Advertisement

Sustained release of locally delivered celecoxib provides pain relief for osteoarthritis: a proof of concept in dog patients

Open AccessPublished:December 02, 2022DOI:https://doi.org/10.1016/j.joca.2022.11.008

      Summary

      Objective

      Drug delivery platforms that allow for gradual drug release after intra-articular administration have become of much interest as a treatment strategy for osteoarthritis (OA). The aim of this study was to investigate the safety and efficacy of an intra-articular sustained release formulation containing celecoxib (CXB), a cyclooxygenase-2 (COX-2) selective inhibitor.

      Methods

      Amino acid-based polyesteramide microspheres (PEAMs), a biodegradable and non-toxic platform, were loaded with CXB and employed in two in vivo models of arthritis: an acute inflammatory arthritis model in rats (n = 12), and a randomized controlled study in chronic OA dog patients (n = 30). In parallel, the bioactivity of sustained release of CXB was evaluated in monolayer cultures of primary dog chondrocytes under inflammatory conditions.

      Results

      Sustained release of CXB did not alleviate acute arthritis signs in the rat arthritis model, based on pain measurements and synovitis severity. However, in OA dog patients, sustained release of CXB improved limb function as objective parameter of pain and quality of life based on gait analysis and owner questionnaires. It also decreased pain medication dependency over a 2-month period and caused no adverse effects. Prostaglandin E2 levels, a marker for inflammation, were lower in the synovial fluid of CXB-treated dog OA patients and in CXB-treated cultured dog chondrocytes.

      Conclusion

      These results show that local sustained release of CXB is less suitable to treat acute inflammation in arthritic joints, while safe and effective in treating pain in chronic OA in dogs.

      Keywords

      Introduction

      The impact of osteoarthritis (OA) has doubled since the mid-20th century; the prevalence and burden is expected to rise in the coming decades
      • Hunter D.J.
      • Bierma-Zeinstra S.
      Osteoarthritis.
      . Recent efforts identify clinical OA phenotypes
      • Dell'Isola A.
      • Allan R.
      • Smith S.L.
      • Marreiros S.S.P.
      • Steultjens M.
      Identification of clinical phenotypes in knee osteoarthritis: a systematic review of the literature.
      but the first line treatment remains similar for all phenotypes. Inflammation in OA has gained much interest for specific therapeutic targeting with inflammatory processes as the common denominator for all clinical phenotypes.
      Pro-inflammatory mediators produced by articular chondrocytes (ACs) and synovial lining cells promote OA progression and severity of the symptoms
      • Sellam J.
      • Berenbaum F.
      The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis.
      . In experimental rat OA, gene expression levels of prostaglandin E synthase (ptges) and prostaglandin-endoperoxide synthase 2 (ptgs2, encoding cyclooxygenase 2), two cyclooxygenase-2 (COX-2) regulated genes, were higher in the knee cartilage than in the control healthy knees
      • Appleton C.T.G.
      • Pitelka V.
      • Henry J.
      • Beier F.
      Global analyses of gene expression in early experimental osteoarthritis.
      . Similarly, synovial fluid (SF) Prostaglandin E2 (PGE2) levels increase in response to anterior cruciate ligament transection
      • Trumble T.N.
      • Billinghurst R.C.
      • McIlwraith C.W.
      Correlation of prostaglandin E2 concentrations in synovial fluid with ground reaction forces and clinical variables for pain or inflammation in dogs with osteoarthritis induced by transection of the cranial cruciate ligament.
      . Additionally, PGE2 levels in SF of human
      • Brenner S.S.
      • Klotz U.
      • Alscher D.M.
      • Mais A.
      • Lauer G.
      • Schweer H.
      • et al.
      Osteoarthritis of the knee – clinical assessments and inflammatory markers.
      and dog OA patients
      • Heikkilä H.M.
      • Hielm-Björkman A.K.
      • Innes J.F.
      • Laitinen-Vapaavuori O.M.
      The effect of intra-articular botulinum toxin A on substance P, prostaglandin E2, and tumor necrosis factor alpha in the canine osteoarthritic joint.
      correlate with pain. Therefore, long-term inhibition of COX-2 to reduce PGE2 secretion may be a solution for effective relief of OA-related pain and inhibit disease progression.
      Celecoxib (CXB) is a selective COX-2 inhibitor suggested to have disease-modifying effects
      • Zweers M.C.
      • de Boer T.N.
      • van Roon J.
      • Bijlsma J.W.J.
      • Lafeber F.P.J.G.
      • Mastbergen S.C.
      Celecoxib: considerations regarding its potential disease-modifying properties in osteoarthritis.
      ,
      • Nakata K.
      • Hanai T.
      • Take Y.
      • Osada T.
      • Tsuchiya T.
      • Shima D.
      • et al.
      Disease-modifying effects of COX-2 selective inhibitors and non-selective NSAIDs in osteoarthritis: a systematic review.
      . Osteophyte formation and bone marrow lesions were reduced in a post-traumatic rat OA model after oral treatment with CXB
      • Panahifar A.
      • Jaremko J.L.
      • Tessier A.G.
      • Lambert R.G.
      • Maksymowych W.P.
      • Fallone B.G.
      • et al.
      Development and reliability of a multi-modality scoring system for evaluation of disease progression in pre-clinical models of osteoarthritis: celecoxib may possess disease-modifying properties.
      . However, oral non-steroidal anti-inflammatory drugs (NSAIDs) can lead to adverse effects, including hypertension (>10%) and gastro-intestinal damage (30–50%), of which 1–2% of the patients experience serious injury
      • Sostres C.
      • Gargallo C.J.
      • Lanas A.
      Nonsteroidal anti-inflammatory drugs and upper and lower gastrointestinal mucosal damage.
      . Moreover, systemic drug administration entails suboptimal local drug exposure, with articular levels reaching only 23–50% of those in the circulation
      • Bertin P.
      • Lapicque F.
      • Payan E.
      • Rigaud M.
      • Bailleul F.
      • Jaeger S.
      • et al.
      Sodium naproxen: concentration and effect on inflammatory response mediators in human rheumatoid synovial fluid.
      , which is below the minimum inhibitory concentration of CXB. Local intra-articular drug concentrations rather than blood plasma levels determine the clinical effect, as illustrated in arthritis patients whose SF levels of orally administered anti-inflammatory drugs correlated with the magnitude of relief of clinical symptoms arthritis
      • Aarons L.
      • Salisbury R.
      • Alam-Siddiqi M.
      • Taylor L.
      • Grennan D.
      Plasma and synovial fluid kinetics of flurbiprofen in rheumatoid arthritis.
      . As such, intra-articular application is an obvious solution. In a post-traumatic rabbit OA model, intra-articularly injected CXB inhibited expression of pro-inflammatory mediators and prevented further structural cartilage damage
      • Jiang D.
      • Zou J.
      • Huang L.
      • Shi Q.
      • Zhu X.
      • Wang G.
      • et al.
      Efficacy of intra-articular injection of celecoxib in a rabbit model of osteoarthritis.
      . However, rapid diffusion and clearance of drugs after direct intra-articular injection require frequent intra-articular injections that are a burden to the patient and not without risk. Extended delivery of NSAIDs from biomaterial carriers would circumvent this problem.
      In vitro release of CXB from amino acid-based polyesteramide microspheres (PEAMs), a biodegradable and non-toxic platform, was demonstrated for >75 days
      • Janssen M.
      • Timur U.T.
      • Woike N.
      • Welting T.J.M.
      • Draaisma G.
      • Gijbels M.
      • et al.
      Celecoxib-loaded PEA microspheres as an auto regulatory drug-delivery system after intra-articular injection.
      . Intra-articularly injected CXB-PEAMs were present in rat OA knee joints >12 weeks and lowered local PGE2 levels
      • Janssen M.
      • Timur U.T.
      • Woike N.
      • Welting T.J.M.
      • Draaisma G.
      • Gijbels M.
      • et al.
      Celecoxib-loaded PEA microspheres as an auto regulatory drug-delivery system after intra-articular injection.
      . Furthermore, intra-articular delivery in a preclinical rat OA model of PEAMs with 2–13× higher CXB loading doses than those employed in the earlier study, attenuated synovial inflammation and inhibited the progression of subchondral bone changes
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Pouran B.
      • de Visserm H.M.
      • Weinans H.H.
      • Thomas R.E.
      • et al.
      Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
      .
      As PEA-mediated CXB delivery appears to be a promising avenue for the treatment of OA, the current study aims to further delineate its applicability in two in vivo arthritis models reflecting two main patient groups. In the first, acute inflammation and pain during severe flares play a central role, which are observed in a subset of OA patients and also in rheumatoid arthritis. The second consists of pet dogs diagnosed with clinical and radiological OA. It is estimated that >10% of the dog population is affected by spontaneous OA
      • Anderson K.L.
      • O’Neill D.G.
      • Brodbelt D.C.
      • Church D.B.
      • Meeson R.L.
      • Sargan D.
      • et al.
      Prevalence, duration and risk factors for appendicular osteoarthritis in a UK dog population under primary veterinary care.
      and OA dog patients are considered a well-accepted model for human OA within the concept of ‘One Medicine’
      • Meeson R.L.
      • Todhunter R.J.
      • Blunn G.
      • Nuki G.
      • Pitsillides A.A.
      Spontaneous dog osteoarthritis — a One Medicine vision.
      .

