Regulation of Chemerin Chemoattractant and Anti-bacterial Activity by Human Cysteine Cathepsins

Abstract

Chemerin, a ligand for the G-protein coupled receptor CMKLR1 (chemokine-like receptor 1), requires C-terminal proteolytic processing to unleash its chemoattractant activity. Proteolytically-processed chemerin selectively attracts specific subsets of immunoregulatory antigen presenting cells, including CMKLR1+ immature plasmacytoid dendritic cells (pDC). Chemerin is predicted to belong to the structural cathelicidin/cystatin family of proteins comprised of antibacterial polypeptide cathelicidins and inhibitors of cysteine proteinases (cystatins). We therefore hypothesized that chemerin may interact directly with cysteine proteases and that it might also function as an antibacterial agent. Here we show that chemerin does not inhibit human cysteine proteases, but rather is a new substrate for cathepsin K and L. Cathepsin K and L-cleaved chemerin triggered robust migration of human blood-derived pDC ex vivo. Furthermore, cathepsin K and L-truncated chemerin also displayed antibacterial activity against Enterobacteriaceae. Cathepsins may therefore contribute to host defense by activating chemerin to directly inhibit bacterial growth and to recruit pDC to sites of infection.

Introduction

Chemerin is a recently characterized chemoattractant protein that serves as a ligand for the seven-pass G-protein coupled receptor (GPCR) CMKLR1 (chemokine-like receptor 1) (1, 2). Two additional heptahelical receptors, GPR1 and CCRL2, have been reported to bind chemerin, although they do not appear to directly support chemotaxis (3, 4). Chemerin circulates as an inactive precursor (pro-chemerin) in blood, and undergoes protease-mediated C-terminal truncation to acquire chemotactic activity. Truncated, bioactive chemerin lacking 6 (chemS157), 8 (chemA155) or 9 (chemF154) amino acids in the C-terminus have been isolated from several bio-fluids, including ascites, serum and hemofiltrate, respectively (reviewed in (5)). Serine proteases of the coagulation, fibrinolytic and inflammatory cascades, including neutrophil elastase and cathepsin G generate bioactive chemerin (6, 7). In addition, chemerin can be activated in a sequential manner by plasma carboxypeptidases following initial cleavage by plasmin (8).

Cells that are critical in linking the innate and adaptive immune responses, such as plasmacytoid dendritic cells (pDC), NK cells and macrophages, express CMKLR1 and respond to chemerin through chemotaxis (1, 2, 9-12). Although the structure of chemerin has not yet been solved, the predicted structural homology between chemerin and inhibitors of cysteine proteinases (cystatins) and antimicrobial cathelicidins (1, 5) suggests that chemerin may inhibit endogenous human cysteine proteases and possibly exhibit antibacterial activity. Alternatively, host cysteine proteases may bind and proteolytically process chemerin. In support of the latter, we recently reported that the cysteine protease staphopain B secreted by the human pathogen S. aureus selectively cleaves and activates chemerin (13).

The papain-like cysteine proteases, including cathepsins B, L and K are well-known degradative enzymes of mammalian cells, participating primarily in intracellular proteolytic pathways (such as antigen processing and presentation) but also extracellular protein turnover. Recent studies show that lysosomal cathepsins can exert their proteolytic activity at extracellular sites (14, 15), where they contribute to a variety of pathophysiological processes, including chronic inflammation associated with obesity (16-18).

Cathelicidins consist of two distinct domains; a highly conserved N-terminal “cathelin-like” domain with homology to the cystatins, and a divergent C-terminal antimicrobial region that varies among species. Only one cathelicidin has been described in humans, cationic antimicrobial peptide of 18 kDa (hCAP18). hCAP18 is cleaved by neutrophil serine proteases such as proteinase 3 to generate a 37 amino-acid antimicrobial peptide LL-37 and a 103 amino-acid cathelin-like domain (19, 20).

