Citrullination alters immunomodulatory function of LL-37 essential for prevention of endotoxin-induced sepsis

Abstract

Cathelicidin LL-37 plays an essential role in innate immunity by killing invading microorganisms and regulating the inflammatory response. These activities depend on the cationic character of the peptide, which is conferred by arginine and lysine residues. At inflammatory foci in vivo, LL-37 is exposed to peptidyl arginine deiminase (PAD), an enzyme released by inflammatory cells. Therefore, we hypothesised that PAD-mediated citrullination of the arginine residues within LL-37 will abrogate its immunomodulatory functions. We found that when citrullinated, LL-37 was at least 40 times less efficient at neutralising the proinflammatory activity of LPS due to a marked decrease in its affinity for endotoxin. Also, the ability of citrullinated LL-37 to quench macrophage responses to LTA and Poly (I:C) signalling via TLR2 and TLR3, respectively, was significantly reduced. Furthermore, in stark contrast to native LL-37, the modified peptide completely lost the ability to prevent morbidity and mortality in a mouse model of D-galactosamine-sensitised endotoxin shock. In fact, administration of citrullinated LL-37 plus endotoxin actually exacerbated sepsis due to the inability of LL-37 to neutralise LPS and the subsequent enhancement of systemic inflammation due to increased serum levels of IL-6. Importantly, serum from septic mice showed increased PAD activity, which strongly correlated with the level of citrullination, indicating that PAD-driven protein modification occurs in vivo. Since LL-37 is a potential treatment for sepsis, its administration should be preceded by a careful analysis to ensure that the citrullinated peptide, is not generated in treated patients.

Introduction

Host defence peptides are evolutionarily ancient components of the innate immune system and act as effector molecules in host defence against pathogens (1). Various families of such peptides have been identified in different mammalian species, including about 30 cathelicidins; although humans express only one. Human cationic protein of 18 kDa (hCAP18) (2) is expressed in epithelial tissues and myeloid cells (3, 4). Thirty-seven carboxyl-terminal amino acid residues of hCAP18 comprise a cationic host defence peptide (LL-37), which has a broad spectrum of antimicrobial activity (5). In addition to its direct microbicidal role, LL-37 mediates other effects including cytotoxicity, chemotaxis, epithelial cell activation, angiogenesis, and epithelial wound repair (4, 6–8). Moreover, LL-37 is a potent regulator of innate immunity since it strongly modulates the responses of myeloid cells to pathogen-associated molecular patterns (PAMPs) (9). The positively charged amphipathic peptide, LL-37 (10, 11), interacts with the negatively charged LPS molecule, thereby inhibiting the binding of LPS to CD14⁺ cells. Thus, LPS-induced cytokine production is reduced (12). Inhibiting LPS is critical for abrogating the deleterious effects of certain infections, including intra-abdominal sepsis and endotoxemia (13). Since the biological activity of LL-37 is dictated by its charge (it is cationic), hydrophobicity, and amphipathicity (14, 15), it is anticipated that a change in these parameters will affect the physiological functions of the peptide. Such modifications can be caused by citrullination of the Arg residues within LL-37 by peptidylarginine deiminases (PAD), enzymes present alongside the peptide at infected/inflammatory sites.

PADs (EC 3.5.3.15) are enzymes that catalyse the deimination (citrullination) of arginine (Arg) residues, leading to post-translational modifications (16). Five different PADs (PAD1, 2, 3, 4, and 6) are present in humans, each showing a different tissue distribution (17–20). Apart from being involved in many physiological processes, PAD-driven citrullination is thought to play a role in some inflammatory states and diseases, including RA, multiple sclerosis, psoriasis, Alzheimer’s disease, primary open-angle glaucoma, and obstructive nephropathy (21, 22). The mechanism(s) underlying these changes in protein function may dependent upon the Arg residues that are essential for protein structure and/or function. The best examples involve chemokines (IL-8/CXCL8, CXCL10, CXCL11, and CXCL12), the citrullination of which profoundly affects their biological activity (23, 24). Significantly, citrullinated chemokines were identified in inflamed tissues in vivo, indicating the involvement of PAD2 and/or PAD4, which are expressed in myeloid cells (23, 24).

LL-37 contains five Arg residues and is therefore easily citrullinated in vitro by PAD2 (25). Since LL-37 is released in large amounts in response to inflammatory stimuli, it is highly plausible that it is a good substrate for PADs, which are secreted into the inflammatory milieu. However, little is known about how citrullination affects those biological activities of LL-37 that are relevant to its role in regulating innate immune responses to TLR agonists. To address this question, we examined how citrullinated LL-37 affected the interaction between LPS and human primary macrophages (hMDMs) or the murine macrophage cell line, RAW 264.7. We found that citrullinated LL-37 was unable to block the LPS-mediated activation of inflammatory macrophages. This was because citrullination led to a marked reduction in the affinity of LL-37 for LPS. This allowed LPS to bind CD14 on the macrophage cell surface and stimulate the expression of proinflammatory mediators. Citrullination of LL-37 also reduced its anti-inflammatory activity against other TLR agonists and host inflammatory mediators. Consistent with this, we found that citrullination of LL-37 abrogated its ability to prevent endotoxic shock in a mouse model.