      Method

      A detailed description of all methods can be found in the online supplementary methods.

      Synthesis and characterization of PEAMs

      PEAMs were synthesized and characterized according to previously published methods
      • Rudnik-Jansen I.
      • Tellegen A.
      • Beukers M.
      • Öner F.
      • Woike N.
      • Mihov G.
      • et al.
      Safety of intradiscal delivery of triamcinolone acetonide by a poly(esteramide)microsphere platform in a large animal model of intervertebral disc degeneration.
      . CXB-loading of the PEAMs resulted in 24.1 wt% CXB. The CXB release profile in phosphate buffered saline (PBS) was determined by HPLC as described previously
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Beukers M.
      • Miranda-Bedate A.
      • Bach F.C.
      • de Jong W.
      • et al.
      Intradiscal delivery of celecoxib-loaded microspheres restores intervertebral disc integrity in a preclinical canine model.
      . The profile confirmed an initial burst release at day one followed by a gradual increase of CXB release in the days thereafter, reaching a cumulative release of 32% at day 14 (Fig. S1).

      Induction of experimental acute arthritis in rats

      The study design [Fig. 1(A)] was approved by the National Commission of animal experiments (AVD108002015282), was supervised by the local Animal Welfare Body (WP#800-15-282-01-003) and is in accordance with the ARRIVE guidelines. To study the long-lasting effects of local controlled CXB delivery, acute arthritis was induced in one knee joint (left) of twelve 8-weeks old female Sprague–Dawley rats (214 ± 12 g; Charles River laboratories, The Netherlands) as described previously
      • Rudnik-Jansen I.
      • Schrijver K.
      • Woike N.
      • Tellegen A.
      • Versteeg S.
      • Emans P.
      • et al.
      Intra-articular injection of triamcinolone acetonide releasing biomaterial microspheres inhibits pain and inflammation in an acute arthritis model.
      . Briefly, local synovitis was induced in all rats by priming the left joint (i.e., experimental joint) with a peptidoglycan-polysaccharide (PGPS) at day −14 by intra-articular injection of PGPS (100P fraction with 5 mg rhamnose/mL, 25 μL PGPS of 0.17 mg/mL, Lee Laboratories). Baseline assessments were performed by IR. Thereafter, the animals were re-randomized based on the degree of pain/swelling, to ensure similar weight-bearing deficiency across treatment groups. Local synovitis was reactivated at day 0, 14, and 28 with intravenous PGPS injections (0.28 mg/mL). The treatment (60 mg/mL empty PEAMs or 80 mg/mL CXB-PEAMs (V = 25 μL, amounting to 402 μg CXB/joint) n = 6 per group, randomly divided) was intra-articularly administered 2.5 h prior to the reactivation at day 0 (IR), whereafter outcome measures were determined (IR/SV, both blinded for the treatment throughout the study). The administered CXB dose was the highest dose shown to be effective in the ACLT rat OA model
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Pouran B.
      • de Visserm H.M.
      • Weinans H.H.
      • Thomas R.E.
      • et al.
      Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
      .
      Fig. 1
      Fig. 1Local delivery of CXB-PEAMs in acute arthritic rat knee joints. (A) Experimental design acute arthritis model: rat knee joints (left) were primed with intra-articular injection of 0.17 mg/mL PGPS at day −14 (blue arrow) and re-activated with 0.28 mg/mL intravenous PGPS injections (red arrows) to stimulate severe flares. CXB-PEAMs or empty PEAMs (control) were administered as single intra-articular dose 2.5 h prior to the first re-activation at day 0 (green dotted line). The second and third reactivations were administered on day 14 and 28. Joint swelling measurements as an indication for synovitis and mechanical hypersensitivity measurements at the hind paw for pain-like behaviors were conducted at all time points, while dynamic weight bearing measurements (scale icon) were conducted at various days throughout the experiment. The animals were euthanized at endpoint on day 42 and the knee joints were histologically analyzed. (B) Swelling of the joint (synovitis) for n = 6 knees per group, n = 5 knees for CXB-PEAMs after 28 days calculated as the increase in knee thickness compared to baseline thickness at day −14 and presented as mean ± SD. One CXB-treated rat reached the humane endpoint which was defined as lameness in combination with severe swelling of both hind paws and no improvement after additional pain medication consisting of 5 mg/kg carprofen. This was reached at day 18 and the rat was excluded from the study. #: P = 0.799 (C) Structural cartilage damage and (D) synovitis were scored on coronal knee joint histological sections at day 42 (n = 10 contralateral; n = 6 empty PEAMs; n = 5 CXB-PEAMs). Data of individual knees are presented as dot plots including the mean (red lines). ns: not significant.

      Outcome measures to evaluate acute inflammation in rat knee joints

      Knee joint swelling was determined by the average of three consecutive measurements using a digital caliper and upon subtraction of the baseline measurements prior to priming (day -14). Mechanical hypersensitivity was assessed by applying von Frey hairs to the hind paw
      • Chaplan S.R.
      • Bach F.W.
      • Pogrel J.W.
      • Chung J.M.
      • Yaksh T.L.
      Quantitative assessment of tactile allodynia in the rat paw.
      ; the 50% threshold was determined using the up-down method
      • Minett M.S.
      • Falk S.
      • Santana-Varela S.
      • Bogdanov Y.D.
      • Nassar M.A.
      • Heegaard A.M.
      • et al.
      Pain without nociceptors? Nav1.7-independent pain mechanisms.
      . A decrease in 50% threshold indicates increased hypersensitivity, and thus increased pain sensation. Weight-bearing and paw surface areas were analyzed using the dynamic weight-bearing (DWB) apparatus (Bioseb, module version 1.4.2.98; Boulogne, France) as described previously
      • Rudnik-Jansen I.
      • Schrijver K.
      • Woike N.
      • Tellegen A.
      • Versteeg S.
      • Emans P.
      • et al.
      Intra-articular injection of triamcinolone acetonide releasing biomaterial microspheres inhibits pain and inflammation in an acute arthritis model.
      .
      Upon termination of the experiment, all hind limbs were collected and processed for histological analysis as described in the supplementary methods. Five μm coronal knee joint sections were stained with Safranin-O/Fast green to evaluate cartilage degeneration using the Mankin score
      • Mankin H.J.
      • Lippiello L.
      Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips.
      (0: “completely healthy” to 14: “total joint destruction”). Synovitis was evaluated using the Krenn score
      • Krenn V.
      • Morawietz L.
      • Burmester G.R.
      • Kinne R.W.
      • Mueller-Ladner U.
      • Muller B.
      • et al.
      Synovitis score: discrimination between chronic low-grade and high-grade synovitis.
      (0: “healthy” to 9: “severe synovitis”) on hematoxylin and eosin-stained sections. Scoring was done independently in a random and blinded fashion (AT/IR). The average score was used for further analysis.

