in neonates and immunocompromised hosts.31 32 The lamina propria, abundant in immunocompetent cells and lymphatic organs, is located beneath the conjunctival epithelium.33 On the other hand, the avascular stroma under the corneal epithelium harbours few immunocompetent cells. The host defence in the conjunctival and corneal component may be different and this may account for the difference in the pathogens that result in keratitis and conjunctivitis.

Immunohistochemical data presented here and elsewhere18 indicate that TLR5 is basolaterally expressed in both conjunctival and corneal epithelium. Similarly, TLR5 is expressed on the basolateral surface of intestinal epithelium and detects invasive flagellated bacteria.29 TLR5 in ocular surface epithelium may be crucial for the sensing of invasive flagellated bacteria. It is possible that the flagellins of P aeruginosa and S marcescens induce an inflammatory reaction only if the integrity of the epithelial barrier is breached, thereby allowing their transepithelial transport to the basolateral site.

Interestingly, the ocular surface epithelium did not respond to flagellin derived from the enteropathogenic, ocular non-pathogenic bacterium S typhimurium. Murine intestinal lamina propria cells can differentiate between pathogenic and non-pathogenic bacteria.34 It remains unknown, however, how TLR can distinguish between pathogenic and non-pathogenic microbes. A collaboration between co-receptors and TLR may be a possible explanation; a receptor called fimbriae cooperates with TLR4 to distinguish between pathogenic and non-pathogenic bacteria in the urinary tract.35 Alternatively, there may be some minor molecular differences at the site of recognition in TLR5 expressed on different cell types. Mouse and human TLR5 discriminates between different flagellins, and the recognition site on TLR5 has been mapped to the diversified extracellular domain.36 Studies are underway in our laboratory to identify the molecular mechanisms that allow HCjEC to discriminate between flagellins derived from ocular pathogenic and non-pathogenic bacteria.

Another group recently reported that primary HCjEC express TLR5 and respond to S typhimurium-derived flagellin.37 Their result differs from the result obtained in our study. The cause of the difference is not evident but it might be due to the dose of flagellin used. That group used an extremely large amount of flagellin in their experiment. Our study confirmed that a much smaller amount (100 ng/ml) of S typhimurium-derived flagellin could upregulate IL-8 production in human colonic carcinoma cell line HT2918 but not HCjEC, and that the same amount of P aeruginosa-derived flagellin could also significantly upregulate IL-6 and IL-8 production in HCjEC.

We document that HCjEC possesses functional TLR5, and that conjunctival epithelial cells respond to exposure to flagellins derived from ocular pathogenic but not non-pathogenic bacteria by secreting pro-inflammatory cytokines. Considering the spatially selective localisation of TLR5 protein on the basolateral side of human conjunctival epithelium, an inflammatory reaction may be induced only when flagellin from ocular pathogenic bacteria cross the conjunctival epithelium to the basolateral side.

Acknowledgments

The authors would like to thank Ms C Mochida for her technical assistance.

REFERENCES

Supplementary materials

  • web only appendices 92/3/411

    Supplemental Figure 1
    Production of IL-6 and IL-8 in human conjunctival epithelial cells was dose dependent. Dose analysis was performed by 0.1ng/ml, 1ng/ml, 10ng/ml and 100ng/ml. The strongest response was achieved with 100ng/ml of P. aeruginosa-, and S. marcescense- derived flagellin. Data represents means ± SEM from the experiment with triplicated dishes.

    Supplemental Figure 2
    The relative expression of IL-6 and IL-8 mRNA in human conjunctival epithelial cells with time-course. Time analysis was performed at 0hr, 1hr, 3hr and 6hr after stimulation with 100ng/ml of P. aeruginosa- derived flagellin. The strongest response was achieved at 1hr after stimulation. Data represents means ± SEM from the experiment with triplicated dishes.

    Files in this Data Supplement:

Footnotes

  • Competing interests: Supported in part by grants-in-aid for scientific research from the Japanese Ministry of Health, Labour and Welfare, the Japanese Ministry of Education, Culture, Sports, Science and Technology, CREST from JST, a research grant from the Kyoto Foundation for the Promotion of Medical Science and the Intramural Research Fund of Kyoto Prefectural University of Medicine.

  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.