Background: In this study, we examined the effect of oxidative stress on cellular energy metabolism and pro- angiogenic/pro-inflammatory mechanisms of primary rheumatoid arthritis synovial fibroblast cells (RASFC) and human umbilical vein endothelial cells (HUVEC). Methods: Primary RASFC and HUVEC were cultured with the oxidative stress inducer 4-hydroxy-2-nonenal (4-HNE), and extracellular acidification rate, oxygen consumption rate, mitochondrial function and pro-angiogenic/pro- inflammatory mechanisms were assessed using the Seahorse analyser, complex I–V activity assays, random mutation mitochondrial capture assays, enzyme-linked immunosorbent assays and functional assays, including angiogenic tube formation, migration and invasion. Expression of angiogenic growth factors in synovial tissue (ST) was assessed by IHC in patients with rheumatoid arthritis (RA) undergoing arthroscopy before and after administration of tumour necrosis factor inhibitors (TNFi). Results: In RASFC and HUVEC, 4-HNE-induced oxidative stress reprogrammed energy metabolism by inhibiting mitochondrial basal, maximal and adenosine triphosphate-linked respiration and reserve capacity, coupled with the reduced enzymatic activity of oxidative phosphorylation complexes III and IV. In contrast, 4-HNE elevated basal glycolysis, glycolytic capacity and glycolytic reserve, paralleled by an increase in mitochondrial DNA mutations and reactive oxygen species. 4-HNE activated pro-angiogenic responses of RASFC, which subsequently altered HUVEC invasion and migration, angiogenic tube formation and the release of pro-angiogenic mediators. In vivo markers of angiogenesis (vascular endothelial growth factor, angiopoietin 2 [Ang2], tyrosine kinase receptor [Tie2]) were significantly associated with oxidative damage and oxygen metabolism in the inflamed synovium. Significant reduction in ST vascularity and Ang2/Tie2 expression was demonstrated in patients with RA before and after administration of TNFi. Conclusions: Oxidative stress promotes metabolism in favour of glycolysis, an effect that may contribute to acceleration of inflammatory mechanisms and subsequent dysfunctional angiogenesis in RA. Keywords: Bioenergetic metabolism, Oxidative stress, Angiogenesis, Rheumatoid arthritis * Correspondence: firstname.lastname@example.org Equal contributors Centre for Arthritis and Rheumatic Diseases, Dublin Academic Medical Centre, St. Vincent’s University Hospital, Dublin, Ireland Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 2 of 15 Background an array of primary lipid peroxidation products, which sub- Angiogenesis is one of the earliest events in the de- sequently decompose and form reactive lipid electrophiles, velopment of rheumatoid arthritis (RA). New blood among which 4-hydroxy-2-nonenal (4-HNE) is the most vessels invade the synovial membrane, resulting in a important signalling molecule . 4-HNE can form cova- self-perpetuating and persistent infiltration of immune lent adducts with DNA, phospholipids and nucleophilic cells into the joint, transforming the synovial tissue amino acids, impairing their structure and biological prop- (ST) into an aggressive, tumour-like ‘pannus’ . New erties. In particular, mitochondria have been reported as a capillaries also facilitate the delivery of sufficient oxy- prominent target of 4-HNE activity . Mitochondrial gen and nutrients to support the proliferating syno- proteins related to mitochondrial energy metabolism, such vium. Although angiogenesis is a prominent feature as adenosine triphosphate synthase subunit β (ATP5B), suc- of RA, the neovascular network is dysfunctional and cinate dehydrogenase flavoprotein subunit and reduced fails to restore tissue oxygen homeostasis, rendering form of nicotinamide adenine dinucleotide (NADH) de- the inflamed ST hypoxic. The increase in metabolic hydrogenase iron–sulphur protein 2 in the electron trans- turnover of the expanding synovial pannus outpaces port chain (ETC), and trifunctional enzyme subunit α in the oxygen supply, resulting in a demand for adeno- the TCA cycle, are highly susceptible to 4-HNE-induced in- sine triphosphate (ATP) and an altered regulation of activation [18–20]. A recent study has also demonstrated cellular metabolic mechanisms [2, 3]. 4-HNE-induced inhibition of sirtuin 3, a major mitochon- Bioenergetics is fundamentally important for all cells drial nicotinamide adenine dinucleotide (NAD )-dependent to enable proliferation, differentiation and maturation, deacetylase, with subsequent up-regulation of vascular with mitochondria being central to biosynthetic and endothelial growth factor (VEGF) expression by breast can- bioenergetic pathways mediated by the tricarboxylic acid cer cells , indicating a close connection between oxida- (TCA) cycle. Thus, alterations to mitochondrial respir- tive stress, mitochondrial function and angiogenesis. ation can play a key role in mediating pathogenic mech- In previous studies, our group assessed levels of syn- anisms in chronic inflammatory diseases [4–6]. One ovial lipid peroxidation in patients with RA and demon- well-known example of mitochondrial dysfunction is the strated a significant inverse correlation between 4-HNE bioenergetic switch in cell metabolism from oxidative expression and oxygen tension in the inflamed joint, phosphorylation (OXPHOS) towards aerobic glycolysis, reflecting mitochondrial damage . Subsequently, we known as the Warburg effect. Although the efficiency of have demonstrated that high synovial lipid peroxidation ATP production per molecule of glucose is much lower positively correlated with clinical disease activity scores, through glycolysis, the yield rate is much faster than that and we have reported reduced 4-HNE levels in patients of OXPHOS, supporting rapid cellular growth. It has with RA who responded to tumour necrosis factor been demonstrated that the Warburg effect is present in (TNF) blocking therapy corresponding with a significant highly proliferating and metabolically active immune increase in partial oxygen pressure in synovial tissue, in- cells in a manner similar to that observed in tumour dicating a reduction in synovial oxidative stress as the cells. In the inflamed joint, an increase in the metabolic joint tissue becomes less hypoxic . In addition, it was state towards glycolysis has been shown in primary observed that increased synovial inflammation and rheumatoid arthritis synovial fibroblasts (RASFC), CD4 angiogenesis was associated with higher oxidative stress T cells, T-helper type 17 (T 17) cells, macrophages and . Given the important role of mitochondrial metab- dendritic cells [7–10]. This is paralleled by elevated lac- olism in the regulation of inflammatory and angiogenic tate levels and diminished glucose in RA synovial fluids responses, in this study we investigated the effect of as well as by increased activity of key glycolytic enzymes oxidative stress on the mitochondrial bioenergetic profile in the RA synovium, indicating that anaerobic glycolysis and the pro-angiogenic/pro-inflammatory mechanisms is favoured in this hypoxic environment [11–13]. More in RASFC and human umbilical vein endothelial cells recently, in vitro studies by our group have shown that (HUVEC). Furthermore, we determined the effects of hypoxia and Toll-like receptor 2 (TLR2)-induced inflam- tumour necrosis factor α inhibitors (TNFi) on the mation promoted mitochondrial dysfunction and oxida- expression of angiogenic markers in RA in relation to tive stress and reprogrammed the nature of cellular synovial oxidative stress in vivo. respiration in RA synovial cells [14, 15]. Oxidative damage occurs through the detrimental effect Methods of hypoxia and is recognised as an important source of gen- Patient recruitment, arthroscopy and sample collection omic instability that leads to respiratory alterations. Hyp- Fifteen patients with active RA were recruited from the oxia promotes overproduction of reactive oxygen species Rheumatology Department of St. Vincent’s University (ROS) that provoke oxidation of polyunsaturated fatty acids Hospital, Dublin, Ireland. All