Materials and Methods

Reagents

Gentamycin, endotoxin (LPS from E. coli O26:B6), lipoteichoic acid (LTA), poly(I:C) and Griess reagents were from Sigma (St. Louis, MO, USA). Fetal calf serum (FBS), RPMI-1640, DMEM, calcium- and magnesium-free PBS (without Ca²⁺ and Mg²⁺), penicillin-streptomycin (PEST) and lymphocyte separation medium were obtained from PAA (Germany). Flagellin, R848 and Pam3CSK4 were purchased from Enzo Life Science (NY, USA). Macrophage-activating lipopeptide (MALP-2) was obtained from Imgenex (San Diego, CA, USA). The purity of TLR agonists was estimated for flagellin 95% (SDS-PAGE), LPS 80%, MALP-2 95% (HPLC), poly(I:C) 99% (thin layer chromatography) and LTA 97% according to the manufacturer’s statement. All agonists, except LPS, were endotoxin free. Recombinant human PAD2 and PAD4 were obtained from Modiquest (The Netherlands).

Peptide synthesis and purification

LL-37 and citrullinated forms of LL-37: LL-37₇ (1 Cit), LL-37_7,29,34 (3 Cit) and LL-37_{all cit} (5 Cit) were assembled using the Fmoc solid-phase peptide synthesis approach using either model 433A (Applied Biosystems, Foster City, CA, USA) or model Liberty (CEM Corporation, Matthews, NC, USA) automated peptide synthesizers followed by cleavage in trifluoroacetic acid (TFA)/phenol/thioanisole/ethanedithiol/water (10 mL: 0.75 g: 0.5 mL: 0.25 mL: 0.5 mL) mixture at 25°C for 90 min, see Barlow et al. for details (26). The peptides were purified by RP-HPLC (>98% purity), and their masses were confirmed by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry. Following lyophilisation, the peptides were obtained in the form of their trifluoroacetate salts. Stock solutions (10 mg/mL) were prepared in phosphate-buffered saline stored in aliquots at −20°C. To confirm the purity of synthetic peptides the LAL (Limulus Amebocyte Lysate) test purchased from Lonza (Germany) was performed. The sequences of LL-37 peptides, which were used in this work, are shown in Table I.

Table I.

The sequence of antimicrobial peptides, which were used in this paper. Citrulline residues (Cit) are bolded.

LL-37 peptide	Sequence
LL-37	LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES
LL-377	LLGDFF(Cit)KSKEKIGKEFKRIVQRIKDFLRNLVPRTES
LL-377,29,34	LLGDFF(Cit)KSKEKIGKEFKRIVQRIKDFL(Cit)NLVP(Cit)TES
LL-37all cit	LLGDFF(Cit)KSKEKIGKEFK(Cit)IVQ(Cit)IKDFL(Cit)NLVP(Cit)TES
LL-37scrbl	RSLEGTDRFPFVRLKNSRKLEFKDIKGIKREQFVKIL

Citrullination by human PAD2 and PAD4

Peptides were diluted to a concentration of 1 mg/mL in PAD assay buffer (100 mM Tris HCl, 10 mM CaCl₂, 5 mM DTT, pH 7.6) and incubated with recombinant human PAD2 or PAD4 (Modiquest, The Netherlands) at a concentration up to 23.3 units/mg peptide, for 2 h at 37°C. Citrullination was terminated by sample dilution in RPMI-1640 with 10% HBSS. Control samples were treated similarly, apart from the addition of PAD.

Cell Culture

PBMCs were isolated from human blood obtained from the Red Cross, Krakow, Poland. The Red Cross de-identified blood materials as appropriate for human subjects confidentiality assurance. Thus, the current paper adheres to appropriate exclusions from human subjects approval. Briefly, PBMC were isolated from EDTA-treated blood using a Lymphocyte Separation Medium (PAA) density gradient yielding the fraction highly enriched in monocytes (90% CD14-positive) as described previously (27). Cells were plated at 3×10⁶/well in 24-well plates (Sarstedt, Germany) in RPMI1640 (PAA) supplemented with 2 mM L-glutamine, 50 μg/mL gentamicin (Sigma), and 10% autological human serum. After 24 h, nonadherent PBMCs were removed by washing with complete medium, and adherent cells were differentiated to hMDMs in this medium for 7 days with fresh medium changed every 2 days. The phenotype of each batch of hMDMs was routinely controlled, after non-enzymatic detachment of cells, by immunofluorescent staining of CD14 (clone: TUK4, DakoCytomation Denmark A/S, Glostrup, Denmark), CD16 (clone: DJ130c, DakoCytomation), CD11b (clone: ICRF44, Becton Dickinson and Co., Franklin Lakes, USA), and CD209 (clone: DCN46, Becton Dickinson) and subsequent flow cytometry analysis (Supplemental Figure 1). The cultures selected for further experiments were positive in at least 90% for the first three markers and less than 1% for CD209. The adherent cells acquired typical macrophage morphology and resting (nonstimulated) cells did not produce inflammatory cytokines: TNF-α, or IL-6.