      Efficacy of local controlled CXB delivery in OA dog patients

      Release behavior and bioactivity of CXB-PEAMs were evaluated in primary dog AC culture in which inflammation was simulated by exogenous tumor necrosis factor (TNF) using a culture system described previously
      • Yang H.Y.
      • Van Dijk M.
      • Licht R.
      • Beekhuizen M.
      • Van Rijen M.
      • Janstål M.K.
      • et al.
      Applicability of a newly developed bioassay for determining bioactivity of anti-inflammatory compounds in release studies - celecoxib and Triamcinolone acetonide released from novel PLGA-based microspheres.
      . Briefly, healthy ACs were enzymatically isolated from cartilage from six experimental dogs from unrelated experiments (AVD108002015282). Cells were expanded till passage 2 in expansion medium (High glucose Dulbecco's modified medium (DMEM) Glutamax, Gibco, Waltham, USA) and refreshed every 3–4 days. Cells were co-cultured with chondropermissive medium (High glucose DMEM, 1% [v/v] penicillin-streptomycin (GE Healthcare, Chicago, USA), 1% [v/v] ITS+ premix (Corning; Corning, NY, USA), 0.04 mg/mL bovine serum albumin (Sigma–Aldrich), 0.1 mM ascorbic acid (Sigma–Aldrich)) in the presence of empty- or CXB-PEAMs that were placed in Transwell® baskets (pore size 0.4 μm, polycarbonate membrane, Corning). CXB-PEAMs containing 7∗10−4 M CXB, from now on referred to as 10−4 M, were used to inhibit COX-2 activity. After 4 h, 10 ng/mL recombinant TNF (R&D Systems, Minneapolis, USA) was provided as a pro-inflammatory stimulus. The transwells with the empty-PEAMs or CXB-PEAMs were transferred to newly plated cells of each donor after 72 h. This was repeated 8 times amounting to a total culture period of 28 days where bioactivity of the released CXB was determined. Every 72 h, medium was collected and stored at −20°C.
      CXB medium levels were measured using enzyme-linked immunosorbent assay (ELISA; Neogen Corporation, Lansing, USA) according to the manufacturer's instructions. PGE2 (Cayman Chemical, Ann Arbor, USA) for n = 3 AC donors and C–C motif chemokine ligand 2 (CCL2; Kingfisher Biotech, Saint Paul, USA) for n = 6 AC donors were measured in the culture medium according to manufacturers' instructions.

      Randomized controlled study in chronic OA dog patients

      The study was conducted upon approval of the departmental Ethical Committee (#AVR17-06). Dogs were considered eligible for the study if they met the inclusion criteria (Table S1) and upon written owner consent. Both owners and assessing veterinarians were blinded to the treatment during the entire study period. A random list for the vials containing CXB-PEAMs (n = 20) or placebo (n = 10) was created in advance (Excel 2016, Microsoft) and dogs were thereby randomly allocated to one of the treatment arms by order of enrollment. For the arthrocentesis and intra-articular injection, dogs were sedated. SF was collected, centrifuged (5 min, 1300g), aliquoted, and stored at −20°C. Suspensions of 70 mg PEAMs/mL for 20 wt% CXB-loaded PEAMs (9.3 mg CXB/mL) and empty-PEAMs were prepared as previously reported
      • Willems N.
      • Mihov G.
      • Grinwis G.C.M.
      • van Dijk M.
      • Schumann D.
      • Bos C.
      • et al.
      Safety of intradiscal injection and biocompatibility of polyester amide microspheres in a canine model predisposed to intervertebral disc degeneration.
      . Dosing of the treatment depended on body weight categories, i.e., 0.5 mL (15–30 kg), 1 mL (30–45 kg), and 1.5 mL (>45 kg) microsphere solution, corresponding to 4.6, 9.3, and 14 mg CXB, respectively, based on results of a preclinical study
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Pouran B.
      • de Visserm H.M.
      • Weinans H.H.
      • Thomas R.E.
      • et al.
      Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
      . Dogs in the placebo group received the same quantity of empty-PEAMs. One- and 2-month follow-up consisted of clinical examination, force plate analysis, owner questionnaires (primary outcome), plain radiographs, SF analysis briefly described below and in detail in supplementary methods. Owners were allowed to use analgesics they were already using prior to the study and were asked to document medication(s) administered during the study (‘additional pain relief medication’) and note adverse reactions. Any pain medication was discontinued 4 days prior to baseline and follow-up measurements to minimize a possible (beneficial) effect of the oral medication on the subjective and objective read out.

      Outcome measures for relief of clinical signs in dog OA

      All dogs underwent full clinical and orthopedic examination prior to, and at the 1- and 2-month follow-up visits. Lameness was recorded on a 4-point scale. Radiographic projections of each joint were used for evaluating radiographic OA (Table S2) prior to inclusion and 2 months after treatment. Osteophyte height was measured as described previously
      • Lavrijsen I.C.M.
      • Heuven H.C.M.
      • Voorhout G.
      • Meij B.P.
      • Theyse L.F.H.
      • Leegwater P.A.J.
      • et al.
      Phenotypic and genetic evaluation of elbow dysplasia in Dutch Labrador Retrievers, Golden Retrievers, and Bernese Mountain dogs.
      .
      Ground reaction forces (GRFs) were measured (9261 Kistler Instrumente)
      • Hazewinkel H.A.W.
      • van den Brom W.E.
      • Theyse L.F.H.
      • Pollmeier M.
      • Hanson P.D.
      Comparison of the effects of firocoxib, carprofen and vedaprofen in a sodium urate crystal induced synovitis model of arthritis in dogs.
      during each visit with a single force plate with the dogs on a leash at a walking gait. Ten valid measurements per limb and per time point were collected, minimally eight when limited by the stamina of the dog and their owners. GRFs in the mediolateral (Fx), craniocaudal (Fy) and vertical (Fz) direction were normalized for body weight (N/kg) whereafter Symmetry indices (SIs) were calculated
      • Fanchon L.
      • Grandjean D.
      Accuracy of asymmetry indices of ground reaction forces for diagnosis of hind limb lameness in dogs.
      .
      To assess treatment outcome on the behavior and function of their dog, owners filled in a questionnaire including some questions from the Canine Brief Pain Inventory (CBPI)
      • Brown D.C.
      • Boston R.C.
      • Coyne J.C.
      • Farrar J.T.
      Ability of the Canine Brief Pain Inventory to detect response to treatment in dogs with osteoarthritis.
      supplemented with questions relating to mobility; owners were asked to focus on their dog's behavior of the last 7 days. To make the scores more intuitive for Dutch owners, scales were flipped (i.e., 1 = lowest; 10 = excellent). Synovial PGE2 and CCL2 levels were measured by ELISA according to the manufacturers' instructions (1:10 diluted) as described before.

      Statistical analysis

      GraphPad 9.0.0 was used for statistical evaluation. Normality of the data was checked by assessing the Q–Q plots, histograms, and Shapiro–Wilks tests. All tests were subjected to Benjamini-Hochberg post-hoc correction for multiple testing and P-values <0.05 were considered significant.
      For the rat study, differences in joint swelling, 50% threshold, and DWB parameters during the priming (day −14 to day 0) and the three reactivation periods (i.e., day 0 (day 1–13), 14 (day 15–27) and 28 (day 29–42)) between treatment groups, were determined with mixed model analysis with “time” and “treatment” as fixed effects and “rat” as random effects. Differences in the area under the curve (AUCs) per period between the treatment groups were evaluated by Welch's t-tests. Differences in histological scores (e.g., Krenn and Mankin score) were evaluated with non-parametric Kruskall–Wallis tests.
      For the in vitro analysis, differences in PGE2 and CCL2 secretion by the ACs in response to empty- and CXB-PEAMs, were evaluated with a two-way ANOVA (factors: treatment; donor).
      For the dog study, the sum score of all questions and δ-sum score (sum score at 1 or 2 months – sum score at baseline) were computed. The sum score was analyzed as repeated measures in a Friedman test to determine the treatment effect over time within the group. To determine efficacy, i.e., difference between treatment arms, the δ-sum score at each time point was subjected to Kruskal–Wallis tests correcting therefore the baseline. Force plate outcomes were analyzed with linear mixed-effects models of non-transformed [peak propulsive force (PPF)] and log transformed data [peak vertical force (PVF), vertical impulse (VI)] with fixed (“time”) and random (“dog”) effects to study differences within the group, while efficacy was evaluated with Brown-Forsythe ANOVA to evaluate differences between groups at the same timepoint.
      Visual lameness scores were analyzed via contingency tables and Fisher's exact tests for lame vs non lame dogs at each time point between treatment arms. SF biomarkers were evaluated with non-parametric Wilcoxon signed-rank tests for differences within groups between time points. To determine efficacy, in the subset of paired SF data, the change in PGE2 (δPGE2) and CCL2 (δCCL2) was computed and subjected to statistical analysis using the unpaired t test for δPGE2 and the Mann Whitney test for δCCL2.
      To evaluate dependency on additional pain relief medication, a log rank test was run to determine differences in the survival distribution (“analgesics free period”) between the treatment- and placebo group. To assist interpretation, the mean difference (MD) with the 95% confidence interval (CI) is provided as MD [lower CI; upper CI] for the continuous clinical data.

