The Burden of Metabolic Syndrome on Osteoarthritic Joints

Bruce M. Dickson; Anke J. Roelofs; Justin J. Rochford; Heather M. Wilson; Cosimo De Bari


Arthritis Res Ther. 2019;21(289) 

In This Article

The Effect of MetS on Macrophage Polarisation

Macrophages are present in metabolic tissues such as fat, liver, and muscle, and their proliferation, plasticity, and polarisation are driven by obesity, with a switch from the M2 to M1 phenotype being observed.[28] Preclinical studies have demonstrated a skewing of macrophages towards the M1 phenotype within synovial and adipose tissues in diet-induced OA.[16] There are several molecular mechanisms through which MetS could promote a pro-inflammatory M1 macrophage phenotype in OA, including metabolic perturbations at the cellular level and changes in the systemic factors such as adipokine levels.

Metabolic Programming of Macrophage Polarisation

Metabolic perturbations, including changes in the levels of oxygen, nutrients, and extracellular metabolites, are perceived by immune cells including macrophages through the activity and levels of the nutrient sensors 5' adenosine monophosphate-activated protein kinase (AMPK), and mammalian target of rapamycin complex 1 (mTORC1). The activity of AMPK plays a key role in metabolic reprogramming in response to nutrient deprivation (Figure 2), via its ability to sense falling intracellular glucose and ATP levels. AMPK activity subsequently increases ATP production whilst reducing anabolic processes to restore cellular energy homeostasis.[29] AMPK activity is reduced by several aspects of MetS including insulin resistance, hyperglycaemia, and elevated circulating pro-inflammatory mediators. A reduction in AMPK activity in macrophages increases aerobic glycolysis by stabilising hypoxia-inducible factor-1α (HIF-1α) via the Warburg effect. Increased glycolysis in macrophages is associated with a pro-inflammatory phenotype as it produces more glucose-6-phosphate (G6P), the main substrate of the pentose phosphate pathway (PPP), allowing the production of NADPH that is used to generate reactive oxygen species (ROS),[30] implicated in immune cell activation and in the damage of chondrocytes. Indeed, G6P-dehydrogenase (G6PD), the first enzyme within the PPP, has been shown to be upregulated in macrophages derived from obese patients and, along with NADPH, to be essential for the activation of NF-κB and ROS formation.[31]

Figure 2.

Metabolic polarisation of macrophages. Circulating monocytes are recruited into the synovium whereby they differentiate into non-activated macrophages. Hyperglycaemia, insulin resistance, and pro-inflammatory cytokines inhibit AMPK activity resulting in HIF-1α stabilisation and increases in aerobic glycolysis. Increases in glycolysis are accompanied by increased PPP activity, and both are involved in M1 macrophage polarisation. Succinate stabilises HIF-1α. Citrate promotes aerobic glycolysis and inflammatory cytokine expression. Obesity and nutrient excess hyperactivate mTORC1 resulting in Akt inhibition and defective M2 polarisation. M2 polarisation is promoted by AMPK activity. AMPK is stimulated by nutrient deprivation, metformin, and adiponectin. Resolvin D1 promotes the re-polarisation of macrophages to the M1 phenotype. AMPK, 5' adenosine monophosphate-activated protein kinase; HIF-1α, hypoxia-inducible factor alpha; PPP, pentose phosphate pathway; mTORC1, mammalian target of rapamycin complex 1; TNF-α, tumour necrosis factor alpha; MMP, matrix metalloproteinase; ROS, reactive oxygen species; IL, interleukin; TGF-β, transforming growth factor beta; VEGF, vascular endothelial growth factor. (A) CD11c, (B) CD14, (C) CD86, and (D) CD206

The nutrient sensor mTORC1 integrates signals from multiple sources, including cellular energy levels, oxygen status, growth factors, and amino acid availability, and is responsible for anabolic processes including protein, lipid, and nucleotide synthesis. Obesity and nutrient excess are known to induce the hyperactivation of mTORC1, which leads to defective M2 polarisation of macrophages via feedback inhibition of the serine-threonine kinase Akt.[32] Akt is responsible for upregulating many of the genes essential in M2 polarisation such as Arg1, Fizz1, and Ym1 whilst at the same time promoting the inhibition of M1 polarisation through downregulating transcription factor FOXO1, essential for PRR, Toll-like receptor 4 (TLR4) production and upregulating IL-1 receptor-associated kinase M (IRAK-M), a TLR4 signalling inhibitor.[32] Similar effects were observed in a murine OA model. Myeloid lineage-specific deletion of tuberous sclerosis complex 1 (TSC1) led to hyperactivation of mTORC1 and was associated with M1 polarisation of synovial macrophages with resultant increases in IL-1, IL-6, and TNF.[33] This skewing to the M1 phenotype was accompanied by worsening of OA. Furthermore, in Rheb1 deletion mice where mTORC1 is constitutively inactive in the myeloid lineage, it led to M2 macrophage polarisation within the synovium accompanied by improvements in OA histological severity. A recent study in rheumatoid arthritis further highlights the detrimental effects that altered AMPK and mTORC1 activity can have on synovial inflammation via effects on T cells. T cells from RA patients were shown to have deficient N-myristoylation, a lipid modification of proteins that changes their physical properties and their subcellular distribution.[34] Defective N-myristoylation of AMPK prevented its activation and instead led to exuberant mTORC1 signalling, stimulating differentiation into pro-inflammatory TH1 and TH17 T cells and promoting inflammation in a humanised mouse model of synovitis.[34] Whether metabolic reprogramming affects T cells in OA remains to be determined.