The murine macrophage cell line RAW 264.7 obtained from American Type Culture Collection was maintained in DMEM (PAA) supplemented with 5% FBS and PEST (100 U/mL penicillin and 100 U/mL streptomycin) at 37°C in a humidified 5% CO₂ atmosphere. Cells were passaged every 2–4 days using cell scraper.

LPS effects on mouse macrophages in vitro

RAW 264.7 cells (3.5 × 10⁵) were seeded in 96-well tissue culture plates in 100 μL of phenol red-free DMEM/5% FBS/PEST. After circa 6 h of incubation to permit adherence, medium was changed and cells were stimulated with 10 ng/mL LPS from E. coli (Sigma), in the presence of 0.01–0.1 μg/ml or absence of native LL-37, LL-37 treated with PAD2 or PAD4 or synthetic citrullinated forms of LL-37. Stimulation was performed in a total volume of 200 μL of DMEM/10% FBS/antibiotics, in triplicates. The level of NO in culture supernatants was determined 20 h after stimulation using the Griess reaction. Nitrite (NO₂⁻), a stable product of NO degradation, was measured by mixing 50 μL of culture supernatants with the same volume of Griess reagent (Sigma) and the absorbance at 540 nm was measured using spectrophotometer. Phenol red-free DMEM with FBS and antibiotics were used as a blank. Standard curve was prepared using 0–80 μM sodium nitrite solutions in dH₂O.

Supernatants from stimulated RAW 264.7 cells were simultaneously analysed for TNF-α and IL-6 content by ELISA (BD Bioscience, San Jose, CA, USA) assay according to the manufacturer’s instructions.

Binding of LPS to RAW 264.7 cells

Effect of LL-37 on LPS binding to cells was determined according to the procedure described by Nagaoka et al. (12). To this end RAW 264.7 cells (10⁶ cells/mL) were incubated with Alexa488-conjugated LPS (100 ng/mL) in DMEM supplemented with 5% FBS at 37°C for 15 min in the absence or presence of LL-37 or citrullinated LL-37 (LL-37_{all cit}). Cells were then washed twice with ice-cold PBS, and the LPS binding was analysed by flow cytometry (FACScan, Becton Dickinson). The mean fluorescence intensity and percentage of cells labelled with Alexa488-conjugated LPS was measured in each group.

Response of macrophages to stimulation with TLR agonists

hMDMs or RAW 264.7 cells were incubated in RPMI or DMEM supplemented with 5% of serum (human serum or FBS) for various lengths of time (6 h for hMDMs, 20 h for RAW 264.7 cells) with the following TLR agonists: LPS (10 ng/mL), lipoteichoic acid (LTA, 10 μg/mL), R848 (10 μM), Pam3CSK4 (1 μg/mL), MALP-2 (10 ng/mL), poly(I:C) (10 μg/mL) or flagellin (100 ng/mL) with or without native LL-37 or full citrullinated LL-37 (LL-37_{all cit}). As cell stimulation read-out we have determined the level of NO (the Griess reaction) and TNFα (ELISA) released into media at 20 h and 6 h post-stimulation by RAW 264.7 and hMDMs cells, respectively.

D-Galactosamine-sensitized endotoxin shock model

A D-galactosamine-sensitized mouse model (28), which is highly susceptible to LPS, was utilized to investigate ability of LL-37 and LL-37_{all cit} to suppress inflammatory reaction in vivo. Male Balb/C mice (8–10 weeks, 22–25 g obtained from Jackson Labs., (USA) were intra-peritoneally injected with 100 μL of D-galactosamine (1.2 mg/g, dissolved in saline) or D-galactosamine + LPS (0.1 μg/g) with or without of native or modified LL-37 (10 μg/g). Survival rates and overall assessment score (OAS) ranging from 0 (normal) to 8 (death) were monitored every 1–2 h for first 24 h post injection and then at 48 h and 72 h later. Reduced motor activity, lethargy, shivering, and piloerection were recorded as symptoms of sepsis. Each of this condition was scored as 0 (no observable symptom), 1 (a noticeable symptom) and 2 (a severe symptom).

For the cytokine assay, CRP quantification and analysis of peritoneal inflammation, mice were sacrificed 6 h post LPS challenge. Blood and peritoneal cells were collected by cardiac puncture and lavage, respectively. The cytokines IL-6, IL-10, MCP-1, INF-γ, TNF-α, and IL-12p70 were measured in plasma using the Cytometric bead array; mouse inflammation kit (Becton Dickinson AB) according to the manufacturer’s instructions.