      Results

      Effect of local delivery of CXB on acute arthritis signs in rat knee joints

      Treatment with CXB-PEAMs did not significantly reduce joint swelling compared to empty-PEAMs during the three reactivation periods (day 1–42) (P = 0.799) although a slight reduction was apparent in the first period [Fig. 1(B)]. The degree of structural cartilage degeneration was not changed by PGPS priming compared to control, nor was it affected by any treatment [Fig. 1(C)]. Synovitis was significantly increased in both the joints treated with empty-PEAMs and joints treated with CXB-PEAMs compared to the contralateral control knees, supporting that the PGPS model induced synovial changes in the primed knees at day 42 [Fig. 1(D)]. Local application of CXB-PEAMs did not ameliorate synovitis compared to empty-PEAMs during the entire experiment.

      Local delivery of CXB does not alleviate pain-like behaviors in rats with arthritic knee joints

      During the priming period (day-14 to day 0), rats developed mechanical hypersensitivity. Despite randomization, mechanical sensitivity of the hind paw during the priming period was higher in rats allocated to CXB-PEAMs, based on the significantly smaller (AUC) of the 50% threshold, compared to those allocated to receive empty-PEAMs at day 0. Each intravenous injection of PGPS increased mechanical sensitivity after each reactivation. Only after the second reactivation, mechanical sensitivity in CXB-treated rats was further increased compared to controls, as seen by a decrease of 50% threshold based on the AUC for the three reactivation periods following intravenous administration of PGPS at day 0, day 14, and 28 [Fig. 2(A)]. All DWB parameters were similar between groups throughout the entire experiment, though the ipsilateral/contralateral-ratio tended to be lower during the 3rd reactivation period for the CXB-treated rats (P = 0.074), indicating that the rats unloaded the experimental limb [Fig. 2(B) and (C)].
      Fig. 2
      Fig. 2CXB-PEAMs do not alleviate pain in acute arthritis in rats. Knees were primed at day −14 (blue arrow) with an intra-articular injection of PGPS and flares were evoked with intravenous injections of PGPS at 0, 14, and 28 days (red arrows). Single intra-articular injections CXB-loaded PEAMs (solid lines) or empty PEAMs (placebo; dotted lines) were given 2.5 h prior to re-activation on day 0 (green dotted line). (A) Mechanical sensitivity of the hind paws and (B–C) dynamic weight bearing was measured at various time points throughout the experiment for n = 6 rats per group (n = 5 rats for CXB-PEAMs after 42 days), presented as mean ± SD and corresponding area under the curve (AUC) are provided for the priming and the three re-activation periods upon intra-venous injection of PGPS. Only P-values <0.05 are annotated in this figure. An increase in mechanical sensitivity indicates that local CXB release did not alleviate arthritis-induced pain.

      Sustained release of CXB from PEAMs alleviates pain in clinical dog OA

      The capacity of PEAM-mediated CXB release to suppress inflammation was evaluated in dog ACs stimulated by TNF. CXB medium levels at day 4 and 14 confirmed sustained release from the PEAMs [Fig. S3(A)]. Culturing without PEAMs or with empty-PEAMs did not affect PGE2 secretion by dog ACs, but this was significantly decreased after exposure to CXB-PEAMs [Fig. S3(B)]. Culturing ACs with empty-PEAMs increased CCL2 production which was prevented by CXB-PEAM- [P = 0.003; Fig. S3(C)].
      Baseline demographics and read outs of the thirty included dogs are provided in Table I and supplementary data (Fig. S4, Tables S3 and S4). There were no major adverse effects during the study period. Peri-articular joint swelling 1–3 days after placebo administration occurred in 2 dogs, which resolved after a few days without additional treatment.
      Table IBaseline demographics for dogs with osteoarthritis (OA) treated with celecoxib-loaded polyesteramide microspheres (CXB-PEAMs) or empty PEAMs (placebo)
      ParameterCXB-PEAMsEmpty PEAMs (placebo)
      Joint distributionhip (n = 5), knee (n = 6), Elbow (n = 9)hip (n = 1), knee (n = 5), elbow (n = 5)
      Age (median (range))5 (9 mo–13 yrs)7.5 (2 yrs–10 yrs)
      Body weight (mean (SD))32 (12)34 (10)
      Gender3 M, 7 MN, 2 F, 9 FN3 M, 2 MN, 5 FN
      Duration of NSAID use in months (mean (SD))12 (12)15 (17)
      Multimodal therapy (i.e., medications in addition to NSAID)Tramadol (n = 3), Tramadol + phenylbutazone (n = 1)
      Previous surgery to the affected joint8/20 (40%)5/10 (50%)
      Sumscore questionnaire (median (range))46 (32–81)49 (20–88)
      Visual lameness score (median (range))1 (1–3)2 (1–3)
      SI – PVF (mean (SD))16.3 (20.6)20.3 (25.8)
      SI – VI (mean (SD))18.3 (23.5)27.5 (29.9)
      SI – PPF (mean (SD))28.8 (24.7)41.8 (35.8)
      Radiographic OA score3 (1–3)3 (2–3)
      Osteophyte score2 (1–3)2 (1–3)
      Mo, months; yrs, years; M, male; MN, male neutered; F, female; FN, female neutered; SI: symmetry index; PVF, peak vertical force; VI, vertical impulse; PPF, peak propulsive force; SD: standard deviation.
      The sum score of the owner's questionnaire was considered the primary patient-oriented read-out. CXB-PEAMs-treated dogs improved in the sum score questionnaire at 1 month (P = 0.0035; rank sum difference −18) and 2 months (P = 0.0150; rank sum difference −15) after treatment [Fig. 3(A)], while the placebo group did not (P > 0.999; rank sum difference 0). The significantly higher δ-sum score at both 1 (P = 0.0127; rank sum difference 16.6) and 2 months (P = 0.0284; rank sum difference 14.7) in the CXB-PEAM-treated group compared to placebo indicated efficacy of the CXB-PEAMs treatment [Fig. 3(B), Table S5].
      Fig. 3
      Fig. 3Sustained release of CXB from PEAMs alleviates pain in OA dog patients. Randomized controlled clinical trial of client-owned dogs suffering from naturally occurring OA. (A) Owner questionnaire sum score and (B) differences in sum score (δ) (sum core at 1 or 2 months – sum score at baseline) provided by the owners of the dogs treated with placebo (n = 10) or CXB-PEAMs (n = 20) at the 1- and 2-month follow-up visits. Means indicated by red bars. (C) Kaplan–Meier curve depicting the additional pain relief medication during the study period and (D) Proportions of the visual lameness scores before and after treatment with placebo or CXB-PEAMs. Symmetry indices (SI) of (E) the peak vertical force (PVF), (F) Peak Propulsive Force (PPF), and (G) the Vertical Impulse (VI) of placebo treated or CXB-PEAMs treated patients 1 and 2 months after administration. Only P-values <0.05 are annotated in this figure.
      Additional pain relief medication was deemed necessary for 9/10 dogs allocated to the placebo group vs 3/20 dogs that received CXB-PEAMs. The survival distribution (“additional analgesia free period”) was significantly higher in the treatment vs placebo group [χ2 = 25.75, P < 0.0001, Fig. 3(C)] with a median medication-free period being undefined for CXB and 7 days for placebo. Radiographic OA severity and osteophyte height did not change during the follow-up period (Table S4; P > 0.15).
      There were significantly less lame CXB-PEAMs-treated dogs at 1 and 2 months compared to placebo (P = 0.012 and P = 0.0016) with dogs receiving placebo having an infinite relative risk to be dependent on oral medication. Kinetic gait parameters improved significantly at 1 and 2 months after intra-articular injection with CXB-PEAMs. Specifically, PVF, VI, and PPF SIs improved at 1- and 2-months post injection of CXB-PEAMs [Fig. 3(E)–(G); Table S6], whereas no statistically significant differences were found between time-points within the placebo group. Only at 1-mo follow-up, PPF and VI SIs were significantly better in CXB-PEAM-treated dogs compared to placebo (P = 0.025; −33.8 [−59.9; −7.69] and P = 0.0031; −0.35 [−0.55; −0.15], respectively).

