Toll-like receptor activation modulates antimicrobial peptide expression by ocular surface cells

Abstract

The ability of the ocular surface to respond to pathogens is in part attributed to toll-like receptors (TLRs) that recognize conserved motifs on various microbes. This study examines TLR expression on various ocular surface cells, if TLR agonists can modulate the expression of antimicrobial peptides (AMPs), human beta defensins (hBD) and cathelicidin (hCAP-18/LL-37) which maybe functionally active against Pseudomonas aeruginosa (PA) and if TLR agonists or AMPs can modulate TLR mRNA expression. TLR1-10 mRNA expression was examined in corneal epithelial, corneal stromal cells and conjunctival epithelial cells by RT-PCR. To confirm protein expression flow cytometry or immunostaining was performed for selected TLRs on some cell cultures. Ocular surface cells were cultured with a range of TLR agonists and then hBD-1, 2, 3, or hCAP-18 mRNA and protein expression was determined by RT-PCR and immunoblotting. In some experiments, cells were cultured with a cocktail of agonists for TLR3, 5 and 6/2 and the antimicrobial activity of the culture media was tested against PA. TLR mRNA expression was also examined in primary human corneal epithelial cells (HCEC) treated with either 3 μg/ml of hBD-2, 5 μg/ml of LL-37 or TLR4, 5 and 9 agonists. Overall, the ocular surface cells expressed mRNA for most of the TLRs but some differences were found. TLR2 was not detected in corneal fibroblasts, TLR4 was not detected in primary cultured or freshly isolated HCEC, TLR5 was not detected in conjunctival epithelial cells (IOBA-NHC) and corneal fibroblasts, TLR7 was not detected in freshly isolated HCEC and TLR10 was not detected in HCEC and IOBA-NHC. TLR8 mRNA was not expressed by any of the samples tested. Immunostaining of cadaver corneas revealed TLR5 and 9 expression throughout the cornea while TLR3 was significantly expressed only in the epithelium. Flow cytometry and immunostaining revealed cultured fibroblasts expressed TLR9 but had no significant TLR3 expression. hBD-2 expression was upregulated by TLR1/2, 3, 4, 5 and 6/2 agonists depending on the cell type, whereas only the TLR3 agonist upregulated the expression of hCAP-18 in primary HCEC. The combination of TLR3, 5 and 6/2 agonists in primary HCEC, upregulated hBD-2 and hCAP-18 mRNA and peptide expression and secretion into the culture media, which significantly killed PA. This antimicrobial activity was primarily attributed to LL-37. TLR agonists did not modulate TLR expression itself, however, LL-37 or hBD-2 downregulated TLR5, 7 and/or 9 mRNA depending on the cell type. TLRs are expressed on the ocular surface and TLR agonists trigger the production of LL-37 and hBD-2, with LL-37 being particularly important for protecting the ocular surface against PA infection.

Figure 1

TLR3 (A), TLR5 (B), and TLR9 (C) expression in the human cornea. Montages encompassing the epithelium (epi), stroma and endothelium (endo) were prepared from central cornea cryosections stained for TLRs. In each section the top panel is the isotype control, the middle panel is tissue stained for the specific TLR and the bottom two images are enlarged to show detail for the epithelium and stroma. Blue fluorescent DAPI was used to stain the nuclei. All images were taken at 200X magnification. Scale bars (shown only for TLR3) represent 40 microns. Results are representative of 3–4 corneas.

Figure 2

TLR3 and 9 protein expression in SV40 HCEC by flow cytometry (A) or in primary cultured fibroblasts by flow cytometry (B) and immunostaining (C). Flow cytometry histograms show the relative fluorescence on the x-axis and the number of events (cell count) on the y-axis. Dashed line represents the isotype control antibody while the solid line represents the specific TLR. Fibroblasts were immunostained for TLR3 (green) and TLR9 (red) with DAPI (blue) being used to stain the nuclei (C). Isotype controls were IgM and IgG for TLR3 and 9 respectively. Images were taken at 200X magnification. The data are representative of 2–3 experiments.

Figure 3

hBD-2 and hCAP-18 mRNA expression in response to TLR agonists in ocular surface cells. SV40 HCEC and IOBA-NHC cells were treated with various TLR agonists: Pam3CSK4 (TLR 1/2), HKLM (TLR 2), PolyI:C (TLR 3), LPS E. coli K12 (TLR 4), Flagellin S. typhimurium (TLR 5), FSL1 (TLR 6/2), Imiquimod (TLR 7), ODN2006 (TLR 9), or 10ng/ml of IL-1β or serum free media (M) for 24 hours. (A) Ethidium bromide stained agarose gel for hBD-2 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Real-time PCR was performed to determine the relative quantity of hBD-2 (B) and hCAP-18 (C) mRNA in primary HCEC. The data are representative of 2–4 independent experiments.

Figure 4

TLR agonists upregulate LL-37 and hBD-2 peptide production as detected by immunoblotting and ELISA respectively. Primary HCEC were treated for 20 hours with either media alone or a combination of TLR 3, 5 and 6/2 agonist which upregulate hBD-2 or LL-37 mRNA. (A) Representative immunoblot for LL-37 secreted in to the culture media (n=2). (B) The pixel intensity of each dot was used for semi-quantitative analysis. (C) hBD-2 protein production was quantitated in the culture media and cell lysate by an ELISA, n=3. A P-value of <0.05 (*) was considered to be statistically significant by Student’s t-test.

Figure 5

Antimicrobial activity against Pseudomonas aeruginosa (PA) of primary HCEC treated with TLR agonist cocktail is dependent on LL-37. (A) The antimicrobial activity of culture media from primary HCEC treated with TLR 3, 5 and 6/2 agonists was determined by a colony count assay. (B) To determine the activity of AMPs in culture media, synthetic LL-37 (3µg/ml) or rhBD-2 (10µg/ml) were incubated with culture media and PA and the percent bacteria killed was calculated with media treated representing no killing. (C) PA were incubated with culture media alone, synthetic LL-37 (3ng/ml) or with media from primary HCEC treated with TLR3, 5 and 6/2 agonists which had then been incubated with preimmune rabbit serum (serum) or a LL-37 blocking antibody (Ab) and the percent bacteria killed was calculated. The figures are representative of 2–6 independent experiments. When comparing two groups, an unpaired Student’s t test was used and a P-value of <0.05 (*) was considered to be statistically significant. For all others, a P-value <0.01 was considered to be significant by ANOVA with Bonferroni’s correction for multiple comparisons.