Immunomodulatory and Anti-Inflammatory Activities of Chicken Cathelicidin-2 Derived Peptides

Host Defence Peptides and derived peptides are promising classes of antimicrobial and immunomodulatory lead compounds. For this purpose we examined whether chicken cathelicidin-2 (CATH-2)-derived peptides modulate the function and inflammatory response of avian immune cells. Using a chicken macrophage cell line (HD11) we found that full-length CATH-2 dose-dependently induced transcription of chemokines CXCLi2/IL-8, MCP-3 and CCLi4/RANTES, but not of pro-inflammatory cytokine IL-1β. In addition, CATH-2 efficiently inhibited IL-1β and nitric oxide production by HD11 cells induced by different sources of lipopolysaccharides (LPS). N-terminal truncated CATH-2 derived peptides maintained the capacity to selectively induce chemokine transcription, but despite their high LPS affinity several analogs lacked LPS-neutralizing capacity. Substitution of phenylalanine residues by tryptophan introduced endotoxin neutralization capacity in inactive truncated CATH-2 derived peptides. In contrast, amino acid substitution of phenylalanine by tyrosine abrogated endotoxin neutralization activity of CATH-2 analogs. These findings support a pivotal role for aromatic residues in peptide-mediated endotoxin neutralization by CATH-2 analogs and were shown to be independent of LPS affinity. The capacity to modulate chemokine production and dampen endotoxin-induced pro-inflammatory responses in chicken immune cells implicates that small CATH-2 based peptides could serve as leads for the design of CATH-2 based immunomodulatory anti-infectives.


Introduction
Worldwide, antibiotics are extensively used by the livestock industry in veterinary therapy and as a feed additive to promote animal growth, especially in the poultry and swine industry, and it is estimated to exceed 50% of the total antibiotics use in some countries [1]. The antibiotics used in veterinary medicine are structurally related to antibiotics commonly used for human therapy [2]. Micro-organisms have demonstrated to develop resistance against conventional antibiotics [3] and infections with multi-drug resistant bacteria are an increasingly common challenge in hospitals while the number of new antimicrobials in development is low. Hence, to reduce the use of conventional antibiotics in veterinary medicine, alternative strategies are needed that suppress outbreaks of infectious diseases in farm animals.
A promising approach is the use of host defense peptides (HDPs) such as cathelicidins and defensins. HDPs are capable of killing a wide variety of micro-organisms, including multidrugresistant bacteria, yeast, fungi, protozoa and viruses. In contrast to conventional antibiotics, HDP-mediated killing of micro-organisms occurs via multiple mechanisms attacking both the cell membrane and intracellular targets [4] and despite millions of years of co-evolution, most micro-organisms remain highly susceptible to HDP-mediated killing [5]. Moreover, HDPs exhibit a diversity of immune-related functions. Some HDPs, such as LL-37 inhibit LPSinduced pro-inflammatory cytokine production [6,7] and protect against development of endotoxin shock in vivo [8]. HDPs may act directly [9] or indirectly by selectively inducing chemokine production in immune cells [8] and recruitment of other immune cells to the site of infection. At the same time, HDPs may enhance antigen uptake and presentation [10] and inhibit apoptosis of neutrophils and macrophages [11,12]. Thus, in addition to direct antimicrobial activities, HDP-derived peptides may boost the innate and adaptive immune system leading to prevention and improved resolution of infectious diseases [13].
CATH-2 has the potential to serve as a paradigm for the development of anti-infectives in poultry with immunomodulatory and/or antibacterial activities. Therefore, we examined the immunomodulatory and anti-inflammatory effects of CATH-2 derived peptides on the avian macrophage-like HD11 cell line. We demonstrate that CATH-2 and several truncated analogs thereof selectively induce chemokine transcription in HD11 cells and inhibit LPS-induced IL-1β and nitric oxide production. We further show that the property to neutralize LPS can be modulated by aromatic amino acid substitution.