patients gave fully in- in plasma and mitochondrial membranes. This generates formed written consent approved by the institutional Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 3 of 15 ethics committee, and the research was performed in deoxyglucose (2-DG; 25 mM) using the Seahorse XF24 accordance with the Declaration of Helsinki. Clinical analyser (Agilent Technologies, Santa Clara, CA, USA). disease activity was assessed with the 28-joint Disease RASFC and HUVEC were seeded at 30,000 cells per well Activity Score (DAS28) using the C-reactive protein in a Seahorse XF96 cell culture microplate (Agilent level. Under local anaesthesia, all patients with RA Technologies) and allowed to adhere for 24 hours. Cells underwent arthroscopy of the inflamed knee joint prior were rinsed with assay medium (unbuffered DMEM sup- to biologic treatment (T0) and a second arthroscopy 3 plemented with 10 mM glucose, 1 mM sodium pyruvate months after commencement of TNFi (T3). ST biopsies and 2 mM L-glutamine, pH 7.4) before incubation with were used for isolation of primary synovial fibroblasts assay medium for 30 minutes at 37 °C in a non-CO in- and histological analyses. cubator. Following incubation, cells were stimulated with 4-HNE (2.5 μM) and vehicle basal medium for 2 hours. RASFC culture Four baseline OCR and ECAR measurements were ob- RASFC biopsies obtained at arthroscopy were digested tained over 28 minutes before injection of specific meta- with 1 mg/ml collagenase type I (Worthington Biochem- bolic inhibitors. Moreover, to challenge the metabolic ical, Lakewood, NJ, USA) in Gibco RPMI 1640 medium capacity of the RASFC and HUVEC, three OCR and (Thermo Fisher Scientific, Paisley, UK) for 4 hours at ECAR measurements were obtained over 15 minutes fol- 37 °C in humidified air with 5% CO . Dissociated cells lowing injection with oligomycin, FCCP, antimycin A were plated in RPMI 1640 medium supplemented with and 2-DG. 10% Gibco FCS (Thermo Fisher Scientific), 20 mM 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (Thermo In vitro mitochondrial dysfunction and mitochondrial Fisher Scientific), penicillin (100 U/ml), streptomycin DNA mutagenesis (100 U/ml) and amphotericin B (Fungizone 0.25 μg/ml; ROS production was assessed using the DCFDA Cellular (Invitrogen, Plymouth, MN, USA). Cells were grown to Reactive Oxygen Species Detection Assay Kit (Abcam, confluence and used between passages 4 and 7. RASFC Cambridge, UK). RASFC were seeded into clear-bottomed, were seeded onto 96-well plates and into T25 flasks and dark-sided 96-well plates at a density of 2.5 × 10 cells/well cultured in the presence of 4-HNE (2.5 μM; Cayman and allowed to attach overnight. Cells were washed in 1× Chemical, Ann Arbor, MI, USA), a highly reactive end buffer and stained with 25 μM2′,7′-dichlorofluorescin product of lipid peroxidation or vehicle basal medium diacetate in 1× buffer for 45 minutes at 37 °C and 5% CO . (0.1% ethanol). The concentration of 4-HNE used in the After staining, cells were washed, treated with 4-HNE and experiments was based on a cell viability assay and pre- incubated at 37 °C in 5% CO . ROS fluorescence signal was viously published studies . Following stimulation, the measured using the SpectraMax Gemini system (Molecular effect of amplified oxidative stress on mitochondrial Devices, Sunnyvale, CA, USA) with excitation and emission function, cellular metabolism and angiogenic responses wavelengths of 485 nm and 538 nm, respectively. Mean was assessed as described below. fluorescence values from four wells for each condition were obtained. To characterise the frequencies of random muta- HUVEC culture tionsinRASFC exposedto4-HNE for24hours,weuseda HUVEC (Lonza, Walkerville, MD, USA) were incubated mitochondrial random mutation capture assay. in MCDB (Thermo Fisher Scientific) supplemented with Mitochondrial DNA (mtDNA) was extracted using a L-glutamine (Thermo Fisher Scientific), 0.5 ml epidermal previously reported protocol . Following extraction, growth factor (Thermo Fisher Scientific), 50 ml FCS 10 μg of mtDNA was digested with 100 U of Taq I (Thermo Fisher Scientific), 0.5 ml of hydrocortisone, peni- restriction enzyme (New England Biolabs, Ipswich, MA, cillin (100 U/ml; Bioscience), streptomycin (100 U/ml; USA), 1× bovine serum albumin, and a Taq I-specific Bioscience) and Fungizone (0.25 μg/ml; Bioscience). Cells digestion buffer (10 mM Tris HCl, 10 mM MgCl 2, were cultured at 37 °C in humidified air with 5% CO and 100 mM NaCl, pH 8.4) for 10 hours, with 100 U of Taq I harvested with trypsin-ethylenediaminetetraacetic acid added to the reaction mixture every hour. PCR (Lonza). Cells were used between passages 20 and 30. amplification was performed in 25-μl reaction mixtures containing 12.5 μl of 2× SYBR Green Brilliant Master Mix Oxygen consumption rate and extracellular acidification (Stratagene, La Jolla, CA, USA), 0.1 μl of uracil DNA gly- rate measured using Seahorse technology cosylase (New England Biolabs), 0.7 μl of forward and re- Oxygen consumption rate (OCR) and extracellular acid- verse primers (10 pM/μl; Integrated DNA Technologies, ification rate (ECAR), reflecting OXPHOS and glycolysis, Skokie, IL, USA), and 6.7 μlofH O. The samples were respectively, were measured before and after treatment amplified using a Roche LightCycler 480 Instrument with oligomycin (2 μg/ml), trifluorocarbonylcyanide phe- (Roche Diagnostics, Indianapolis, IN, USA), according to nylhydrazone (FCCP; 5 μM), antimycin A (2 μM) and 2- the following protocol; 37 °C for 10 minutes, 95 °C for 10 Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 4 of 15 minutes, followed by 45 cycles of 95 °C for 15 seconds progress of the coupled reaction in the presence of 4- and 60 °C for 1 minute. Samples were kept at 72 °C for 7 HNE was monitored as a decrease in absorbance at OD minutes and following melting-curve analysis were imme- 340 nm. Results were calculated using SoftMax Pro 5.3 diately stored at − 80 °C. The primer sequences used were microplate analysis software (Molecular Devices). The as follows: for mtDNA copy number, 5′-ACAGTTTATG activity of complexes I, II, IV and V is proportional to TAGCTTACCTCC-3′ (forward) and 5′-TTGCTGCG the decrease in absorbance, and the linear rate of reduc- TGCTTGATGCTTGT-3′ (reverse); for random muta- tion in absorbance over time was calculated. The activity tions, 5′-CCTCAACAGTTAAATCAACAAAACTGC-3′ of complex III is proportional to the increase in absorb- (forward) and 5′-GCGCTTACTTTGTAGCCTTCA-3′ ance, and the linear rate of increase in absorbance over (reverse). time was calculated. For each complex, results are graphically demonstrated as the percentage of enzymatic Examination of mitochondrial complexes I–V activity activity in the presence of 4-HNE relative to the percent- Mitochondrial complexes I–V OXPHOS activity assay age of basal activity. kits (Abcam) were used to screen the direct effect of 4-HNE on all complexes of the mitochondrial respira- Quantification of pro-angiogenic mediators in RASFC tory chain. These assays are performed using whole To assess the effects of oxidative stress on secretion of bovine heart mitochondria, a rich source of OXPHOS VEGF, angiopoietin 2 (Ang2), platelet-derived growth complexes. The activity of mitochondrial complexes factor subunit B (PDGF-B), basic fibroblast growth fac- I–V was measured as per the manufacturer’sinstruc- tor (bFGF), interleukin (IL)-8, regulated on activation, tions. Briefly, OXPHOS complex I (NADH ubiquin- normal T cell expressed and secreted (RANTES) and one oxidoreductase) catalyses electron transfer from intercellular adhesion molecule (ICAM), RASFC were NADH to the electron carrier, ubiquinone, concomi- seeded into 96-well plates. Confluent RASFC were tantly pumping protons across the inner mitochon- serum-starved for 24 hours and then cultured with 4- drial membrane. The progression of this reaction was HNE for 24 hours. Supernatants were harvested, and monitored following the oxidation as a decrease in protein secretion levels were quantified using MSD as- absorbance at optical density (OD) 340 nm. OXPHOS says (Meso Scale Discovery, Rockville, MD, USA) or spe- complex II (succinate-coenzyme Q reductase) cataly- cific enzyme-linked immunosorbent assays (ELISAs) ses electron transfer from succinate to the electron (R&D Systems, Minneapolis, MN, USA). carrier, ubiquinone. The product, ubiquinol, is used by complex III in the respiratory chain, and fumarate Induction of pro-angiogenic mechanisms of HUVEC in is necessary to maintain the TCA cycle. The produc- response to oxidative stress-activated RASFC tion of ubiquinol in the presence of 4-HNE was mon- To examine if oxidatively activated RASFC could further itored at OD 600 nm. To examine OXPHOS complex affect pro-angiogenic mechanisms of HUVEC, RA fibro- III activity, succinate (electron donor of complex II) blast cells were stimulated with 4-HNE for 24 hours, and oxidised cytochrome c (electron acceptor of com- and conditioned media (CM) were harvested. As a basal plex III) were added to the mitochondria to start the medium, we used fibroblast-conditioned media from electron transfer reaction that takes place during RASFC cultured in the absence of 4-HNE. Next, the cul- OXPHOS. ture of HUVEC was supplemented with 10% fibroblast- The rate of coupled complex II + III reaction was mea- conditioned media. To ensure that the effects on sured by monitoring the conversion of oxidised cyto- HUVEC function were not due to residual 4-HNE in the chrome c into reduced form, observed as an increase in 10% fibroblast-conditioned media, HUVEC were also absorbance at OD 550 nm. OXPHOS complex IV (cyto- cultured with RPMI 1640 medium containing 4-HNE at chrome c oxidase) transfers electrons from reduced the same concentration (0.25 μM), which is the same cytochrome c to molecular oxygen and concomitantly concentration as that in the 10% RASFC CM. Following pumps protons across the inner mitochondrial mem- 24-hour exposure of HUVEC to fibroblast-conditioned brane. The progression of this reaction was monitored media, pro-angiogenic responses of endothelial cells following the oxidation as a decrease in absorbance at were assessed as described in the subsections that OD 550 nm. OXPHOS complex V makes about 95% of a follow. cell’s ATP using energy generated by the proton-motive force and can also function in the reverse direction in HUVEC transwell invasion chambers the absence of a proton-motive force, hydrolysing ATP BD BioCoat Matrigel invasion chambers (BD Biosciences, to generate adenosine diphosphate (ADP) and inorganic Wokingham, UK) were used to examine HUVEC invasion. phosphate. The production of ADP by ATP synthase can Cells were seeded at a density of 2.5 × 10 per well in the be coupled to the oxidation of NADH to NAD , and the migration chamber on 8-μm membranes pre-coated with Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 5 of 15 Matrigel. HUVEC media containing 10% fibroblast- number was measured with a microplate reader at a conditioned media was placed in the lower well of the wavelength of 550 nm. chamber, and cells were allowed to migrate for 48 hours. Non-migrating HUVEC were removed from the upper Quantification of pro-angiogenic mediators in HUVEC surface by gentle scrubbing. Cells that had invaded were HUVEC were seeded into 96-well plates and left over- attached to the lower membrane and fixed with 4% para- night at 37 °C and 5% CO . The following day, cells were formaldehyde (PFA) and stained with 0.1% crystal violet. stimulated with 10% fibroblast-conditioned media for To assess the average number of invading HUVEC, cells 24 hours. Next, supernatants were harvested, and pro- were counted in five random high-power fields. tein secretion levels of Ang2 and PDGF-B were quanti- fied by using a specific ELISA (R&D Systems). HUVEC tube formation Matrigel (50 μl; BD Biosciences, San Jose, CA, USA) was Immunofluorescence staining of RASFC and synovial plated in 96-well culture plates after thawing on ice and tissue allowed to polymerise for 30 minutes at 37 °C in humidi- Single-immunofluorescence staining was performed on fied air with 5% CO . HUVEC were removed from culture, RASFC following 24-hour cell stimulations with 4-HNE. trypsinised and resuspended at a concentration of 4 × 10 To visualise immunoexpression of VEGF, cells were fixed cells/ml in endothelial cell growth medium. Five hundred in 4% PFA and stained with primary rabbit antibody against microliters of cell suspension was added to each chamber VEGF (Abcam). To demonstrate ST co-expression of in the presence of 10% fibroblast-conditioned media and markers of angiogenesis, oxidative stress and bioenergetics, cultured for 8 hours. The tube analysis was determined dual-immunofluorescence staining was performed on cryo- from five sequential fields (magnification × 10) with a stat synovial sections. ST sections were fixed with acetone focus on the surface of the Matrigel by two blinded ob- for 10 minutes and co-incubated with primary mouse anti- servers and a connecting branch between two discrete body against human 4-HNE (GENTAUR, Kampenhout, endothelial cells was counted as 1 tube. Belgium) and with primary rabbit antibodies against VEGF, Ang2, Tie2, ATP5B and glucose transporter 1 (GLUT1) (all HUVEC wound repair assay from Abcam), glyceraldehyde 3-phosphate dehydrogenase HUVEC were seeded onto 24-well plates and grown to (GAPDH) (Trevigen, Gaithersburg, MD, USA) and pyru- confluence. A single scratch wound was induced through vate kinase isozyme 2 (PKM2) (Abgent, San Diego, CA, the middle of each well with a sterile pipette tip. Cells were USA). Following overnight incubation in a humidified subsequently stimulated for 24 hours with 10% fibroblast- chamber, RASFC and ST samples were incubated with conditioned media. HUVEC migration across the wound Invitrogen Alexa Fluor 488-conjugated goat Invitrogen margins from 8 hours was assessed and photographed Superclonal™ anti-mouse secondary antibody (Thermo using a phase-contrast microscope. Semi-quantitative ana- Fisher Scientific) and Cy™3–conjugated goat anti- lysis of cell repopulation of the wound was assessed. Briefly, rabbit secondary antibody (Jackson ImmunoResearch, images of the scratch wound assays were taken at × 10 West Grove, PA, USA) for 60 minutes and counterstained magnification. The mean closure of the wound was manu- with 4′,6-diamidino-2-phenylindole (DAPI) nuclear stain ally calculated from the average of three individual mea- (Sigma-Aldrich) for 10 minutes. Samples were mounted surements from each wound. This process was repeated for with Molecular Probes antifade mounting medium all technical replicates. Measurement of scratches at time 0 (Thermo Fisher Scientific) and assessed by immunofluor- were designated as 100% open. From this, the percentage of escence microscopy (Olympus BX51; Olympus, Hamburg, closure for all scratches was calculated. Germany). HUVEC proliferation IHC and scoring of synovial tissue A crystal violet cell proliferation assay was used to assess IHC was performed using 7-μm cryostat ST sections HUVEC proliferation in the presence of RASFC- and the DAKO ChemMate EnVision kit (Dako/Agi- conditioned media. HUVEC were seeded into 96-well lent Technologies, Glostrup, Denmark). Sections were culture plates at a density of 5000 cells/well and left defrosted at room temperature for 20 minutes, fixed overnight at 37 °C and 5% CO . Next, cells were stimu- in acetone for 10 minutes and washed in PBS for 5 mi- lated with 10% fibroblast-conditioned media for 24 hours. nutes. Non-specific binding was blocked using 1% casein Following cell culture, cells were washed with PBS, fixed in PBS for 20 minutes. The sections were incubated with in 4% PFA and stained with 1% crystal violet solution. rabbit monoclonal primary antibodies against human Plates were washed with tap water and then dried VEGF, Ang2, Tie2, ATP5B (all from Abcam), GAPDH overnight. Cells were resuspended in 1% Triton X-100 (Trevigen) and mouse monoclonal antibodies against hu- solution (Sigma-Aldrich, St. Louis, MO, USA), and cell man 4-HNE (GENTAUR). Immunoglobulin G control Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 6 of 15 antibodies were used as negative controls. Following 1- Representative HUVEC OCR and ECAR profiles be- hour incubation with primary antibody, endogenous per- fore and after injections of oligomycin, FCCP, antimy- oxidase activity was blocked using 0.3% hydrogen perox- cin A and 2-DG are shown on Fig. 2a. Similarly to ide for 5 minutes. Slides were incubated for 30 minutes RASFC, 4-HNE inhibited basal mitochondrial respir- with secondary antibody/horseradish peroxidase (Dako/ ation, maximal mitochondrial respiration, ATP synthesis Agilent Technologies). 