Identification of Novel Chondroprotective Mediators in Resolving Inflammatory Exudates Free

This article is featured in In This Issue, p.2517

There is an emerging appreciation of the complex, integrated cellular and tissue processes that regulate the end phase of the acute inflammatory response (1). Timely resolution of a localized inflammatory response requires engagement of specific pathways and mediators to remove immune cells and reprogram tissue resident macrophages, yet ultimately to enact tissue reparative processes (2). This latter aspect of the resolution program is the most important in avoiding a maladaptation of the tissue (3): efficient repair in the absence of scarring and/or fibrosis prevents chronicity of the local inflammatory reaction and favors the regain of tissue functionality and homeostasis (4).

Persistent inflammation of the joint structures and progressive alteration/destruction of articular elements with absent or inadequate tissue repair are typical of rheumatoid arthritis and osteoarthritis (OA) (5). Although the primary risk factor is age, OA frequently results from traumatic joint injuries, such as those resulting from a sports injury or traffic accidents (5, 6). Surgical restoration of joint stability following such injuries seldom prevents future development of OA (6), which often leads to depression and social isolation of the affected persons, thus presenting a significant social and healthcare burden. Currently available treatment options rely heavily on conservative pain management strategies, through the use of analgesics and nonsteroidal anti-inflammatory drugs. These methods offer only temporary, modest pain relief, cause numerous side effects and comorbidities, and fail to address the underlying causes of articular tissue inflammation and degeneration. Given the near complete lack of reparative therapies for cartilage defects in routine clinical practice, innovative approaches to drug discovery are essential to advance this field.

Recent studies have indicated that endogenous peptides, such as the melanocortins, could limit chondrocyte inflammation and cartilage degradation, an effect linked to prevention of cell death (7). Furthermore, the delivery of platelet-rich plasma into the joint seems to ameliorate cartilage defects; however, the molecules responsible for these chondroprotective effects remain elusive (8–10).

Interrogation of the inflammatory environment has led to the discovery of endogenous proresolving molecules, such as resolvin E1 and protectin D1 (11). Similarly, we hypothesized that tissue-protective mediators might be present in resolving pleural exudates. We used acute pleurisy (12) as a source of exudates and a combination of in vitro and proteomic analyses, to identify molecules in resolving exudates that afford chondroprotection. Delivery of these mediators during ongoing murine experimental arthritis in vivo limited cartilage destruction in the knee joint. Altogether, this study identifies novel protein activities that promote chondrocyte anabolism providing novel cues to cartilage repair in joint disease.

Animals were maintained on a standard chow pellet diet with free access to water, and a 12 h light-dark cycle. All animal experiments were approved by the local Animal Use and Care Committee in accordance with the U.K. Animals (Scientific Procedures) Act, 1986.

Male Wistar rats (n = 6; 120–150 g; Charles River, U.K.) were anesthetized with isoflurane and 1% (w/v) λ-Carrageenan (Sigma-Aldrich, Poole, U.K.) solution was injected into the pleural cavity (12, 13). Pleural exudates were harvested at 6 h (inflammatory exudate) and 24 h (resolving exudate) postinjection by washing each cavity with 1 ml of 100 U/ml Heparin in PBS as an anticoagulant. Exudates contaminated with blood were discarded. Rat exudates were centrifuged at 1500 rpm for 10 min and the resulting cell-free supernatants were pooled and filtered using 0.22 μm filters (Amicon Ultrafree-MC; Millipore) for 7 h at 4°C to remove any residual carrageenan.

Male C57BL/6 mice (n = 4–6; ∼30 g body weight) were purchased from Charles River and administered 100 μl K/BxN serum on day 0 and day 2 of arthritis. On day 3, α₁-antitrypsin (AAT) (100 ng in 5 μl per mouse) or gelosin (GSN) (30 ng in 5 μl per mouse) was administered via intra-articular (i.a.) injection into the knee joint. The contralateral knees received saline (5 μl). On day 5, joints were collected in formalin for 48 h before decalcification in formic acid (10% w/v). Following processing and paraffin embedding, samples were prepared for histology. Coronal sections of the knee joint were taken (6 μm thickness) and stained with toluidine blue for the analysis of cartilage integrity. Images were captured using EVOS XL Core Imaging System (Thermo Fisher Scientific, Paisley, U.K.); percentage area toluidine blue positive was measured via ImageJ (National Institutes of Health) by splitting each image into its RGB channels and quantifying the positive area after applying a threshold.

