Introduction
Mitochondria produce ATP through oxidative metabolism to provide cells with energy under physiological conditions. The mitochondrial electron transport chain (ETC) is also a major cellular source of reactive oxygen species (ROS) as some of the electrons passing to molecular oxygen are prone to leakage from the chain and get trapped by oxygen, which converts to superoxide [
1]. Hypoxia characterised by an inadequate supply of molecular oxygen, can trigger mitochondria dysfunction through ineffective functioning of respiratory complexes of ETC [
2,
3].
Free oxygen radicals are highly active molecules and increased mitochondrial ROS generation promotes cellular oxidative stress resulting in oxidative mitochondrial DNA (mtDNA) damage and lipid peroxidation. Moreover, ROS mediate the stress signalling pathways involving nuclear factor-kappa B (NF-κB) [
4]. mtDNA is in the proximity of ROS generation site and has relatively limited repair capacity, which makes it vulnerable to high mutation rates [
5]. Mutations and deletions of the mitochondrial genome in genes encoding proteins for subunits of mitochondrial respiratory chain complexes I-V, rRNA and tRNA have been linked to a variety of degenerative human diseases and high levels of mtDNA mutations have been also found in many tumours and cancer cells [
5,
6].
Oxidative stress, which arises from an imbalance between ROS production and antioxidant defences, results also in lipid peroxidation of cell membrane polyunsaturated fatty acids [
7]. The primary products of free-radical attack of biological membranes are lipid hydroperoxides, which can decompose to highly reactive, cytotoxic secondary end products, such as 4-hydroxy-2-nonenal (4-HNE) [
8]. 4-HNE is an endogenously generated α,β unsaturated aldehyde, which is not only a marker of extensive oxidative stress but also can modulate cellular metabolism, inflammatory responses and apoptosis via its effects on transcriptional regulation and protein modification [
9]. 4-HNE-induced mitochondrial protein modifications include those involved in the ETC, cellular respiration and Krebs cycle [
10]. Moreover, 4-HNE can form adducts on DNA bases and modifies mtDNA thus measurement of such modifications may reflect the level of mitochondrial alterations [
11].
Inflammatory arthritis (IA) is a chronic, progressive disorder associated with joint inflammation, synovial tissue hypertrophy, joint effusions and degradation of articular cartilage and bone. The normal synovial tissue is a relatively acellular structure with a lining layer (one to two cells thick) comprised of macrophages and fibroblasts. The morphology of IA synovium is strikingly different. There is a significant increase in the number of blood vessels that are associated with differential vascular morphology. Furthermore, the early vascular changes are accompanied by increased recruitment of macrophages and synovial fibroblast cells in the lining layer, along with infiltration of T, B and plasma cells. The precise mechanisms involved in regulation of persistent synovial infiltration and invasion are unclear, but high levels of TNF-α may be crucial in mediating the pathogenesis of IA. TNF-α is a proinflammatory cytokine, activating the NF-κB pathway, leading to a downstream cascade of other proinflammatory cytokines [
12,
13]. Moreover, it is known to increase mitochondrial ROS production [
14,
15] and induce the formation of lipid-derived aldehydes [
16]; however TNF-α-induced mitochondrial mutagenesis has not yet been examined in patients with IA. Current targeted biologic therapies, including anti-TNF-α inhibitors result in greater disease improvement and prevention of joint erosion, although clinical studies on the efficacy of TNF-α blocking agents clearly show that about 40% of patients receiving this therapy are non-responders.
Recently, we demonstrated that successful biologic therapy significantly improves
in vivo synovial hypoxia and it is strongly associated with improvement of joint inflammation [
17]. In this study we investigate if successful anti-TNF-α treatment alters the levels of early mitochondrial genome alterations, which can play a crucial role in governing clinical response or resistance. Furthermore, we determine if TNF-α blocking therapy changes the levels of synovial 4-HNE, further confirming the relation between hypoxia, oxidative damage and mitochondrial mutagenesis.
Discussion
Chronic inflammatory arthropathies, such as RA and PsA, are characterised by complex chronic inflammatory processes. Oxygen metabolism is important in synovitis and joint destruction [
23]. ROS stimulates synovial fibroblasts to secrete matrix metalloproteinases, inhibits cartilage proteoglycan synthesis and accelerates bone resorption [
24,
25]. Previously, we have demonstrated profoundly hypoxic synovial environment of the inflamed joint (approximately 3%) [
26]. Furthermore, we have shown that biologic anti-TNF-α therapy significantly increased the synovial
in vivo tpO
2 levels only in those patients who respond to anti-TNF-α therapy [
17]. In this study we examine the effect of TNF-blocking therapy on mitochondrial mutagenesis and synovial oxidative stress profiles. We report for the first time that the increase in tpO
2 levels observed in responders is associated with significant decrease and strong inverse correlation of synovial lipid peroxidation. In addition, increases in tpO
2 significantly reduces the levels of random mitochondrial mutations, presumably as a result of decreased oxidative stress profile.