In this work, we found that while chemerin does not inhibit the proteolytic activities of cathepsin L or cathepsin K, these cysteine proteases are potent activators of chemerin. Cathepsins L and K initially and efficiently cleave pro-chemerin to release a six amino acid peptide from the carboxyl-terminus, generating chemS157; the enzymes can also cleave chemerin to release a 38-residue carboxyl-terminal peptide, generating chemR125. The activated chemerin S157 is a potent attractant for CMKLR1+ cells, including human blood pDC. In addition, we demonstrate that although the smaller chemerin fragment generated by both cathepsins (chemerin R125) does not support chemotaxis of CMKLR1+ cells, both chemS157 and chemR125 display comparable antimicrobial activity against Enterobacteriaceae.

Material and methods

Materials

Recombinant cathepsin B, L and K, and chemS157 were purchased from R&D Systems. Anti-CD3, -CD14, -CD16, -CD19, -CD20, -CD56 biotin-linked mAbs, as well as FITC-labeled CD123 and APC-labeled BDCA-2 were obtained from BD Pharmingen, Miltenyi Biotec, eBioscience and Biolegend. Recombinant chemerin isoforms, full-length prochemerin, chemerin serum from (chemA155) and chem/SspB (chemS157) were produced as previously described (2, 13, 21). Recombinant Fc-chemerin proteins were produced and purified from CHO cells via transient transfection and Protein A purification. DNA fragments corresponding to the desired chemerin proteins were amplified by PCR and cloned in-frame downstream of human IgG1 Fc domain, which is downstream of a secretion signal peptide in mammalian expression vector pLEV113 (LakePharma). There is a 9 amino acid glycine-rich linker between the Fc and chemerin domains. Plasmid DNA was transfected into CHO cells using Lafectine transfection reagent (LakePharma), and cell culture supernatant was collected 3 days post transfection. Fc fusion proteins were purified with Protein A resins (Mab Select SuRe GE Healthcare), and final proteins were formulated in 100 mM Tris, 150 mM NaCl and 0.45% NaOAc.

Inhibition assays

Inhibitory activity against cathepsins B, K and L was assayed fluorometrically using a Molecular Devices Gmini XS system, with an excitation of 350 nm and emission wavelength of 460 nm. Cathepsins were pre-activated in a buffer containing 100 mM sodium acetate, pH 5.5, 100 mM NaCl, 10 mM DTT, 1mM EDTA and 0.01% Tween20 for 10 minutes at 25°C. Cathepsins (0.2 nM) were then incubated with 10 or 100 molar excess of recombinant pro-chemerin or the non-specific cysteine proteinase inhibitor E64 for 30 min at 37°C in the same buffer containing the fluorogenic substrates Z-Phe-Arg-AMC (10 μM, Sigma) for cathepsin B and L or Boc-VLK-AMC (10 μM, Peptides International) for cathespin K, in a total volume of 100 μL. The progress of the reaction was monitored by fluorescence spectroscopy and the data was plotted vs. time. V_max values were calculated by linear fit to the time-dependent curve. The residual cathepsin activity present in each treatment group is presented as a percentage of the V_max determined in the absence of inhibitor (cathepsin + substrate only).

Mass spectrometry

MALDI-TOF MS/MS was performed by the Stanford Protein and Nucleic Acid Biotechnology Facility (Stanford University, Stanford, CA, USA). Following trypsin digest of the cat K and cat L truncated-chemerin band, MALDI-TOF MS was performed and the mass values were used in a Mascot search (www.matrixscience.com) of public peptide databases. PeptideCutter was used to predict the mass values of tryptic chemerin fragments (www.expasy.org). Mass values obtained following collision induced dissociation (CID) of the predicted carboxyl-terminal peptide (f) were compared with the predicted CID mass values for the peptide corresponding to residues 141-157 of chemerin (sequence:AGEDPHSFYFPGQFAFS), using Mascot software (Matrix Science).