3,3'-Diaminobenzidine (1:50) was and reserve capacity (all p < 0.01) with concomitant eleva- used to visualise staining, and Mayer’shaematoxylin tion of basal glycolysis (p < 0.01) and glycolytic reserve (p (BDH Laboratories, Poole, UK) was incubated for 30 sec- < 0.05) in HUVEC exposed to oxidative stress (Fig. 2b). onds as a counterstain prior to mounting in DPX mount- ing media. Slides were scored separately for lining layer Examination of mitochondrial mutagenesis and activity of (LL), sublining layer (SL) and vascular region (BV) using a enzymes of mitochondrial OXPHOS complexes under 4- well-established and validated semi-quantitative scoring HNE-induced oxidative stress method , where the percentage of cells that were posi- We have previously shown that increased mtDNA muta- tive for a specific marker was compared with the percent- tion frequency and mitochondrial dysfunction in the RA age of cells that were negative. Percentage positivity was joint were strongly associated with synovial inflammation graded using a 0–4 scale, where 0 = no stained cells, 1 = and hypoxia [27, 28]. We have also reported, at a func- 1–25%, 2 = 25–50%, 3 = 50–75 and 4 = 75–100% stained tional level, induction of pro-angiogenic responses of cells. Images were captured using an Olympus DP50 light endothelial cells in the presence of oxidative stress . In microscope and AnalySIS software (Olympus Soft Imaging the present study, we assessed the frequency of mtDNA Solutions, Lakewood, CO, USA). mutations and mitochondrial dysfunction in RASFC sub- jected to 4-HNE. We observed increases in ROS produc- Statistical analysis tion and mtDNA point mutations in RASFC in the IBM SPSS Statistics version 20 for Windows software presence of 4-HNE compared with basal cells (p <0.001 (IBM, Armonk, NY, USA) was used for statistical ana- and p = 0.06, respectively) (Fig. 3a). 4-HNE protein lysis. Wilcoxon’s signed-rank test, Spearman’s rank- adduction may alter protein activity; therefore, we next ex- correlation coefficient and the Mann-Whitney U test amined the activity of the individual proteins of mitochon- were used for analysis of non-parametric data. Paramet- drial OXPHOS complexes I–V. 4-HNE significantly ric data were analysed using one-way analysis of vari- reduced the activity of complex III by 8% and complex IV ance. All p values were two-sided, and p values less than by 70% compared with basal values (both p <0.01). Lower 0.05 were considered statistically significant. enzymatic activity following 4-HNE stimulation was also detected for complex I by 9%, complex II by 22% and Results complex V by 12% (all p =0.2) (Fig. 3b). Oxidative stress alters cellular bioenergetics in RASFC and HUVEC in vitro In vitro secretion of pro-angiogenic and pro-inflammatory Previous studies by our group demonstrated altered cel- mediators under oxidative stress conditions lular bioenergetics in RASFC in the presence of hypoxia Because we found a close association of redox state with , and we have also demonstrated high oxidative energy metabolism in RASFC, we next examined the ef- stress in the inflamed synovium . Therefore, in this fect of oxidative stress on angiogenic and inflammatory study, we further investigated whether oxidative stress in mediators from RASFC. Figure 4 demonstrates increased the inflamed joint is involved in metabolic reprogram- VEGF immunofluorescence staining in RASFC cultured ming of RASFC and HUVEC. Figure 1a demonstrates in the presence of 4-HNE compared with the basal cells. representative OCR and ECAR profiles before and after In addition, 4-HNE significantly increased secretion of injections of oligomycin, FCCP, antimycin A and 2-DG key pro-inflammatory and pro-angiogenic mediators in basal and 4-HNE-stimulated RASFC. We show, for compared with basal RASFC (VEGF, Ang2, bFGF, IL-8 the first time to our knowledge, that inhibition of OCR [all p < 0.05], PDGF-B, RANTES, ICAM [all p < 0.01]). following 4-HNE-induced oxidative stress was associated These findings, along with our previously published in with a shift in RASFC metabolism towards glycolysis. 4- vitro study showing TNF-α-induced mitochondrial dys- HNE reduced basal mitochondrial respiration (p<0.05), par- function , further support the concept of the com- alleled by a reduction in maximal mitochondrial respiration plex interplay between oxidative damage, oxygen (p < 0.001), ATP synthesis (p = 0.1) and reserve capacity (p < metabolism and angiogenesis in RA. Therefore, we next 0.01) (Fig. 1b). This metabolic reprogramming was further determined angiogenic in vivo responses following TNFi accompanied by increased levels of basal glycolysis (p<0.01), in 15 patients with RA at baseline (T0) and 3 months glycolytic capacity (p < 0.01) and glycolytic reserve (p =0. after the commencement of biologic treatment (T3). 2) in RASFC subjected to oxidative stress (Fig. 1b). Additional file 1: Figure S1A shows changes of Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 7 of 15 Fig. 1 Bioenergetic metabolism in primary rheumatoid arthritis synovial fibroblast cells (RASFC) subjected to 4-hydroxy-2-nonenal (4-HNE)-induced oxidative stress. a Representative oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) Seahorse analyser profiles before and after injections of oligomycin, trifluorocarbonylcyanide phenylhydrazone (FCCP), antimycin A and 2-deoxyglucose (2-DG) in RASFC in the presence and absence of 4-HNE. b Bar graphs demonstrate quantification of basal mitochondrial (Mt) respiration, maximal Mt respiration, adenosine triphosphate (ATP) synthesis, reserve capacity, basal glycolysis, glycolytic capacity and glycolytic reserve in RASFC (n = 5) subjected to oxidative stress. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, significant differences from basal level macroscopic vascularity and ST expression of VEGF, demonstrates the effect of basal or 4-HNE RASFC-CM on Ang2 and Tie2 from T0 to T3. Additional file 1: Figure invasion, the formation of tube-like structures and migration S1B graphically illustrates decreases in ST VEGF (p =0. of HUVEC. Figure 5b graphically illustrates markedly in- 1), Ang2 (p < 0.005) and Tie2 (p < 0.005) after TNFi duced invasion (p < 0.001), proliferation (p < 0.05), therapy. number of formed tube-like structures (p < 0.001), cell migration across the wound (p < 0.001) and secretion Oxidative stress-activated RASFC promote pro-angiogenic of Ang2 and PDGF-B (both p values < 0.05) in mechanisms in HUVEC HUVEC in response to basal or 4-HNE RASFC-CM. RASFC are known to be strongly involved in regulating To confirm that the increase in pro-angiogenic re- pathological angiogenesis in the inflamed joint . sponses of HUVEC was due to oxidatively activated Therefore, we next examined if the observed alterations in RASFC and not to residual 4-HNE present in the CM, cellular bioenergetics and pro-inflammatory processes in additional experiments were performed, consisting of RASFC in response to oxidative stress could subsequently RPMI 1640 media supplemented with 4-HNE (0.25 μM; influencepro-angiogenicmechanismsinHUVEC.Westim- 4-HNE RPMI 1640 control), which would be at the same ulated RASFC in the presence or absence of 4-HNE and har- concentration of 4-HNE in the 10% RASFC-CM. A sig- vested the supernatants, termed conditioned media.Fig. 5a nificant increase in invasion (p < 0.001), number of formed Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 8 of 15 Fig. 2 Bioenergetic metabolism in human umbilical vein endothelial cells (HUVEC) subjected to 