The immortalized chondrocytes C-28/I2 were kindly provided by Dr. M. Goldring (14). Cells were cultured in complete medium: DMEM/Ham’s F12 (1:1; Life Technologies-Invitrogen, Paisley, U.K.), supplemented with 10% nonheat-inactivated FCS (Life Technologies-Invitrogen) and maintained at 5% CO₂. High-density three-dimensional (3D) micromass (MM) cultures were generated as previously described (15). The rat chondrocyte-restricted RCJ3.1C5.18 cell line (16–18) was maintained in DMEM supplemented with 10% NI-FCS, 1% dexamethasone (1 ×10⁻⁷ M), 100 U/ml penicillin, 100 μg/ml streptomycin (Omega Scientific, Tarzana, CA), and 5% CO₂. For experiments, the RCJ3.1C5.18 cells were plated in 24-well plates at 5 × 10⁴ cells per cm² and maintained for 24 h in complete medium without dexamethasone.

All chondrocyte cultures were serum starved for 24 h in phenol red-free DMEM/Ham’s F12 (1:1) supplemented with 1% insulin-transferrin-selenium G supplement (ITS; Invitrogen) to allow for collagen type 2 and aggrecan (ACAN) production. Chondrocytes were then stimulated as indicated in the individual figure legends.

The filtered resolving exudate was subjected to size-exclusion chromatography using a pre-equilibrated HiLoad Superdex-200 16/60 column with a flow rate of 1 ml/min. A total of two column volumes (equivalent to 240 ml) of buffer were used for isocratic elution and 1 ml fractions collected. Fractions corresponding to specifically chosen peak regions were pooled together into nine reconstituted fractions (see Fig. 2A) and later used for differential stimulation of chondrocytes.

The resulting pooled rat pleural exudate fractions were reduced in Laemmli sample buffer as previously described (19). After electrophoresis and silver staining, bands were subjected to in-gel digestion with trypsin using an Investigator ProGest (DIGILab) robotic digestion system. Tryptic peptides from the digests were separated on a reverse-phase nanoflow HPLC system (UltiMate 3000 RSLCnano; Thermo Fisher Scientific, Hertfordshire, U.K.) and eluted with a 40 min gradient (2–30% B in 35 min, 30–40% B in 5 min, 99% B in 10 min, and 2% B in 20 min, where A = 2% acetonitrile, 0.1% formic acid in HPLC-grade H₂O and B = 80% acetonitrile, 0.1% formic acid in HPLC-grade H₂O). The column was coupled to a PicoView nanospray source (New Objective). Spectra were collected from an ion trap mass analyzer (LTQ Orbitrap XL; Thermo Fisher Scientific) using full mass spectrometry (MS) over the mass-to-charge range 400–1600. MS/MS was performed on the top six ions in each full MS scan using the data-dependent acquisition mode with dynamic exclusion enabled. MS/MS peak lists were generated by extract_msn.exe software and matched to a human database (UniProtKB/Swiss-Prot_2013_08) using Mascot 2.3.01 (MatrixScience, London, U.K.). Carboxyamidomethylation of cysteine was chosen as fixed modification and oxidation of methionine as variable modification. The mass tolerance was set at 10 ppm for the precursor ions and at 0.8 Da for fragment ions. Two missed cleavages were allowed. Scaffold (v4.0.6; Proteome Software, Portland, OR) was used to calculate the spectral counts and to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at >95% probability as specified by the Peptide Prophet algorithm. Protein identifications were accepted if established at >99% probability with at least two independent peptides.