Purification of pDC

The Institutional Review Board at Jagiellonian University approved all human subject protocols. Human blood was collected and peripheral blood mononuclear cells (PBMC) were harvested following LSM1077 (PAA Laboratories) gradient separation, as described by the manufacturer. pDC were enriched from PBMC using negative selection with biotinylated mAbs directed against CD3, CD14, CD16, CD19, CD20, CD56 and anti-biotin MACS Microbeads (Miltenyi Biotech), according to the manufacturer's recommendations. Cells were blocked with 50-80% autologous plasma and then stained for flow cytometry analysis using mAbs against CD123 and BDCA-2 to identify pDC. Stained cells were analyzed on a LSRII flow cytometer (Becton Dickinson).

Chemotaxis assay

Cathepsins were incubated with recombinant human pro-chemerin for 10 min at 37°C and then tested in an in vitro chemotaxis assay using murine pre-B lymphoma L1.2 cells stably transfected with human recombinant CMKLR1 (CMKLR1/L1.2) or purified human blood pDC. In each case, enzymatic digestion was stopped by placing samples on ice and diluting with chemotaxis medium (RPMI 1640 containing 10% FBS). Where indicated recombinant chemerin isoforms, Fc-chemerin fusion proteins or Fc alone were used. One hundred μl of cells (2.5×10⁵ cells/well) were added to the top well of 5 μm pore transwell inserts (Costar) and test samples were added to the bottom well in a 600 μl volume. Migration was assayed for 2 h at 37°C. The inserts were then removed and cells that had migrated through the filter to the lower chamber were collected and counted by flow cytometry (FACSCalibur, BD Biosciences). The results are presented as % input migration. CXCL12 served as a positive control.

Antimicrobial assay

Escherichia coli (HB101, a conventional laboratory strain) and Klebsiella pneumoniae (a clinical isolate from human bronchoalveolar lavage fluid) were used in this study. The antimicrobial activity of the indicated chemerin forms was estimated using a microtitre broth dilution assay (22). A single colony of bacteria was inoculated into 20 ml Mueller Hinton Broth (MHB) (Difco) and incubated overnight at 37°C, sub-cultured once at 1:100 dilution in MHB, and then grown for 2-3 h to mid-logarithmic phase. Cell numbers were calculated using previously determined standard curves, and for subsequent experiments bacteria was used at 2-7 × 10⁵ colony-forming units (CFU)/ml. Bacterial suspensions (90 μl) in MHB were mixed with 10 μl of diluent (10mM HEPES, 100 mM Tris, 150 mM NaCl and 0.45% NaOAc, or MHB) (control) or 10 μl of different concentrations of chemerin, Fc-chemerin fusion proteins, Fc alone or synthetic LL-37 (Emory Microchemical Facility) and incubated at 37°C for the indicated times. After serial dilutions with MHB, the diluted mixture was plated on MHB agar plates and incubated at 37°C overnight for enumeration of CFU. In selected experiments, samples of the bacteria/peptide mixtures were also analyzed by flow cytometry and by spectrophotometry. These methods produced comparable results to the colony-forming assay (data not shown).

Results

Cathepsin L and cathepsin K activate chemerin chemoattractant activity

We initially tested chemerin for cystatin-like activity. Substrate hydrolysis by cathepsins B (cat B), L (cat L), or K (cat K) was not significantly inhibited by pro-chemerin, serum-form bioactive chemerin (chemA155), or SspB-activated chemerin (chemS157), even at 10:1 and 100:1 molar ratios of chemerin-to-cathepsin (Table I and Fig. 1). The general cysteine proteinase inhibitor E-64 efficiently abolished the activity of all cathepsins examined (Fig. 1). Thus chemerin does not appear to function as a cystatin.

Table I.

Comparison of chemerin isoforms used in this study.