Analysis of the content of citrulline proteins and PAD activity

The citrulline presence (both free and incorporated in polypeptide chains) and PAD activity in blood serum were determined using the modified Boyde and Rahmatullah method, based on the chemical modification of the citrulline side-chain and colorimetric detection of the derivative (29). Briefly, for the activity measurement the 1 μL of serum was mixed with 9 μL of PBS in a 96-well microtiter plate followed by addition of 40 μL of 10 mM Benzoyl-Arg-Ethyl Ester (Sigma) in 0.1 M Tris, 10 mM CaCl₂, 5 mM DTT. A plate was incubated for 1 h at 55°C. After incubation, the enzymatic reaction was stopped by the addition of 10 μL of 5 M of HClO₄. The colour was developed by adding 150 μL of freshly prepared 1:2 mixture of solution A (0.5% diacetyl monoxime, 0.01% thiosemicarbazide in water) and B (0.25 mg/mL FeCl₃, 24.5% H₂SO₄, 17% H₃PO₄ in water) to each well, followed by incubation of the plate at 110°C for 17 min and OD_{535 nm} was measured using SpectraMax microplate reader (Molecular Devices, CA, USA) against a sample blank. Sample blanks were incubated in parallel and contain serum and all reagents except the substrate (Benzoyl-Arg-Ethyl Ester). In the same time the read-out of the sample blanks against a reagent blank (samples containing all reagents but without the substrate and serum) reflects the basal level of citrullination (free and peptidyl citrulline) in each serum sample. Free L-citrulline in the concentration range of 1–100 nmol per well was used to prepare a calibration curve. The enzymatic activity was defined as production of 1 μmol of citrulline within 1h incubation at 55 °C.

Ethics statement

All studies were performed in accordance with European Union regulations for the handling and use of laboratory animals. The protocols were approved by the institutional Animal Care and Use Committees (Jagiellonian University, Krakow, Poland, permit No. 11/2013).

Statistical analyses

Statistical comparisons were performer with Prism 4.0 Software (Graph-Pad), using two-tailed Student’s t-test for comparisons of two data sets, and ANOVA for multiple comparisons. The survival curve was analysed using the Mantel–Cox test. A p-value of <0.05 was considered to be significant.

Results

The in vitro citrullination of LL-37 by PAD2 and PAD4 abrogates its anti-inflammatory properties

The cationic character of LL-37 facilitates electrostatic interactions with negatively charged molecules, such as the anionic lipid A domain of endotoxin (12). LL-37 binds to LPS and prevents it from binding to receptors expressed by monocytes and macrophages, a process that initiates inflammatory responses. Therefore, we tested whether the conversion of cationic Arg residues to neutral peptidyl citrulline residues, a reaction catalysed by PADs, influences the anti-inflammatory properties of LL-37. Human MDMs and RAW 264.7 cells were stimulated for 6 and 20 h with LPS (10 ng/mL) in the presence of either native LL-37 or the PAD2- or PAD4-treated peptide. As shown in Fig. 1, pre-incubating LL-37 with PAD2 or PAD4 removed its ability to neutralise LPS and quench the release of TNF-α (Fig. 1A) and NO (Fig. 1B) by macrophages. Neither of the PADs alone, nor LL-37 alone, had any effect on TNF-α and NO production at the concentrations used here.

Fig. 1. Citrullination abolishes the ability of LL-37 to quench the proinflammatory activity of LPS.

Human MDMs (A) and mouse RAW 264.7 (B) macrophages were stimulated with 10 ng/mL LPS in the presence of native or citrullinated LL-37 at the indicated concentrations (0.1–10 μg/mL). Citrullinated LL-37 was obtained by treatment of the native peptide with human PAD2 or PAD4 at 23.3 U/mg peptide). The level of TNF-α (A) and NO (B) in the culture supernatants was determined using ELISA or the Griess assay at 6 or 20 h post-stimulation, respectively. Since neither the LL-37 nor the PAD enzymes alone induced the release of NO or TNF-α, for the sake of clarity these controls are not shown in the figure. Data represent the mean ± SD of three independent experiments. ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001.

The degree to which the ability of LL-37 to quench the proinflammatory activity of LPS is impaired depends on the level of citrullination

Depending on the incubation time and enzyme concentration, deimination of the Arg residues within LL-37 generates different levels of citrullination as shown by amino acid sequence analysis of the peptide incubated with PAD2, PAD4 and rabbit PAD (Supplemental Figure 2). In concordance with observation by Kilsgard and colleagues Arg residues in position 7 was the most susceptible to citrullination followed by Arg29 and Arg34 (25). Since the level of citrullination may variably affect the ability of the peptide to neutralise LPS we verify whether abrogating of the anti-inflammatory activity of LL-37 depends on the degree of citrullination. To this end we stimulated RAW 264.7 cells with LPS alone (10 ng/mL) or with LPS in the presence of native LL-37 or peptides harbouring different numbers of citrullinated Arg residues: Arg7Cit (LL-37₇), Arg7,29,34Cit (LL-37_7,29,34) or Arg7,19,23,29,34Cit (LL-37_{all cit}) (Table I). We found that both LL-37 and LL-37₇ completely quenched NO release [Fig. 2A] and inhibited TNF-α and IL-6 secretion [Fig. 2B, C] by LPS-stimulated murine macrophages. By contrast, LL-37_7,29,34 and LL-37_{all cit} only partially blocked the proinflammatory activity of LPS. The strongest ablation of LPS-neutralising activity was observed with the fully citrullinated peptide [Fig. 2A]. A scrambled peptide (LL-37_scrb), which was used as a control, showed no effect on LPS-mediated secretion of proinflammatory factors. Complete citrullination of LL-37 also attenuated its ability to interfere with the LPS-stimulated secretion of TNF-α by hMDMs [Fig. 2D]. The effect was more significant in the presence of autologous serum [Fig. 2E] and was also observed in the presence of serum obtained from human AB blood [Fig. 2F]. This suggests that post-translational modification by PADs can ablate the anti-inflammatory activity of LL-37 against human cells. To further investigate effect of citrullination on LL-37 functions, we used the fully citrullinated form of the peptide (LL-37_{all cit}).