      Effects of CXB-PEAMs on joint inflammation in dogs with OA

      In joints injected with CXB-PEAMs, SF PGE2 levels were significantly lower than at baseline [Fig. 4(A); P = 0.025; −474 [−1820; 24]]. CCL2 levels had significantly increased compared to baseline in the placebo-treated, but not in the CXB-PEAMs-treated joints [Fig. 4(B); P = 0.031; 562 [0–4486]]. The δPGE2 tended to be lower in the CXB-PEAMs-treated joints compared to placebo (P = 0.0768; −599 [−1270; 72]), while δCCL2 did not differ (P = 0.8369; 59 [−4065; 2606]).
      Fig. 4
      Fig. 4Pain relief after local controlled release of CXB is associated with inhibition of Prostaglandin E2 secretion. (A) Prostaglandin E2 (PGE2) and (B) CCL2 levels in the synovial fluid of dog OA patients treated with placebo (P; empty PEAMs) or CXB-PEAMs at baseline post treatment. Data of individual dogs are presented as dot plots including the mean. There were n = 8 placebo and n = 16 CXB-treated, and n = 9 placebo and n = 14 CXB-treated synovial samples at baseline and at 2 months follow up available. Of those, n = 7 placebo and n = 12 CXB-treated paired samples were used to compute and analyze the change (δ) in synovial fluid levels over time (C) δPGE2 and (D) δCCL2. Only P-values <0.05 and trending differences are annotated in this figure.

      Discussion

      Drug delivery platforms that gradually release drugs after intra-articular administration have become of much interest as a treatment strategy for OA
      • Tryfonidou M.A.
      • de Vries G.
      • Hennink W.E.
      • Creemers L.B.
      “Old Drugs, New Tricks” – local controlled drug release systems for treatment of degenerative joint disease.
      . The present study shows that controlled release of CXB from PEAMs is not suitable for treating acute arthritis but was safe and effective to treat pain related to chronic OA in dogs for at least 2 months after injection.
      It has become more evident that also inflammation contributes to pain in OA
      • Philpott H.T.
      • O'Brien M.
      • McDougall J.J.
      Attenuation of early phase inflammation by cannabidiol prevents pain and nerve damage in rat osteoarthritis.
      ,
      • McDougall J.J.
      • Muley M.M.
      • Philpott H.T.
      • Reid A.
      • Krustev E.
      Early blockade of joint inflammation with a fatty acid amide hydrolase inhibitor decreases end-stage osteoarthritis pain and peripheral neuropathy in mice.
      . We therefore evaluated the efficacy of CXB-PEAMs in the PGPS rat model, a model characterized by severe flares of inflammation. Despite the total dose of ∼400 μg CXB-PEAMs shown to be effective in inhibiting inflammation in OA
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Pouran B.
      • de Visserm H.M.
      • Weinans H.H.
      • Thomas R.E.
      • et al.
      Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
      , which theoretically would result into at least 20% being released within the first week of injection and hence exceeds the reported IC50 for COX-2
      • Chan C.-C.
      • Boyce S.
      • Brideau C.
      • Charleson S.
      • Cromlish W.
      • Ethier D.
      • et al.
      Rofecoxib [Vioxx, MK-0966; 4-(4′-Methylsulfonylphenyl)-3-phenyl-2-(5H)-furanone]: a potent and orally active cyclooxygenase-2 inhibitor. Pharmacological and biochemical profiles.
      by 2000–8000 times this did not reduce joint swelling nor lowered inflammation-induced mechanical hypersensitivity in the paws over the course of 42 days. Noteworthy, in vitro release profiles are not predictive for the in vivo situation, where many factors are at play that influence release of the drug and clearance from the joint
      • Tryfonidou M.A.
      • de Vries G.
      • Hennink W.E.
      • Creemers L.B.
      “Old Drugs, New Tricks” – local controlled drug release systems for treatment of degenerative joint disease.
      . There may be several reasons why CXB-PEAMs were unable to treat pain caused by the severe inflammatory state of the joint, including (a) dosing of CXB was too low for this type of arthritis and hence the release of the CXB-PEAMS did not match the needs of a fulminantly inflamed joint; (b) rapid degradation of the PEAMs due to increased levels of protease activity typically present in heavily inflamed joints like in rheumatoid arthritis
      • Solau-Gervais E.
      • Zerimech F.
      • Lemaire R.
      • Fontaine C.
      • Huet G.
      • Flipo R.M.
      Cysteine and serine proteases of synovial tissue in rheumatoid arthritis and osteoarthritis.
      limiting the duration of the therapeutic effect of the released drug, and (c) the generally limited potency of NSAIDs in fulminant inflammation. PGE2 may not be the main driver in the PGPS arthritis model, where Th2-dependent monocyte infiltration is dominating
      • Kimpel D.
      • Dayton T.
      • Kannan K.
      • Wolf R.E.
      Streptococcal cell wall induced arthritis: leukocyte activation in extra-articular lymphoid tissue.
      . Employing a similar drug delivery system releasing the glucocorticoid triamcinolone acetonide we demonstrated major improvements on joint inflammation in the same rat arthritis model
      • Rudnik-Jansen I.
      • Schrijver K.
      • Woike N.
      • Tellegen A.
      • Versteeg S.
      • Emans P.
      • et al.
      Intra-articular injection of triamcinolone acetonide releasing biomaterial microspheres inhibits pain and inflammation in an acute arthritis model.
      . Corticosteroids exert stronger inhibitory effects on leukocytes (e.g., Th2-lymphocytes) compared to NSAIDs
      • Maślanka T.
      Effect of dexamethasone and meloxicam on counts of selected T lymphocyte subpopulations and NK cells in cattle - in vivo investigations.
      . Furthermore, the mode of action between glucocorticoids and NSAIDS is rather different. Glucocorticoids can affect inflammatory responses pre-and post-transcriptionally, for example, inhibiting transcription factors that induce cytokines or affect cytokine signaling in recipient cells
      • Barnes P.J.
      How corticosteroids control inflammation: Quintiles Prize Lecture 2005.
      ,
      • Cashman J.N.
      The mechanisms of action of NSAIDs in analgesia.
      , while COX-2 inhibitors target specific production of prostaglandins
      • Knights K.M.
      • Mangoni A.A.
      • Miners J.O.
      Defining the COX inhibitor selectivity of NSAIDs: implications for understanding toxicity.
      . Pathways downstream of PGE2 may be modulated by NSAIDs, while glucocorticoids have a broader action which may have translated to the differences seen in pain inhibition in our previous study. These observations emphasize the importance of patient stratification to determine the target population, or OA phenotype, that will respond and benefit most from the treatment.
      We utilized the dog patient serving as target species and as a natural disease model for human patients
      • Meeson R.L.
      • Todhunter R.J.
      • Blunn G.
      • Nuki G.
      • Pitsillides A.A.
      Spontaneous dog osteoarthritis — a One Medicine vision.
      to evaluate safety and efficacy of the CXB-PEAMs platform for OA pain management. In vitro, the presence of CXB-PEAMs lowered the secretion of PGE2 and CCL2 by ACs, indicating an anti-inflammatory effect. The present study did not specifically study biocompatibility in the OA joint. The increased secretion of CCL2 by ACs upon exposure to empty-PEAMs may relate to the increased phagocytic capacity of OA ACs and the adoption of a pro-inflammatory phenotype
      • Caramés B.
      • Taniguchi N.
      • Otsuki S.
      • Blanco F.J.
      • Lotz M.
      Autophagy is a protective mechanism in normal cartilage, and its aging-related loss is linked with cell death and osteoarthritis.
      . Whether the synovial lining macrophages in OA joints become activated upon exposure to the PEAM degradation products, as known for other materials
      • Utomo L.
      • Boersema G.S.A.
      • Bayon Y.
      • Lange J.F.
      • Van Osch G.J.V.M.
      • Bastiaansen-Jenniskens Y.M.
      In vitro modulation of the behavior of adhering macrophages by medications is biomaterial-dependent.
      , remains to be determined. It was reported that microspheres are entrapped in the synovium and surrounded by mononuclear inflammatory and giant cells; CD68+ macrophages were detected in the synovial lining of OA rat knees that received empty-PEAMs
      • Janssen M.
      • Timur U.T.
      • Woike N.
      • Welting T.J.M.
      • Draaisma G.
      • Gijbels M.
      • et al.
      Celecoxib-loaded PEA microspheres as an auto regulatory drug-delivery system after intra-articular injection.
      . Nonetheless, it remains to be determined whether the presence of macrophages is biomaterial-driven cell infiltration or relates to the OA joint environment wherein macrophages have a well-described role
      • Fernandes T.L.
      • Gomoll A.H.
      • Lattermann C.
      • Hernandez A.J.
      • Bueno D.F.
      • Amano M.T.
      Macrophage: a potential target on cartilage regeneration.
      .
      In the population of patient dogs suffering from OA, significant improvements in owner-perceived pain and lameness were evident in the CXB-PEAMs but not the placebo treatment, indicating a minimal caregiver placebo effect in this veterinary clinical trial. This seemingly contradicts Conzemius and Evans
      • Conzemius M.G.
      • Evans R.B.
      Caregiver placebo effect for dogs with lameness from osteoarthritis.
      reporting ∼40% caregiver and veterinarian placebo effects in an FDA-approved dog study. Noteworthy, the latter was only based on global and subjective evaluation of lameness. Several new owner questionnaires, inspired from tools applied to human patients, have been developed and adopted for veterinary use. Recent large randomized veterinary clinical studies in the OA field employing the CBPI, reported placebo effects ranging between 16%
      • Corral M.J.
      • Moyaert H.
      • Fernandes T.
      • Escalada M.
      • Kira S Tena J.
      • Walters R.R.
      • et al.
      A prospective, randomized, blinded, placebo-controlled multisite clinical study of bedinvetmab, a canine monoclonal antibody targeting nerve growth factor, in dogs with osteoarthritis,.
      to 28%
      • Robertson-Plouch C.
      • Stille J.R.
      • Liu P.
      • Smith C.
      • Brown D.
      • Warner M.
      • et al.
      A randomized clinical efficacy study targeting mPGES1 or EP4 in dogs with spontaneous osteoarthritis,.
      . Importantly, these beneficial analgesic effects of the CXB-PEAMs were corroborated by objective measures. Dogs treated with CXB-PEAMs did not depend on oral pain medication compared to placebo over time and showed a decrease in asymmetrical limb loading during the follow-up period, indicating improvement in gait due to pain relief.
      To further understand what could have contributed to clinical improvement in the dog RCT, SF PGE2 and CCL2 levels were determined. The reduction in PGE2 levels, but not in CCL2, suggests that relief of clinical signs was likely mediated by COX-2 inhibition in the OA dog population studied, and not via the CCL2/CCR2 signaling axis as others have reported
      • Li L.
      • Jiang B.E.
      Serum and synovial fluid chemokine ligand 2/monocyte chemoattractant protein 1 concentrations correlates with symptomatic severity in patients with knee osteoarthritis.
      ,
      • Stock J.L.
      • Shinjo K.
      • Burkhardt J.
      • Roach M.
      • Taniguchi K.
      • Ishikawa T.
      • et al.
      The prostaglandin E2 EP1 receptor mediates pain perception and regulates blood pressure.
      . Despite the strong reduction in pain outcomes in the dog population, CXB-PEAMs did not cause structural changes within the 2-month follow up period. Here, osteophytes were not affected, which contrasts with the effects seen previously in an OA rat model, where this, along with other structural changes, was reduced
      • Tellegen A.R.
      • Rudnik-Jansen I.
      • Pouran B.
      • de Visserm H.M.
      • Weinans H.H.
      • Thomas R.E.
      • et al.
      Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
      . However, such experimental models are based on the induction of disease, hence osteophyte formation was most likely prevented rather than reverted. Furthermore, in the clinical dog study, the follow-up period might have been too: an osteophyte size increase rate of only 1 mm/3 years has been reported in the dog (i.e., ∼0.028 mm/2 months)
      • Greene L.M.
      • Marcellin-Little D.J.
      • Lascelles B.D.X.
      Associations among exercise duration, lameness severity, and hip joint range of motion in labrador retrievers with hip dysplasia.
      . The use of more advanced and sensitive imaging modalities such as CT or MRI might provide more information.
      In conclusion, local sustained release of CXB with CXB-PEAMs is safe and effective to treat chronic OA pain in dogs suffering from natural disease, but was unsuitable to treat acute inflammation in arthritic joints with flares in rats. In OA dog patients, limb function and quality of life were improved, and additional pain medication dependency over a 2-month period was decreased without any substantial adverse effects. Follow-up studies with a larger sample size, longer follow-up time, and predefined rescue analgesia protocols are needed to investigate whether prolonged local exposure to CXB results in less osteophyte formation in dog patients with early OA. Considering the promising data in the current study, this CXB delivery platform could offer a suitable strategy for long-term OA pain management and further advance this concept towards first-in-man strategies.