Total RNA was purified in quadruplicate from C28/I2 MMs (0.5 × 10⁶ chondrocytes per MM) using RNeasy Plus Mini Kit (Qiagen, Manchester, U.K.). cDNA was synthesized using SuperScript III First-Strand Synthesis System (Invitrogen). OligodT primers were used to synthesize the first strand cDNA (Promega, Southampton, U.K.). Following spectrophotometric quantification, RNA was normalized to 1 μg total RNA per cDNA reaction.

cDNA, diluted 1:5, was used as template to determine relative amounts of mRNA by real-time PCR (ABI Prism 7900 Sequence Detection System; Applied Biosystems) using specific primers (QuantiTect Primer Assay; Qiagen) and 2× Power SYBR Green Mastermix (Applied Biosystems, Thermo Fisher Scientific). The expression of COL2A1, ACAN, SOX9, matrix metalloproteinase (MMP)-13, and IL6 was interrogated. Cycle threshold (Ct) values were measured and calculated by Sequence Detector Software v2.4. Relative amounts of mRNA were normalized to endogenous control (GAPDH) and to vehicle. Relative mRNA contents were calculated as x = 2^−ΔΔCt, where ΔΔCt = ΔE − ΔC and ΔE = Ct_sample − Ct_GAPDH and ΔC = Ct_control − Ct_GAPDH (20).

Quantitation of cartilage-specific sulfated glycosaminoglycans (sGAGs) deposition was performed as previously described (15). Washed MMs were fixed with 4% glutaraldehyde solution (v/v in double distilled H₂O), and submerged in Alcian blue (AB) 8GS dye (48 h at 1% w/v) in 0.1 N HCl (pH < 1; Carl Roth, Karlsruhe, Germany) at room temperature. AB dye was extracted in guanidine-HCl (Sigma-Aldrich) for 48 h at RT. A₆₃₀ of the extracted dye was measured and concentration was quantified by interpolation with AB standard curve and normalized to DNA content (nanogram per microgram). DNA content was measured in the extracted dye solution by fluorescence (485/535 nm).

Levels of IL-6, IL-8, and MMP13 in cell-free supernatants were measured using a FlowCytomix multiple analyte detection system (eBioscience, Hatfield, U.K.). ACAN release by the chondrocyte MMs was measured by DuoSet ELISA Development Systems (R&D Systems, Minneapolis, MN).

All PCR data are reported as mean ± SEM unless otherwise indicated in individual figure legends. Significant differences in PCR experiments were determined with the nonparametric Kruskal–Wallis ANOVA test, followed by Dunn multiple comparison post hoc test. AB-staining experiments were analyzed using one-way ANOVA, followed by Dunnett multiple comparison post hoc test; histology-staining experiments were analyzed via two-way ANOVA, followed by Bonferroni post hoc correction for multiple comparisons. All statistical analyses were performed using GraphPad Prism 5.0, (GraphPad Software, CA). Values were considered significant for p < 0.05.

Acute and resolving pleurisy was produced by injection of carrageenan into the pleural cavity of rats. Thus, as expected, immune cell migration and exudate formation peaked at 6 h (12.6 ± 2.5 × 10⁶ cells; n = 6), whereas by 24 h the inflammation had subsided with reduced cell numbers (4.8 ± 1.3 × 10⁶ cells; p < 0.05), alongside published data (12, 13). The 6 h exudate (onset of inflammation) hereafter referred to as the inflammatory exudate, and the 24 h exudate, coincident with the initiation of resolution, hereafter referred to as the resolving exudate, were used.

Serial dilutions of cell-free exudates were tested on RCJ3.1C5.18 cells to monitor COL2A1, ACAN, and MMP13 gene expression (Fig. 1). The inflammatory exudates displayed typical catabolic activity: at its highest concentration (dilution 1:10), the exudate provoked ∼20-fold inhibition of COL2A1 and ∼8-fold inhibition of ACAN gene products, an effect paralleled by a significant increase in MMP13 gene transcription (Fig. 1; p < 0.05). The resolving exudate was mostly inactive on its own, except for a modest modulation of ACAN expression (p < 0.05; Fig. 1). Gene modulation was time-dependent with 48 h being optimal (Fig. 1B); this incubation period was selected for subsequent experiments. Stimulation of RCJ3.1C5.18 rat chondrocytes with both the inflammatory (at 1:100 fixed dilution) and resolving exudate significantly inhibited MMP13 transcription by 20%, alongside augmented expression of COL2A1 and ACAN mRNAs (p < 0.05; Fig. 1). Thus, the resolving exudate significantly reverted the catabolic effect of the inflammatory exudate.