4-hydroxy-2-nonenal (4-HNE)-induced oxidative stress. a Representative oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) Seahorse analyser profiles before and after injections of oligomycin, trifluorocarbonylcyanide phenylhydrazone (FCCP), antimycin A and 2-deoxyglucose (2-DG) in HUVEC in the presence and absence of 4-HNE. b Bar graphs demonstrate quantification of basal mitochondrial (Mt) respiration, maximal Mt respiration, adenosine triphosphate (ATP) synthesis, reserve capacity, basal glycolysis, glycolytic capacity and glycolytic reserve in HUVEC (n = 3) subjected to oxidative stress. Data are presented as mean ± SEM. *p < 0.05 and **p < 0.01, significant differences from basal level tube-like structures (p < 0.01) and cell migration across immunofluorescence images demonstrating co-localisation the wound (p < 0.001) in HUVEC in response to 4-HNE of 4-HNE with angiogenic factors (VEGF, Ang2, Tie2) RASFC-CM compared with 4-HNE RPMI 1640 control , as well as with mitochondrial (ATP5B) and glycolytic media further supports the direct effect of 4-HNE on (GAPDH, PKM2, GLUT1) proteins, is demonstrated in RASFC-induced angiogenesis in the inflamed joint (Add- Fig. 6. Additional file 3: Figure S3 and Additional file 4: itional file 2: Figure S2). Figure S4 show single images of VEGF, Ang2, Tie2, GAPDH, PKM2, GLUT1 and ATP5B (all in red), single Association between ST angiogenesis, oxidative stress images of 4-HNE immunofluorescence (in green), as well and bioenergetics as single images of DAPI (in blue), along with their con- Finally, the correlation of angiogenic factors with previ- trols with isotype-matched antibodies. ously assessed markers of oxidative stress and metabol- ism in this patient cohort was examined . ST 4-HNE Discussion expression was associated with increased expression of In this study, we demonstrate, for the first time to our VEGF (r = 0.63; p = 0.015) and Tie2 (r = 0.56; p = 0.029), knowledge, that oxidative stress reprograms cellular bio- GAPDH (r = 0.60; p = 0.03) and with reduced levels of energetics of RASFC and HUVEC by downregulating ATP5B (p = − 0.52, p = 0.017). Furthermore, representative OXPHOS and promoting glycolysis. This change was Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 9 of 15 Fig. 3 Mitochondrial mutagenesis and activity of enzymes of mitochondrial oxidative phosphorylation (OXPHOS) complexes under 4-hydroxy-2-nonenal (4-HNE)-induced oxidative stress. a Bar graphs demonstrate increased production of reactive oxygen species (n = 7), paralleled by the greater frequency of mitochondrial DNA mutation (n = 5) in primary rheumatoid arthritis synovial fibroblast cells (RASFC) in response to 4-HNE. b Activity of mitochondrial OXPHOS complexes I–V in the presence of 4-HNE. 4-HNE reduces the activity of complex I by 9%, complex II by 22%, complex III by 8%, complex IV by 70% and complex V by 12% (all complexes measured in triplicate). For each complex, results are graphically demonstrated as the percentage of enzymatic activity in the presence of 4-HNE relative to the percentage of basal activity. Data is represented as Mean ± SEM, **p<0.01; ***p<0.001 significantly different to basal reflected by a decrease in mitochondrial maximal and Hypoxia is a fundamental metabolic change in ST of RA ATP-linked respiration and reserve capacity, whereas associated with elevated mitochondrial ROS production glycolytic capacity and glycolytic reserve were elevated and lipid peroxidation. Covalent modifications of mtDNA, in the presence of 4-HNE. A bioenergetic switch was lipids and proteins by 4-HNE have been reported to com- coupled with higher ROS production and mtDNA promise mitochondrial integrity and function, including re- mutations, in addition to the reduced enzymatic activity spiratory metabolism, protein transportation, mitochondrial of mitochondrial complexes III and IV. Oxidative stress dynamics and quality control through fission, fusion and also induced secretion of pro-angiogenic and pro- mitophagy . We have previously shown that increased inflammatory mediators by RASFC. CM from 4-HNE- mtDNA mutation frequency and mitochondrial dysfunc- activated RASFC potentiated pro-angiogenic mecha- tion in the RA joint correlated with greater hypoxia, oxida- nisms in HUVEC, as reflected by elevated cell invasion, tive stress, vascularity and pro-inflammatory cytokines [27, proliferation, migration, the formation of tube-like struc- 28]. Our present in vitro findings using RASFC further tures and secretion of pro-angiogenic mediators. In vivo demonstrate high susceptibility of the mitochondrial gen- co-expression of angiogenic markers, oxidative damage ome to oxidative damage. A mitochondrial random muta- and oxygen metabolism was demonstrated in ST. Finally, tion capture assay was used to quantify the frequency of a decrease in ST angiogenesis was observed in patients random mitochondrial point mutations in RASFC follow- with RA following TNFi therapy. ing 4-HNE stimulations. This methodology relies on single- Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 10 of 15 Fig. 4 4-Hydroxy-2-nonenal (4-HNE) induces pro-angiogenic and pro-inflammatory mechanisms in primary rheumatoid arthritis synovial fibroblast cells (RASFC). Increased vascular endothelial growth factor (VEGF) immunofluorescence in RASFC subjected to 4-HNE compared to the basal cells and quantification of VEGF, angiopoietin 2 (Ang2), basic fibroblast growth factor (bFGF), interleukin (IL)-8, platelet-derived growth factor subunit B (PDGF-B), regulated on activation, normal T cell expressed and secreted (RANTES), intercellular adhesion molecule (ICAM) in RASFC supernatants (n = 7) following cell culture with 4-HNE. Data are presented as mean ± SEM. *p < 0.05 and **p < 0.01, significant differences from basal level. Red = VEGF; blue =4′,6-diamidino-2-phenylindole–; magnification of photomicrographs × 40 molecule amplification to screen a large number of mtDNA presence of 4-HNE, RASFC and HUVEC switched their molecules for the presence of unexpanded mutations that bioenergetic profile from OXPHOS to anaerobic glycoly- may appear following oxidative stress. Elevated mitochon- sis to respond to an increased energy demand. This was drial mutagenesis detected in RASFC exposed to oxidative coupled with reduced maximal and ATP-linked respir- stress supports theevidenceofthemutagenicnatureof 4- ation and reserve capacity, in contrast to elevated glyco- HNE. 4-HNE-guanine adducts have been detected in the lytic capacity and glycolytic reserve. In addition, the p53 tumour suppressor gene in a human lymphoblastoid enzymatic activities of mitochondrial complexes were cell line, causing gene mutation and affecting cell cycle ar- decreased by oxidative stress. This compensatory reli- rest, apoptosis, DNA repair and differentiation . Ele- ance on anaerobic glycolysis may provide a short-term vated mtROS levels are considered a primary source of solution; however, prolonged dependence may result in a mitochondrial mutagenesis. Our findings show high pro- severe energy deficiency that ultimately creates a bio- duction of ROS by RASFC exposed to 4-HNE, indicating energetic crisis, most likely supporting abnormal angio- the ability of 4-HNE to further exacerbate ROS generation, genesis, cellular invasion and pannus formation. thereby creating the vicious cycle of oxidative stress- Our data are consistent with previous studies demon- induced alteration to the mitochondrial genome. strating 4-HNE-induced mitochondrial respiration defi- We investigated if mitochondrial genome instability ciency in cardiac and small airway epithelial cells [32, driven by products of lipid peroxidation is accompanied 33]. Inhibition of mitochondrial respiration following 4- by defects in respiratory metabolism. The two major en- HNE stimulation could be due to reduced functionality ergy pathways were measured to find that in the from 4-HNE protein-adducts of proteins associated with Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 11 of 15 Fig. 5 Effects of primary rheumatoid arthritis synovial fibroblast cell (RASFC)-conditioned