Fig. 2. The degree of impairment in the anti-inflammatory properties of LL-37 depends on the level of citrullination.

The effect of citrullination on the ability of LL-37 to quench the proinflammatory effects of endotoxin was estimated by measuring NO release and IL-6 and TNF-α secretion by RAW 264.7 cells and TNF-α from human (hMDMs) macrophages. Cells were stimulated with 10 ng/mL LPS in the presence of native LL-37 (10 μg/mL) or citrullinated synthetic forms of LL-37 (10 μg/mL; LL-37_7, LL-37_7,29,34, or LL-37_{all cit}). A scrambled peptide (LL-37_scrb; comprising 37 amino acids residues found in LL-37 but randomly arranged) was used as a control. The NO level in the culture medium of RAW 264.7 cells at 20 h post-stimulation was determined using the Griess assay (A). The amounts of IL-6 (B) and TNF-α (C) secreted by RAW 264.7 cells and the amount of TNF-α secreted by human macrophages (D–F) were measured by ELISA at 6 h post-induction. Since none of the LL-37 peptides alone induced the release of NO or TNF-α, for the sake of clarity these controls are not shown in the figure. Data represent the mean ± SD of three independent experiments. ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001.

Citrullination affects the ability of LL-37 to inhibit endotoxin-mediated cell stimulation

To examine the different abilities of native and modified LL-37 to quench the proinflammatory activity of LPS, we measured the inhibition (IC₅₀) of LPS-induced macrophage stimulation. Human MDMs or RAW 264.7 cells were stimulated with either LPS alone or with LPS plus different concentrations of native LL-37 or LL-37_{all cit}. LL-37 strongly inhibited endotoxin-mediated TNF-α or NO secretion in a dose-dependent manner (IC₅₀ = 5.21 μg for hMDMs and IC₅₀ = 3.4 μg for RAW 264.7 cells) [Fig. 3 A, B]. By contrast, the citrullinated peptide was far less effective (IC₅₀ = 25.4 μg for hMDMs and IC₅₀ = 143 μg) [Fig. 3 A, B]. Even at very high non-physiological concentrations of LL-37_{all cit} (200 μg/mL = 44.4 μM), the level of NO release by control cells was still 36%. Taken together, these data indicate that, compared with the native peptide, citrullinated LL-37 is at least 5 (in case of hMDMs) or 40 (in case of RAW 264.7) times less potent at inhibiting LPS-mediated macrophage stimulation.

Fig. 3. Native and citrullinated LL-37 show a marked difference in their ability to inhibit endotoxin activity.

hMDMs (A) or RAW 264.7 cells (B) were stimulated with 10 ng/mL LPS plus different concentrations of the native (LL-37) or citrullinated (LL-37_{all cit}) peptides (1–200 μg/mL). The level of TNF-α (A) or NO (B) in the culture supernatants at 6 h or 20 h post-stimulation was determined using ELISA or the Griess assay. Since none of LL-37 peptides alone induced the release of NO or TNF-α, for the sake of clarity these controls are not shown in the figure. Data represent the mean ± SD of three independent experiments.

Citrullination impairs the ability of LL-37 to prevent LPS binding to the cell surface

The reduced ability of LL-37 to block the proinflammatory activity of LPS appears to be related to a reduction in the affinity of the citrullinated peptide for endotoxin; thus endotoxin can still interact with cell surface receptors (12). We next examined the effects of LL-37_{all cit} on the binding of FITC-conjugated LPS to CD14⁺ RAW 264.7 cells by flow cytometry. When used alone, FITC-LPS clearly bound to the cell surface. This interaction was not affected by low concentrations of LL-37 (0.4 μg/mL), but was inhibited by 2 μg/mL LL-37 and was almost completely blocked by 10 μg/mL LL-37 [Fig. 4A]. Conversely, the binding of FITC-LPS was only slightly inhibited by the citrullinated peptide [Fig. 4A]. Even at the highest concentration tested (10 μg/mL LL-37_{all cit}), 75% of cells were FITC-LPS+. By contrast, when cells were incubated with FITC-LPS in the presence of native LL-37, only 7.5% cells were positive [Fig. 4B]. Taken together, these results reveal the mechanism underlying the observed reduction in the anti-inflammatory activity of citrullinated LL-37.

Fig. 4. Citrullination of LL-37 prevents the peptide from blocking the binding of FITC-conjugated LPS to RAW 264.7 cells.

RAW 264.7 cells (5 × 10⁵ cells/mL) were incubated with 100 ng/mL FITC-conjugated LPS (FITC-LPS) in the absence or presence of LL-37 or LL-33_{all cit} peptide (0.4–10 μg/mL) in RPMI 1640 containing 10% FBS for 20 min at 37°C. After washing, the binding of FITC-LPS was analysed by flow cytometry. Background fluorescence was determined using RAW 264.7 cells incubated without FITC-LPS. (A) Representative histograms from one of three independent experiments are shown. (B) The mean percentage of macrophages that bound LPS. Data are expressed as the mean ± SD of three separate experiments. ns, not significant; ***, p<0.001.