Chemerin form	Abbreviations	C-term amino acid sequence with residue position in Italics. Non native aa are shown in bold.	Ref.
Pro-chemerin (full-length chemerin)	Pro-chem	111 VLGRLVHCPIETQVLREAEEHQETQCLRVQRAGEDPHSFYFPGQFAFSKALPRS 163	(7)
Serum form	chemA155	111 VLGRLVHCPIETQVLREAEEHQETQCLRVQRAGEDPHSFYFPGQFAPH 155	(7, 21)
Staphopain B-treated chemerin	Chem/SspB chemS157 minor form chemS148	111 VLGRLVHCPIETQVLREAEEHQETQCLRVQRAGEDPHSFYFPGQFAFS 157 111 VLGRLVHCPIETQVLREAEEHQETQCLRVQRAGEDPHS 148	(13)
Cat L –treated chemerin Cat K-treated chemerin	chemS157 chemR125	111 VLGRLVHCPIETQVLREAEEHQETQCLRVQRAGEDPHSFYFPGQFAFS 157 111 VLGRLVHCPIETQVLR 125	This work

Fig. 1. Chemerin does not inhibit the activity of the cysteine cathepsins.

Cathepsins were incubated with 10 (10x) or 100 molar excess (100x) of recombinant pro-chemerin (pro-chem), serum chemerin (chemA155), chem/SspB (chemS157) or 100 molar excess of the general cysteine proteinase inhibitor E64 for 30 min at 37°C with fluorogenic substrates. * p<0.05 by Student's t-test. Data are shown as mean±SD of 3 independent measurements done in duplicate. Control contained no inhibitor.

To test if pro-chemerin is a substrate for human cysteine proteases, pro-chemerin was incubated with purified cathepsin B, L or K and tested for attractant activity. A controlled digest of pro-chemerin, using low concentrations of either cat L or cat K (1000-fold less than pro-chemerin) generated a single primary proteolytic product (Fig. 2A). Two apparent cleavage products were generated when cat L and cat K concentrations were increased 10-50 times, suggesting further digestion of pro-chemerin. Under similar conditions, cat B did not cleave pro-chemerin (data not shown). Interestingly, the two specific >10 kD chemerin cleavage products generated by cat L and cat K can be distinguished by polyacrylamide gel electrophoresis only under reducing conditions (Fig 2A). Since pro-chemerin contains three disulfide bonds, these data suggest that under non-reducing conditions the dual-cleaved chemerin products remain associated with the holo-molecule through SS bond(s).

Fig. 2. Cathepsin L and K-activated chemerin triggers CMKLR1/L1.2 cell and human blood pDC chemotaxis.

(A) Various concentrations of cat L or cat K were incubated with 2.9 uM chemerin and the resulting products were separated by SDS-PAGE under either reducing (left panel) or non-reducing (right panel) conditions. Arrowheads indicate the two predominant chemerin cleavage products. Representative of n=3 performed with similar results. (B-C) Cat K or cat L activated chemerin was tested in in vitro transwell chemotaxis using CMKLR1/L1.2 transfectants (B) or human blood pDC (C). The chemerin samples described in A were diluted to a final concentration of 1 nM and tested in chemotaxis assays. CXCL12 (10 nM) was used as a positive control for pDC chemotaxis (C). The mean from duplicate wells of 3 or 4 experiments ± SD is shown. * p<0.05 by Student's t-test comparing cathepsin treated chemerin vs. pro-chemerin, or CXCL12 vs. (-) control.

CMKLR1-transfected L1.2 cells migrated significantly to cat L or cat K-treated pro-chemerin (Fig. 2B). The chemotactic response of CMKLR1/L1.2 transfectants was dependent on cathepsin concentration, with cat L and cat K eliciting maximal effects on chemerin-mediated migration at a 1:100 pro-chemerin/cathepsin ratio (Fig. 2B). Although chemerin treated with cat L appeared to elicit slightly higher chemotactic response compared to cat K-treated chemerin (Fig. 2B), the difference was not statistically significant. Freshly isolated CMKLR1+ human blood pDC also migrated in response to cathepsin-cleaved chemerin, suggesting that these enzymes may be involved in pDC recruitment (Fig. 2C). No cell migration was detected in the absence of chemerin, or when pro-chemerin or the cathepsins were tested alone (Fig. 2). Compared to cathepsins L and K, cat B had negligible effects on chemerin chemoattractant activity (data not shown). Taken together, these data suggest that incubation of cat L and cat K with pro-chemerin results in generation of bioactive chemerin chemoattractant.

