Alpha- 1- antitrypsin reduces inflammation and exerts chondroprotection in arthritis

While new treatments have been developed to control joint disease in rheumatoid arthritis, they are partially effective and do not promote structural repair of cartilage. Following an initial identification of α- 1- Antitrypsin (AAT) during the resolu-tion phase of acute inflammation, we report here the properties of this protein in Ex vivo downregulated adamts5 In vitro studies chondrocytes SERPINA1 rAAT chondrogenic activation PKA- dependent inhibition of Wnt/β- catenin


| INTRODUCTION
Arthritis is a common health problem in the global population, affecting more than 350 million people, causing more disability than any other condition, including heart disease and diabetes. Rheumatoid arthritis (RA) is mainly characterized by joint inflammation and destruction of cartilage and the underlying bone tissues, leading to loss of motion in the affected joint. Despite the use of anti-inflammatory agents and disease modifying anti-rheumatic drugs, there are no treatments able to halt or reverse the ongoing destruction of the joint.
First isolated in 1955 and named after its ability to inhibit trypsin, 1,2 alpha-1-Antitrypsin (AAT) was identified as a potent serine protease inhibitor, 3 with highest affinity for neutrophil elastase (NE). 4,5 Produced by hepatocytes and secreted in the blood, AAT concentrations rise ~4-fold during infection and inflammation, 6 remaining elevated up to 7 days, 7 suggesting a function in host protection. It is now reported that several cell types can produce AAT [7][8][9][10][11] implying tissue-specific functions beyond AAT roles in acute inflammation.
Currently, intravenous infusions of human plasmapurified AAT preparations are used for the treatment of lung diseases associated with severe inherited AAT deficiency. But the pleiotropic properties of AAT provide a rationale for using this therapy outside of this disease. AAT therapy is beneficial in several experimental settings including transplant rejection, ischemia-reperfusion injury, graft-vs-host disease, autoimmune encephalomyelitis, preeclampsia, and inflamed pancreatic islets. [12][13][14][15][16] Thus, AAT may be a more complex mediator of inflammation and host response than initially proposed.
In collagen-induced arthritis in mice, AAT as well as AAT gene therapy, significantly impacted disease onset and progression 14,16,17 while in a model of gouty arthritis, AAT-Fc fusion protein reduced joint swelling and cellular infiltration. 18 We have recently identified AAT as an abundant protein in resolving (end-of-acute inflammation) exudates, suggesting that it could be exerting tissue-protective and resolving actions. Initial analyses in experimental arthritis and in vitro on 3D chondrocyte cultures showed AAT to reduce joint inflammation and inhibit cartilage catabolism. 19 However, the molecular mechanisms responsible for these effects were not studied and, more broadly, whether such macroscopic actions could be secondary to NE inhibition remains unclear. Here, we have expanded our knowledge on the biology of AAT and established at least some of the mechanisms underpinning its effects on chondrocytes, in conditions where NE and other serine proteases are unlikely to be present.

| MATERIALS AND METHODS
2.1 | Animals and models of inflammatory arthritis C57BL/6 mice (22-30g; 10-12 weeks old) or male Wistar rats (250-450g; 8-15 weeks old), purchased from Charles River Laboratories, were maintained on a standard chow pellet diet with access to water ad libitum, and a 12-hour light-dark cycle. All animal experiments were approved by the local Animal Use and Care Committee in accordance with the United Kingdom Animals (Scientific Procedures) Act, 1986 and Canadian Council for Animal Care guidelines (http://www.ccac.ca/) and were approved by the Dalhousie University Committee on Laboratory Animals.

| KBxN serum-transfer arthritis
Male C57BL/6 mice (n = 5-6 per group) were administered 100 µL (ip) of arthritogenic K/BxN serum at day 0 and day 2 of arthritis, as described. 19 On day 3, mice were administered either with vehicle (saline), or AAT (100 µg/mouse, ip, Abcam, ab91136, Cambridge, UK). Disease development was assessed through clinical score (0-3) of arthritis, where 0 equates no evidence of inflammation; 1, subtle inflammation of metatarsal phalangeal joints, individual phalanx evaluated; 2, edema on dorsal and ventral surface of paws; 3, major edema on all aspects of paws (maximum score 3 per paw, 12 per animal). Paw edema was measured by water plethysmometry (Ugo Basile SRL, Milan, Italy) and together with the cumulative disease incidence were also recorded daily. Mice were weighed daily for signs of cachexia.