MetS can also impact on crucial metabolites involved in macrophage polarisation and activity. One of these metabolic intermediates is succinate. It increases not only due to the Krebs cycle stalling in M1 macrophages but also in response to hyperglycaemia and obesity. Succinate has been shown to compete with prolyl hydroxylase resulting in the stabilisation of HIF-1α within macrophages with subsequent sustained production of IL-1β through directly binding to the Il1b promoter.[35] The stalled Krebs cycle causes the accumulation of a further intermediate, citrate, within the mitochondria that is crucial to M1 effector function. Citrate is exported out of the mitochondria and is further metabolised to acetyl-CoA, vital in the acetylation of histones regulating not only the transcription of glycolytic enzymes, needed to increase energy production in the M1 macrophage, but also of inflammatory cytokines such as IL-6.[36]

Macrophage Polarisation Induced by AGEs and FFAs

In addition to affecting key nutrient sensors and metabolic intermediates that polarise macrophages, the MetS can influence macrophage function via advanced glycation end-products (AGEs) and free fatty acids (FFAs) that act directly on macrophages. Chronic hyperglycaemia non-enzymatically glycates proteins and lipids and thus produces advanced glycation end-products (AGEs). AGEs are recognised by receptors for AGEs (RAGEs) expressed upon macrophages and their activation results in M1 polarisation and increased transcription of TNF and IL-1β via NF-κB.[37] A similar effect occurs due to FFAs. Prolonged periods of overnutrition initially lead to healthy adipose expansion, but when this capacity becomes exceeded, adipocytes are no longer able to safely store lipids and protect other tissue from their deleterious effects as excess lipids remain acellular in the form of FFAs. FFAs bind to TLR4, resulting in M1 macrophage activation and pro-inflammatory cytokine production.[38]

The Influence of Adipokines on Macrophage Polarisation

Leptin, the first adipokine discovered, plays a critical role in controlling food intake through central mechanisms. In addition, it is now considered to have an inflammatory role. Leptin activates the JAK2-STAT3 and PI3K-AKT-mTOR pathways in macrophages to promote a pro-inflammatory phenotype with the secretion of TNFα and IL-1β.[39] Leptin concentrations in the synovial fluid of OA patients correlate with BMI.[40] In addition to adipose tissue, leptin is produced locally within the joint by the cartilage, IFP, and synoviocytes,[40] and leptin levels are significantly higher in the synovial fluid than in the serum of OA patients.[41] Expression in cartilage is upregulated in OA[40] and correlates with BMI of the patient,[41] suggesting an important role for locally increased leptin production by joint tissues. In support of the clinical relevance of leptin in OA development, leptin serum levels 10 years prior to MRI assessment were associated with cartilage defects, bone marrow lesions, osteophytes, meniscal abnormalities, synovitis, and joint effusion in a population of middle-aged women.[42] These findings provide a strong indication for a role of leptin in the pathophysiology of OA.

Adiponectin, another adipokine produced by adipose tissue, has also been shown to influence macrophage polarisation state. Macrophages activated by M2 stimulants, IL-4 and IL-13, were shown to have increased AMPK activity and fatty acid oxidation when exposed to adiponectin. This resulted in increased levels of IL-10—a hallmark of M2 macrophage effector function. However, adiponectin also appeared to promote TNF, IL-6, and IL-12 production when macrophages were exposed to M1 polarising conditions.[43] In contrast, in a series of in vitro experiments, adiponectin was shown to promote re-polarisation of M1 macrophages towards an M2 phenotype, indicating a possible role in the resolution of inflammation.[44] Accordingly, a longitudinal study reported that OA progressed more slowly in patients with higher levels of adiponectin within their synovial fluid. Interestingly, adiponectin levels were inversely proportional to patients BMI.[45] This inverse relationship between adiponectin levels and BMI may be explained by adiponectin production being sensitive to both oxidative stress and fibrosis occurring in unhealthy adipose tissue expansion associated with obesity.[46] Thus, obesity and MetS downregulate one of the adipokines that may confer protection against OA via its effects on the innate immune system. However, another study showed that plasma adiponectin levels and adiponectin production by OA cartilage positively correlated with OA severity in a cohort of 35 patients undergoing total knee replacement surgery.[47] The role of adiponectin in OA pathophysiology thus remains to be clarified.