Citrullination abolishes the ability of LL-37 to prevent the mortality and morbidity associated with septic shock

Neutralisation of LPS by LL-37 prevents the activation of CD14/toll-like receptor-4-expressing cells and protects mice from endotoxic shock. We used the well-established D-galactosamine-sensitised mouse model (28) to examine the ability of citrullinated LL-37 to prevent the lethal effects of LPS. Intraperitoneal (i.p.) injection of D-galactosamine sensitised mice to the lethal effects of LPS; indeed, 90% of the sensitised mice died within 17 h post-injection (hpi) of LPS (0.1 μg/g) [Fig. 5A]. However, the administration of LL-37 (10 μg/g) completely prevented LPS-induced mortality (100% of animals survived). By stark contrast, LL-37_{all cit} provided no protection at all, and all animals were dead within 16 h pi [Fig. 5A]. An analysis of the overall assessment scores (OAS) at 6 h and 10 h post-infection revealed a significant greater deterioration in the health of animals injected with LPS plus LL-37_{all cit} then with LPS alone [Fig. 5B]. There were no significant differences between these two groups of mice at later time points. Remarkably, animals co-injected with native peptide (10 μg/g) showed no signs of morbidity (according to the OAS). The OAS for mice injected with LPS and native LL-37 was very low and not significantly different from that for control animals (mice injected with native LL-37 alone or with vehicle (D-galactosamine) alone) over the 16 h study period [Fig. 5B].

Fig. 5. Citrullinated LL-37 increases the susceptibility of mice to endotoxic shock.

D-galactosamine (1.2 mg/g)-sensitised Balb/C mice were injected intraperitoneally (i.p.) with LPS (0.1 μg/g) without or with LL-37 or LL-37_{all cit} (10 μg/g). Mice injected with D-galactosamine alone or with each LL-37 peptide alone were used as controls. The survival rate and overall assessment scores (OAS) were estimated at 6–17 h post-injection. Data are expressed as the mean value for each group (n=8–10/mice group). (A) Mortality is expressed as a percentage. The mortality in the control groups was zero. The survival of mice injected with LPS plus LL-37 or with LPS plus LL-37 _{all cit} was compared. Survival statistics were calculated using the Mantel–Cox test. **, p<0.01; (n = 10 mice per tested group). (B) Overall assessment scores (OAS) expressed as the mean ± SEM (n = 10 mice per tested group). In a separate experiment, mice were sacrificed 6 h after i.p. injection of LPS and/or the LL-37 peptides or buffer. Peritoneal lavage was then performed and blood samples collected (n = 8 per group). (C) Influx of inflammatory cells into the peritoneum. Data are expressed as the mean ± SEM, ***, p<0.001; (n = 10 per group). (D) The levels of C-reactive protein in mouse serum. Data are expressed as the mean ± SEM. *, p<0.05 (n = 10 per group). (E) Serum levels of TNF-α, IL-6, IL-10, MCP-1, IL-12p70 and IFN-γ measured using a cytokine bead array system. Values were compared between animals injected with LPS, LPS/LL-37, or LPS/LL-37_{all cit}. The levels of each individual cytokine in the control mice are illustrated by the dotted grey lines. Data are expressed as the mean ± SEM. ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001 (n = 10 per group).

Examination of the peritoneal cavity revealed a significant increase in inflammatory cell infiltration into the peritoneum of animals treated with LPS plus LL-37 compared with that in animals injected with LPS plus LL-37_{all cit}. In the latter group, the number of inflammatory cells was similar to that in non-treated controls and mice injected with LPS alone [Fig. 5C].

Serum analyses revealed that animals injected with LPS showed a significant increase in the levels of CRP [Fig. 5D] and five different cytokines [Fig. 5E] compared with control mice (the mean level of cytokines in control animals is marked by dotted grey lines). The exception was IL-12p70, which showed no difference in blood concentration between the groups. Importantly, the injection of LPS plus native LL-37 led to a significant reduction in CPR and cytokine levels. This was not observed upon injection of LPS plus the citrullinated peptide, clearly indicating the lack of any ability to prevent LPS-induced increases in CRP, IL-6, IFN-γ and IL-10 [Fig. 5D, E]. The results clearly demonstrate that citrullination of LL-37 abrogates its anti-endotoxic effects in vivo, which are maintained both locally and systemically by the native peptide (13).

Both PAD activity and the level of citrullination are higher in mouse serum under septic conditions

The data we have gathered thus far strongly argue that citrullination of LL-37 alters the immunomodulatory function of the peptide which are essential for prevention of endotoxin-induced sepsis. To further support this theory, we next determined the levels of PAD activity and the content of citrulline (both free and encompassed in proteins/peptides) in mouse serum. We found that serum from mice exposed to endotoxin showed significantly (p<0.01) higher PAD activity than that from healthy control animals [Fig. 6A]. Consistent with this, the level of citrullination in serum collected from septic mice was markedly higher than that in control mice [Fig. 6B], and the levels of both showed a positive reciprocal correlation [Fig. 6C]. Collectively, these data clearly show that the increases in PAD activity and in the amount of citrullinated proteins/peptides in the blood are associated with sepsis.