| NE-induced knee inflammation
Mice (n = 8 per group) were anesthetized using isoflurane (2%-4%; 100% oxygen at 1 L/min), the right knee joint was shaved, and baseline knee joint diameter was recorded using a digital micrometer (Control Company, Friendswood, TX, USA), as described. 20 Briefly, mice were injected with a mixture of saline (5 µL):NE (5 µg/5 µL), or AAT:NE (100 ng:5 µg in 10 µL), into the right knee. Post-injection, the knee was extended and flexed for 30 seconds to disperse the mixture throughout the joint. Vascular processes were monitored by intravital microscopy and laser speckle contrast analysis at 4 hours post-treatment. A separate cohort of mice was used to monitor mechanical hypersensitivity using von Frey hair algesiometry and weight-bearing deficits using dynamic weight-bearing apparatus.

| Histological and microscopy analyses
Mice were anesthetized and sacrificed by cervical dislocation at different times, related to the specific model applied, thus day 5 post-serum, 4 hours post-NE, or 21 days post-CFA injections. Knee joints were collected and fixed in 10% Neutral Buffered Formalin (v/v) and decalcified in 10% Formic Acid (v/v), followed by paraffin embedding. Coronal sections (5 µM) of the knee joints were stained with Hematoxylin and Eosin, or Toluidine Blue, according to the standard methods, to visualize the levels of joint inflammation and cartilage damage, respectively. Standard light microscopy was used to determine degree of synovitis, pannus formation, and cartilage damage; both rated from 0 (no disease) to 3, severe joint destruction, by blinded examiners. Percentage area toluidine blue positive was measured by ImageJ (NIH) by splitting each image into its RGB channels and quantifying the positive area after applying a threshold.

| Intravital microscopy
Intravital microscopy was used to assess leukocyteendothelial interactions within the knee joint microvasculature, as described. [21][22][23] Briefly, mean arterial pressure was measured on a differentially amplified BP monitor (BP-1; World Precision Instruments, Sarasota, FL, USA) which was connected to carotid artery cannula through an in-line pressure transducer (Kent Scientific Corporation, Torrington, CT, USA). Leukocytes were stained using Rhodamine 6G (0.05%; mice-0.06 mL, rats-0.12 mL) injected via the jugular vein. The knee joint microvasculature was accessed by removing a small piece of overlying skin (1.0 × 0.5 cm) and visualized using a Leica DM2500 microscope (Leica Microsystems Inc, Richmond Hill, ON, Canada; magnification 200×). Straight, branchless, postcapillary venules (d = 20-50 μm) were located on the knee joint capsule and three 1-min recordings were made for each time point using a BC-71 AVT camera (Horn Imaging, Aahen, Germany). Rolling and adherent leukocytes numbers were recorded and averaged.

| Nociception assays
2.3.1 | Measurement of mechanical hypersensitivity von Frey hair filaments were used to assess ipsilateral hind paw mechano-sensitivity in mice. 23 The 50% withdrawal threshold was calculated using Dixon's up-down method. 24,25 A filament was applied to the plantar surface of the hind paw for 3s-a positive response was considered when the animal licked/shook the paw post-application of the von Frey hair filament. After a positive response, the next lower filament was applied. However, if an animal did not respond to a filament the next filament with a higher bending force was applied. The maximum cut-off for a mouse was 4 g and for a rat was 15 g bending force. After notice of the first response, four more responses were recorded. The 50% withdrawal threshold was calculated using the formula: 10[Xf + kδ]/10 000; where Xf = value (in log units) of the final von Frey hair used, δ = mean difference (in log units) between stimuli, and k = tabular value for the pattern of the last six positive/ negative responses.

| Arthritic nociception
Dynamic weight bearing was used to measure spontaneous pain behavior in freely moving animals, as described. 26 Briefly, animals were placed on a sensor pad (Bioseb DWB Mouse/Rat Sensor Pad, Vitrolles, France) inside a Plexiglass chamber with a video camera attached at the top (DFK22AUC03 camera, Imaging Source, Charlotte, NC, USA). A 4-5 min recording was made, during which hind limb weight-bearing changes were continuously monitored. Hind paw weight distribution was calculated using Bioseb software (v1.4.2.92) using the following formula: % weight on ipsilateral paw = (weight borne by the ipsilateral paw / weight borne by the ipsilateral paw + weight borne by the contralateral paw) × 100.