Size-exclusion chromatography of the resolving exudate yielded eight fractions (Fig. 2A). A ninth fraction was included, representing a trough with very low protein content. Fractionated bioactivities of the resolving exudate were tested on human C-28/I2 chondrocytes grown in high-density 3D MMs. Initially samples were pooled into high m.w. (HMW) and low m.w. (LMW) fractions (Fig. 2A). Incubation of chondrocytes with HMW sample (1:30 dilution) reduced proteoglycan deposition (>60%; p < 0.05). This effect was paralleled by downregulation of COL2A1 and ACAN gene products (p < 0.05). Conversely, treatment with the LMW sample resulted in a significant increase in proteoglycan deposition (∼50%, p < 0.05) and modest anabolic effects on gene expression (Fig. 2B). The pooled LMW fraction was tested against IL-1β (20 ng/ml): in these settings, the anabolic properties of the LMW fractions were magnified with augmented COL2A1 (∼6-fold) and ACAN (∼5-fold) mRNA levels compared with IL-1β (p < 0.01; Fig. 2C).

The individual fractions (1:30 dilution) were tested against the catabolic stimulus IL-1β observing that the first three fractions were mostly inactive (Fig. 2D). HMW fractions 5 and 6 potentiated the cellular response to IL-1β with significant upregulation of catabolic genes IL6 (p < 0.001) and MMP13 (p < 0.05, compared with IL-1β stimulation alone). HMW fractions 4 and 5 had a moderate effect on COL2A1 transcription (p < 0.05; Fig. 2D), whereas fractions 8 and 9 counteracted IL-1β with a significant stimulation of the anabolic response, as evident from increased COL2A1 (p < 0.01) and ACAN gene products. These effects prompted identification of the protein content of these chondromodulatory fractions.

Tandem gel-LC-MS-MS proteomic analysis was performed on all fractions. The proteomic analysis identified a total of 61 proteins across all fractions, which are listed in Supplemental Table I alongside the number of proteolytic peptides sequenced for each hit. Based on their relative abundance, and detailed literature review, we selected and tested >7 candidate proteins (data not shown) that had indications for functions relevant to chondroprotection. Of those, AAT, GSN, and hemopexin (HX) were brought forward (see 20Discussion) for more in-depth functional analysis.

Addition of these three proteins to unstimulated chondrocyte MMs displayed minimal changes in the cellular response: as an example, AAT triggered 2-fold upregulation of COL2A1 (p < 0.05), and a concentration-dependent downregulation of MMP13, optimal at 30 μg/ml (p < 0.001; Fig. 3A–D). However, bioactions were more acute in the presence of IL-1β; in these settings, addition of AAT improved COL2A1 and ACAN gene expression by ∼4-fold and SOX9 transcription was enhanced by ∼3-fold, compared with IL-1β stimulation alone (p < 0.01; Fig. 4A, 4B). Simultaneously, incubation with this glycoprotein reduced (∼70%) 1) IL-1β–induced MMP13 gene expression and 2) secretion of IL-6 and IL-8 (Supplemental Table II).

On its own, GSN afforded minimal modulation on COL2A1 and ACAN expression, with a modest effect on SOX9 (p < 0.01; Fig. 3A–C). Again, in the presence of IL-1β, GSN prevented the catabolic response by upregulating the anabolic genes (p < 0.05; Fig. 4A–C), although concomitantly inhibiting MMP13 transcription in a concentration-dependent manner (p < 0.01; Fig. 4D). IL-1β–induced IL-6 secretion was also inhibited by GSN (maximum effect at 0.1 μg/ml; p < 0.05; Supplemental Table II).

Varied results were obtained for HX, which on its own produced a more marked yet balanced response, with increased expression of anabolic and catabolic gene products (Fig. 3). In the presence of IL-1β, HX significantly reverted the catabolic response on all anabolic genes tested (p < 0.05; Fig. 4A–C), although only being effective counteracting IL-1β-induced MMP13 transcription at the lowest concentration tested (p < 0.05; Fig. 4D). HX significantly reduced IL-1β–induced IL-6 production (Supplemental Table II).

