Chemerin 156F, generated by chymase cleavage of prochemerin, is elevated in joint fluids of arthritis patients

A cDNA clone encoding human chemerin (NM_002889) was obtained from Open Biosystems (Huntsville, AL, USA). A cDNA fragment encoding chem156F was amplified by PCR using Platinum Pfx DNA polymerase (Invitrogen, Carlsbad, CA, USA) with termination codons inserted into the primers (CHEM-5 Ngo Primer 5′-AGCCGGCCACCATGCGACGGCTGCTGAT used as sense primer, and hCHEM-3 Nhe-F Primer 5′-AGCTAGCTTAGAAGGCGAACTGTCCAGGGA used as anti-sense primer) [13]. The cDNA encoding chem156F was cloned into a ubiquitous chromatin opening element (UCOE) vector, pCET-1019AS-puro, provided by Millipore. The plasmid encoding chem156F was transfected into CHO-S cells and subsequent selection was performed as described previously [15]. Plasmids encoding cDNAs for mouse chemerin forms, homologous to human chem156F (mouse mchem155F) and human chem157S (mouse chem156S), were produced in a similar fashion.

The selected CHO-S clone expressing chem156F was grown in suspension culture in CD Opti-CHO medium. To avoid cleavage of chemerin, the CHO-S cells were expanded in 225-cm² T-flasks for 3 days before switching to CD-CHO medium. Static culture was then used to produce chem156F in 225-cm² T-flasks for 3 days at 37 °C and purified as described previously [15]. The purified protein was characterized by SDS-PAGE. Mouse mchem156S and mchem155F were produced, purified, and characterized by the same methods.

L1.2 cells stably expressing human CMKLR1 were provided by Dr Brian Zabel and Dr Eugene C. Butcher (Stanford University School of Medicine and Veterans Affairs Palo Alto Health Care System) and cultured in RPMI 1640 with 10% FBS and 1 mg/ml geneticin [15]. Cells were labeled with Quest Fluo-8 AM dye, and then stimulated with recombinant chemerin forms in a 96-well plate calcium mobilization assay [15]. Each assay was repeated at least three times for each protein or peptide.

Anti-human chemerin 156F (anti-chem156F) IgY was raised in chickens against the peptide sequence CZ¹⁵¹PGQFAF¹⁵⁶ conjugated to KLH (Aves Labs, Tigard, OR, USA) and anti-chem156F IgY from eggs was purified by affinity chromatography to its cognate peptide bound to sepharose. To eliminate cross-reactivity with the chem157S form, negative selection affinity chromatography was performed by adsorption on an amino-linked (AminoLink kit; Thermo Scientific, Rockford, IL, USA) column coupled with the noncognate chemerin peptide KC¹⁵²GQFAFS¹⁵⁷. Rabbit polyclonal anti-human chemerin 163S (anti-chem163S), anti-human chemerin 158 K (anti-chem158K), anti-human chemerin 157S (anti-chem157S), and chicken polyclonal anti-human chemerin 155A (anti-chem155A) were prepared as described previously [14].

Purified recombinant chem163S, chem158K, chem157S, chem156F, and chem155A (200 ng each) were separated by SDS-PAGE under reducing conditions followed by western blot analysis with 500 ng/ml of the purified anti-chem156F IgY. The blot was developed with peroxidase-conjugated goat anti-chicken IgY antibody (100 ng/ml; Aves Labs) and detected using ECL (GE Healthcare, Sunnyvale, CA, USA).

A rat monoclonal anti-human chemerin antibody (4 μg/ml; R&D Systems, Minneapolis, MN, USA) in PBS buffer was coated onto 96-well ELISA plates, and nonspecific binding sites were blocked with 1% BSA in PBS for 1 h before sample addition. Purified recombinant chem163S, chem158K, chem157S, chem156F, and chem155A were used as standards to construct calibration curves. Samples and standards were diluted with 1% BSA in PBS and incubated in the wells for 2 h. After washing with 0.05% Tween 20 in PBS, the samples were incubated with specific cognate antibodies (500 ng/ml) in PBS with 1% BSA for 1 h. In the case of anti-chem163S, chem158K peptide (10 μg/ml) was added to remove residual cross-reactivity with chem158K. In the case of anti-chem158K, chem163S peptide (10 μg/ml) was added to remove residual cross-reactivity with chem163S. In the case of anti-chem156F, chem155A peptide (10 μg/ml) was added to remove residual cross-reactivity with chem155A. For anti-chem155A, 10 μg/ml each of chem157S and chem158K peptides was added to remove the residual cross-reactivity with chem157S and chem158K. After washing with 0.05% Tween 20 in PBS, the samples were incubated with peroxidase-conjugated goat anti-rabbit IgG antibody (100 ng/ml) or peroxidase-conjugated goat anti-chicken IgY antibody (100 ng/ml) in PBS with 1% BSA for 1 h. After washing, tetramethylbenzidine substrate (Alpha Diagnostic International, San Antonio, TX, USA) was incubated for 10 min followed by the addition of Stop Solution (Alpha Diagnostic International) and absorbance at 450 nm measured. The concentrations of human chemerin forms were calculated from the calibration curves of the purified chemerin standards.