| Cell culture and transfections
The immortalized chondrocyte C-28/I2 lines were purchased from Merck Millipore, cultured in EmbryoMax DMEM supplemented with 4500 mg/L of Glucose, 2.25 g/L of Sodium Bicarb and L-Glut, without Sodium Pyruvate, 10% FCS (Merck Millipore, Hertfordshire, UK) and maintained at 5% CO 2 . Human Articular Chondrocytes (HACs) were purchased from Cell-Applications Inc and maintained in Chondrocyte Growth Medium (Cell-Applications; Sigma-Aldrich, Poole, UK) at 5% CO 2 . HACs were used after three passages maximum.
High-density micromass cultures were generated, as described. 19 In some cases, prior to stimulations (as indicated in individual figure legends), the micromass were serumstarved for 24 hours in Phenol Red-Free DMEM/Ham's F12 (1:1; Sigma-Aldrich, Poole, UK), supplemented with 1% of Insulin-transferrin-selenium G supplement (ITS; Invitrogen, Paisley, UK) to allow for collagen type 2 and aggrecan transcription. Some chondrocytes were transfected with JetPrime transfection reagent (Polyplus-transfection, Illkirch, France) and either SERPINA1-His-Bio (Addgene, Watertown, MA, USA) or eGFP (Addgene, Watertown, MA, USA) for 24 hours in the presence of 10% FCS and then, grown in 3D micromass cultures for additional 24 hours prior to experimentation, as indicated in individual figure legends.
HEK293 cells were maintained in DMEM supplemented with 10% of FCS and maintained at 5% CO 2 .

| Luciferase reporter assays
Sub-confluent HEK293 cells, grown in monolayer in the presence of serum (10% FCS), were transfected with JetPrime with > 90% transfection efficiency. For CREB signaling interrogation cells were transfected with a luciferase reporter driven by reiterated consensus CRE-binding sites (CREfirefly luciferase, Promega, Hampshire, UK), henceforth referred to as CREB Reporter, plus CMV-Renilla luciferase plasmid (in a ratio 1:100). For Wnt/β-catenin signaling interrogation, cells were transfected with TCF/LEF-firefly luciferase reporter vector (SUPER8XTOPFlash, 27 referred to as TOPFlash) plus CMV-Renilla luciferase plasmid (in a ratio 1:100). Twenty-four hours after transfection, the medium was replaced, and the cells were treated for 24 hours as specified in individual figure legends. Luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Southampton, UK).

| Biochemical analyses
2.6.1 | Alcian Blue staining of sulfated glycosaminoglycans Cartilage-specific sulfated glycosaminoglycans deposition was quantified, as previously described. 19 Briefly, micromasses were fixed with 4% glutaraldehyde solution (v/v in ddH 2 O) and submerged in Alcian Blue 8GX dye (1% in 0.1N HCl w/v; pH 0.2; Atom Scientific, Cheshire, UK) for 24 hours at room temperature (RT). Alcian Blue dye was extracted in guanidine-HCl (8 M; Sigma-Aldrich, Poole, UK) for 48 hours at RT. Concentration was quantified by interpolation of A 630 of the extracted dye with Alcian Blue standard curve and normalized to DNA content (ng/µg). DNA content was measured in the extracted dye solution by fluorescence (485/535nm) using the SYBR Green method, according to the manufacturer's instructions (OriGene Technologies GmbH, Herford, Germany).

| Quantification of cytokine release
Plasma levels of mouse cytokines in cell-free supernatants were measured using a Ready-Set-Go ELISA Kits (eBioscience, Hatfield, UK).

| Western blot analyses
Western blot analyses were conducted for signaling pathways interrogation in chondrocytes. Antibodies and dilutions used are listed in Table 1. Briefly, stimulated cell monolayers were washed in ice-cold PBS and lysed in ice-cold RIPA Lysis buffer, supplemented with EDTA-free Protease Inhibitor cocktail (Sigma-Aldrich, Poole, UK) and Phosphatase Inhibitor Cocktail 2/3 (Sigma-Aldrich, Poole, UK) for 30 minutes on ice. Protein concentrations in cleared cell lysates were determined by bicinchoninic acid protein assay (Pierce BCA Protein Assay Kit, Thermo Fisher Scientific, Paisley, UK). Samples (20 µg total protein) were prepared for SDS-PAGE on 10% (w/v) NuPAGE Bis-Tris gels (Invitrogen, UK) and transferred onto Nitrocellulose membranes (BioTrace, Pall Corporation, Port Washington, NY, USA). Blots were then blocked with 5% BSA in 0.1% TBS-Tween (TBS-T) and incubated with primary antibodies at the concentrations listed in Table 1 overnight at 4°C. For α-Tubulin immunoblotting, His-AAT or NE expression, membranes were blocked in 5% fat-free milk protein in 0.1% TBS-T (Blotto, Santa Cruz Biotechnology, Heidelberg, Germany). After four 5-minute washes in 0.1% TBS-T, some membranes were incubated for 45 minutes with secondary HRP-conjugated Immunoglobulins, and after further 3-5 washes, protein bands were visualized chemiluminescence (Luminata Forte; Merck Millipore, Hertfordshire, UK) using FluorChem E imaging system (Protein Simple, CA, USA). Measurements of band densitometry and quantification of protein expression were conducted using ImageJ (NIH). 29 Phosphorylated protein expression was normalized to total protein levels and to α-Tubulin (loading control).