      CRediT author contributions

      Anna Tellegen: Methodology, Formal analysis, Data curation, Investigation, Writing- original draft preparation, Visualization. Imke Rudnik-Jansen: Methodology, Investigation, Data curation and analysis, Writing-reviewing. Lizette Utomo: Formal analysis, Data interpretation, Writing- original draft preparation, reviewing, and editing, Visualization. Sabine Versteeg: Investigation, Data curation. Martijn Beukers: Methodology, Investigation. Roelof Maarschalkerweerd: Investigation. Dick van Zuilen: Investigation. Nicoline van Klaveren: Investigation. Kaat Houben: Investigation. Erik Teske: Investigation. René van Weeren: Writing-reviewing and editing. Nina Karssemakers-Degen: Methodology, Formal analysis, Investigation. George Mihov: Conceptualization, Methodology, supervision. Jens Thies: Conceptualization, supervision. Niels Eijkelkamp: Methodology, supervision. Laura Creemers: Conceptualization, Methodology, Supervision, Data interpretation, Writing-reviewing and editing, Funding acquisition. Björn Meij: Methodology, Investigation, Supervision, Data curing and interpretation, Writing-reviewing and editing. Marianna Tryfonidou: Conceptualization, Methodology, Investigation, Supervision, Data analysis and interpretation, Writing- original draft preparation, reviewing, and editing, Project administration, Funding acquisition.

      Conflict of interest

      The authors from DSM Biomedical have proprietary and commercial interest in the materials discussed in this article. None of the authors has any other financial or personal relationships that could inappropriately influence or bias the content of the paper.

      Funding

      This work was financially supported by a research grant from Life Sciences Health (LSH) Impulse, (ArIADNE; project #40-43100-98-022). The financial contribution of the Dutch Arthritis Society is also gratefully acknowledged (LLP22 and LLP12).

      Acknowledgments

      We wish to thank Arie Doornenbal and Willem de Jong for their technical assistance.

      Appendix A. Supplementary data

      The following are the Supplementary data to this article:
      Fig. S1
      Fig. S1Characteristics of celecoxib-loaded polyesteramide microspheres (CXB-PEAMs). A) Scanning electron microscopy (SEM) images of CXB-PEAMs at two magnifications. Scale bar: 50 μm (left); 10 μm (right). B) absolute (μg) and relative (%) cumulative release of CXB from the PEAMs measured in vitro in PBS. Data shown as mean ± SD for n = 3 technical replicates.
      Fig. S2
      Fig. S2Examples of radiographic data indicating the locations of osteophyte measurements. Radiographs obtained prior to or two months after intra-articular injection with celecoxib-loaded or unloaded polyesteramide microspheres. Radiography was employed to confirm OA and assessment of OA progression was done based on osteophyte height measurements at indicated locations (arrows). (A) Mediolateral and (B) craniocaudal radiographic projections obtained of the elbow joint of dog no. 19 prior to inclusion in the study. Osteophytes were measured at the cranial aspect of the radial head, the caudal surface of the lateral condylar ridge, the medial contour of the humeral condyle and the medial contour of the medial coronoid process. (C) Mediolateral and (D) caudocranial radiographs of the knee joint of dog 1 prior to inclusion in the study. Osteophytes were measured at the proximal trochlear edge, the distal patella, and the lateral and medial tibial plateau. (E) Ventrodorsal radiographs of the pelvis of dog no. 28 prior to and (F) 2 months after treatment of the left hip joint with empty microspheres, respectively. Osteophytes on the cranial and caudal edge of the acetabulum and at the femoral neck were recorded.
      Fig. S3
      Fig. S3Effects of celecoxib-loaded polyesteramide microspheres (CXB-PEAMs) on dog articular chondrocyte (AC) culture. (A) Schematic overview of the experimental design used for evaluating the biological effects of CXB released from CXB-PEAMs on primary dog ACs stimulated with 10 ng/mL recombinant human tumor necrosis factor alpha (TNF-α; control) to simulate inflammation. The released CXB in the culture medium was measured via ELISA at days 4 and 14 (n = 3 AC donors). (B) PGE2 (n = 3 AC donors) and (C) CCL2 (n = 6 AC donors) concentrations in the medium of the ACs in response to the presence of PEAMs and released CXB from CXB-PEAMs.
      Fig. S4
      Fig. S4Inclusion flowchart of dog patients with osteoarthritis treated with celecoxib-loaded polyesteramide microspheres (CXB-PEAMs) or empty PEAMs (placebo).