Fig. 6. PAD activity and citrullination are increased in septic mouse serum.

D-galactosamine (1.2 mg/g)-sensitised Balb/C mice were injected intraperitoneally with LPS (0.1 μg/g). Six hours later, the animals were sacrificed and blood samples taken. Mice injected with D-galactosamine alone were used as controls. Data are expressed as the mean values for ten mice per group. Serum (1 μL) was tested for PAD enzyme activity (A) and the presence of citrulline (B) as described in Materials and Methods. Data are expressed as the mean ± SEM. **, p<0.01;***, p<0.001. (C) There is a positive correlation between PAD activity and the level of citrullinated proteins in the serum. The correlation coefficient (r) and p-values are indicated.

Citrullination abolishes the ability of LL-37 to down-regulate cellular responses to TLR ligands and host inflammatory mediators

In addition to the effects of LPS, LL-37 also suppress the proinflammatory effects of other bacterial TLR ligands, such as LTA, and those of endogenous inflammatory mediators, including IFN-γ (30). Therefore, we next investigated the effects of LL-37 citrullination on cellular responses to TLR agonists and IFN-γ. Human macrophages, which are sensitive to all types of TLR agonist (31), were treated with seven different TLR agonists in the presence of native and citrullinated LL-37 and the secretion of TNFα was measured. We found that citrullination inhibited the anti-inflammatory effects of LL-37 against LTA and Poly I:C [Fig. 7A], but had no effect on signalling elicited by others agonists. Moreover, compared with the native peptide, the ability of LL-37_{all cit} to inhibit the production of TNFα by hMDMs and NO by RAW 264.7 macrophages stimulated with IFN-γ alone or with IFN-γ plus LPS was markedly inhibited [Fig. 7B, C]. Taken together, these findings suggest that the citrullination of LL-37 not only eliminates its anti-endotoxin activity but also prevents it from neutralising other inflammatory mediators.

Fig. 7. Citrullination abolishes the ability of LL-37 to quench the proinflammatory activity of both human and murine macrophages induced by bacterial antigens and host mediators.

(A) hMDMs were stimulated with TLR ligands (LPS (100 ng/mL), Pam3CSK4 (1 μg/mL), LTA (10 μg/mL), MALP-2 (10 ng/mL), poly(I:C) (10 μg/mL), flagellin (100 ng/mL), and R848 (10 μM)) either alone or in the presence of LL-37 or LL-37_{all cit} (both at 10 μg/mL) for 6 h. The level of TNF-α in the cell supernatants was then measured. PBS alone was used as a control. Results are representative of three independent experiments performed in triplicate (for each experiment, the hMDMs were derived from different donors). Data are expressed as the mean ± SD. ns, not significant. *, p<0.05; **, p<0.01. (B, C) hMDMs or RAW 264.7 cells were stimulated for 6 h or 20 h with IFN-γ (1 ng/mL) and LPS (10 ng/mL) in the absence or presence of LL-37 or LL-37_{all cit} (10 μg/mL). The level of TNF-α or NO in the culture supernatants was determined using ELISA or the Griess assay. Data represent the mean ± SD of three independent experiments. *, p<0.05; **, p<0.01; ***, p<0.001.

Discussion

The large numbers of neutrophils that infiltrate a site of infection constitute the main source of cathelicidin LL-37, an important component of the host defence system. Apart from its direct bactericidal activity, LL-37 released from neutrophils (along with that locally produced by activated epithelial cells) effectively blocks the inflammatory response induced upon exposure to microbial antigens and viable pathogens (6, 13, 32). However, both the bactericidal and immunomodulatory effects of LL-37 are absent from a synthetic LL-37 peptide in which the Arg residues are replaced by citrulline residues (25). Here, we showed that citrullination of LL-37 by human PAD2 and PAD4, which are co-expressed at sites of inflammation, completely inhibited the ability of LL-37 to neutralise the proinflammatory activity of this ubiquitous PAMP. These findings were confirmed using synthetic citrullinated LL-37 peptides. In contrast to the study by Kilsgard et al. (25), we found that partially citrullinated LL-37 showed a significant decrease in immunomodulatory activity. This may be of physiological significance since it is plausible that several citrullinated forms of LL-37 (harbouring different levels of citrullination) are generated in vivo.

In contrast to the native peptide, citrullinated LL-37 was unable to block the LPS-stimulated release of NO and TNF-α by macrophages (Fig. 2). Quantitative analysis revealed that modified LL-37 was at least 40 times less potent at inhibiting LPS-mediated macrophage activation (Fig. 3). One may speculate that this could be compensated for by increasing the concentration of the modified peptide; however, this is unlikely, since citrullinated LL-37 was significantly less effective at inhibiting LPS than the native peptide even when used at a concentration of 200 μg/mL. Furthermore, such a high concentration of LL-37 is toxic to some cells (e.g., neutrophils) (11) and is not expected to occur in vivo. To summarise, our data indicate that citrullinated LL-37 (at physiological concentrations) is unable to prevent the LPS-induced activation of macrophages.