| Statistical analysis
All qPCR data are reported as Mean ± SEM unless otherwise indicated in individual figure legends. Significant differences in qPCR experiments were determined with the nonparametric Kruskal-Wallis ANOVA test, followed by Dunn's multiple comparison test. AB staining and histology-staining experiments were analyzed using one-way ANOVA, followed by Dunnett's multiple comparison posttest; Joint inflammation data (knee diameter, vascular conductance, and leukocyte trafficking) were analyzed using Student's unpaired t test. Joint pain (von Frey, dynamic weight bearing) data were analyzed using one-way ANOVA test. All statistical analyses were performed using GraphPad Prism 5.0, (GraphPad Software, CA, USA). Values were considered significant, if P < .05.

| Systemic AAT ameliorates cartilage damage in murine KBxN inflammatory arthritis
To complement our initial study with intra-articular delivery of AAT, 19 we tested if systemic AAT could protect the arthritic joint. Following intra-peritoneal administration of AAT, while we observed minimal modulation of the clinical parameters of arthritis ( Figure S1), histological analysis of knee sections revealed discernible proteoglycan loss from cartilage in arthritic joints: sulfated glycosaminoglycan content was reduced by > 30%, compared to vehicle-injected mice, an effect partially reversed by AAT ( Figure 1A, B).

T A B L E 1 Antibodies used for interrogation of signaling pathways
A marked accumulation of leukocytes occurred in the joints alongside pannus formation and synovial hypertrophy, all reduced by AAT ( Figure 1C, D).
To test indications of potential phenotypic alterations in the joints of AAT-treated mice, transcription of key cartilagerelevant genes was monitored. On day 5 of arthritis, sox9 was downregulated in the arthritic paws, with col2a1 and acan expression being unaffected ( Figure 2). The cytokine il1 and cartilage-destructive proteases adamts5 and mmp13 gene products were all upregulated (Figure 2A). Following systemic delivery of AAT, sox9 was upregulated alongside its transcriptional targets, col2a1 and acan, and over-expression of mmp13 and adamts5 was blocked. AAT reduced il1 gene expression and circulating serum levels of both IL-1β and IL-6, while IL-10 remained unaffected ( Figure 2B).

| AAT ameliorates joint inflammation and pain in models of arthritis
We expanded these data by testing AAT in three other models. Persistent inflammation was observed within the knee joint microvasculature of CFA-injected rats during a 3-week time course (not shown). Prophylactic treatment with AAT decreased knee swelling at day 21 ( Figure 3A). Intravital microscopy revealed that the number of rolling and adherent leukocytes was significantly reduced by AAT ( Figure 3B, C). This treatment also affected other parameters including knee swelling, leukocyte kinetics, and mechanical hypersensitivity ( Figure 3A-D). CFA injection induced a late allodynic response, causing a decrease in withdrawal threshold on days 7-21 ( Figure 3D, F) and considerable weight-bearing deficits on days 1-17 ( Figure 3E). AAT (100 µg ip; every 2 days) prevented mechanical hypersensitivity over the whole treatment (Day 0-7; Figure 3F). Although AAT treatment marginally improved weight-bearing deficits, the changes recorded did not reach statistical significance ( Figure 3E).
Kaolin/Carrageenan i.a. injection increased joint swelling, vascular conductance, the number of rolling, and adherent leukocytes, all markers reduced by AAT ( Figure S2A-D).
We and others 30 have detected large amounts of NE complexed with AAT in synovial fluids of rheumatoid arthritis, and to a lesser extent in osteoarthritic samples ( Figure S3A; Table S2). To establish if AAT anti-arthritic effects observed here related to NE-inhibition, AAT was pre-mixed with NE at a marked molar deficit, insufficient to cause enzymatic inhibition (100 ng AAT-5 µg NE). Following injection into the mouse knee joint, a much lower degree of proteoglycan loss was observed in the group receiving AAT (Figure 4A, B). Following NE, ~75% of joints presented mild synovitis, compared to ~25% of the mice receiving NE + AAT ( Figure 4C,  D). NE increased hind paw mechanical sensitivity by 75%, F I G U R E 1 Systemic administration of AAT ameliorates cartilage destruction in inflammatory arthritis. Inflammatory arthritis was induced with arthritogenic serum K/BxN (200 µL; days 0 and 2); mice (n = 5-6) were treated with vehicle (saline) or a single injection of AAT (100 µg ip) on day 3 (arrow on graphs). On day 5, knees were harvested, paraffin embedded, and multiple coronal sections (5-µm) were stained with H&E or Toluidine blue; A, Representative images (40×) of knee joint microstructure with evident signs of proteoglycan depletion are shown (red arrowheads). B, Cartilage integrity calculated as proportion of cartilage positive for Toluidine blue staining; C, Representative images (40× magnification) of histological sections from naive and arthritic joints stained with H&E. D, Histomorphometric analyses of joint sections. P < .05 vs respective control. Scale Bar: 200 µm. PF, pannus formation; S, synovitis; m, meniscus; F, femur; T, tibia. Data presented as Mean ± SEM of n = 5-6. *P < .05; one-way ANOVA followed by Bonferroni's multiple comparison posttest and this was partially rescued by AAT ( Figure 4F). Joint nociception following NE-injection was abrogated by AAT ( Figure 4J), alongside 10% reduction in NE-induced knee swelling ( Figure 4G), and ~ 60% in leukocyte rolling and adhesion ( Figure 4D, E).
These results suggest that AAT prevents the development of pain behavior, joint inflammation, and cartilage degradation in arthritis at least partially through mechanisms outside of NE-inhibition.

