Osteoarthritis (OA) is the most prevalent arthritic disease affecting the joints. While mainly related to aging, it is also associated with a diversity of risk factors including genetic predisposition, epigenetic factors, gender, obesity, exercise, work-related injury, and trauma. Irreversible and gradual loss of the articular cartilage remains the fundamental feature of OA pathophysiology ^[1][2][3][4][1,2,3,4]. However, in the course of the disease, all the joint tissues including synovium and bone undergo physical, functional, and metabolic alterations that comprise different cellular types as well as components of the extracellular cellular matrix (ECM). The ECM is a complex and specialized three-dimensional macromolecular network, present in nearly all tissues, which also interacts with cell surface receptors on joint resident cells. It is secreted, assembled, and modeled by the surrounding cells, providing physical support and organization to tissues. ECM is involved in many cell functions, providing cells with chemical and mechanical signals to regulate cell proliferation, survival, migration, and differentiation. However, changes in the composition and physical properties of the ECM lead to the development of many diseases, including cancer and rheumatic diseases, such as OA and rheumatoid arthritis (RA), among others ^[5][6][5,6]. Moreover, cell stress and ECM degradation promote maladaptive healing reactions, including inflammatory pathways of innate and adaptive immunity [7]. The loss of the biomechanical properties of cartilage induced by prior injuries or increased loading is the most significant feature in the initiation and progression of OA. Among the fibrillary components of ECM, the glycoprotein fibronectin (Fn) is an important member that acts as a bridging molecule in matrix assembly and cell–matrix interfaces. During the development of OA, tissue proteolysis and injury induce ECM degradation generating Fn fragments (Fn-fs), among other catabolic mediators, that promote inflammation and degradation by the induction of cytokine [8] and proteinase expressions [5].

Fibronectin is an adhesive glycoprotein, widely distributed in most ECM that regulates different cellular functions such as adhesion, motility, growth, differentiation, and opsonization ^[9][10][9,10]. It is a dimeric protein formed by two polypeptide chains with multiple repeated modular structures joined by two anti-parallel S-S bonds at the C-terminus. Each monomer of the Fn dimer has a molecular weight of 230–270 kDa, and it is composed of three different types of modules: twelve Fn type I (FnI), two Fn type II (FnII) and fifteen to seventeen Fn type III (FnIII) ^{[11][12][13][14][15]}[11,12,13,14,15] (Figure 1A). FnI, FnII, and FnIII domains are made up of 40, 60, and 90 amino acids, respectively. Twenty different Fn proteins are observed in humans even though Fn is encoded by a single 75 kb gene [16]. This protein diversity is obtained by alternative splicing of two type FnIII exons, called extra domains A and B (EIIIA and EIIIB) and by a segment connecting two other FnIII repeats, FnIII14 and FnIII15, called the type III connecting segment (IIICS) or V domain, that can be assembled in four different ways or fully omitted. The multimodular structure and intermodular regions permit flexibility of the Fn molecule, which is involved in regulating its functions ^[13][15][16][13,15,16]. The different specific domains of Fn can interact with multiple binding partners, for example cell-surface receptors or other ECM components such as heparin, collagen, and proteoglycans. Fn is structured into four functional domains including a N-terminal domain (FnI1-9 plus FnII1-2), which binds Fn to heparin, collagen or fibrin; a central binding domain (FnII1-12) is responsible in part for Fn binding with the cells through the interaction with different integrins; and a C-terminal binding domain (FnIII12-14 plus IIICS plus FnI10-12) that also has other cell-binding sites, heparin- or fibrin-binding sites and the disulfide bridges responsible for Fn dimerization. For all of these reasons, Fn is considered to be a key molecule in the control of cellular regulatory processes and an essential scaffolding protein to preserve and maintain tissue organization [15].

There are two forms of Fn based on its solubility: plasma Fn, the soluble form, and cellular Fn, the water-insoluble form. The latter appears on cell surfaces and in the ECM of various tissues such as the synovial membrane and cartilage, but also in the synovial fluid. Moreover, plasma Fn has been reported to be integrated into the ECM with cellular Fn. Plasma Fn and cellular Fn have different structures and forms of assembly in three-dimensional networks (Figure 1B). Whereas plasma Fn lacks the alternatively spliced EIIIA and EIIIB regions, cellular Fn has different parts of these domains. In addition, only one subunit of plasma Fn possesses an IIICS segment. Taking these data into consideration, the number of different cellular Fn isoforms that can be generated is much higher than plasma Fn, due to the presence of alternative variants in the EIIIA, EIIIB, and IIICS segments ^[15][16][15,16].

