**3. Discussion**

In this study we demonstrated that YP, a meroterpenoid isolated from the brown alga *C. usneoides* [26], is able to decrease inflammation, cell differentiation and extracellular matrix (ECM) degradation in different osteoarthritis in vitro/ex vivo OA model systems. By using a pipeline with increasing complexity, from 2D monolayer cultures of THP-1 macrophages, primary chondrocytes and synoviocytes, to ex vivo culture of human cartilage explants and a newly developed OA explant-based co-culture model, YP consistently promoted a protective effect under pro-inflammatory and mineralizing stimuli. These results bring new evidences on the health benefits of YP as a protective OA agen<sup>t</sup> by attenuating cartilage degrading processes under known OA promoting stimuli, with consequent cartilage maintenance promoting effects, with potential therapeutic application.

In OA, cartilage loss and synovial membrane inflammation are two major factors responsible for disease progression and associated outcomes. Complex and interconnected molecular events of cartilage homeostasis disruption associated to inflammation known to fuel cartilage degradation, are recognized as crucial for disease development and important targets for therapeutic approaches [32]. Cartilage degradation is associated with chondrocytes differentiation leading to apoptosis and deposition of a mineralized extracellular matrix, which in turn contributes to loss of ECM integrity and inflammation [33]. In fact, although the pathways involved in the crosstalk between inflammation and cartilage degradation are still not completely clarified, mineralizing and inflammatory events occur in a close related manner during OA progression [30]. BCP and calcium pyrophosphate (CPP) crystals, consistently associate with the early stage of OA and have a pathogenic role in the development and rapid progression to end-stage OA [11,12]. BCPs have been found in the synovial fluid and membrane, and cartilage from OA patients [34], and associated with the activation of macrophages, synovial fibroblasts and articular cells, resulting in increased cell proliferation and production of pro-inflammatory cytokines and MMPs [35,36]. In concordance, our results show an inflammatory response to hydroxyapatite stimulation in all tested OA models, similar to those obtained with the classical inflammatory cytokine IL-1β, and to previously reported effects in OA cell models [30,37]. Of particular relevance, at cartilage level, treatment with hydroxyapatite induced overexpression of Col10 and Runx2, indicative of triggered chondrocyte differentiation towards hypertrophy and calcification. In addition, up-regulation of COX-2 and IL-6, widely known to be associated with joint inflammation, and MMP3, a major responsible for ECM degradation, clearly demonstrate the detrimental potential of calcification in OA. This is in line with recent data showing that BCP upregulate IL-6 in in vivo murine OA models, which in turn induced the expression of genes involved in calcification, promoting BCP formation and potentiating a vicious cycle [38]. Increased levels of BCP and IL-6 were also associated with cartilage degradation through the induction of matrix-degrading enzymes activity in chondrocytes [38]. In another study, calcium-phosphate complexes were shown to induce MMP3 and MMP13, which in turn, promoted the release of calcium and phosphate through degradation of the ECM calcified cartilage, in a positive loop [39]. Additionally, the effect of IL-1β on cartilage is known to reflect not only the catabolic effect of aggrecanases and MMPs upregulation, but also the downregulation of chondrogenic extracellular matrix synthesis [40,41]. In concordance, our results showed that IL-1β induced an overexpression of MMP3. Overall, our results clearly demonstrated the potentialities of the developed ex vivo explant-based co-culture OA model to study the interplay between cartilage degradation and inflammation, reflecting early molecular events leading to subsequent phenotypic cartilage alterations, of critical value in drug development for potential anti-osteoarthritic compounds such as YP.

YP has previously shown to have anti-inflammatory properties associated with the inhibition of TNF α in LPS-activated human macrophages [26]. In the present study, we demonstrated that YP was able to counteract inflammation, cell di fferentiation and ECM degradation, induced not only by IL-1β but also by hydroxyapatite, in all OA models, including primary articular cells, cartilage explants and ex vivo explant-based co-culture systems. These e ffects were demonstrated at multiple levels. Through downregulation of master players involved in pro-inflammatory reactions, such as NF-kB, COX-2 and IL-6, and the ECM catabolic marker MMP3, YP is directly contributing to preserve cartilage homeostasis, by avoiding ECM disruption and cartilage collapse. Similarly, the capacity to downregulate crucial genes involved in chondrocyte di fferentiation such as Col10 and Runx2, suggests YP as an inhibitor of chondrocytes hypertrophic di fferentiation. The resulting decrease of apoptosis and ECM mineralization, indirectly contributes to a consequent decrease of pro-inflammatory reactions, ultimately preserving cartilage homeostasis. Although our studies were not directed to evaluate the effect of YP as a structural cartilage-modifying drug, its capacity to inhibit early molecular events leading to joint deterioration, suggests YP as a potential disease modifying OA drug, worth to be further investigated.

Additionally, this YP protective role might represent a promising alternative to the anti-inflammatory drugs commercially available to manage symptomatology associated with OA and chronic autoimmune and inflammatory diseases, mostly based on NSAIDs target to inactivate COX enzymes (COX-1 and COX-2) [42], or biologics targeting crucial pro-inflammatory cytokines such as TNF α and IL-1β [43,44]. Although some e ffectiveness has been shown in slowing inflammatory reactions, the growing list of adverse side e ffects and the high percentage of patients presenting no response to these treatments, clearly demonstrate the urgen<sup>t</sup> need for safer and more e ffective anti-inflammatory drugs. In this field, natural derived products, such as YP, have been considered as promising and valid alternatives. Some examples are the tetracyclic triterpenoid glycoside Ginsenoside Rb1 (G-Rb1) and curcumin, which have shown both in vitro and in vivo the capacity of targeting the production of several pro-inflammatory species and promoting the synthesis of anti-inflammatory mediators, with cartilage protective e ffects [45–48].

Considering the pivotal role of NF-kB as a major regulator of inflammation, many strategies have been developed to block NF-kB signaling in a variety of inflammatory disease settings [49]. Although in the context of OA these strategies are still in their infancy, the crucial role of NF-kB signaling mediating inflammatory responses, but also the hypertrophic conversion of articular cartilage chondrocytes, leading to ECM damage and cartilage destruction, is of paramount importance in the disease context [50,51]. Our results demonstrate that YP is able to downregulate NF-kB expression and decrease IkB α phosphorylation in chondrocytes, strongly suggesting that YP cartilage protective properties are associated, at least in part, with the inhibition of NF-kB nuclear translocation and consequent decreased activation of catabolic pathways, including expression of cytokines and chemokines, inflammatory mediators, matrix degrading enzymes, and regulators of chondrocytes di fferentiation. In agreement, YP treatments consistently decreased levels of COX-2 and IL-6, MMP3, Col10 and Runx2 in cartilage tissue, clearly demonstrating the potential of YP in ameliorating cartilage homeostasis and integrity, a good rationale for the exploitation of YP in the treatment of OA.

### **4. Materials and Methods**

### *4.1. Isolation of Amentadione (YP)*

The meroditerpenoid amentadione (YP) was isolated from the brown alga *Cystoseira usneoides* collected off the coast of Tarifa (Spain) as previously described [26]. Briefly, the frozen alga was extracted with methanol and after evaporation of the solution under reduced pressure, the aqueous residue was extracted with diethyl ether. The resulting extract was subjected to column chromatography (CC) on silica gel (70–230 mesh) (Merck KGaA, Darmstadt, Germany) eluting with a mixture of *n*-hexane/diethyl ether (50:50, *v*/*v*), then diethyl ether, mixtures of chloroform/methanol (90:10 and 80:20, *v*/*v*), and finally methanol. The fraction that eluted with chloroform/methanol (90:10, *v*/*v*) was further separated by CC on silica gel using as eluents mixtures of *n*-hexane/ethyl acetate (50:50 to 30:70, *v*/*v*), then ethyl acetate, and finally methanol. The compound YP was obtained by reversed phase HPLC separation of selected subfractions using as eluent methanol/water (70:30, *v*/*v*). HPLC separations were performed on a LaChrom-Hitachi apparatus (Merck), equipped with Kromasil 100-5C18 columns (250 × 10 mm, 5 μm or 250 × 4.6 mm, 5 μm) (Hichrom, Reading, UK), using an RI-71 differential refractometer or L-7400 UV detector (Merck). The pure compound YP was identified as amentadione [52] by NMR and HRMS analysis (Figure S1) and the negative optical rotation. NMR spectra were recorded on an Agilent 500 spectrometer (Agilent Technologies, Santa Clara, CA, USA), HRMS spectra were obtained on a Waters SYNAPT G2 spectrometer (Waters, Milford, MA, USA), and optical rotation was measured on a JASCO P-2000 polarimeter (JASCO, Tokyo, Japan).