| AAT suppresses Wnt signaling and activates CREB signaling in chondrocytes
Next, we examined if AAT supported chondrogenic differentiation of chondrocytes in the absence of inflammatory stimuli. Unstimulated chondrocytes do not express endogenous NE ( Figure 5A, Figure S3B), therefore, any function exerted by AAT in chondrocytes must be independent of its anti-inflammatory/anti-elastolytic effects. Indeed, both rAAT and AAT gene transfer promoted deposition of Alcian Blue-positive ECM ( Figure 6A, E), and this coincided with upregulation of SOX9, COL2A1, and ACAN gene products ( Figure 6B, F).
To gain information on the pathways directly engaged by AAT, we interrogated the Smad, Wnt/β-catenin, and protein kinase A(PKA)/ cAMP response element-binding protein (CREB) signaling networks, all of which may have a role in chondrogenic differentiation. Initial experiments for Smad2/3 phosphorylation indicated lack of modulation upon AAT application ( Figure S3).
Serpins, including AAT, have been reported to modulate Wnt signaling [31][32][33] in various cell types. Using the Wnt reporter assay TOPFlash, we discovered that AAT potently inhibited the capacity of Wnt3A to induce β-catenin accumulation in chondrocytes, even in a polymerized (non-inhibitory) form ( Figure 5B, C, F). The nodal point of this inhibition occurred downstream of the Wnt receptors, because AAT still suppressed the Wnt reporter assay when activated in a ligand-and receptor-independent manner, by application of the glycogen synthase kinase 3 beta (GSK-3β) inhibitor 6-bromoindirubin 3′-oxime (BIO) ( Figure 5C, D).
Through interrogation of the PKA/CREB signaling pathway (Figure 6C, D, G, H), we could reveal that both rAAT and SERPINA1 transfection activated CREB signaling in chondrocytes: augmented CREB phosphorylation and activation of the CREB reporter assay were detected, with a boosted chondrocyte anabolism ( Figure 6A, B, E, F). PKA inhibition abrogated AAT-induced CREB phosphorylation and activation of CREB-reporter assay ( Figure 7B, C), and blocked the anabolic properties of AAT on chondrocytes, as quantified by normalized Alcian Blue staining ( Figure 7A).
Overall, these findings demonstrate (i) a dissociation between AAT's anti-inflammatory effects, potentially linked to NE inhibition, and its chondroprotective anabolic and anti-catabolic functions, and (ii) an induction of cartilageprotective pathways through a possible receptor-mediated signaling mechanism.

| DISCUSSION
Rheumatoid arthritis is a debilitating disease characterized by joint inflammation, pain, and structural damage. Patients affected by RA experience loss of function and pain, two major symptoms which markedly impact on quality of life. Despite recent successes in therapeutic developments, there is still a lack of new therapies that can target cartilage repair or prevent further damage if not reversing the disease.
We have recently identified AAT as a pivotal anti-arthritic molecule with anti-inflammatory and cartilage-protective effects when administered locally into murine arthritic joints. 19 The results presented herein expand on these findings by demonstrating that systemic administration of AAT suppressed joint inflammation and cartilage damage in distinct models of rodent inflammatory arthritis and provided significant reduction in arthritic nociception. Crucially, AAT displayed independent anabolic effects on cartilage: it increased ECM production and upregulated expression of key chondrogenic genes in the absence of inflammatory stimuli or NE. These effects were associated with downstream inhibition of Wnt/β-catenin and activation of PKA/CREB signaling F I G U R E 3 AAT ameliorates CFAinduced joint inflammation and pain. Rats received CFA (50 μl) into the right knee joint. AAT (100 µg:1 mL saline ip) was administered on day 1, 15 minutes before injection on day 0, and on days 2, 4, and 6 post-CFA injections. Joint inflammation was assessed on day 21 post-CFA injection including A, knee diameter and the number of rolling B, and adherent leukocytes C, and mechanical hypersensitivity D, Weight-bearing deficits E, and mechanical hypersensitivity F, were tested at baseline and on days 1, 3, 7, 10, 14, 17, and 21 post-CFA injection. Data presented as Mean ± SEM, n = 8-14. *P < .05, **P < .01, P < .0001. Student's unpaired t test A-D, two-way ANOVA followed by Bonferroni's multiple comparison posttest (E, F)