LPS binding affinity
Dansyl-polymyxin B was prepared and quantitated as described by Schindler and Teuber [17]. Subsequently, the capacity of peptides to displace dansyl-labeled polymyxin B complexed to S. minnesota LPS was determined according to Moore et al. [18].
Cell culture HD11 cells [19] were a kind gift of dr. Jos van Putten (department of Infectious diseases and Immunology, Utrecht University, The Netherlands). HD11 cells were seeded in 96 wells tissue culture treated plates (1×10 5 cell/mL) and incubated for 18 h at 37°C (5% CO 2 ) before treatment. For peptide-induced effects cells were treated with RPMI-1640 medium containing 20 μM peptide during 4 h or 24 h. In LPS neutralization experiments, final concentrations of Anti-Inflammatory Chicken CATH-2 Derived Peptides 50-100 ng/ml LPS were pre-incubated with or without 20 μM peptide for 30 min at 37°C (5% CO 2 ), applied to the cells and incubated for 4 h. To examine neutralization of LPS-primed cells HD11 cells were exposed for 30 min to 100-1000 ng/ml S. minnesota LPS, washed once and incubated during 4 h in the absence or presence of 20 μM Peptide. Supernatants were collected for the determination of nitric oxide production (24 h incubations only). Remaining cells were lysed in lysis buffer and total RNA was isolated and purified using a High Pure RNA isolation kit (Roche, Mannheim, Germany) according the manufacturer's recommendations. RNA quantity and purity were tested using a Nanodrop ND-1000 spectrophometer (Nanodrop Technologies, Wilmington, DE). Cytotoxicity of peptides was determined using the cell viability reagent WST-1 (Roche) as described previously [13], and by measuring LDH release after 24 h exposure. LDH activities in supernatant cell fractions were measured using the Cytoxicity Detection Kit PLUS (Roche) according the manufacturers recommendations and expressed as the % of released LDH relative to the total LDH activity.

Real time PCR
RNA (250 ng) was reverse transcribed using an iScript cDNA synthesis kit (Bio-rad laboratories (Hemel Hempstead, UK) according to the manufacturers' instructions. Primers and probes were designed and produced by Eurogentec (Seraing, Belgium) ( Table 2). Quantitative real time PCR was performed on a Bio-rad MyiQ system using a qPCR Mastermix (Eurogentec) and 400 nM of each primer and probe. Reactions were performed as follows: 3 min at 95°C; 40 cycles: 10 s at 95°C, 30 s at 60°C and 30 s at 72°C. Relative gene expression levels were normalized against the expression levels of the house keeping genes GAPDH and 28S.

Nitric oxide production
Nitric oxide production was assessed as the accumulation of nitrite (NO 2 − ) in cell supernatants during a 24 h incubation period. Nitrite concentrations were determined using a colorimetric reaction with the Griess reagent using (from 1 to 50 μM) sodium nitrite dissolved in water as standards. Briefly, cell culture supernatants were mixed with an equal volume of 1% sulfanilamide (dissolved in 2.5% phosphoric acid) and incubated for 5 min. The same volume of 0.1% N-(1-naphthyl) ethylenediamine dihydrochloride was added and incubated for 5 min. The absorbance was measured at 520 nm using a 96-well microplate reader (FLUOstar Omega, BMG labtech).

Statistical analysis
Statistical analyses were performed using SPSS version 16.0 statistical software (SPSS inc., Chicago, IL) with one-way analysis of variance (ANOVA) and Dunnett post hoc tests.