The mechanism underlying the reduced ability of citrullinated LL-37 to neutralise LPS was examined using the CD14⁺ murine macrophage cell line RAW264.7. Flow cytometry analysis revealed that, in contrast to the native peptide, citrullinated LL-37 was significantly less efficient at blocking the binding of FITC-conjugated LPS to RAW264.7 cells (Fig. 4). This appears to be due to its markedly reduced affinity for LPS. Such a mechanism agrees with the results of the in vitro experiments; however, we must bear in mind that the proinflammatory effects observed in vivo may also depend on the activation of various intracellular signalling pathways by citrullinated LL-37. Such a mixed mode of action may explain why sepsis is more severe in mice injected with citrullinated LL-37 than in animals treated with LPS alone (see below).

The changes of anti-inflammatory activity induced by citrullination of LL-37 were associated with peptide inability to prevent mortality in a mouse model of D-galactosamine-sensitised endotoxic shock (Fig. 5A). In fact, during the initial stages of sepsis development, mice inoculated with LPS plus LL-37_{all cit} showed significantly higher morbidity than those injected with LPS alone (Fig. 5B). The injection of citrullinated LL-37 into septic mice induced higher levels of circulating inflammatory cytokines (particularly IL-6, IL-10, and IFN-γ), which are indicative of a massive inflammatory reaction. This is supported by the high CRP levels in the sera of septic mice compared with those in animals injected with the native peptide. Collectively, while citrullination of LL-37 compromised the peptide capacity to suppress the release of pro-inflammatory cytokines, it has no effect of chemokine responses. This suggests that some of the immunomodulatory functions of the peptide indirectly involved in leukocyte recruitment are maintained. This is in contrast to direct chemotactic activity of LL-37, which is clearly compromised by citrullination (Fig. 5 C). Taken together the data presented herein show that the ability of the citrullinated LL-37 to control inflammatory reaction is significantly altered (Fig. 5 C–E) what may even contribute to severity of experimental sepsis (Fig. 5 A, B).

Citrullinated LL-37 is most likely generated within the inflammatory milieu, which is rich in neutrophil extracellular traps (NETs) (33). Here, the peptide is exposed to neutrophil-derived PAD4, which catalysis citrullination of histones essential for NET formation (34). Colocalization of histones and LL-37 on DNA strings (35) argues that NETs can be a source of citrullinated LL-37 at least in endotoxemia and sepsis in mice, conditions associated with abundant NETosis within the liver sinusoids (36). We anticipated that such extensive NETosis as well as cells dying of necrosis and apoptosis might serve as both a local and systematic source of PADs, which can also citrullinate proteins in the inflammatory milieu outside NETs. Indeed, we found increased levels of PAD activity and citrulline in the serum of septic mice compared with control animals. Importantly, the PAD activity showed a strong correlation with the level of citrullination in the serum (Fig. 6). Taken together, these data argue that extracellular PAD activity in septic mice citrullinates not only proteins/peptides associated with the NETs structure, but also other proteins in the circulation, thereby affecting their biological activity (23, 24).

The results showing that injection of LL-37_{all cit} in contrast to the native peptide can not protect against the endotoxin-induced sepsis in mice may explain a puzzling clinical observation: rather than protecting patients from sepsis-induced death, increased levels of AMPs, exacerbate the condition (37, 38). It is likely that, in such patients, LL-37 is citrullinated by PADs released from neutrophils. To verify this hypothesis, the levels of citrullinated LL-37 need to be measured in septic patients; unfortunately, there is no reliable method of quantifying (or even detecting) citrullinated LL-37. Significantly, in stark contrast to other neutrophil-derived peptides/proteins, including α-defensins, lactoferrin, and heparin-binding protein (all of which are elevated in sepsis patients), the level of LL-37 (as detected by ELISA) remains unchanged (39). In this context it is tempting to speculate that a significant amount of LL-37 is citrullinated and is, therefore, not recognised by the monoclonal antibodies used in the ELISA. This idea is supported by our observation that citrullination of LL-37 prevents detection by mAbs in Western blotting and dot-blot assays. Taken together, the data suggest that it is very likely that the blood concentrations of native LL-37 versus citrullinated LL-37 may be a potential diagnostic biomarker for human sepsis.

The immunomodulatory role of LL-37 is not limited to its interaction with LPS. The peptide also exerts a strong influence on cellular signalling pathways triggered by other TLR ligands and host inflammatory mediators such as TNF-α and IFN-γ. LL-37 inhibits the release of proinflammatory cytokines by human monocytic cells stimulated with TLR2 and TLR4 agonists, modulates the activation of DCs by TLR ligands (32), and affects cytokine expression mediated by TLR-2, -3 and -4 signalling in human gingival fibroblasts (40). Most of these activities will likely be lost after citrullination. Here, we showed that citrullination reduced the ability of LL-37 to quench the signalling pathways triggered by LTA, Poly (I:C), and IFN-γ [Fig. 7]. These data clearly indicate that the citrullination of LL-37 affects cellular responses to a broad range of microbial structural components and host-derived inflammatory mediators. Therefore, citrullination of LL-37 may also be relevant to the pathogenicity of sepsis caused by Gram-positive bacteria and fungi.

Abbreviations

LL-37

cathelicidin

PAD

peptidyl arginine deiminase

hMDMs

human monocyte-derived macrophages