| 9 of 16
KANEVA Et Al pathways in chondrocytes. Taken together, these data indicate that AAT is a versatile anti-arthritic agent that has the potential to promote cartilage repair and dampen down joint inflammation and pain. Although the exact mechanisms underlying the therapeutic effect of AAT in arthritis in vivo remain to be further scrutinized, several hypotheses were considered here. The first one is centered on inhibition of inflammatory cytokine production. Various cytokines, including IL-1β, and TNF-α, play major roles in the pathogenesis of RA 34 and strategies which block their activity have proven to be effective treatments. 35 AAT-deficient individuals lack control over inflammatory mediators, including IL-1β, IL-6, IL-8, and TNF-α, 36 and stimulation of peripheral blood mononuclear cells with AAT reduces their pro-inflammatory signature. 37 Furthermore, Janciauskiene and colleagues reported that AAT's inhibition of endotoxin-stimulated TNF-α and enhancement of IL-10 in human monocytes, are mediated by elevation of cAMP and activation of cAMP-dependent protein kinases. 38,39 Additionally, in a model of gouty arthritis, AAT-Fc fusion protein was shown to reduce joint swelling and cellular infiltration through inhibition of protease-dependent processing/release of IL-1β. 18 Building on our previous study with intra-knee injection of AAT, 19 we observed that a single, low F I G U R E 4 AAT protects against NE-induced cartilage degradation, knee inflammation, and pain. C57BL/6 mice were injected with either PBS (5 µL) and NE (5 µg/5µL), or AAT (100 ng/5 µL) and NE (5 µg/5 µL) into the right knee joint. After 4 hours, the knees were harvested, paraffin embedded, and multiple coronal sections (5-µm) were stained with H&E or Toluidine blue. A, Representative images (40 x) of knee joint microstructure with evident signs of proteoglycan depletion shown (red arrowheads); B, Cartilage integrity calculated from percentage area of cartilage positive for toluidine blue staining; C, Representative images (40× magnification) of histological sections from naive and arthritic joints stained with H&E; D, Histomorphometric analyses of joint sections. Scale bars: 200 µm. m, meniscus; F, femur; T, tibia. In separate experiments, treatment with AAT reduced G, knee diameter, H, vascular conductance, and the number of (E) rolling and (F) adherent leukocytes at 4 hours post-injection, compared to NE alone (yellow arrows indicate stained leukocytes). NE i.a. injection caused a significant secondary allodynia I, and weight-bearing deficits J, counteracted by AAT co-administration. Data presented as Mean ± SEM. *P < .05, **P < .01, ***P < .0001; Student's unpaired t test; n = 8-14 (A-H), one-way ANOVA followed by Bonferroni's posttest; n = 6-12 (I, J) dose systemic injection of recombinant AAT is sufficient to reduce joint inflammation and cartilage proteoglycan depletion. AAT inhibited local IL-1β gene expression, circulating levels of the cytokine and reduced expression of Adamts5 and MMP-13 genes within the joint. Simultaneously, AAT upregulated expression of the chondro-specific master transcription regulator Sox9 and its downstream targets collagen type II and aggrecan. AAT also inhibited proliferative synovitis and neutrophil infiltration into the synovial fluid and pannus development in the KBxN poly-arthritis model. Since the major histological characteristics of inflammatory arthritis are proliferative synovitis and leukocyte infiltration, it is plausible that neutrophil proteases contribute to cartilage degradation and overall joint injury. Inflammatory stimuli activate neutrophils to release the contents of azurophilic granules, including NE, cathepsin-G, and proteinase-3, that proteolytically modify chemokine and cytokine activity, interact with cell-surface receptors and contribute to neutrophil migration by cleaving adhesion molecules. 40 In particular, NE degrades elastin, collagens, proteoglycans, and other ECM components 41 and its physiological activity is regulated by endogenous inhibitors, 5 with AAT being the most potent and with highest affinity. 42,43 It is well documented that NE activity and expression are increased in synovial fluids and synovia of patients with RA. 30,43 Experimentally, the incidence of arthritis is reduced through the use of NE inhibitors, confirming its contributory role to disease development. 44 To test this hypothesis, we applied AAT at a dose well below levels needed for enzymatic inhibition of NE: for complete NE neutralization to occur, a molar ratio of AAT:NE > 2:1 is needed. 45 Consistent with previous studies where both recombinant NE and NE-rich RA synovial fluids degraded cartilage proteoglycans, 46,47 here NE injection into murine knee joints caused rapid proteoglycan loss from the cartilage, accompanied by a generalized knee inflammation. NE provoked mechanical allodynia and weight-bearing deficits. 23 Treatment with AAT at a dose less than 1/100 than that necessary to block NE, still prevented leukocyte infiltration, improved cartilage integrity, reduced synovial hyperaemia, and significantly ameliorated joint pain.
To investigate further whether the beneficial properties of AAT were distinct from NE inhibition, experiments were conducted with isolated chondrocytes, which do not express NE. These cells responded to both recombinant AAT stimulation and SERPINA1 transfection by activating biosynthetic pathways resulting in augmentation of anabolic gene expression, including the chondrogenic factor Sox9. 48 Major F I G U R E 5 AAT inhibits canonical Wnt-signaling. A, Neutrophil elastase (NE) expression was analyzed by Western blotting in lysates (25 µg) from unstimulated C28/I2s and primary neutrophils; rNE (2 µg) was used for positive control (full blot in Supplementary Figure S3B). B, D, Chondrocytes were stimulated for 6 hours with vehicle, Wnt3A (100 ng/mL) or BIO (20 µM) in presence or absence of AAT (10 µg/mL). Western blots show representative experiments out of three performed; α-tubulin was tested as loading control. C, F, HEK293 cells transiently transfected with TOPFlash and Renilla reporter plasmids were stimulated with either vehicle, Wnt3A (100 ng/mL) alone or in the presence of native or polymerized AAT (0.1-10 µg/mL). E, TOPFlash and Renilla transfected HEK293 cells were transfected with GFP or SERPINA1-His-Bio for 24 hours and then, stimulated with either vehicle, or BIO (20 µM) for further 24 hours. Luciferase activity standardized to Renilla luciferase activity was calculated to yield relative luciferase activity (RLU) then expressed as fold change over vehicle-treated cells as control. Statistical analyses were conducted with one-way ANOVA, followed by Bonferroni multiple-comparison posttest with *P < .05, **P < .001, ***P < .0001 vs respective control