Antibacterial and cytotoxic activities of CATH-2-derived peptides
To determine the effect of truncation and amino acid substitution on the antibacterial activity minimal inhibitory concentrations (MIC) were determined using colony count assays. All truncated and substituted peptides exhibited antibacterial activity against avian pathogenic E. coli O78 (Table 1); bacterial survival was reduced to below the detection limit in the presence of 5-20 μM peptide. To determine the toxicity of CATH-2 analogs HD11 cells were exposed for 24 h to different peptide concentrations. Most peptides did not affect HD11 metabolic activity up to peptide concentrations of 20 μM (Fig 1). At 40 μM, HD11 metabolic activities decreased considerably for all peptides except peptide C (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)[F12Y]. The peptide-induced LDH release from HD11 cells during 24 h exposure corresponded well with the metabolic activities of HD11 cells (Fig 2); less than 5% LDH release occurred for most peptides at 20 μM.

Full-length CATH-2 peptide selectively induces chemokine transcription in HD11 cells
To examine if CATH-2 analogs induce chemokine production in chicken macrophages, HD11 cells were exposed to different doses of peptide. CATH-2 was found to dose-dependently
doi:10.1371/journal.pone.0147919.g003 transcription (Fig 4C). Phe/Trp residues did not consistently alter peptide-induced chemokine expression of HD11 cells (Table 3). Substitution by tyrosine increased overall cytokine transcriptional levels during short exposure, and this effect disappeared when exposure was prolonged.
Nitric oxide (NO) production by macrophages plays a dual role; NO can rapidly react with superoxide to produce peroxynitrate, a reactive oxygen species with potent antimicrobial activity, whereas sustained high levels may lead to tissue damage and cell death [21]. To determine the nitric oxide levels produced by HD11 cells in the presence of LPS, HD11 cells were stimulated for 24 h with different concentrations of LPS derived from different bacterial species. HD11 cells were most susceptible to stimulation by N. meningitidis H44/76 LPS, i.e. at 10 ng/ ml LPS already induced 83.1% of the maximal NO production (Fig 5A). A moderate response was observed for LPS derived from C. jejuni, S. minnesota and S. enteritidis LPS, e.g. 50 ng/ml of LPS was needed to produce 61.7%, 36.2% and 31.2%, respectively of the maximal NO production. HD11 production of NO was least responsive to E. coli O111:B4 LPS.
Next, neutralization of LPS-induced NO production by full-length CATH-2 peptide was examined. For this purpose HD11 cells were stimulated during 24 h with pre-incubated mixtures of LPS (100 ng/ml) and CATH-2 peptide (20 μM). CATH-2 significantly reduced LPSinduced NO production for LPS of all sources except Neisseria LPS (Fig 5B). The lack of Neisseria LPS inhibition can be explained by the fraction of free LPS. Under the conditions used, CATH-2 most potently inhibited S. minnesota LPS (85%) of which the residual NO production (10 μM) corresponded to10 ng/ml unbound LPS. A similar unbound fraction of N. meningitidis LPS, would induce approx. 90 μM NO. Thus, the most likely explanation is that under these conditions the high potency of the unbound fraction of N. meningitidis LPS was sufficient to obtain a maximal NO production.

Structure of LPS-neutralizing peptides
To visualize the possible mode of CATH-2 analog binding to lipid A, the 3D structures of CATH-2 and C(1-21) were predicted using iTASSER [22] and compared with the configuration of known LPS-binding peptides paradaxin and human lactoferrin (hLF11) when in complex with LPS (Fig 6). Lactoferrin-derived peptide LF11 possesses three basic residues (Arg5, Lys9, Arg11) that match the distance between both glucosamine phosphate groups in lipid A and is known to adopt a T-shaped configuration when binding to LPS [23]. Similarly, the Lys8, Lys16 residues and Phe2, Phe3 residues of paradaxin were shown to be in close proximity of lipid A [24]. The generated models of CATH-2 and C(1-21) indicate that in a hydrophobic environment, e.g. a biological membrane, all Phe residues in these peptides align in the same plane where they may interact with lipid A acyl chains and suggest that N-terminal basic residues Arg 1 , and Arg 4 Arg 7 are the most likely residues to interact with lipid head phosphate groups.