Plasma Fn is synthesized by hepatocytes and directly released into circulation in a soluble, compact and inactive form. Blood plasma Fn levels increase after inflammation, major trauma, or pathologies such as atherosclerosis, ischaemic heart disease, and stroke ^{[18][19][20][21]}[18,19,20,21]. Many cell types produce cellular Fn, including fibroblasts, endothelial cells, myocytes, chondrocytes, and synovial cells [15]. The different isoforms of cellular Fn are tissue-dependent, temporally regulated, and cell-type-specific. These isoforms regulate the properties of the ECM and consequently modulate different cellular processes. Under certain pathological conditions, some of the isoforms are exceptionally synthesized by cells or undergo a considerable increase in their synthesis. For instance, the isoform EIIIA is increased in synovial fluid of RA joints and correlates with the progression of joint destruction [22]. Another example is the presence of different synovial IIICS (+) isoforms, which vary their expression levels according to the degree of cartilage degeneration in OA [23]. An additional important fact to take into account is that the presence or absence of the different EIIIA, EIIIB, or IIISC regions affects the orientation and flexibility of the rest of the FnIII modules, altering the three-dimensional structure of the Fn and therefore its interactions during matrix assembly that can modulate Fn-cell signaling ^[15][24][15,24].

The functional form of Fn in vivo is its fibrillary state; thus, Fn molecules must be assembled into supermolecular fibers that form an interconnected network [14]. The presence of cells is necessary for Fn-matrix assembly, which occurs in several stages. First connections between Fn and cells are established through surface receptors such as integrins. Then, Fn is unfolded resulting in the exposure of Fn binding sites that allow Fn-Fn intermolecular interactions. Finally, the Fn-matrix assembly is formed. Both the expression of Fn and its assembly is regulated by a multitude of molecules in a cell-specific manner [15]. The Fn matrix assembly occurs at times of dynamic tissue remodeling, formation or repair and is essential during embryonic development [14].

High levels of Fn have been described in the superficial layer of OA articular cartilage. Moreover, elevated levels of Fn accumulate in the inflamed synovial tissue and in the articular cartilage of RA joints. Enhanced levels of Fn are also observed in the synovial fluid, being 55% higher than levels in normal plasma ^{[9][25][26][27]}[9,173,174,175]. In the course of these rheumatic diseases, in addition to the enhanced amounts of Fn in the joints, an exacerbated activation of Fn-degrading proteases has been described. In fact, increased levels of Fn-fs have been reported in the synovial fluid of OA and RA patients ^{[9][17][28][29]}[9,17,176,177]. These fragments are included within the term matrikines, peptides originated from the fragmentation of ECM proteins that play an important role in both health and disease ^[30][178]. The activity of proteases is crucial during infection and inflammation, being responsible for the generation of different Fn-fs.

Fn-fs have properties not present in native Fn, and are the main candidates for the maintenance of cartilage destruction and synovial tissue inflammation in OA. Thus, 29-, 45-, 120-, and 200 kDa Fn-fs derived from Fn have been found in OA cartilage and synovial fluid, where they stimulate the production of various inflammatory cytokines, such as TNFα and IL-1β ^[31][114]. In addition, increased release of 29 kDa N-terminal heparin-binding, 50 kDa N-terminal gelatin-binding, and 110-140 kDa C-terminal heparin-binding Fn-fs have been detected in bovine injured cartilage explants compared to controls ^[32][179]. More recently, it has been reported that 29 kDa N-terminal heparin-binding Fn-fs inhibited autophagy through modifying localization of HMGB1 in human articular chondrocytes ^[33][180].

Along with OA, another musculoskeletal pathology with a huge worldwide prevalence is low back pain, which is generated by the degenerative disc disease, among other causes ^[34][181]. The degenerative progression in intervertebral discs is generally identified as intervertebral disc degeneration (IVD) ^[35][182]. IVD and OA share characteristics and properties such as cell physiology and ECM of the nucleus pulposus and articular cartilage. It has been suggested that, in acute stages, IVD is accompanied by loss of joint space, subchondral sclerosis, and osteophytes, comparable to OA in the articular joint. As occurs in OA, in IVD local inflammation is the result of mechanical overloading or low grade systemic inflammation, being characterized by increased levels of inflammatory cytokines. Moreover, this inflammatory cascade produces the degradation of ECM with the involvement of different proteinases including MMP-1–3, -7–10, and -12–14, as well as ADAMTS4 and -5. As a result, in both diseases, a chronic inflammation loop is established that degenerates both intervertebral disc and articular joint ^[36][183]. An additional element of similarity is the role that Fn plays in both pathologies. In this sense, an increased expression of Fn and Fn-fs has been reported in human spontaneously degenerative discs ^[37][184]. Moreover, the high levels of Fn-fs have been related with disc degeneration mediated by upregulation of MMP-3 and MMP-13 ^[38][29]. Ruel et al. have described that adult IVD tissues contains many Fn-fs that were absent from the infant disc tissue (6). The highest amount of FN-fs was present in the moderately degenerative discs, as well as in the initial phases of disc degeneration, before observable changes to the disc tissue arise, thus suggesting that Fn-fs play a significant role in the initiation and evolution of disc degeneration ^[39][185].

Baker et al. proposed a mathematical model showing that positive and negative feedback monitoring the degree of cytokine production can regulate either the pre-disposition to OA or the initiation of OA. Moreover, they proposed that manipulation of cytokines, proteinases and Fn-fs levels could be used to treat OA and other related pathologies, suggesting that multiple treatment targets may be essential to break or slow disease progression ^[40][186].

In the next section, we describe the multi-target effect of different Fn-fs in the modulation of ECM-degrading enzymes in OA.