media on angiogenic responses of human umbilical vein endothelial cells (HUVEC). a Representative images demonstrating invasion, the formation of tube-like structures and migration of HUVEC cultured in the presence of basal or 4-hydroxy-2-nonenal (4-HNE)-supplemented conditioned media. Magnification × 10 of photomicrographs demonstrating invasion, tube formation (arrows indicate connecting branches) and cell migration. b Bar graphs demonstrate an increase in the number of invading, proliferating and migrating HUVEC, a higher number of connecting branches formed between HUVEC, and greater angiopoietin 2 (Ang2) and platelet- derived growth factor subunit B (PDGF-B) release from HUVEC exposed to 4-HNE-supplemented conditioned media (n = 6). Data are presented as mean ± SEM. *p <0.05 and ***p < 0.001, significant differences from basal level. hpf High-power field the ETC and ATP synthase, or it could be due to a di- including endothelial cell invasion, proliferation and minished ability of RASFC to detoxify 4-HNE because migration; the formation of tube-like structures; and se- this process requires energy. A study using a proteomic cretion of pro-angiogenic mediators. These findings pro- approach identified several 4-HNE-modified mitochon- vide evidence for direct and indirect pro-angiogenic drial proteins in cardiac mitochondria from mice treated effects in response to 4-HNE within the inflamed joint. with doxorubicin, one of the most widely used chemo- Our findings are in agreement with other studies show- therapeutic drugs . Identified proteins were related ing 4-HNE-induced expression of COX-2, IL-1β, IL-18 to mitochondrial energy metabolism, including subunits and NF-κB and activation of the NLRP3 inflammasome of the ETC such as NDUFS2 (complex I), SDHA (com- [36–38]. In inflammatory conditions, oxidative stress plex II) and ATP5B (complex V), as well as dihydroli- may mediate angiogenic mechanisms through VEGF- poamide dehydrogenase, a component of the TCA cycle. independent pathways involving ROS-induced lipid oxi- Subsequently, 4-HNE adduction reduced enzymatic ac- dation. Redox upregulated angiogenic responses of tivity of the mitochondrial proteins, declined OCR and RASFC observed in our study were also reported by increased ECAR profiles. Other studies identified the 4- others in HUVEC, keratinocytes and epithelial lung and HNE modification of proteins involved in metabolism, retinal cells [29, 39, 40]. Blocking glycolysis with glyco- adhesion, cytoskeletal reorganisation and anti-oxidation lytic inhibitors reduces pro-inflammatory responses of in human platelets . RASFC and HUVEC as well as the severity of arthritis in In this study, increased secretion of pro-angiogenic K/BxN mice [7, 14]. Furthermore, hypoxic activation of and pro-inflammatory mediators by RASFC was ob- the glycolytic enzyme glucose-6-phosphate isomerase served in the response to 4-HNE. Furthermore, RASFC- up-regulated VEGF secretion, proliferation and invasion CM potentiated pro-angiogenic processes of HUVEC, in RASFC and HUVEC . Several glycolytic proteins, Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 12 of 15 Fig. 6 Synovial tissue (ST) angiogenesis, oxidative stress and cellular bioenergetics. To support the concept that oxidative stress, angiogenesis and energy metabolism are interconnected processes that co-exist during the inflammation milieu, double-immunofluorescence staining was performed. ST slides were co-incubated with primary mouse antibody against human 4-hydroxy-2-nonenal (4-HNE) and with primary rabbit antibodies against angiogenic factors (vascular endothelial growth factor [VEGF], angiopoietin 2 [Ang2], tyrosine kinase receptor [Tie2]), glycolytic proteins (glyceraldehyde 3-phosphate dehydrogenase [GAPDH], pyruvate kinase isozyme 2 [PKM2], glucose transporter 1 [GLUT1]) and a mitochondrial marker (adenosine triphosphate synthase subunit β [ATP5B]). Representative merged immunofluorescence images demonstrate examples of co-localisation (yellow) of 4-HNE with VEGF, Ang2, Tie2, GAPDH, PKM2, GLUT1 and ATP5B. Cells stained green are positive for 4-HNE only; cells stained red are positive only for VEGF, Ang2, Tie2, GAPDH, PKM2, GLUT1 and ATP5B. Arrows indicate examples of co-localisation. Magnification of photomicrographs × 20, insets show high-power magnification of co-localisation. Representative images show single immunofluorescence of 4-HNE, VEGF, Ang2, Tie2, GAPDH, PKM2, GLUT1 and ATP5B along with their controls. Isotype-matched antibodies are shown in Additional file 3: Figure S3 and Additional file 4: Figure S4 including PKM2, GAPDH, fructose bisphosphate aldol- commencement of treatment. Following TNFi treatment, ase A (aldolase A) and phosphoglycerate kinase 1 have reduced ST expression of VEGF, Tie2 receptor and its been shown to be adducted by lipid electrophiles [42– Ang2 ligand was observed, which further supports the 44]. Subsequently, this covalent modification is sug- strong link between angiogenesis and TNF-α.These findings gested to impair glucose metabolism and result in the are in agreement with those of other studies showing re- accumulation of glycolytic intermediates. This is consist- duced expression of angiogenic markers and endothelial cell ent with studies showing significant increases in lactate activation following TNFi treatment [46, 47]. Additionally, levels and ECAR by human platelets cultured with 4- previous studies have demonstrated positive effects of TNFi, HNE  and increased F-fludeoxyglucose uptake and including etanercept and infliximab, on oxidative glycolytic metabolism by oxidised low-density lipopro- damage in RA, showing significantly reduced serum tein (oxLDL) through upregulation of GLUT1 expres- and urinary levels of oxidative DNA damage and lipid sion and hexokinase activity . This response was peroxidation with corresponding decreases in DAS28 mediated by hypoxia-inducible factor 1α activation and score following TNFi therapy [48, 49]. Similarly, the dependent on ROS overproduction. In turn, this meta- serum level of oxidative stress markers was remark- bolic effect of oxLDL was completely abrogated by Src ably suppressed in patients with RA treated with toci- (PP2) and phosphatidylinositol-3 kinase inhibitors, sup- lizumab IL-6-blocking therapy compared with those porting the regulatory role of this pathway in glucose treated with anti-TNF antibodies . metabolism and immune cell activation. TNF-α promotes angiogenesis and may regulate capil- Conclusions lary formation via VEGF, Ang1 and Ang2 and their recep- In this study, we examined the interplay of synovial tors. In this study, we investigated whether TNFi therapy cellular bioenergetics, oxidative stress and angiogen- alters levels of angiogenic markers 3 months after the esis in RA. We have demonstrated that oxidative stress Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 13 of 15 switched bioenergetic profiles from OXPHOS to anaer- after starting therapy; TCA: Tricarboxylic acid cycle; T 17: T-helper type 17 cells; Tie2: Tyrosine kinase receptor; TLR2: Toll-like receptor 2; TNF-α: Tumour obic glycolysis in response to an increased energy demand necrosis factor α; TNFi: Tumour necrosis factor α inhibitor; VEGF: Vascular in the inflamed joint. This creates a bioenergetic crisis that endothelial growth factor may contribute to dysfunctional angiogenesis to further Funding promote inflammatory mechanisms in RA. In addition, This study was funded by the Health Research Board of Ireland and Arthritis ST upregulation of the angiopoietin/Tie2 system can be Ireland. altered following TNFi therapy. Availability of data and materials The datasets analysed in the present study are available from the Additional files corresponding author on reasonable request. Authors’ contributions Additional file 1: Figure S1. Synovial tissue expression of angiogenic EB and MB conducted most of the experiments and analysed data. TMG markers in patients with RA. A Representative images demonstrating performed some of the experiments. EB, DJV, UF and MB participated in the macroscopic vascularity and ST