Identification of the cathepsin chemerin cleavage sites

We used MALDI-TOF mass spectrometry to determine the cathepsin chemerin cleavage sites. The mass value for the larger chemerin cleavage product initially and efficiently generated by cat L and cat K (experimental mass [M+H]⁺ 16,152 Da; calculated mass [M+H]⁺ 16,155 Da; Δ mass 3 Da) corresponds to chemerin residues 18-157 (ADPELT......GQFAFS, chem157S). To confirm the processing site, the larger protein band (indicated by arrowheads in Fig. 2A) was isolated, digested with trypsin, and analyzed by mass spectrometry. A peptide fragment with a mass value of 1903.8 was identified, corresponding to a non-tryptic peptide comprising amino acids 141-157 from the C-terminus of chemerin (Fig. 3A, B). This peptide confirmed the initial and predominant cat K- and cat L-mediated chemerin cleavage site as NH₂..AFS↓KAL..COOH. Further microsequencing of this peptide by collision-induced dissociation (tandem MS/MS) confirmed the sequence (Fig. 3 C). Interestingly, this cleavage site is identical to a previously identified endogenous active human chemerin isoform isolated from ascites fluid (1), and to a main chemerin isoform generated by the S. aureus-secreted cysteine protease SspB (chem/SspB) (Table I).

Fig. 3. Identification of the chemerin cat K and cat L processing site.

Pro-chemerin was incubated with cat L and cat K and the resulting peptide fragments were resolved by SDS-PAGE. The larger chemerin band (indicated by an arrowhead in Fig. 2) was digested with trypsin and subjected to MALDI-TOF MS analysis. (A) Peptide fragments corresponding to predicted tryptic chemerin mass values are labeled with their mass values (a-f). * indicates carbamidomethyl modification of cysteine. (B) The peptides corresponding to the chemerin fragments identified by MALDI-TOF MS are indicated by underlining and lettered (a-f). The vertical bars indicate predicted trypsin cleavage sites. The amino-terminal signal peptide is indicated by italics, as is the carboxyl-terminal peptide released upon cathepsin cleavage. The carboxyl-terminus of truncated chemerin is non-tryptic carboxyl-terminal peptide (f). (C) Collision induced dissociation (CID) mass spectra (tandem MS/MS analysis) of peptide (f) confirms the sequence identity of the non-tryptic carboxyl-terminal peptide. The major intensity peaks are predicted b or y series ions.

The mass value for the smaller chemerin cleavage product generated by cat L and cat K (experimental mass [M+H]⁺ 12,442 Da; calculated mass [M+H]⁺ 12,437 Da; Δ mass 5 Da) corresponds to chemerin residues 18-125 (ADPELT......ETQVLR, chemR125). MALDI-TOF MS and CID analysis of the tryptic digests confirmed that the most distal tryptic or non-tryptic fragment was “LVHCPIETQVLR” (fragment “d” in Fig. 3B, data not shown). Thus, prolonged incubation (data not shown) or incubation with 10-50x higher concentrations of cat K or cat L (Fig. 2A) cleaves chemerin at position NH₂..QVLR↓EAEE..COOH and the released C-term peptide likely remains linked to the holo-molecule through a disulfide bond, as indicated by Fig 2A.