LPS neutralization by CATH-2 analogs does not correlate with LPS binding affinity
Dansyl-polymyxin B (DPmB) experiments were performed to get insight in the relationship between the LPS binding affinity of peptides and their capacity to neutralize LPS. Compared to unlabeled polymyxin B (I 50 = 4.19 μM), full-length CATH-2, C(1-21) and C(4-21) displaced DPmB from S. minnesota LPS at a 20-fold lower concentration (Table 4). However, similar high binding affinities were found for C(7-21) (I 50 = 0.33 μM) and C(1-15) (I 50 = 0.12 μM) that lacked LPS-neutralization capacity. Further peptide truncation gradually decreased LPS binding capacity and was not affected by aromatic amino acid substitution.
To determine if LPS-neutralization occurred via other mechanisms than direct binding, HD11 cells were primed during 30 minutes with different doses of LPS, washed once to remove exogenous LPS and incubated for 4 h in the absence or presence of peptide. IL-1β transcription induced by LPS priming proved to be partially inhibited by CATH-2 and C(1-21) (p<0.01) and strongly inhibited by LL-37 (p<0.001), whereas no inhibition was observed with C(4-21) (Fig 7).

Discussion
Cathelicidins [25,26] and cathelicidin-derived peptides, such as IDR-1 [27] and IDR1002 [28], have demonstrated in vivo protection against infections by Gram-positive and Gram-negative bacteria. In the case of cathelicidin-derived peptides that lacked direct antimicrobial activity this was found to be associated with the selective induction of chemokines, subsequent recruitment of monocytes and neutrophils, and to be monocyte/macrophage dependent [27]. In this work the immunomodulatory and anti-inflammatory activities of chicken CATH-2 analogs were examined using HD11 cells, a chicken macrophage-like cell line. Full-length CATH-2 peptide and truncated analogs were found to selectively induce CXCLi2/IL-8 and MCP-3 transcription in HD11 cells. These findings corroborate with observations by others for LL-37; induction of MCP-1 expression in murine RAW264.7 cells [8] and induction of MCP-1, MCP-3 and IL-8 via the ERK1/2 and p38 pathways, but not of pro-inflammatory cytokines IL-6, TNF-α or IL-1β, in human monocytes [29]. HD11 cells responded similarly to LL-37 by primarily inducing the transcription of CXCLi2 and MCP-3, suggesting that they may share common signaling pathways.
Next we endeavored to enhance the anti-inflammatory properties of smaller CATH-2 analogs while maintaining other immunomodulatory properties. The prominent presence of aromatic moieties in small immunomodulatory peptides [30] and their frequently observed proximity to lipid A acyl chains in peptide-LPS complexes [23,27,31] led us to believe that the position and type of aromatic amino acid may be key factors in enhancing or altering immunomodulatory properties. Indeed, aromatic amino acid residue substitution was found to greatly affect anti-inflammatory activities of truncated CATH-2 analogs, without substantially changing their potency to induce chemokine transcription in HD11 cells. LPS neutralizing capacity could be enhanced in peptide C(4-21) and introduced in inactive peptides C(7-21) and C(10-21) by substitution of the single Phe residue by Trp. Surprisingly, peptide C(1-15) failed to induce chemokine transcription. In addition, C(1-15), previously shown to moderately [13] or weakly [32] inhibit LPS-induced cytokine expression, was under the conditions used not able to block LPS-induced cytokine transcription in HD11 cells (data not shown). Thus, in nonsubstituted CATH-2 analogs the C-terminal segment (VTITIQ) is pivotal for both immunomodulation and LPS neutralization. However, Phe/Trp substitution of C(1-15) has been shown to introduce significant LPS neutralizing capacity [32]. This can be explained by the behavior of tryptophan in lipid bilayers. Tryptophan side chains and indol groups prefer to localize at the interface of water-associated lipid head groups and acyl chains, probably stabilized by dipole interactions [33,34], suggesting that Phe/Trp-substituted CATH-2 analogs are prone to adopt a different orientation when binding to lipid A. Among aromatic residues Trp is the most polarizable residue and able to form a hydrogen bond via its indol N-H moiety with acyl carbonyl as well as phosphate oxygen groups [35]. Hence, most likely the Phe/Trp substitution in peptide C(10-21) improved LPS neutralization capacity by stabilization of the peptides hydrophobic core to the lipid acyl region. In contrast, substitution of Phe residues by Tyr residues was found to abrogate the LPS neutralizing capacity of 'active' CATH-2 analogs. This is comparable to the reported effects of Phe/Trp and Phe/Tyr substitution of Polymyxin B nonapeptide; i.e. LPS-binding and-neutralization were maintained when Phe was substituted by Trp, but strongly abrogated upon substitution with Tyr [36].
It has been shown that cationic peptides in complex with LPS adopt a configuration in which basic residues interact with lipid A glucosaminoglycan phosphate groups while hydrophobic residues interact with its acyl chains. In the case of the polymyxin B-LPS complex, two positively charged α,γ-diaminobutyric acid (DAB) residues and one Phe residue are necessary for LPS-binding [37]. Structure overlay of the LPS-LF11 complex and E. coli iron uptake receptor FhuA-LPS complex revealed a conserved structural LPS-binding motif consisting of three basic residues (Arg or Lys) and a single Phe residue [38]. We reasoned that CATH-2 analogs, possessing multiple basic and hydrophobic residues may adopt a similar configuration. Moreover, arginine residues are flanking each aromatic residue, which may aid to stabilize the conformation through formation of intramolecular cation-π interactions between aromatic and arginine residues [39]. Based on the homology found between the deduced C(1-21) peptide configuration and known lipid A binding peptides we consider it likely that this CATH-2 analog interacts with lipid A via its Arg1, Arg4 and Arg7 residues and multiple Phe residues.
However, our results also indicate that LPS neutralization by CATH-2 analogs cannot be explained by LPS binding affinity alone. CATH-2, C(1-21) and LL-37 could partially or completely neutralize IL-1β transcription in LPS-primed HD11 cells, indicating that LPS neutralization by LL-37 and some CATH-2 analogs is in part independent of binding to LPS.
Despite its capacity to neutralize LPS when co-incubated, peptide C(4-21) was not able to neutralize LPS-priming induced IL-1β transcription, suggesting a direct binding mode of action for this peptide. Although LPS-binding has been found to be correlated to inhibition of LPSinduced pro-inflammatory cytokine production for several structurally different cationic peptides [6], LPS signaling may be inhibited via alternative mechanisms. For instance, LL-37 and LL-37-derived peptides were shown to partially inhibit LPS signaling through binding to cell surface CD14 [40]. Furthermore, LPS-induced gene expression profiles were found to be differentially altered in the presence of LL-37 and BMAP-27: strongly inhibiting one set of proinflammatory genes while the expression levels of other genes, including negative regulators of NF-κB and certain chemokines, were not substantially affected [41]. For LL-37, BMAP-27 and polymyxin B it has also been described that they can inhibit the nuclear translocation of NF-κB subunits p50 and p65, which is pivotal to LPS-induced pro-inflammatory cytokine production [41]. Additionally, LL-37 is known to selectively and directly affect chemokine production via activation of MAPK pathways [29] and it has been suggested that MAP kinase ERK1/2-mediated CREB phosphorylation could inhibit NF-κB activation by competing with CBP/p300 binding to NF-κB subunit p65 [42].
Overall, the immunomodulatory and anti-inflammatory properties of CATH-2 derived peptides demonstrated towards avian macrophages suggest that treatment of poultry with CATH-2 derived peptides may lead to a selective recruitment and activation of avian immune cells while dampening excessive immune responses. We believe that CATH-2 derived peptides, due to their unique properties, could serve as leads for the design of veterinary animal-specific therapeutics.