Functional data on sGAG deposition were obtained to complement gene expression studies. MMs were incubated for 48 h with a concentration range of AAT, GSN, and HX (0.1–30 μg/ml) in the presence of IL-1β. Fig. 5A and 5B shows that AAT counteracted the action of IL-1β by bringing sGAG deposition levels back to those of vehicle-treated MMs. For GSN (0.1–10 μg/ml) and HX (0.1–30 μg/ml) a bell-shaped concentration-response curve was produced, with optimal effects at 0.3 and 1.0 μg/ml, respectively (>3-fold increase; p < 0.001; Fig. 5B).

Next, we assessed the functions of the polypeptides in more complex settings like those promoted by chondrocyte stimulation with OA synovial fluid (OASF), rich in variety of proinflammatory cytokines and cartilage degrading enzymes (21). AB staining revealed that OASF-stimulated chondrocyte MMs (1:100 dilution as selected from preliminary analyses, data not shown) markedly inhibited sGAG deposition (p < 0.01, Fig. 5C, 5D). Treatment of MMs with AAT, GSN, or HX abrogated the catabolic effect of the fluids. Similar counteracting properties were evident when we quantified OASF-induced release of IL-6, IL-8, and MMP13 (Fig. 5E).

To complete these in vitro discoveries we selected an animal model of inflammatory arthritis, where immune cell and exudation are also present but are additionally coupled to significant cartilage destruction, complementing the initial studies with the acute inflammatory exudates. The serum-transfer model of inflammatory arthritis has been applied to unveil chondroprotective mechanisms as evoked by extracellular vesicles (22) or calcitonin nanomedicines (23, 24). GSN or AAT were administered by i.a. injection and cumulative data for both GSN and AAT are reported in Fig. 6. Toluidine blue staining of multiple focal sections of each joint showed widespread loss of staining in sham arthritic mice, where articular cartilage was eroded by ∼60% (sGAG content; p < 0.05 compared with naive mice). Importantly, therapeutic i.a. injection of either AAT (100 ng) or GSN (30 ng) recovered cartilage integrity by 52 and 34%, respectively, compared with vehicle-injected contralateral joints (p < 0.05; Fig. 6).

We applied an unbiased approach to identify proteins with chondroprotective activity in resolving exudates. Using a combination of in vitro and in vivo techniques, complemented by analyses with human synovial fluids, we describe the presence of the anabolic factors AAT, GSN, and HX in resolving exudates and characterize their chondroprotective properties. This approach could become paradigmatic for identifying novel therapeutic leads.

Acute inflammation is tightly regulated in space and time, with an onset phase followed by a resolution phase that paves the way to restoring the homeostatic balance. This second phase of inflammation has gained interest in the last decade, with the characterization of specific proresolving mediators (1) and receptors (25) and definition of fundamental proresolving processes, which until recently have been predominantly studied in the context of immune cell reactivity and behavior (4, 26). However, the end point of resolution is the repair of the affected tissue with recovery of normal physiological functions. As reasoned in the 1Introduction, few studies have addressed the pharmacological modulation of skin repair, applying models of wound healing (27, 28). However, several other tissues would undergo reparative processes at the end of inflammatory episodes and, equally, could be amenable to therapeutic exploitation using approaches that emerge from the science of resolution. This overarching hypothesis guided the current investigation by employing a model of resolving inflammation that has previously allowed identification of novel tissue-protective pathways, including those centered on hemoxygenase-1 (13), cyclo-oxygenase type 2 (12), and more recently, lipidomics approaches, which have led to the identification of novel lipid mediators in biological fluids (29), e.g., resolvin D1, endowed with cartilage-protective actions during ongoing inflammatory arthritis (30). This approach is especially pertinent for the attempted repair of cartilage, as cartilage erosions and sport injuries often lead to osteoarthritis with no treatable parameters.