Chemerin S157 but not chemerin R125 is chemotactively active

Since separation of the smaller cleavage product by HPLC required reducing the S-S bonds, which would likely alter its secondary structure and possibly bioactivity, to determine the relative biological activities of the two cathepsin-generated chemerin products, we generated recombinant Fc-chemerins, abbreviated further as Fc-chemS157 (the larger cleavage product) and Fc-chemR125 (the smaller cleavage product). The Fc-chemerins have a glycine linker on the N-term that is connected to the Fc domain of human IgG1, thus the C-term of the Fc-chemerin fusion proteins remained native. Fc alone or Fc-chemR125 failed to trigger CMKLR1+ cell migration at every concentration tested (up to 50 nM) (Fig. 4). On the other hand, chemS157 was the most potent attractant for CMKLR1+ cells, with 1 nM eliciting a maximum 42±8% cell migration. Recombinant, commercially available chemerin S157 and HPLC-purified chemerin cleavage product of SspB (13) were used interchangeably as chemS157, as both gave similar results in chemotaxis assays (data not shown). In previous work, we and others have shown that chemS157 is a chemoattractant for pDC, NK cells and monocyte-derived DC and macrophages (1, 11, 13).

Fig. 4. Chemerin S157 is the most active chemoattractant chemerin.

CMKLR1/L1.2 cell migration to the indicated doses of the indicated chemerin isoforms was assessed by in vitro transwell chemotaxis. Migration to media only or to Fc only (tested at molar concentrations equivalent to the Fc-chemerin fusion proteins), are shown as negative controls. The mean from 3 experiments ± SD is shown. Statistically significant differences between media or Fc-control and the chemerin isoforms are indicated by asterisks (p<0.05, Student's t test).

The Fc-chemS157 fusion protein also triggered statistically significant CMKLR1/L1.2 cell migration at concentrations of 5, 10 and 50 nM, although the response was not as robust as the unmodified form. ChemA155 also triggered CMKLR1+ cell migration, although it should be noted that this recombinant protein has a non-native “PH” carboxyl-terminal dipeptide following the native terminal alanine (residue 155) (Table I), which may alter its activity.

Chemerin displays antibacterial activity

The predicted homology between chemerin and the antimicrobial cathelicidins, coupled with the expression of chemerin in the skin and other epithelial cell surfaces continually exposed to bacterial challenge (9), led us to evaluate possible antibacterial activities of chemerin. Different chemerin isoforms were first tested for antibacterial activity against a laboratory strain of Escherichia coli. Human cathelicidin LL-37 (3 μM) was used as a positive control peptide inhibitor of bacterial growth (20). As demonstrated in Fig. 5A, full length chemerin (pro-chemerin, 2 μM) significantly inhibited the growth of E. coli when incubated for 24 h, leading to survival of 59±13% of bacteria compared to vehicle-treated E. coli set as 100%. Notably, truncated chemerin was an even more effective inhibitor of bacterial growth. The primary chemerin cleavage product generated by cat K and cat L (chemS157) (Table I) was significantly more effective than pro-chemerin in inhibiting bacterial growth, resulting in 33±15% E. coli survival (Fig. 5A). Truncated chemA155 also significantly reduced E. coli survival (39±19%) vs. pro-chemerin (Fig. 5A). We also tested chemerin (2μM) for antibacterial activity against a clinical isolate of another genus within the family of Enterobacteriacae, Klebsiella pneumoniae. As demonstrated in Fig. 5A, incubation of K. pneumoniae with pro-chemerin resulted in 58±14% viable bacteria compared to control, whereas chemA155 and chemS157 reduced K. pneumoniae survival to 32±11% and 46±13, respectively. Thus, similar to E. coli, the truncated form of chemerin was more effective at inhibiting K. pneumoniae growth than pro-chemerin. However, it should be noted that 24h incubation of pro-chemerin with either E. coli or K. pneumoniae also resulted in some truncation of pro-chemerin (data not shown), suggesting that native protein requires removal of inhibitory C-terminal sequence in order to display full antibacterial activity. Taken together, these data suggest that proteolytic cleavage increases the antimicrobial activity of chemerin.

Fig. 5. Chemerin exhibits antibacterial activity.