| 11 of 16
KANEVA Et Al cartilaginous extracellular matrix components, aggrecan and collagen type II were also upregulated, indicating that in the absence of inflammation, AAT activates pathways involved in chondrogenic differentiation and might directly contribute to the cartilage protection observed in experimental arthritis.
Based on these results, we hypothesized that AAT could be interacting with some of the signaling networks controlling chondrogenic differentiation, including Smad2/3, PKA/CREB, and Wnt/β-catenin signaling. Early studies have indicated that impaired ALK5/Smad2/3 signaling, is implicated in cartilage destruction in arthritis 49,50 ; however, F I G U R E 6 AAT activates PKA/CREB signaling in chondrocytes. A, C28/I2 chondrocytes grown in 3D micromass cultures were serum starved for 24 hours and stimulated with vehicle (DMSO), AAT (10 µg/mL) or Fsk (30 µM) for 24 hours. E, F, G, HACs transfected with GFP or SERPINA1-His-Bio for 24 hours were grown in micromasses cultures for 48 hours. Micromasses were stained with Alcian Blue for the detection of sulfated glycosaminoglycans and spectrophotometric quantification of guanidine-HCL-extracted Alcian Blue dye was normalized to DNA content (µg/ng) (n = 6). B, F, ACAN, COL2A1, and SOX9 expression was determined by qPCR. C, G, CREB-phosphorylation was analyzed by Western blotting after 30-min stimulation of C28/I2s with Fsk (30 µM), or AAT at the indicated concentrations (C), or after 24 hours transfection of HACs with GFP or SERPINA1-Bio-His plasmids (G). D, H, CREB-related gene activation was analyzed by reporter assay 24 hours after stimulations of HEK293 cells transiently transfected with CREB and Renilla reporter plasmids. Luciferase activity standardized to Renilla luciferase activity was calculated to yield relative luciferase activity (RLU) then expressed as fold change over DMSO-treated controls. Data are Mean ± SEM (n = 3 experiments) and statistical analyses were conducted with Kruskal-Wallis test followed by Dunn's multiple comparison post hoc test (B, F), one-way ANOVA, followed by Bonferroni posttest (C, D) or Unpaired student's t test as appropriate with *P < .05, **P < .01, ***P < .001 vs respective control (A, E, G, H) in the current study AAT did not activate either transcription factor. In contrast, the canonical Wnt pathway regulates multiple biological and pathological processes including angiogenesis and inflammation. Relevantly, it regulates chondrocyte differentiation in health and disease, and while moderate Wnt activity is essential for chondrocyte proliferation and maintenance of their typical characteristics, 46 excessive activation results in chondrocyte hypertrophy, and expression of cartilage-degrading MMPs. 47,51,52 Recently, the involvement of canonical Wnt/β-catenin signaling pathway in the pathophysiology of cartilage degenerative disease has attracted much attention: the arthritic joint is well known to express an abundance of Wnt ligands and overactivation of Wnt-signaling pathway is a major contributing factor toward the progression of both rheumatoid 51,53 and osteo-arthritis. 52,54 Therefore, Wnt inhibition may represent a promising strategy for disease modification in osteoarthritis 47,52,55 and possibly RA and this notion is supported by successful early clinical studies. 54 In the absence of Wnt ligands, cytoplasmic β-catenin is degraded by a complex composed of the constitutively active kinase GSK-3β, casein kinase I (CKI), adematous polyposis coli (PCI), and Axin. Upon Wnts engaging with their co-receptor complex (Frizzled and low-density lipoprotein-related protein (LRP5/6)) GSK-3β is inactivated. In this condition, βcatenin accumulates, translocates into the nucleus, and together with TCF/LEF transcription factors activates transcription of Wnt target genes. 56 Here, we report that both the native and polymeric, non-inhibitory 38,57 AAT potently inhibited Wnt3A-mediated β-catenin accumulation in chondrocytes. The nodal point of this inhibition occurred downstream of the Wnt receptors: AAT still suppressed the Wnt reporter assay when activated by the GSK-3β inhibitor BIO to by-pass receptor engagement. This is particularly important because we could postulate that in the arthritic joint environment abundance of Wnt ligands will not override the AAT-driven inhibition of Wnt-signaling. This scenario is supported by a link between Wnt/β-catenin signaling and IL1β-induced cartilage degradation: IL1β, which is upregulated in arthritic joints (Figure 2), induces expression of Wnt ligands in chondrocytes, 58 and IL1β and Wnt3A together induce greater loss of proteoglycans from cartilage ECM than either one individually. 59 Moreover, Wnt-signaling activation by Wnt3A or the GSK-3β inhibitor BIO was shown to induce MMP13 mRNA expression and proteolytic activity in murine cartilage explants. 60 The facts that (i) cartilage degradation induced by pro-inflammatory cytokines involves Wnt/β-catenin pathway, and (ii) increased