Distinct biological activities were detected in pleural exudates harvested at onset and peak of inflammation, the latter coinciding with the onset of resolution (31). We applied a streamlined experimental strategy, where rat chondrocytes were rapidly substituted with 3D cultures of human cells. Relevant gene expression patterns were measured together with deposition of extracellular matrix, culminating with proteomic analysis coupled with mass spectrometry—a powerful tool for the exploration and discovery of novel biomarkers, and an approach well known in the identification of OA-driving players (32). A comprehensive review of the proteomics hits was conducted, taking into account their relative abundance in the various fractions in parallel with a comprehensive literature review of the identified proteins, which was cross-referenced against any possible anabolic or anti-inflammatory properties that they exhibit in various tissue systems. The candidate proteins fitting the criteria were taken forward for further testing in our chondrocyte assays, based on the hypothesis that such effects might be translatable to chondrocytes and/or cartilage and/or experimental arthritis.

Using this approach, we identified several candidate mediators and selected three of them, namely AAT, GSN, and HX, for which we define new properties in the context of chondrocyte biology. All three factors were detected by Western blotting in both the resolving and inflammatory exudates, with levels more pronounced in the resolving exudate (data not shown). Herein, we progressed our investigation by utilizing a dual translational remit: 1) testing the identified factors against human arthritic synovial fluids, which allowed for assessment of their effects against the numerous catabolic and proinflammatory mediators, contained in the complex arthritic synovial fluids, which drive cartilage destruction in arthritis (21), and 2) in vivo administration of those proteins in a model of inflammatory arthritis. Of interest, and in support of our approach, recent proteomic analysis of human arthritic synovial fluids identified these proteins though without postulating relevant bioactions (33).

AAT, GSN, and HX are acute phase proteins released by the liver in settings of systemic stress. AAT is a 52-kDa glycoprotein functioning as the major natural inhibitor of serine proteases (34) with potent effects on neutrophil elastase (35, 36). It is produced mainly by hepatocytes, yielding circulating steady-state levels of 1–3 mg/ml; these levels rise during infection and inflammation (33) and remain elevated for 7–10 d (37). Local inflammation-driven production of AAT by endothelial and myeloid cells is also reported (37). The property of AAT to inactivate protease activity allows both direct and indirect reduction of inflammation as some of these proteases activate protease-activated receptors, thus affecting the release of proinflammatory cytokines (38). In addition, AAT possesses a broad spectrum of anti-inflammatory (37, 39) and immunomodulatory properties, unrelated to protease inhibition (37, 40), including elevation of cAMP in target cells (41) and inhibition of the aggrecanase ADAMTS-4 (42). A recent study reported the antiarthritic properties of an ATT-Fc fusion protein in murine acute gouty arthritis with reduced release of IL-1β and induction of IL-1 receptor antagonist (43). Equally pertinent is the observation that AAT treatment reduces autoimmunity and delays arthritis development in a mouse model of collagen-induced arthritis (44). Our data add to the current wealth of AAT biology showing that at low nanomolar concentrations (1–10 μg/ml equivalent to 19–192 nM), this peptide partially blocked the catabolic response evoked by IL-1β in chondrocytes both for anabolic genes and MMP13. In the AB assay, where IL-1β was used as a stimulus, AAT was active at concentrations as low as 0.1 μg/ml (1.9 nM), whereas full protection against the catabolic status induced by complex OASF was evident at 10 μg/ml (192 nM), probably due to the presence of multiple catabolic and proinflammatory mediators in the inflamed synovial fluid (21).

GSN and HX displayed similar, yet not overlapping, profiles to AAT. GSN, an 82-kDa Ca²⁺-regulated actin filament severing, capping, and nucleating protein, seems to be involved in the control of biological processes beyond filament remodeling. There are two isoforms of GSN: plasma or pGSN, and cytosolic or cGSN (45). Circulating levels of pGSN in the blood of healthy individuals are ∼200 ± 50 mg/l and may be a key component of an extracellular actin scavenging system operating during tissue damage (46). pGSN decreases during acute injury and inflammation, and rGSN treatment of animals improves outcomes following major trauma (47, 48), sepsis (49, 50), and burn injuries (51). Interestingly, circulating levels of pGSN are decreased in the plasma of patients suffering from rheumatoid arthritis (52) and analyses of the synovial fluids of those patients show consumption of pGSN into the inflamed joints (52). The modulatory functions of GSN we identified on chondrocytes are further supported by the findings that GSN^−/− mice experience exacerbated arthritis (53). Additionally, GSN is expressed by both normal and hypertrophic chondrocytes (54) yet its functions have been scarcely investigated in the context of the biology of this cell type. In analogy to ATT, abundant GSN has been identified in human inflammatory fluids (33, 52, 55).