The antimicrobial activity of full-length (pro-chem), chemA155, chemS157, Fc-chemS157 (HPLC-purified chem/SspB, Table I), Fc-chemR125, and Fc alone were tested against E. coli (A-D) and K. pneumoniae (A) using the microtitre broth dilution assay. Chemerin isoforms were tested at 2 μM (A-B), 0.5 μM (D) or at the indicated concentrations (C). Synthetic LL-37 (3 μM) was used as a positive control (A). Bacteria were incubated with the peptides for 24 h (A, C, D) or for indicated times (B). The results are expressed as the mean ± SD of nine (A, E. coli), four (A, K. pneumoniae, B, C,), or two (D) independent experiments done in duplicate. Statistically significant differences between pro-chemerin and the truncated chemerin isoforms are indicated by asterisks and p values (Student's t test) in A. Statistically significant differences comparing the indicated protein preparations with vehicle or Fc alone-treated E. coli are shown by asterisks, * p <0.05, ** p<0.001, (Student's t test) in B-D.

The inhibition of E. coli growth by chemerin was detectable after just 8 hours (Fig. 5B). Treatment with pro-chemerin, chemA155 or chemS157 diminished the survival of E. coli to 65±7%, 53±10% and 60±5% of control, respectively (Fig. 5B). The chemerin-mediated decrease in bacterial viability grew even more pronounced at prolonged incubation times; after 24 hours, for example, the percentage of viable bacteria decreased to 53±9%, 35±9% and 35±4%, respectively (Fig. 5B). With the exception of 4h, when pro-chemerin seemed to be more effective compared to chemA155 and chemS157 (87±5, 95±3, 94±4), the truncated chemerin forms demonstrated stronger antibacterial activity against E. coli (Fig. 5B). Interestingly, chemerin-mediated inhibition of E. coli survival was detectable over a relatively wide range of concentrations, from 2 μM (the highest concentration tested leading to statistically significant growth inhibition) to as little as 0.125 μM-0.0625 μM for chemA155 and chemS157 (Fig. 5C). This was in contrast to LL-37, which was highly effective at inhibiting E. coli growth when used at 3 μM (Fig. 5A), but had almost no effect when tested at 1.5 μM (data not shown). The enhanced antibacterial activity of the truncated chemerin isoforms compared to pro-chemerin suggests an inhibitory role for the pro-chemerin C-terminal peptide. Addition of chemically synthesized C-terminal peptide KALPRS that is released from pro-chemerin by SspB, catK and catL, however, did not reduce the antibacterial activity of the larger truncated chemerin forms (data not shown). These data suggest that following release from the core protein, the C-terminal peptide no longer plays an inhibitory role. Interestingly, both chemerin Fc fusion proteins, Fc-chemS157 and Fc-chemR125 displayed comparable anti-bacterial activity against E. coli (Fig. 5D). These data suggest that in contrast to chemotactic activity, the antibacterial properties of chemerin appear to be localized closer to the N-terminus, since removal of 38 residues from the C-term did not abrogate its antibacterial activity.

Discussion

Here we identify a novel antimicrobial activity associated with chemerin, and show that host-derived cat L and cat K can cleave and activate the leukocyte attractant activity of chemerin, as well as enhance its antibacterial effects.

Various serine proteases have been reported to effectively convert chemerin to a potent chemoattractant in vitro. There is also a single example of a cysteine protease, S. aureus-derived staphopain B (SspB) that can efficiently activate human chemerin. In addition, host-originating cathepsin S and calpains have been reported to process mouse chemerin, although in this case the proteolysis of the C-terminus generates chemerin variants equipped primarily with anti-inflammatory properties (13, 23). Cysteine cathepsins of the papain-like family are normally confined to the endosomal/lysosomal network. However, there is evidence that certain cathepsins are also active extracellularly, either in association with the cell surface or in soluble form (14). Some cells such as macrophages and fibroblasts constitutively secrete cysteine cathepsins as zymogens (14). Moreover, macrophages have been reported to deploy enzymatically active cathepsin B, L and S and exhibit an elastin-degrading phenotype, indicating that macrophages can mobilize cysteine cathepsins to participate in the pathophysiologic remodeling of the extracellular matrix (17). Massive amounts of extracellular cathepsins, probably released from macrophages, are found in the bronchial tree of patients suffering from acute pulmonary inflammation (24). In addition, cat K is strongly implicated in maintaining the homeostasis of the extracellular matrix in the lung (25). Since chemerin mRNA is abundantly produced in lung (1, 2), collectively these data suggest that either cat L or cat K may be involved in chemerin processing in this organ. Alternatively, significant expression and/or activity of cat K and cat L in the joints of patients with rheumatoid arthritis and skin dermatoses, respectively (26, 27), together with reported chemerin immunoreactivity and/or bioactivity in psoriasis skin and inflamed synovial fluid (1, 9, 28) suggests that these cathepsins may play a role in chemerin cleavage in joints and skin. Since cathepsin-mediated processing releases chemerin attractant activity, these enzymes may have an important regulatory role in immune cell migration. Notably, the presence of pDC in lung as well as the inflamed joints and psoriatic skin (1, 9, 28, 29) supports the notion that cat K and cat L, through the generation of active chemerin, may contribute to pDC recruitment to these sites.