Human synovial fluid samples were obtained under a protocol for discarded specimens approved by Stanford University School of Medicine or Partners Healthcare Institutional Review Boards. The samples were stored frozen at − 80 °C until the time of analysis. After thawing, 30 μl of a synovial fluid sample diluted to 400 μl with PBS was mixed with 50 μl of heparin-agarose (Sigma, St. Louis, MO, USA) and Complete Protease Inhibitor (Roche Applied Science, Pleasanton, CA, USA) [14], and the eluates were assayed by the specific ELISAs.

Thawed synovial fluid samples were diluted 2-fold to 5-fold and assayed by human CMA1/Mast Cell Chymase ELISA kit (Life Span Biosciences Inc., Seattle, WA, USA).

Peptides (1–3 μM) representing the C-terminus of chem163S (chem163S-15mer) and chem156F (chem156F-8mer) were treated with chymase (30 nM) in PBS for 30 min at 37 °C before terminating the reaction with 100 nM PPACK. Then 100 μl of each reaction mixture was loaded onto a Zorbax Eclipse Plus C18 (4.6 mm × 150 mm) column (Agilent, Santa Clara, CA, USA) and separated with a 0–40% acetonitrile gradient in 0.1% trifluoroacetic acid (v/v) by HPLC. The concentration of peptides present in the reaction mixture was determined from the area under the curve for that peptide by comparison to a standard curve (1 μM–3 mM) of that peptide. The values for K_M and k_cat were determined by fitting to the Michaelis–Menten equation by nonlinear regression analysis using Prism v6 (Graphpad, San Diego, CA, USA). The experiments were performed three times independently, and the data were pooled for analysis.

Comparison of two samples was by Student’s t test; multigroup comparisons were by ANOVA followed by post hoc Kruskal–Wallis analysis. The analysis was carried out using Prism v7 (GraphPad, La Jolla, CA, USA). p < 0.05 was considered significant.

DNA encoding chem156F was cloned into a plasmid, pCET-1019AS-puro, that has a UCOE element upstream of a guinea pig CMV promoter before transfection into CHO-S cells [15]. Two days after transfection, stable clones were selected by limiting dilution in 10 μg/ml puromycin, and 90 clones were screened for production of chem156F. A clone was identified that produced chem156F with a productivity of 4.74 pg/cell per day. Recombinant chem156F was purified by single-step cation exchange chromatography (Additional file 1: Figure S1A) and its C-terminus was confirmed by mass spectroscopy as described for purification and characterization of other recombinant chemerin forms [15]. The purified chem156F together with other purified recombinant human chemerins, chem163S, chem158K, chem157S, and chem155A, were shown to be > 95% pure by SDS-PAGE (Additional file 1: Figure S1B). Antibody raised against chem156F reacted only with chem156F (Additional file 1: Figure S1C).

We evaluated the potency of the purified chem156F in induction of calcium mobilization using CMKLR1-transfected L1.2 cells. Purified chem163S and chem157S were used as controls (Fig. 1a). The EC₅₀ values for chem163S and chem156F were 252 ± 141 nM and 133 ± 41.5 nM respectively, while the EC₅₀ value of chem157S was 5.83 ± 2.48 nM which is > 20-fold more potent than chem156F and > 40-fold more potent than chem163S. Thus, chem156F was about twice as active as the precursor, chem163S, but very much less active than chem157S, the most active form of chemerin based on the CMKLR1-transfected cell assay. The relative potencies of chem163S, chem156F, and chem157S were similar to that of their equivalent C-terminal peptides [15].

Unlike in mouse where mouse mchem156S, the homolog of human chem157S, and mouse mchem155F, the homolog of human chem156F, have approximately the same activity (data not shown), in humans chem156F is significantly less potent than chem157S. One explanation is that species differences between human and mouse CMKLR1 lead to chem156F being active on mouse but much less so on human CMKLR1. We tested this possibility by determining the potency of mouse mchem156S and mouse mchem155F in chemotaxis on L1.2 cells transfected with human CMKLR1. As on the L1.2 cells transfected with mouse CMKLR1, mouse mchem156S and mchem155F were equipotent showing that the difference in potency is not due to species-specific activity of chemerin on CMKLR1 cells (Fig. 1b).

In order to specifically detect chem156F in human samples, we prepared chicken anti-chemerin 156F (anti-chem156F) by immunization with a peptide representing the C-terminal sequence of chem156F. The specificity of the affinity-purified anti-chem156F IgY was demonstrated by western blot analysis in which anti-chem156F IgY only recognized its cognate protein but not the other four chemerin forms (Additional file 1: Figure S1C).

Antibodies directed against the different forms of chemerin were used to develop sandwich ELISAs specific for the different chemerins, chem163S, chem158K, chem157S, chem156F, and chem155A. Recombinant chem163S, but not recombinant chem158K, chem157S, chem156F, or chem155A, was specifically detected in a dose-dependent manner by anti-chem163S, with a lower limit of detection of 0.5 ng/ml (Fig. 2). Likewise, anti-chem158K specifically recognized recombinant chem158K and not the other forms, whereas anti-chem157S and anti-chem155A specifically recognized recombinant chem157S and chem155A, respectively (Fig. 2). Anti-chem156F recognized recombinant chem156F, with some cross-reactivity against recombinant chem157S (Fig. 2). In order to measure chem156F accurately, taking into account the antibody’s cross-reactivity with chem157S, we used the signal from the chem156F ELISA to give the total concentration of [chem156F + chem157S] from which we then subtracted the chem157S concentration, as measured by the specific chem157S ELISA, to give the concentration of chem156F. Using this panel of ELISAs, the levels of different forms of human chemerin can be determined specifically.

We have previously observed extensive proteolytic cleavage of chemerin (~ 75%) in synovial fluid samples from inflammatory arthritis patients and shown that chem158K is the dominant form [14]. To determine whether chem156F is present in the synovial fluid of arthritis patients, we investigated the levels of the five chemerin forms in synovial fluid samples from 25 patients with OA and 25 patients with RA. In OA synovial fluid samples, the levels of chem163S, chem158K, chem157S, chem156F, and chem155A were 24.28 ± 4.85 ng/ml, 77.04 ± 12.25 ng/ml, 1.31 ± 0.77 ng/ml, 24.06 ± 5.51 ng/ml, and 0.57 ± 0.43 ng/ml (mean ± SEM, n = 25) respectively (Fig. 3a). In RA synovial fluid samples, the corresponding levels were 3.06 ± 0.97 ng/ml, 40.57 ± 10.47 ng/ml, 0.29 ± 0.29 ng/ml, 20.35 ± 5.19 ng/ml, and 0 ± 0 ng/ml (mean ± SEM, n = 25) (Fig. 3b). In OA synovial fluid samples, when the fractions of individual chemerins from the sum of chemerin forms of each sample were averaged, 52% was chem158K, 20% was chem156F, while uncleaved chem163S was 23% (Fig. 3c). In RA joint fluid samples, the majority of chemerin was found to be chem156F (76.7%), with another 21.3% of chemerin presented as chem158K (Fig. 3c). The fraction of cleaved chemerins (chem158K + chem157S + chem156F + chem155A) was 75% in OA synovial fluid samples and 100% in RA synovial fluid samples (Fig. 3d), demonstrating that significant cleavage and processing of chemerin occurs in this extravascular compartment. Only minimal levels of chem157S and chem155A were detected in synovial fluids from either type of arthritis.

One enzyme that might generate chem156F from chem163S is chymase [8]. Therefore, we investigated chymase levels in OA and RA synovial fluid samples of arthritis patients. In OA and RA synovial fluid samples, the levels of chymase were 349.7 ± 200.2 pg/ml (n = 10) and 942.6 ± 366.8 pg/ml (n = 8) (mean ± SEM), respectively (Fig. 4a).

To determine the kinetic constants for cleavage of chem163S by chymase, chem163S peptides consisting of the last 15 amino acids of the C-terminus of chemerin (chem163S-15mer) were subjected to chymase cleavage followed by HPLC analysis. Hydrolysis of chem163S-15mer by chymase showed K_M of 556.1 ± 174 μM, k_cat of 436.4 ± 44.96/min (mean ± SEM), and k_cat/K_M of 0.78/μM/min (Fig. 4b and Table 1).