VEGF, Ang2 and Tie2 immunostaining at data analysis and manuscript preparation and provided final approval of the baseline (T0) and 3 months after the commencement of biologic version to be published. DJV, ZS, UF and MB participated in the study design treatment (T3). Magnification of photomicrographs × 20. B Baseline and and supervised the research. DJV, CO and CTN recruited all patients, 3 months post-TNFi quantification of ST VEGF, ST Ang2 and ST Tie2 in performed the arthroscopies and provided all clinical information. All authors patients with RA (n = 15). Data are presented as mean ± SEM. *p < 0.05; read and approved the final manuscript. **p < 0.01. (TIF 5466 kb) Additional file 2: Figure S2. Effects of oxidatively activated RASFC on Ethics approval and consent to participate angiogenic responses of HUVEC. To confirm that the increase in pro- All of this research was carried out in accordance with the Declaration of angiogenic responses of HUVEC is due to oxidatively activated RASFC and not Helsinki, and approval for this study was granted by the St. Vincent’s residual 4-HNE present in the conditioned media, HUVEC were cultured in the University Hospital Medical Research and Ethics Committee. All patients gave presence of RPMI 1640 media supplemented with 4-HNE (0.25 μM; 4-HNE fully informed written consent approved by the institutional ethics RPMI 1640 control), which was at the same concentration of 4-HNE in the committee. 10% RASFC conditioned media. Representative images and bar graphs demonstrate higher invasion, greater number of formed tube-like structures Consent for publication and greater cell migration across the wound in HUVEC in response to 4-HNE Written consent was obtained from all the participants in and authors of this RASFC-conditioned media (4-HNE RASFC-CM; n = 6) than in response to study. 4-HNE RPMI 1640 control. Data are presented as mean ± SEM. **p <0.01 and ***p < 0.001, representing significant differences from control. Magnification of Competing interests photomicrographs demonstrating invasion, tube formation (arrows show The authors declare that they have no competing interests. connecting branches) and cell migration × 10. (TIF 8263 kb) Additional file 3: Figure S3. ST angiogenesis and oxidative stress. Representative images of immunofluorescent staining between markers Publisher’sNote of angiogenesis and 4-HNE in inflamed synovial tissue of patients with Springer Nature remains neutral with regard to jurisdictional claims in RA: VEGF, Ang2 and Tie2 (red); 4-HNE (green); DAPI (blue); and merged published maps and institutional affiliations. images (yellow). Insets show negative control staining with isotype-matched antibodies. Magnification of photomicrographs × 20. Author details (TIF 9793 kb) Department of Rheumatology, University of Debrecen Medical and Health Additional file 4: Figure S4. ST cellular bioenergetics and oxidative Science Centre, 98. Nagyerdei krt, Debrecen, Hungary. Centre for Arthritis stress. Representative immunofluorescence images show co-localisation and Rheumatic Diseases, Dublin Academic Medical Centre, St. Vincent’s of the oxidative stress marker 4-HNE with glycolytic proteins (GAPDH, University Hospital, Dublin, Ireland. Molecular Rheumatology, Trinity PKM2, GLUT1) and a mitochondrial marker (ATP5B) in inflamed ST of Biomedical Sciences Institute Trinity College Dublin, Dublin, Ireland. patients with RA: GAPDH, PKM2, GLUT1, and ATP5B (red); 4-HNE (green); Department of Rheumatology and Immunology, Singapore General DAPI (blue); merged images (yellow). Insets show negative control staining Hospital, Singapore, Singapore. Duke-NUS Medical School, Singapore, with isotype-matched antibodies. Magnification of photomicrographs × Singapore. 20. (TIF 9536 kb) Received: 7 December 2017 Accepted: 12 April 2018 Abbreviations 2-DG: 2-Deoxyglucose; 4-HNE: 4-Hydroxy-2-nonenal; ADP: Adenosine References diphosphate; Ang2: Angiopoietin 2; ATP: Adenosine triphosphate; 1. Tas SW, Maracle CX, Balogh E, Szekanecz Z. Targeting of proangiogenic ATP5B: Adenosine triphosphate synthase subunit β; bFGF: Basic fibroblast signalling pathways in chronic inflammation. Nat Rev Rheumatol. 2016; growth factor; CM: Conditioned media; DAS28: 28-Joint Disease Activity 12(2):111–22. Score; DAPI: 4′,6-Diamidino-2-phenylindole; ECAR: Extracellular acidification 2. Bodamyali T, Stevens CR, Billingham ME, Ohta S, Blake DR. Influence of rate; ETC: Electron transport chain; FCCP: Trifluorocarbonylcyanide hypoxia in inflammatory synovitis. Ann Rheum Dis. 1998;57(12):703–10. phenylhydrazone; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase; 3. Stevens CR, Blake DR, Merry P, Revell PA, Levick JR. A comparative study by GLUT1: Glucose transporter 1; HUVEC: Human umbilical vein endothelial cells; morphometry of the microvasculature in normal and rheumatoid synovium. ICAM: Intercellular adhesion molecule; IL-8: Interleukin 8; Mt: Mitochondrial; Arthritis Rheum. 1991;34(12):1508–13. mtDNA: Mitochondrial DNA; OCR: Oxygen consumption rate; OD: Optical 4. Taylor RW, Turnbull DM. Mitochondrial DNA mutations in human disease. density; oxLDL: Oxidised low-density lipoprotein; OXPHOS: Oxidative Nat Rev Genet. 2005;6(5):389–402. phosphorylation; PDGF-B: Platelet-derived growth factor subunit B; 5. Ospelt C, Gay S. Somatic mutations in mitochondria: the chicken or the PFA: Paraformaldehyde; PKM2: Pyruvate kinase isozyme 2; RA: Rheumatoid egg? Arthritis Res Ther. 2005;7(5):179–80. arthritis; RANTES: Regulated on activation, normal T cell expressed and 6. Sena LA, Li S, Jairaman A, Prakriya M, Ezponda T, Hildeman DA, Wang CR, secreted; RASFC: Primary rheumatoid arthritis synovial fibroblast cell; Schumacker PT, Licht JD, Perlman H, et al. Mitochondria are required for RMC: Random mutation capture assay; ROS: Reactive oxygen species; antigen-specific T cell activation through reactive oxygen species signaling. ST: Synovial tissue; T0: Time point 0 or baseline; T3: Time point 3 months Immunity. 2013;38(2):225–36. Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 14 of 15 7. Garcia-Carbonell R, Divakaruni AS, Lodi A, Vicente-Suarez I, Saha A, 28. Harty LC, Biniecka M, O’Sullivan J, Fox E, Mulhall K, Veale DJ, Fearon U. Cheroutre H, Boss GR, Tiziani S, Murphy AN, Guma M. Critical role of glucose Mitochondrial mutagenesis correlates with the local inflammatory metabolism in rheumatoid arthritis fibroblast-like synoviocytes. Arthritis environment in arthritis. Ann Rheum Dis. 2012;71(4):582–8. Rheumatol. 2016;68(7):1614–26. 29. Biniecka M, Connolly M, Gao W, Ng CT, Balogh E, Gogarty M, Santos L, 8. ShiLZ, Wang R, HuangG,Vogel P, NealeG,Green DR,Chi H. Murphy E, Brayden D, Veale DJ, et al. Redox-mediated angiogenesis in HIF1α-dependent glycolytic pathway orchestrates a metabolic the hypoxic joint of inflammatory arthritis. Arthritis Rheumatol. 2014; checkpoint for the differentiation of T 17 and T cells. J Exp Med. 66(12):3300–10. H reg 2011;208(7):1367–76. 30. Juarez M, Filer A, Buckley CD. Fibroblasts as therapeutic targets in 9. Krawczyk CM, Holowka T, Sun J, Blagih J, Amiel E, DeBerardinis RJ, Cross JR, rheumatoid arthritis and cancer. Swiss Med Wkly. 2012;142:w13529. Jung E, Thompson CB, Jones RG, et al. Toll-like receptor-induced changes 31. Hu W, Feng Z, Eveleigh J, Iyer G, Pan J, Amin S, Chung FL, Tang MS. in glycolytic metabolism regulate dendritic cell activation. Blood. 2010; The major lipid peroxidation product, trans-4-hydroxy-2-nonenal, 115(23):4742–9. preferentially forms DNA adducts at codon 249 of human p53 gene, a 10. Weyand CM, Goronzy JJ. Immunometabolism in early and late stages of unique mutational hotspot in hepatocellular carcinoma. Carcinogenesis. rheumatoid arthritis. Nat Rev Rheumatol. 2017;13(5):291–301. 2002;23(11):1781–9. 11. Naughton D, Whelan M, Smith EC, Williams R, Blake DR, Grootveld M. An 32. Galam L, Failla A, Soundararajan R, Lockey RF, Kolliputi N. 4-Hydroxynonenal investigation of the abnormal metabolic status of synovial fluid from regulates mitochondrial function in human small airway epithelial cells. patients with rheumatoid arthritis by high field proton nuclear magnetic Oncotarget. 