Our data also uncover a novel role for chemerin as a host-expressed antibacterial agent in host defense. Despite low primary sequence homology between chemerin and antibacterial cathelicidins, the conserved positioning of key cysteine residues leads to a predicted shared similar tertiary structure, although recent NMR assignment of human chemerin does not exclude a different fold (30). LL-37, the 37-aa C-terminal derivative of human cathelicidin hCAP18 is well known for its potent and broad-spectrum bacterial killing activity. However, chemerin is structurally similar to the cathelin-like N-terminal region. Interestingly, the cathelin-like domain of hCAP18 has been reported to possess antimicrobial activity, although the mechanism by which it inhibits bacterial growth is not known (20).

Chemerin may exert antimicrobial activity on the surface of skin and/or lung where it is locally expressed (1, 9). For example, the respiratory surface is continually exposed to pathogenic organisms, such as K. pneumoniae, which, as shown in this report, might be a direct chemerin target. Although either pro-chemerin or the C-term truncated chemerin forms displayed antibacterial activity, C-terminal processing augmented the inhibitory effect of chemerin on the growth of Enterobacteriaceae. However, our data suggest that pro-chemerin is also processed by bacterial proteases during incubation, although the protease(s) responsible remain to be identified. It will be interesting to map the specific chemerin domains/regions responsible for its anti-microbial activity. Our preliminary data suggest that most of the anti-bacterial activity is associated with the chemerin region(s) located within 65-115aa (data not shown). This is consistent with our data showing that Fc-chemS157 and Fc-chemR125 have similar antibacterial activity, although the inhibitory C-term peptide must be removed for full antibacterial effects.

Although the antimicrobial effects of chemerin on E. coli and K. pneumoniae were less potent compared with the classical antibacterial peptide LL-37, chemerin showed bactericidal properties at much lower concentrations. In general, pore-forming antimicrobial peptides, such as LL-37, require micromolar concentrations for activity. However, some antibacterial peptides, such as Lactocococcus-derived nisin operate in the nanomolar range (31). This ability is attributed to docking to a specific component on the bacteria cell wall for subsequent pore formation, or to the dual killing mechanisms of the peptide, which in addition inhibits bacterial cell wall biosynthesis (32). Chemerin might employ a similar strategy to exert its antimicrobial activity in the nanomolar range. However, since antimicrobial properties can be sensitive to pH and ionic composition of the peptide environment (31), it will be important to determine whether chemerin operates in conditions similar to those found in the skin and/or bronchial tree.

Thus, our present work uncovers a novel antibacterial property of chemerin and characterizes the activation of chemerin by host-derived cysteine proteases of the cathepsin family, and adds a new dimension to the ways chemerin may modulate and augment immunity.

Abbreviations

Cat B, L and K

cathepsin B, L and K respectively

CMKLR1

chemokine-like receptor 1

pDC

plasmacytoid dendritic cells

pro-chem, pro-chemerin, chemA155, S157 R125

chemerin truncated at indicated residue position, chemS, chemerin seum form, chem/SspB, staphopain B-truncated chemerin

SspB

staphopain B