Increased synovial lipodystrophy induced by high fat diet aggravates synovitis in experimental osteoarthritis

Osteoarthritis (OA) is the most common joint disorder worldwide, characterized by joint pain, impaired mobility and structural changes in the joints. Although cartilage destruction is the main feature of the disease, every joint structure such as the synovium, bone, meniscus or muscle is affected, leading to the recognition of OA as a whole-organ disease [1]. Synovial inflammation is present in a substantial population of patients with OA and has been associated with different signs and symptoms of the disease, including increased pain and joint effusion, which could promote more rapid cartilage degeneration [2, 3]. Elevated thickness of the lining layer and greater presence and activation of synovial macrophages have been identified in cartilage degradation and osteophyte formation in both human and experimental OA [4‐6].

OA is not merely a local disease, but there are different systemic processes that determine its progression. The concept of metabolic OA, coined in recent years, identifies a syndrome whereby the contribution of metabolic dysregulation and low-grade systemic inflammation to the progression of the disease has been firmly pointed out [7, 8]. The metabolic syndrome (MetS) comprises a cluster of conditions, including glucose intolerance, high blood pressure, hypercholesterolemia and hypertriglyceridemia, and obesity [9, 10]. The accumulation of the different components of MetS has been related to both the occurrence and progression of knee OA [11, 12]. However, little is known about the specific contribution of each of these metabolic alterations in OA progression, and specifically in the synovial damage associated with OA.

The contribution of obesity to OA progression is probably the most extensively studied association [7, 13]. In fact, obesity has been pointed out as the main contributing factor for the association between OA and MetS, in studies showing a markedly attenuated association after adjustment for body mass index [14]. However, OA is also common in non-weight bearing joints of obese persons, suggesting a systemic mechanism rather than a simply mechanic phenomenon [7].

The possible role of hyperlipidemia in mediating obesity-related effects on OA has been explored in different studies. Contradictory results have been published on the relationship between serum lipids and OA incidence in humans, probably due to the presence of obesity and being overweight as confounding factors [15]. In turn, different experimental studies have suggested that hypercholesterolemia could be mainly associated with osteophyte generation rather than to aggravation of cartilage lesions [16, 17]. Macrophages, endothelial cells and fibroblasts are dominant cells within the synovium, together with abundant adipose tissue that constitute the synovial stroma, and every component is sensitive to changes in lipid levels [17, 18].

Adipokines have been considered at least partially responsible for the link between systemic metabolic alterations and OA [19‐21]. Adipokines are essentially released by adipocytes and exhibit pleiotropic functions both in central and peripheral systems, including blood pressure control, hemostasis, food intake, energy expenditure, cell metabolism and inflammation, among others [19‐21]. They are also synthesized by joint cells during OA, mainly by the synovium, cartilage and intra-articular fat tissue, and have been demonstrated to play proinflammatory and catabolic or anabolic roles in OA pathophysiology. It has been hypothesized that the altered circulating patterns of adipokines induced by obesity could be responsible for the deleterious effect of this disease on OA. However, it is not known whether expression and release of adipokines in the joint could be modulated by metabolic factors during OA, thus contributing to disease progression.

Therefore, this work aimed to study the effect of hypercholesterolemia, without any other component of the MetS, on synovial inflammation in an experimental model of knee OA. We have also determined the synovial expression and systemic concentration of adiponectin and leptin, two adipokines involved in joint deterioration associated with metabolic OA.

In this study, we have shown that HFD aggravated OA synovitis, by inducing severe tissue architecture disorganization of the synovium, along with remarkable intensification of the proinflammatory cytokines IL-1β, IL-6 and TNF, and extensive infiltration of macrophages. However, HFD did not have any effect on the aggravation of the pathologic change in cartilage associated with OA. A relevant histological synovial alteration was the significant loss of synovial adipose tissue content, in correlation with decreased leptin and adiponectin gene expression.

In order to isolate the effect of hyperlipidemia, we employed an experimental model of HFD intake that was not associated with weight gain [28]. The lack of significant weight gain in the HFD group has been previously reported and attributed to the animal self-regulation of caloric intake [36]. In fact, animals in both the HFD and OA-HFD groups had a significant decrease in weight gain, which was probably related to the increase in systemic inflammation induced by the diet [28]. Different studies using lipid-rich diets have not been able to adequately apportion the contribution of added mechanical load and hyperlipidemia in OA, a factor that was avoided in our experiments.

Different patterns of synoviopathy have been described in patients with OA, both in late and early disease, such as those with an increased fibrotic component or those essentially characterized by augmented inflammatory parameters [37]. In our rabbits, OA synovitis was associated with a significant increment of fibrotic tissue and partial loss of adipose tissue, and scarce presence of macrophages. A HFD induced both qualitative and quantitative changes in the SM in the rabbits with OA. However, HFD did not significantly aggravate cartilage damage in either the HFD group or the OA-HFD group. These data are in line with previously published results [38, 39], and suggest that the aggravation in synovial inflammation induced by HFD is not a secondary event induced by more severe pathological change in the cartilage.

The higher grade of synovial inflammation in the OA-HFD group was characterized by the remodeling of adipose tissue and by adipocyte loss. The remaining adipocytes had heterogeneous morphology and were significantly smaller in comparison to the OA and control groups, confirmed by the decreased presence of PLIN. HFD further increased synovial macrophages, most of them with the appearance of foam cells, whereas the fibrotic component was similar to that observed in the OA group. To our knowledge, this is the first report of correlation between synovial inflammation and loss of adipose tissue in this localization. Contradictory reports have been published on the contribution of the volume or area of the intra-articular fat tissue to joint deterioration and OA symptoms [40‐42]. However, the alterations in adipocyte size, morphology, loss of adipose tissue with increased fibrosis and inflammatory content have been well-described in inflamed adipose tissue in other anatomic localizations, and described as lipodystrophy [43, 44]. The study of the synovial fat pad as an independent adipose intra-articular tissue has contributed to its identification as a crucial player in OA progression [45], although lack of recognition of the synovium as a whole, integrated, functional and structural unit hampers the understanding of the mechanisms involved in the synovial alterations in OA. Histologically, synovial lining, adipose sub-lining and synovial fat pad represent a continuum. Sub-lining adipose tissue and the fat pad seem to share a common inflammatory state, both in cell content and cell phenotype, induced by the disease process more than by tissue-specific signals [24, 46].

The mechanisms by which adipose tissue can be replaced by fibrotic tissue in the OA synovium have not been fully elucidated. However, the increase in the hypoxia-associated mediators, induced by biomechanical alterations and proinflammatory cytokines, could be at least partially responsible for this phenomenon. An increase in hypoxia-induced factor-1 (HIF-1)α has been described in the OA synovium, in correlation with greater joint destruction [47, 48]. In adipose tissue, with a similar structure and cellular component to that observed in the stroma of the SM, HIF-1α induces tissue fibrosis and inhibits pre-adipocyte differentiation [49]. Furthermore, in inflamed adipose tissue from mice fed a HFD, HIF-1α-stimulated macrophages form highly hypoxic structures called crown-like structures (CLS), comprising macrophages encircling dead or dying adipocytes [50]. Indeed, we have previously identified CLS in the OA synovium in both human and hypercholesterolemic rabbits with synovial inflammation [5, 28].

Hyperlipidemia in rabbits in the OA group did not seem to enhance the presence of fibrosis-associated proteins, such as col I. However, it evoked a dramatic increase in macrophage infiltration in the synovium and greater decrease in adipose tissue content. Dyslipemia has been directly related to macrophage infiltration and inflammation in the synovium and adipose tissue [51]. In hyperlipidemic mice with OA, the synovial proinflammatory macrophage subset was identified as responsible for an increase in TNF synthesis and extracellular matrix remodeling in the synovial membrane [51]. In line with these data, we identified greater TNF expression in the synovium in the OA-HFD group that paralleled the increased macrophage density in this tissue. Although little is known about the metabolic regulation of synovial macrophages, prolonged lipid exposure could result in failure of the lipid-handling mechanisms, leading to different lipotoxic events, such as those described in obesity-associated insulin resistance, atherosclerosis and other inflammatory diseases related to MetS [52]. Thus, hyperlipidemia could drive M1 macrophage polarization in the OA synovium, resulting in a major presence of proinflammatory cytokines, as has been described in adipose tissue [52, 53]. Furthermore, adipocyte apoptosis and impaired adipogenesis have been also associated with the increased lipolysis induced by over-nutrition or HFD feeding [52]. Although hyperlipidemia could aggravate OA synovial inflammation, increasing macrophage density and adipose tissue destruction, the presence of hyperlipidemia per se could only have limited effects on SM alterations, as recently reported in HFD-fed mice [54].

Leptin and adiponectin gene expression diminished in the SM in the OA and OA-HFD group in comparison to control animals. These results appear to correlate with the amount of intra-articular adipose tissue rather than with the presence of a proinflammatory milieu. Furthermore, circulating leptin was significantly increased in HDF-fed animals, probably due to the effect of the diet on the extra-articular fat tissue [55]. Our data are in line with previous reports indicating that hyperlipidemia could be an aggravating factor for OA through the stimulation of systemic proinflammatory mediators [56]. In the OA group we also found increased circulating leptin as previously described in human and experimental OA, related to joint damage [19]. Therefore, our data do not support the hypothesis that hyperlipidemia could be an aggravating factor in metabolic OA, stimulating adipokine expression within the intra-articular adipose tissue. Different joint cells, such as chondrocytes or bone cells, could be responsible for adipokine synthesis in response to biomechanical or proinflammatory stimuli [19, 20].

In summary, these data show that HFD aggravates the inflammation in the SM of rabbits with OA by inducing an increase in the infiltrating macrophages in the synovium, together with macrophage and metabolic-mediated remodeling of adipose tissue, and further elevation of proinflammatory cytokines. The lipotoxic effects induced by dyslipemia in adipocytes and macrophages could play a decisive role in the joint deterioration of patients with OA and MetS, supporting the hypothesis of a plausible link between obesity and OA going beyond mechanical loading.