      References

        • Hunter D.J.
        • Bierma-Zeinstra S.
        Osteoarthritis.
        Lancet. 2019; 393: 1745-1759https://doi.org/10.1016/S0140-6736(19)30417-9
        • Dell'Isola A.
        • Allan R.
        • Smith S.L.
        • Marreiros S.S.P.
        • Steultjens M.
        Identification of clinical phenotypes in knee osteoarthritis: a systematic review of the literature.
        BMC Musculoskelet Disord. 2016; 17: 1-12https://doi.org/10.1186/s12891-016-1286-2
        • Sellam J.
        • Berenbaum F.
        The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis.
        Nat Rev Rheumatol. 2010; 6: 625-635https://doi.org/10.1038/nrrheum.2010.159
        • Appleton C.T.G.
        • Pitelka V.
        • Henry J.
        • Beier F.
        Global analyses of gene expression in early experimental osteoarthritis.
        Arthritis Rheum. 2007; 56: 1854-1868https://doi.org/10.1002/art.22711
        • Trumble T.N.
        • Billinghurst R.C.
        • McIlwraith C.W.
        Correlation of prostaglandin E2 concentrations in synovial fluid with ground reaction forces and clinical variables for pain or inflammation in dogs with osteoarthritis induced by transection of the cranial cruciate ligament.
        Am J Vet Res. 2004; 65: 1269-1275https://doi.org/10.2460/ajvr.2004.65.1269
        • Brenner S.S.
        • Klotz U.
        • Alscher D.M.
        • Mais A.
        • Lauer G.
        • Schweer H.
        • et al.
        Osteoarthritis of the knee – clinical assessments and inflammatory markers.
        Osteoarthr Cartil. 2004; 12: 469-475https://doi.org/10.1016/j.joca.2004.02.011
        • Heikkilä H.M.
        • Hielm-Björkman A.K.
        • Innes J.F.
        • Laitinen-Vapaavuori O.M.
        The effect of intra-articular botulinum toxin A on substance P, prostaglandin E2, and tumor necrosis factor alpha in the canine osteoarthritic joint.
        BMC Vet Res. 2017; 13https://doi.org/10.1186/s12917-017-0990-y
        • Zweers M.C.
        • de Boer T.N.
        • van Roon J.
        • Bijlsma J.W.J.
        • Lafeber F.P.J.G.
        • Mastbergen S.C.
        Celecoxib: considerations regarding its potential disease-modifying properties in osteoarthritis.
        Arthritis Res Ther. 2011; 13https://doi.org/10.1186/ar3437
        • Nakata K.
        • Hanai T.
        • Take Y.
        • Osada T.
        • Tsuchiya T.
        • Shima D.
        • et al.
        Disease-modifying effects of COX-2 selective inhibitors and non-selective NSAIDs in osteoarthritis: a systematic review.
        Osteoarthr Cartil. 2018; 26: 1263-1273https://doi.org/10.1016/j.joca.2018.05.021
        • Panahifar A.
        • Jaremko J.L.
        • Tessier A.G.
        • Lambert R.G.
        • Maksymowych W.P.
        • Fallone B.G.
        • et al.
        Development and reliability of a multi-modality scoring system for evaluation of disease progression in pre-clinical models of osteoarthritis: celecoxib may possess disease-modifying properties.
        Osteoarthr Cartil. 2014; 22: 1639-1650https://doi.org/10.1016/j.joca.2014.06.013
        • Sostres C.
        • Gargallo C.J.
        • Lanas A.
        Nonsteroidal anti-inflammatory drugs and upper and lower gastrointestinal mucosal damage.
        Arthritis Res Ther. 2013; 15: S3https://doi.org/10.1186/ar4175
        • Bertin P.
        • Lapicque F.
        • Payan E.
        • Rigaud M.
        • Bailleul F.
        • Jaeger S.
        • et al.
        Sodium naproxen: concentration and effect on inflammatory response mediators in human rheumatoid synovial fluid.
        Eur J Clin Pharmacol. 1994; 46: 3-7https://doi.org/10.1007/BF00195907
        • Aarons L.
        • Salisbury R.
        • Alam-Siddiqi M.
        • Taylor L.
        • Grennan D.
        Plasma and synovial fluid kinetics of flurbiprofen in rheumatoid arthritis.
        Br J Clin Pharmacol. 1986; 21: 155-163https://doi.org/10.1111/j.1365-2125.1986.tb05170.x
        • Jiang D.
        • Zou J.
        • Huang L.
        • Shi Q.
        • Zhu X.
        • Wang G.
        • et al.
        Efficacy of intra-articular injection of celecoxib in a rabbit model of osteoarthritis.
        Int J Mol Sci. 2010; 11: 4106-4113https://doi.org/10.3390/ijms11104106
        • Janssen M.
        • Timur U.T.
        • Woike N.
        • Welting T.J.M.
        • Draaisma G.
        • Gijbels M.
        • et al.
        Celecoxib-loaded PEA microspheres as an auto regulatory drug-delivery system after intra-articular injection.
        J Control Release. 2016; 244: 30-40https://doi.org/10.1016/j.jconrel.2016.11.003
        • Tellegen A.R.
        • Rudnik-Jansen I.
        • Pouran B.
        • de Visserm H.M.
        • Weinans H.H.
        • Thomas R.E.
        • et al.
        Controlled release of celecoxib inhibits inflammation, bone cysts and osteophyte formation in a preclinical model of osteoarthritis.
        Drug Deliv. 2018; 25: 1438-1447https://doi.org/10.1080/10717544.2018.1482971
        • Anderson K.L.
        • O’Neill D.G.
        • Brodbelt D.C.
        • Church D.B.
        • Meeson R.L.
        • Sargan D.
        • et al.
        Prevalence, duration and risk factors for appendicular osteoarthritis in a UK dog population under primary veterinary care.
        Sci Rep. 2018; 8: 1-12https://doi.org/10.1038/s41598-018-23940-z
        • Meeson R.L.
        • Todhunter R.J.
        • Blunn G.
        • Nuki G.
        • Pitsillides A.A.
        Spontaneous dog osteoarthritis — a One Medicine vision.
        Nat Rev Rheumatol. 2019; 15: 273-287https://doi.org/10.1038/s41584-019-0202-1
        • Rudnik-Jansen I.
        • Tellegen A.
        • Beukers M.
        • Öner F.
        • Woike N.
        • Mihov G.
        • et al.
        Safety of intradiscal delivery of triamcinolone acetonide by a poly(esteramide)microsphere platform in a large animal model of intervertebral disc degeneration.
        Spine J. 2019; 19: 905-919https://doi.org/10.1016/j.spinee.2018.10.014
        • Tellegen A.R.
        • Rudnik-Jansen I.
        • Beukers M.
        • Miranda-Bedate A.
        • Bach F.C.
        • de Jong W.
        • et al.
        Intradiscal delivery of celecoxib-loaded microspheres restores intervertebral disc integrity in a preclinical canine model.
        J Control Release. 2018; 286: 439-450https://doi.org/10.1016/j.jconrel.2018.08.019
        • Rudnik-Jansen I.
        • Schrijver K.
        • Woike N.
        • Tellegen A.
        • Versteeg S.
        • Emans P.
        • et al.
        Intra-articular injection of triamcinolone acetonide releasing biomaterial microspheres inhibits pain and inflammation in an acute arthritis model.
        Drug Deliv. 2019; 26: 226-236https://doi.org/10.1080/10717544.2019.1568625
        • Chaplan S.R.
        • Bach F.W.
        • Pogrel J.W.
        • Chung J.M.
        • Yaksh T.L.
        Quantitative assessment of tactile allodynia in the rat paw.
        J Neurosci Methods. 1994; 53: 55-63https://doi.org/10.1016/0165-0270(94)90144-9
        • Minett M.S.
        • Falk S.
        • Santana-Varela S.
        • Bogdanov Y.D.
        • Nassar M.A.
        • Heegaard A.M.
        • et al.
        Pain without nociceptors? Nav1.7-independent pain mechanisms.
        Cell Rep. 2014; 6: 301-312https://doi.org/10.1016/j.celrep.2013.12.033
        • Mankin H.J.
        • Lippiello L.
        Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips.
        J Bone Joint Surg Am. 1970; 52: 424-434https://doi.org/10.2106/00004623-197052030-00002
        • Krenn V.
        • Morawietz L.
        • Burmester G.R.
        • Kinne R.W.
        • Mueller-Ladner U.
        • Muller B.
        • et al.
        Synovitis score: discrimination between chronic low-grade and high-grade synovitis.