2015;6(39):41508–21. resonance spectroscopy. FEBS Lett. 1993;317(1-2):135–8. 33. Hill BG, Dranka BP, Zou L, Chatham JC, Darley-Usmar VM. Importance of the 12. Ciurtin C, Cojocaru VM, Miron IM, Preda F, Milicescu M, Bojinca M, Costan O, bioenergetic reserve capacity in response to cardiomyocyte stress induced Nicolescu A, Deleanu C, Kovacs E et al. Correlation between different by 4-hydroxynonenal. Biochem J. 2009;424(1):99–107. components of synovial fluid and pathogenesis of rheumatic diseases. Rom 34. Zhao Y, Miriyala S, Miao L, Mitov M, Schnell D, Dhar SK, Cai J, Klein JB, J Intern Med 2006, 44(2):171-181. Sultana R, Butterfield DA, et al. Redox proteomic identification of HNE- 13. Hitchon CA, El-Gabalawy HS, Bezabeh T. Characterization of synovial tissue bound mitochondrial proteins in cardiac tissues reveals a systemic from arthritis patients: a proton magnetic resonance spectroscopic effect on energy metabolism after doxorubicin treatment. Free Radic investigation. Rheumatol Int. 2009;29(10):1205–11. Biol Med. 2014;72:55–65. 14. Biniecka M, Canavan M, McGarry T, Gao W, McCormick J, Cregan S, 35. Ravi S, Johnson MS, Chacko BK, Kramer PA, Sawada H, Locy ML, Wilson LS, Gallagher L, Smith T, Phelan JJ, Ryan J, et al. Dysregulated Barnes S, Marques MB, Darley-Usmar VM. Modification of platelet proteins bioenergetics: a key regulator of joint inflammation. Ann Rheum Dis. by 4-hydroxynonenal: potential mechanisms for inhibition of aggregation 2016;75(12):2192–200. and metabolism. Free Radic Biol Med. 2016;91:143–53. 15. McGarry T, Biniecka M, Gao W, Cluxton D, Canavan M, Wade S, Wade S, 36. Kauppinen A, Niskanen H, Suuronen T, Kinnunen K, Salminen A, Kaarniranta Gallagher L, Orr C, Veale DJ, et al. Resolution of TLR2-induced inflammation K. Oxidative stress activates NLRP3 inflammasomes in ARPE-19 through manipulation of metabolic pathways in rheumatoid arthritis. Sci cells—implications for age-related macular degeneration (AMD). Immunol Rep. 2017;7:43165. Lett. 2012;147(1-2):29–33. 16. Xiao M, Zhong H, Xia L, Tao Y, Yin H. Pathophysiology of mitochondrial lipid 37. Park S, Sung B, Jang EJ, Kim DH, Park CH, Choi YJ, Ha YM, Kim MK, Kim ND, oxidation: role of 4-hydroxynonenal (4-HNE) and other bioactive lipids in Yu BP, et al. Inhibitory action of salicylideneamino-2-thiophenol on NF-κB mitochondria. Free Radic Biol Med. 2017;111:316–27. signaling cascade and cyclooxygenase-2 in HNE-treated endothelial cells. 17. Zhong H, Yin H. Role of lipid peroxidation derived 4-hydroxynonenal Arch Pharm Res. 2013;36(7):880–9. (4-HNE) in cancer: focusing on mitochondria. Redox Biol. 2015;4:193–9. 38. Zarrouki B, Soares AF, Guichardant M, Lagarde M, Geloen A. The lipid 18. Benderdour M, Charron G, DeBlois D, Comte B, Des Rosiers C. Cardiac peroxidation end-product 4-HNE induces COX-2 expression through mitochondrial NADP -isocitrate dehydrogenase is inactivated through p38MAPK activation in 3T3-L1 adipose cell. FEBS Lett. 2007;581(13): 4-hydroxynonenal adduct formation: an event that precedes hypertrophy 2394–400. development. J Biol Chem. 2003;278(46):45154–9. 39. Vatsyayan R, Lelsani PC, Chaudhary P, Kumar S, Awasthi S, Awasthi YC. The expression and function of vascular endothelial growth factor in retinal 19. Humphries KM, Yoo Y, Szweda LI. Inhibition of NADH-linked mitochondrial pigment epithelial (RPE) cells is regulated by 4-hydroxynonenal (HNE) respiration by 4-hydroxy-2-nonenal. Biochemistry. 1998;37(2):552–7. and glutathione S-transferaseA4-4. Biochem Biophys Res Commun. 2012; 20. Echtay KS, Esteves TC, Pakay JL, Jekabsons MB, Lambert AJ, Portero-Otin M, 417(1):346–51. Pamplona R, Vidal-Puig AJ, Wang S, Roebuck SJ, et al. A signalling role for 4-hydroxy-2-nonenal in regulation of mitochondrial uncoupling. EMBO J. 40. Bochkov VN, Philippova M, Oskolkova O, Kadl A, Furnkranz A, Karabeg E, 2003;22(16):4103–10. Afonyushkin T, Gruber F, Breuss J, Minchenko A, et al. Oxidized 21. Li YP, Tian FG, Shi PC, Guo LY, Wu HM, Chen RQ, Xue JM. 4-Hydroxynonenal phospholipids stimulate angiogenesis via autocrine mechanisms, promotes growth and angiogenesis of breast cancer cells through HIF-1α implicating a novel role for lipid oxidation in the evolution of stabilization. Asian Pac J Cancer Prev. 2014;15(23):10151–6. atherosclerotic lesions. Circ Res. 2006;99(8):900–8. 22. Biniecka M, Kennedy A, Fearon U, Ng CT, Veale DJ, O’Sullivan JN. Oxidative 41. Lu Y, Yu SS, Zong M, Fan SS, Lu TB, Gong RH, Sun LS, Fan LY. damage in synovial tissue is associated with in vivo hypoxic status in the Glucose-6-phosphate isomerase (G6PI) mediates hypoxia-induced arthritic joint. Ann Rheum Dis. 2010;69(6):1172–8. angiogenesis in rheumatoid arthritis. Sci Rep. 2017;7:40274. 42. Camarillo JM, Ullery JC, Rose KL, Marnett LJ. Electrophilic modification of 23. Biniecka M, Kennedy A, Ng CT, Chang TC, Balogh E, Fox E, Veale DJ, Fearon PKM2 by 4-hydroxynonenal and 4-oxononenal results in protein cross- U, O’Sullivan JN. Successful tumour necrosis factor (TNF) blocking therapy linking and kinase inhibition. Chem Res Toxicol. 2017;30(2):635–41. suppresses oxidative stress and hypoxia-induced mitochondrial mutagenesis 43. Tsuchiya Y, Yamaguchi M, Chikuma T, Hojo H. Degradation of in inflammatory arthritis. Arthritis Res Ther. 2011;13(4):R121. glyceraldehyde-3-phosphate dehydrogenase triggered by 24. Poli G, Schaur RJ, Siems WG, Leonarduzzi G. 4-Hydroxynonenal: a 4-hydroxy-2-nonenal and 4-hydroxy-2-hexenal. Arch Biochem Biophys. membrane lipid oxidation product of medicinal interest. Med Res Rev. 2005;438(2):217–22. 2008;28(4):569–631. 25. Vermulst M, Bielas JH, Loeb LA. Quantification of random mutations in the 44. Martinez A, Dalfo E, Muntane G, Ferrer I. Glycolitic enzymes are targets of mitochondrial genome. Methods. 2008;46(4):263–8. oxidation in aged human frontal cortex and oxidative damage of these proteins is increased in progressive supranuclear palsy. J Neural Transm. 26. Youssef PP, Kraan M, Breedveld F, Bresnihan B, Cassidy N, Cunnane G, Emery 2008;115(1):59–66. P, Fitzgerald O, Kane D, Lindblad S, et al. Quantitative microscopic analysis 45. Lee SJ, Thien Quach CH, Jung KH, Paik JY, Lee JH, Park JW, Lee KH. Oxidized of inflammation in rheumatoid arthritis synovial membrane samples low-density lipoprotein stimulates macrophage F-FDG uptake via hypoxia- selected at arthroscopy compared with samples obtained blindly by needle inducible factor-1α activation through Nox2-dependent reactive oxygen biopsy. Arthritis Rheum. 1998;41(4):663–9. species generation. J Nucl Med. 2014;55(10):1699–705. 27. Biniecka M, Fox E, Gao W, Ng CT, Veale DJ, Fearon U, O’Sullivan J. Hypoxia 46. Paleolog EM,Hunt M,Elliott MJ,FeldmannM,Maini RN,Woody JN. induces mitochondrial mutagenesis and dysfunction in inflammatory Deactivation of vascular endothelium by monoclonal anti-tumor arthritis. Arthritis Rheum. 2011;63(8):2172–82. Balogh et al. Arthritis Research & Therapy (2018) 20:95 Page 15 of 15 necrosis factor α antibody in rheumatoid arthritis. Arthritis Rheum. 1996; 39(7):1082–91. 47. Canete JD, Pablos JL, Sanmarti R, Mallofre C, Marsal S, Maymo J, Gratacos J, Mezquita J, Mezquita C, Cid MC. Antiangiogenic effects of anti-tumor necrosis factor α therapy with infliximab in psoriatic arthritis. Arthritis Rheum. 2004;50(5):1636–41. 48. Kageyama Y, Takahashi M, Ichikawa T, Torikai E, Nagano A. Reduction of oxidative stress marker levels by anti-TNF-α antibody, infliximab, in patients with rheumatoid arthritis. Clin Exp Rheumatol. 2008;26(1):73–80. 49. Kageyama Y, Takahashi M, Nagafusa T, Torikai E, Nagano A. Etanercept reduces the oxidative stress marker levels in patients with rheumatoid arthritis. Rheumatol Int. 2008;28(3):245–51. 50. Hirao M, Yamasaki N, Oze H, Ebina K, Nampei A, Kawato Y, Shi K, Yoshikawa H, Nishimoto N, Hashimoto J. Serum level of oxidative stress marker is dramatically low in patients with rheumatoid arthritis treated with tocilizumab. Rheumatol Int. 2012;32(12):4041–5.
Arthritis Research & Therapy – Springer Journals
Published: May 29, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud