Cathelicidin antimicrobial peptide LL-37 in psoriasis enables keratinocyte reactivity against TLR9 ligands

Increased LL-37 is associated with elevated TLR9 in psoriatic epidermis

We hypothesized that the expression and function of TLRs by epidermal keratinocytes might be a critical element in amplifying the immune response. Measurement of TLR mRNA abundance in skin biopsies of psoriatic lesion revealed that TLR9, but not the other TLRs, was significantly elevated in psoriatic lesional skin compared to psoriatic non-lesional skin or lesional skin of atopic dermatitis (Figure 1a–1d), which are consistent with a previous report (Miller et al., 2005). TLR3 expression was significantly decreased in lesional skin of atopic dermatitis compared to psoriatic lesional skin. The increase of TLR9 expression in psoriatic lesional skin corresponded with increased TLR9 production in keratinocytes as immunohistochemical staining revealed that TLR9 protein expression was greatly enhanced in keratinocytes (Figure 1e–1g). TLR9-positive cells are also seen in the dermis of psoriatic lesional skin. Co-immunostaining revealed that cathelicidin was similarly located with TLR9 in the psoriatic epidermis (Figure 1h–1j).

Since immunohistochemical identification of cathelicidin in this study and prior studies (Frohm et al., 1997) has not defined the peptide form of cathelicidin present in psoriasis, and these various proteolytically processed forms are critical to the function of the peptide (Braff et al., 2005; Yamasaki et al., 2007), we performed surface-enhanced laser desorption-ionization time-of-flight mass spectrometry (SELDI-TOF-MS) on whole extracts of psoriatic lesional skin to define the peptide mass and thus its sequence. Unlike normal skin or skin from other inflammatory disorders such as rosacea (Yamasaki et al., 2007), all cathelicidin detectable in psoriasis was in the form of LL-37 (Figure 1k).

LL-37 induces TLR9 expression in keratinocytes

We next evaluated if the presence of LL-37 could influence the expression of TLR9 in keratinocytes. Cultures of normal human epidermal keratinocytes (NHEKs) were exposed to LL-37 at concentrations within the physiologic range of cathelicidin previously measured in psoriatic skin (Ong et al., 2002). Keratinocytes cultured with LL-37 increased expression of mRNA for TLR9, but not TLR1, TLR2 or TLR3 (Figure 2a), a pattern similar to the expression of TLRs found in psoriatic epidermis. TLR5 and TLR6 were not significantly increased, and TLR4, TLR7 and TLR8 were not consistently detectable in cultured keratinocytes (data not shown). We also examined the effect on TLR9 expression by inflammatory cytokines including IL-22, but under these culture conditions interferon-γ, IL-4, IL-17A, IL-22 or TNFα did not induce TLR9 mRNA in keratinocytes (Figure 2b). LL-37 induced TLR9 expression in a dose-dependent manner (Figure 2c), and a peak increase of TLR9 mRNA was observed 12 hours after the stimulation (Figure 2d). TLR9 protein increase paralleled TLR9 mRNA induction by LL-37 (Figure 2e–2g).

LL-37 enhances type I IFN expression from keratinocytes exposed to DNA

To evaluate the functional significance of keratinocyte TLR9, IFNA2 and IFNB1 mRNA abundance was measured in the presence of the TLR9 ligand CpG DNA. As previously reported (Miller et al., 2005), CpG by itself had minimal effect on keratinocytes but greatly increased type I IFNs (IFNA2 and IFNB1) mRNA in the presence of LL-37 (Figure 3a and 3b). Whole genomic DNA had a similar effect to simulate IFN expression in the presence of LL-37 (Figure 3c). The induction of IFNB1 expression was dose-dependent of LL-37 (Figure 3d). A corresponding increase of IFN-β protein was also confirmed by ELISA (Figure 3e, 3f) and the induction was observed 24 hours after the stimulation (Figure 3g). This effect was significantly inhibited by the TLR9 antagonist G-ODN (Figure 3h).