        Histopathology. 2006; 49: 358-364https://doi.org/10.1111/j.1365-2559.2006.02508.x
        • Yang H.Y.
        • Van Dijk M.
        • Licht R.
        • Beekhuizen M.
        • Van Rijen M.
        • Janstål M.K.
        • et al.
        Applicability of a newly developed bioassay for determining bioactivity of anti-inflammatory compounds in release studies - celecoxib and Triamcinolone acetonide released from novel PLGA-based microspheres.
        Pharm Res. 2015; 32: 680-690https://doi.org/10.1007/s11095-014-1495-z
        • Willems N.
        • Mihov G.
        • Grinwis G.C.M.
        • van Dijk M.
        • Schumann D.
        • Bos C.
        • et al.
        Safety of intradiscal injection and biocompatibility of polyester amide microspheres in a canine model predisposed to intervertebral disc degeneration.
        J Biomed Mater Res Part B Appl Biomater. 2017; 105: 707-714https://doi.org/10.1002/jbm.b.33579
        • Lavrijsen I.C.M.
        • Heuven H.C.M.
        • Voorhout G.
        • Meij B.P.
        • Theyse L.F.H.
        • Leegwater P.A.J.
        • et al.
        Phenotypic and genetic evaluation of elbow dysplasia in Dutch Labrador Retrievers, Golden Retrievers, and Bernese Mountain dogs.
        Vet J. 2012; 193: 486-492https://doi.org/10.1016/j.tvjl.2012.01.001
        • Hazewinkel H.A.W.
        • van den Brom W.E.
        • Theyse L.F.H.
        • Pollmeier M.
        • Hanson P.D.
        Comparison of the effects of firocoxib, carprofen and vedaprofen in a sodium urate crystal induced synovitis model of arthritis in dogs.
        Res Vet Sci. 2008; 84: 74-79https://doi.org/10.1016/j.rvsc.2007.02.005
        • Fanchon L.
        • Grandjean D.
        Accuracy of asymmetry indices of ground reaction forces for diagnosis of hind limb lameness in dogs.
        Am J Vet Res. 2007; 68: 1089-1094https://doi.org/10.2460/ajvr.68.10.1089
        • Brown D.C.
        • Boston R.C.
        • Coyne J.C.
        • Farrar J.T.
        Ability of the Canine Brief Pain Inventory to detect response to treatment in dogs with osteoarthritis.
        J Am Vet Med Assoc. 2008; 233: 1278-1283https://doi.org/10.2460/javma.233.8.1278
        • Tryfonidou M.A.
        • de Vries G.
        • Hennink W.E.
        • Creemers L.B.
        “Old Drugs, New Tricks” – local controlled drug release systems for treatment of degenerative joint disease.
        Adv Drug Deliv Rev. 2020; 160: 170-185https://doi.org/10.1016/j.addr.2020.10.012
        • Philpott H.T.
        • O'Brien M.
        • McDougall J.J.
        Attenuation of early phase inflammation by cannabidiol prevents pain and nerve damage in rat osteoarthritis.
        Pain. 2017; 158: 2442-2451https://doi.org/10.1097/j.pain.0000000000001052
        • McDougall J.J.
        • Muley M.M.
        • Philpott H.T.
        • Reid A.
        • Krustev E.
        Early blockade of joint inflammation with a fatty acid amide hydrolase inhibitor decreases end-stage osteoarthritis pain and peripheral neuropathy in mice.
        Arthritis Res Ther. 2017; 19https://doi.org/10.1186/s13075-017-1313-1
        • Chan C.-C.
        • Boyce S.
        • Brideau C.
        • Charleson S.
        • Cromlish W.
        • Ethier D.
        • et al.
        Rofecoxib [Vioxx, MK-0966; 4-(4′-Methylsulfonylphenyl)-3-phenyl-2-(5H)-furanone]: a potent and orally active cyclooxygenase-2 inhibitor. Pharmacological and biochemical profiles.
        J Pharmacol Exp Ther. 1999; 290
        • Solau-Gervais E.
        • Zerimech F.
        • Lemaire R.
        • Fontaine C.
        • Huet G.
        • Flipo R.M.
        Cysteine and serine proteases of synovial tissue in rheumatoid arthritis and osteoarthritis.
        Scand J Rheumatol. 2007; 36: 373-377https://doi.org/10.1080/03009740701340172
        • Kimpel D.
        • Dayton T.
        • Kannan K.
        • Wolf R.E.
        Streptococcal cell wall induced arthritis: leukocyte activation in extra-articular lymphoid tissue.
        Inflammation. 2003; 27: 59-70https://doi.org/10.1023/A:1023215515617
        • Maślanka T.
        Effect of dexamethasone and meloxicam on counts of selected T lymphocyte subpopulations and NK cells in cattle - in vivo investigations.
        Res Vet Sci. 2014; 96: 338-346https://doi.org/10.1016/j.rvsc.2014.02.006
        • Barnes P.J.
        How corticosteroids control inflammation: Quintiles Prize Lecture 2005.
        Br J Pharmacol. 2006; 148: 245-254https://doi.org/10.1038/sj.bjp.0706736
        • Cashman J.N.
        The mechanisms of action of NSAIDs in analgesia.
        Drugs. 1996; 52: 13-23https://doi.org/10.2165/00003495-199600525-00004
        • Knights K.M.
        • Mangoni A.A.
        • Miners J.O.
        Defining the COX inhibitor selectivity of NSAIDs: implications for understanding toxicity.
        Expert Rev Clin Pharmacol. 2010; 3: 769-776https://doi.org/10.1586/ecp.10.120
        • Caramés B.
        • Taniguchi N.
        • Otsuki S.
        • Blanco F.J.
        • Lotz M.
        Autophagy is a protective mechanism in normal cartilage, and its aging-related loss is linked with cell death and osteoarthritis.
        Arthritis Rheum. 2010; 62: 791-801https://doi.org/10.1002/art.27305
        • Utomo L.
        • Boersema G.S.A.
        • Bayon Y.
        • Lange J.F.
        • Van Osch G.J.V.M.
        • Bastiaansen-Jenniskens Y.M.
        In vitro modulation of the behavior of adhering macrophages by medications is biomaterial-dependent.
        Biomed Mater. 2017; 12https://doi.org/10.1088/1748-605X/aa5cbc
        • Fernandes T.L.
        • Gomoll A.H.
        • Lattermann C.
        • Hernandez A.J.
        • Bueno D.F.
        • Amano M.T.
        Macrophage: a potential target on cartilage regeneration.
        Front Immunol. 2020; : 111https://doi.org/10.3389/FIMMU.2020.00111
        • Conzemius M.G.
        • Evans R.B.
        Caregiver placebo effect for dogs with lameness from osteoarthritis.
        J Am Vet Med Assoc. 2012; 241: 1314-1319https://doi.org/10.2460/JAVMA.241.10.1314
        • Corral M.J.
        • Moyaert H.
        • Fernandes T.
        • Escalada M.
        • Kira S Tena J.
        • Walters R.R.
        • et al.
        A prospective, randomized, blinded, placebo-controlled multisite clinical study of bedinvetmab, a canine monoclonal antibody targeting nerve growth factor, in dogs with osteoarthritis,.
        Vet Anaesth Analg. 2021; 48: 943-955https://doi.org/10.1016/j.vaa.2021.08.001
        • Robertson-Plouch C.
        • Stille J.R.
        • Liu P.
        • Smith C.
        • Brown D.
        • Warner M.
        • et al.
        A randomized clinical efficacy study targeting mPGES1 or EP4 in dogs with spontaneous osteoarthritis,.
        Sci Transl Med. 2019; 11https://doi.org/10.1126/SCITRANSLMED.AAW9993/SUPPL_FILE/AAW9993_SM.PDF
        • Li L.
        • Jiang B.E.
        Serum and synovial fluid chemokine ligand 2/monocyte chemoattractant protein 1 concentrations correlates with symptomatic severity in patients with knee osteoarthritis.
        Ann Clin Biochem. 2015; 52: 276-282https://doi.org/10.1177/0004563214545117
        • Stock J.L.
        • Shinjo K.
        • Burkhardt J.
        • Roach M.
        • Taniguchi K.
        • Ishikawa T.
        • et al.
        The prostaglandin E2 EP1 receptor mediates pain perception and regulates blood pressure.
        J Clin Invest. 2001; 107: 325-331https://doi.org/10.1172/JCI6749
        • Greene L.M.
        • Marcellin-Little D.J.
        • Lascelles B.D.X.
        Associations among exercise duration, lameness severity, and hip joint range of motion in labrador retrievers with hip dysplasia.
        J Am Vet Med Assoc. 2013; 242: 1528-1533https://doi.org/10.2460/javma.242.11.1528