KANEVA Et Al
Wnt-signaling has been implicated in the initiation and progressive deterioration of cartilage degeneration in arthritis, suggests that AAT inhibition of the canonical pathway in vivo could counteract cytokine-induced cartilage degradation. 61 This hypothesis is partially supported by the present study, where AAT significantly suppressed il1 transcription and IL1β plasma levels, and at the same time abrogated mmp13 transcription. Further investigations are required to confirm this mechanism of action. cAMP is another important regulator of chondrocyte differentiation. 62 Higher intracellular cAMP levels promote transcription of type II collagen and aggrecan. Extracellular stimuli activating adenylyl cyclase regulate the levels of cAMP, which activates PKA, which controls the expression of cAMP-inducible genes via phosphorylation of nuclear transcription factors like CREB. 63 Elevation of cAMP activates pro-resolving pathways and reduces pro-inflammatory cytokine signaling in various cell types, including chondrocytes. For example, melanocortin receptor agonists, 64 calcitonin, 65 and the direct activator of adenylate cyclase Forskolin 62 reduce pro-inflammatory cytokines and promote chondrogenic processes through the elevation of intracellular cAMP. Our results demonstrate that AAT activates CREB in chondrocytes through a PKA-dependent pathway. Such an effect is in keeping with previous reports. 39,66 Phosphorylated CREB then undergoes nuclear translocation and binding to CRE sites on target genes, 67 promoting deposition of matrix by the chondrocytes, as shown elsewhere. 39 The engagement of the PKA/CREB pathway by AAT was confirmed with the PKA inhibitor H-89, which impaired downstream CREB signaling and AB-positive proteoglycan deposition. Although it is plausible that AAT-induced SOX9 upregulation is CREB-dependent, as CREB is a transcription factor regulating SOX9 promoter activation and Sox9 phosphorylation, 68 the involvement of other pathways cannot be excluded at present.
The simultaneous CREB-activation and Wnt-inhibition, which was recently described to be a potent driver of chondrogenesis and cartilage regeneration, 69 is a discovery central for understanding AAT's mechanism of action in chondrocytes and one that, being the focus of ongoing endeavors, could prove to be central in driving AAT's antiarthritic properties.
In conclusion, we show here that in addition to its known anti-proteolytic and broad anti-inflammatory properties, AAT can independently function as an acute phase mediator of anti-nociception and chondroprotection. We partly elucidated the specific signaling pathways responsible for these functional outcomes post-AAT application, in vitro and in vivo. As such, these findings reveal the potential dual benefits of using AAT to preserve articular cartilage and alleviate arthritic nociception and inflammation.