Metabolomics and metabolic function analysis of the secretome of articular cartilage and isolated chondrocytes in response to pro-inflammatory cytokines

Abstract

Purpose: Chondrocytes rely primarily on glycolysis to meet their energy requirements, but possess the metabolic flexibility to support cell survival and matrix synthesis during periods of nutrient stress, by enhancing glycolysis with mitochondrial respiration. Accessing this ‘spare respiratory capacity’ requires optimal mitochondrial function, but since impaired mitochondrial function is implicated in osteoarthritis (OA), this mechanism may be deficient or attenuated in joint disease. Metabolic adaptation is evident in early-stage OA, however cartilage from late-stage disease does not seem to have this flexibility. A deeper understanding of these complex metabolic pathways may identify new metabolic markers of disease stage, and support therapeutic strategies for treating OA. Metabolomics has the potential to identify underlying metabolic changes, reveal pathological pathways, novel biomarkers and therapeutic targets. The aim of this study was to identify metabolic processes involved in early stage disease by analysis of metabolites and metabolic function in two inflammatory models of cartilage degradation. Methods: Macroscopically normal articular cartilage was obtained from equine and bovine metacarpophalangeal joints. Six-millimetre diameter equine cartilage explants (n=6) and isolated primary equine chondrocytes (n=4), seeded at high density (105,000/cm2), were cultured for 7 days in serum-free low or high glucose Glutamax DMEM (Gibco), respectively, with or without 10ng/ml equine interleukin-1β (IL-1β) and 10ng/ml equine tumour necrosis factor-α (TNF-α). Spent media (secretome) was collected and subjected to metabolomic analysis. Secretome metabolite levels were measured using the AbsoluteIDQ® p180 targeted metabolomics kit (Biocrates), and a Waters Xevo TQ-S mass spectrometer coupled to an Acquity UPLC system. Multivariate analysis was performed by principal component analysis and orthogonalized partial least squares discriminant analysis using SIMCA-P v12.0 software (Umetrics). Metabolic function of primary equine (n=9) and bovine chondrocytes (n=3) was determined using Seahorse XFp and XFe24 analyzers (Agilent). Cells were seeded at high density, treated with species-specific 10ng/ml IL-1β and/or 10ng/ml TNF-α for 18 h, and metabolically challenged with the Mito Stress Test during analysis. Metabolite levels, and oxygen consumption rates, were normalised to total cell protein, and differences between means determined by one-way analysis of variance with Tukey’s multiple comparison post-tests. Results: Secretome metabolites which decreased with pro-inflammatory cytokine treatment were proline, ornithine and alpha-aminoadipic acid (p<0.0001). Citrulline increased with cytokine treatment (p<0.0001) and glutamate, present in DMEM, was elevated with cytokine treatment (p<0.0001). Metabolomic analysis of the chondrocyte secretome showed that glutamine decreased (p<0.02) with cytokine treatment whereas citrulline was elevated (p<0.003). Metabolic analysis showed that combined cytokine treatment reduced basal respiration, ATP production and negated spare respiratory capacity in chondrocytes (p<0.01), and the effect was due to IL-1β alone. Conclusions: Cytokines decreased proline, ornithine, alpha-aminoadipic acid and glutamine, which are all downstream metabolites of glutamate. Along with elevated glutamate, this suggests that cytokine treatment inhibits glutamate uptake and metabolism. A marked increase in citrulline observed in both cell and explant models may be attributed to cytokine induction of nitric oxide synthase, resulting in elevated citrulline production via the urea cycle. IL-1β alone prevented chondrocytes from accessing their spare respiratory capacity, and remained glycolytic. In conclusion, glutamate and citrulline metabolism, which are both tightly regulated by mitochondrial metabolism, are altered by cytokines, and IL-1β alone is responsible for the metabolic switch observed in this model. These metabolic pathways could provide markers of early-stage disease, and targets for therapeutic intervention.