HX (60 kDa) is primarily produced by hepatocytes, with moderate amounts also synthesized by the nervous system, skeletal muscle, and kidney (56). HX binds heme with high affinity controlling heme-iron availability in tissues and cells. In the context of inflammation, HX can inhibit granulocyte recruitment and—in systemic sepsis—the production of proinflammatory cytokines (57). A negative modulatory function on Th17 T-cell differentiation has been reported (58), with clear potential for antiarthritic benefits (59).

In our experimental settings, (p)GSN promoted anabolic gene modulation with significant changes in vehicle (e.g., SOX9) and cytokine-stimulated (COL2A1, ACAN, and MMP13) conditions. HX was quite potent on IL-1β–stimulated MMs, with significant effects at 3–10 μg/ml, (50–167 nM). Both GSN and HX—in analogy to AAT—promoted proteoglycan deposition against IL-1β and OASF stimulation with optimal concentrations of 0.3 and 10 μg/ml, respectively.

A common thread is emerging here, with the notion that acute phase proteins—previously categorized passive regulators of inflammation—are endowed with specific and relevant properties at different cellular levels, including those relevant to joint diseases, from skeletal muscle to T cells and, from our data, to chondrocytes. In our rationale of identification of novel effectors of chondroprotection, it was important to establish efficacy against the human synovial fluids, thus indicating a translational potential for the chondromodulatory proteins identified in the resolving exudates, because both contain a variety of stimuli, rather than a single cytokine (e.g., IL-1β).

Equally important was to expand these in vitro analyses and establish proof-of-concept activity by monitoring efficacy of AAT and GSN on knee cartilage during ongoing arthritis. Local administration of a low dose of either AAT or GSN, calculated to fall within the range that afforded in vitro chondroprotection, prevented the loss of cartilage integrity. A therapeutic approach was used, where the peptides were injected after initial signs of arthritis were evident: it was important to establish pharmacological efficacy against internal controls, hence the vehicle-injected contralateral knees. The exact mechanism of action of both GSN and AAT is currently being further investigated. We can postulate that for GSN we might have replenished the protein, which would be naturally lacking in the synovial fluids of arthritic mice, thus restoring chondroprotection. For AAT, it is plausible that along with IL-1β reduction and IL-1RA augmentation, this protein can reduce neutrophil infiltration and modify the activity of neutrophil elastase in the inflamed joints. In both cases, the in vivo mechanisms are currently being investigated in follow-up studies.

In conclusion, we propose that innovation in the identification and/or development of novel antiarthritic strategies may benefit from the exploitation of endogenous pathways of the resolution of inflammation. Reasoning that the end point of resolution is tissue repair, healing, and regaining of physiological functions, we have analyzed experimental inflammatory exudates and focused our analyses on proteins identified in human synovial fluids. The properties of these proteins were then characterized on human chondrocytes, detailing partial over-lapping but also distinct biology for AAT, GSN, and HX on basal or activated cells. Importantly, an in vivo proof-of-concept study held true on the in vitro data produced for AAT and GSN. Collectively, the data presented here add AAT, HX, and GSN to the list of arthritic disease modifiers with a potential therapeutic value. Of note, i.a. injection of glucocorticoids is still a common therapeutic option (see Ref. 60 for a recent review), and we reason that this route of administration could be applied—at least initially—as a delivery method of these chondroprotective proteins. However, further deciphering on how they operate and identification of their pharmacophores or of their receptor targets on chondrocytes, as well as exploiting novel drug delivery methods, could inform the development of innovative antiarthritic and regenerative therapeutic strategies.

The authors have no financial conflicts of interest.