Psoriatic keratinocytes express IFN-β

The in vitro experiments predicted that epidermal keratinocytes affected by LL-37 would produce type I IFNs. As expected, keratinocytes in lesional skin from psoriatic patients showed abundant expression of IFN-β compared to keratinocytes from normal skin that had small amounts of LL-37 (Figure 4a–4c). IFN-β-positive cells were also observed in the upper dermis (Figure 4d–4f). A corresponding increase of IFNB1 mRNA was confirmed by real-time PCR of skin biopsies (Figure 4g).

Keratinocytes respond to sequential exposure to LL-37 and CpG, but not to LL-37 complexed to CpG

LL-37 and DNA have been previously proposed to form a complex between the peptide and nucleic acids to activate TLR9 in pDCs, thus LL-37 protects DNA from degradation and facilitates DNA entry to the early endosome (Lande et al., 2007). The methods used in Figure 3 relied on the addition of LL-37 to keratinocytes followed by sequential addition of CpG. This approach may have enabled formation of a complex between the cationic peptide and anionic DNA. Alternatively, LL-37 may have directly interacted with the keratinocyte membrane to enhance entry of CpG. To distinguish these events, LL-37 and CpG were first mixed together to allow complex formation. Assembly of these two molecules into a complex was confirmed by HPLC. CpG alone was seen as the expected single peak (Figure 5a). Once CpG was pre-incubated with LL-37 the assembly into a multimeric complex could be directly visualized by the disappearance of the single peak of CpG and the appearance of several broad peaks (Figure 5b). However, the CpG/LL-37 complex was not able to enhance type I IFNs production from keratinocytes. In contrast, sequential addition of LL-37 followed immediately by CpG at the same concentrations as used in the complex was required to stimulate keratinocyte IFN expression (Figure 5c).

We also examined if other cathelicidin antimicrobial peptides have similar capacity to increase IFNs production from keratinocytes. KS-30, a derivative of LL-37 that has previously shown the capacity to activate IL-8 expression in keratinocytes (Braff et al., 2005), showed a similar ability to increase type I IFN when added before CpG (Figure 5d). Several other cationic peptides including the murine cathelicidin peptide mCRAMP (GL-33), other human cathelicidin peptides such as FA-29 and DI-27, or the cationic but unstructured proline-rich antimicrobial porcine cathelicidin peptide PR-39 did not enhance type I IFNs induction. Although these cationic AMPs could not enhance type I IFNs in keratinocytes they were still able to bind and form complexes with CpG as observed by HPLC analysis (data not shown).

Overexpression of TLR9 enhances type I IFN induction by the ligands in keratinocytes

We further investigated the role of TLR9 in keratinocyte recognition of DNA by transient overexpression of TLR9 in keratinocytes. Plasmid DNA encoding human TLR9 (pUNO-hTLR9-HA) was transfected into keratinocytes. These cells (or controls) were then exposed to CpG or LL-37 then CpG. Transfection of pUNO-hTLR9 increased TLR9 mRNA expression several fold (Figure 6a). Addition of CpG alone increased IFNB1 approximately 3-fold in keratinocytes transfected with TLR9 (pUNO-TLR9) but not the control (pUNO-mcs) (Figure 6b). As seen previously, LL-37 treatment of control keratinocytes greatly enhanced IFNB1 expression by the sequential stimulation with CpG. (Figure 6c). This increase occurred to a significantly larger extent when cells were transfected with pUNO-TLR9-HA. Genomic DNA had a similar effect in the presence of LL-37 (Figure 6d).

Skin Biopsies

This study was approved by the Human Research Protection Program at the University of California, San Diego and performed in accordance with the Declaration and Helsinki Principles. Skin was collected from patients with psoriasis or atopic dermatitis and from healthy volunteers. For all procedures, informed consent was obtained. After the injection of local anesthesia, 2 or 4 mm punch biopsies were taken from the untreated lesional skin of individuals with psoriasis or atopic dermatitis and from healthy volunteers. All samples were stored at −80 °C until use.

Peptide synthesis

Cathelicidin peptides, LL-37, GL-33, PR-39, KS-30, FA-29 and DI-27 were commercially prepared by Synpep (Dorschner et al., 2001; Murakami et al., 2004; Yamasaki et al., 2007).

Cell culture and stimuli

NHEKs were cultured as previously described (Morizane et al., 2010). Cells were stimulated with synthetic cathelicidin peptides (LL-37, GL-33, PR-39, KS-30, FA-29, DI-27, 0.1–10 μM), IFN-γ (2 ng/ml, R&D Systems, Minneapolis, MN), IL-4, IL-17A, IL-22 (50 ng/ml, R&D Systems), TNF-α (50 ng/ml, Chemicon, Temecula, CA), type C CpG ODN M362 (2 μM, Invivogen), genomic DNA (10 μg/ml), or G-ODN (4 μM, Invivogen) in 24-well flat bottom plates (Corning Incorporated Life Sciences, Lowell, MA) for 3–72 hours. Genomic DNA was purified from HaCaT cells by Wizard Genomic DNA Purification Kit (Promega, Madison, WI). Pertussis toxin (100 ng/ml, Sigma-Aldrich) and AG1478 (2 μM, Calbiochem) were used to pre-incubate for 1 hour.

Plasmid DNA transfection

Transfection was performed by Amaxa Human Keratinocytes Nucleofector Kit (Lonza Cologne AG). NHEKs at 60–70 % confluency were harvested by trypsinization and 1 × 10⁶ cells were transfected with 3 μg plasmid DNA, pUNO-mcs (control plasmid) or pUNO-hTLR9-HA (Invivogen), following the protocol. Transfected cells were incubated in 6-well plates at 37 °C under standard tissue culture conditions for 24 hours. The cells were then stimulated with CpG or LL-37 then CpG or genomic DNA for 24 hours.

Quantitative real-time PCR (qRT-PCR)

Total RNA from skin samples and cultured keratinocytes was extracted, and cDNA was synthesized as previously written (Buchau et al., 2009). TaqMan Gene Expression Assays (Applied Biosystems ABI, Foster City, CA) were used to analyze mRNA expression as described by manufacturer instruction. GAPDH and 18s ribosomal RNA were examined as previously written (Morizane et al., 2010; Taylor et al., 2007). mRNA expression was calculated as relative expression to GAPDH mRNA (cultured cell samples) or 18S ribosomal RNA (skin samples), and data are presented as fold change against each control (mean of non-stimulated cells or mean of psoriatic non-lesional skins).

ELISA

IFN-β in culture media was measured by a commercial sandwich ELISA following the manufacturer’s instructions (PBL InterferonSource, Piscataway, NJ).

Immunostaining

For immunohistofluorescence, frozen sections (6 μm) were fixed with 4% paraformaldehyde, blocked with 3% BSA in PBS, and incubated with mouse monoclonal anti-TLR9 antibody (IMGENEX, San Diego, CA), rabbit anti-LL-37 antibody, or mouse monoclonal anti-IFN-β antibody at 4 °C overnight. Normal mouse and rabbit IgG were used as negative control. After washing with PBS, FITC-conjugated goat anti-mouse IgG antibody (Sigma) for TLR9 or IFN-β staining or FITC-conjugated goat anti-rabbit IgG antibody (Jackson Immuno Research LABORATORIES, INC) and AlexaFluor568-conjugated goat anti-mouse IgG antibody (Molecular Probes/Invitrogen) for double staining of TLR9 and LL-37 were used as second antibody. For immunocytofluorescence, keratinocytes were grown on chamber slides. After 4% paraformaldehyde fixation and blocking with 3% BSA in PBS, TLR9 staining was performed as described above. Sections were mounted in ProLong Gold Anti-Fade reagent with DAPI (Molecular Probes/Invitrogen). Images were obtained using an Olympus BX41 fluorescent microscope (Scientific Instrument Company, Temecula, CA).

Surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS)

Cathelicidin peptides were analyzed by SELDI-TOF-MS as previously written (Yamasaki et al., 2007).

Size-exclusion HPLC

Superdex^™75 Columns (Amersham Biosciences) and AKTA purifier HPLC system (Amersham Biosciences) were used for analysis. CpG (500 pmol) was pre-incubated with LL-37, GL-33 or PR-39 (2.5 nmol) in 100 μl PBS overnight. The samples were eluted with PBS at a flow rate of 1.0 mL/min. The detection wavelength was 214 nm.

Statistics

Results are expressed as the mean ± SEM. Student’s t test was used to determine significance. Mann-Whitney U test was used to compare IFN-β protein expression and mRNA expression in skin samples between each group. P < 0.05 was considered significant.