TNF-α affects many cellular processes, such as activation of phospholipases [
27], proteases [
28] and DNA damage [
29]. Mitochondrially derived ROS are strongly implicated in TNF-α cytotoxicity and may mediate the activation of transcriptional factor NF-κB, which in turn can stimulate mitochondrial NADPH oxidase [
15,
30]. Inhibition of ETC complex III by antimycin A increases ROS and inhibits TNF-α triggered NF-κB activation, highlighting the importance of the ETC in TNF-α cytotoxicity [
31]. Recently, we have shown that hypoxia is an important stimulus of TNF-α secretion, where higher levels of synovial fluid TNF-α were detected in patients with synovial tpO
2 less than 20 mmHg than in those with tpO
2 more than 20 mmHg [
26].
Oxidative stress arising from overproduction of ROS leads to formation of reactive aldehydes such as 4-HNE. Mitochondrial are primed for attack by 4-HNE and formation of adducts between 4-HNE and mitochondrial components. Detection of 4-HNE-mitochondrial protein adducts can reflect mitochondrial dysfunction and oxidative stress [
32]. We have previously assessed the expression of synovial lipid peroxidation in IA patients and demonstrated a significant inverse correlation between 4-HNE expression and oxygen tension of the inflamed join, probably reflecting mitochondrial damage [
20]. Mitochondrial membrane components are targets for 4-HNE modification and the adenine nucleotide translocator in the inner mitochondrial membrane is affected by lipid peroxidation [
33]. This study in the first to show that patients who respond to TNF-blocking therapy show a significant increase in tpO
2 and this is associated with reduced 4-HNE levels. In contrast, in non-responders there is no change in
in vivo oxygen levels and subsequently no change in 4-HNE levels. These data suggest that as the joint tissue becomes less hypoxic, a corresponding reduction in oxidative stress is affected. Previous studies have demonstrated positive effects of anti-TNF-α treatment on oxidative damage in RA, where urinary levels of oxidative DNA damage and lipid peroxidation were significantly reduced at three months therapy [
34]. However, our study considerably extends the above reports and shows direct evidence of a significant reduction of oxidative stress in relation to
in vivo hypoxia measurements.
We have recently demonstrated that increased tpO
2 levels after successful anti-TNF biologic therapy is associated with reduced disease activity and macroscopic vascularity [
17]. Furthermore, we have also reported that high synovial 4-HNE levels positively correlated with clinical disease activity scores in patients prior to receiving TNF-α blocking therapy [
20]. In this study the same parameters were assessed in patients after anti-TNF-α treatment and we found significant positive association between synovial 4-HNE expression and clinical measures of arthritis.
Several cellular and environmental sources of synovial oxidative stress have been proposed, including activated neutrophils, monocytes and macrophages, hypoxia and vascular changes. Furthermore, studies by Remans et al. indicated synovial T lymphocytes as the main generators of intracellular free radicals in RA patients [
35]. We demonstrate a correlation between oxidative stress, inflammation and angiogenesis, where increase in tpO
2 and reduce oxidative stress observed in responders is associated with lower microscopic scores of T-cells (CD4 and CD8), B-cells (CD20), macrophages (CD68) and angiogenesis (VEGF). Experiments using 4-HNE-modified antigens of T and B cells showed rapid autoimmune response, suggesting that B and T cell modification by 4-HNE may result in the onset of autoimmune reactions or even autoimmune disease processes [
36]. The link between oxidative lipid modifications and activation of the inflammatory potential of macrophages has been also suggested [
37]. In human osteoarthritic chondrocytes 4-HNE induces prostaglandin E release and cyclooxygenase-2 (COX-2) expression, providing evidence for the role of 4-HNE as redox-sensitive signalling mechanisms of inflammatory response [
38]. Furthermore, 4-HNE elevated VEGF secretion has been shown in retinal pigment epithelial cells [
39] and vascular smooth muscle cells [
40]. This correlation of VEGF expression and 4-HNE supports our current findings.
RA has many features of autoimmune disease; however, some studies suggest inflammation-independent joint destruction [
41]. It has been shown that elevated production of ROS at the sites of chronic inflammation has genotoxic effects and increases the likelihood of mutagenic events. In RA, local exposure to oxidative stress was found to induce genetic changes and was proposed as a mechanism that permanently alters and imprints synovial cells [
42,
43]. Furthermore, oxidative stress can suppress expression of DNA repair enzymes in inflamed synovium such as DNA mismatch repair system that might potentially limit the accumulation of mutations [
44]. Other extensive studies demonstrated synovial
p53 mutations, which are characteristic DNA damage caused by oxidative stress. High expression of
p53 was found in synovial tissue from longstanding RA patients and lower in early RA patients, osteoarthritis (OA) and reactive arthritis patients [
45]. This oxidative DNA damage of
p53 gene is likely to promote neoplastic transformation of synovial cells that may subsequently contribute to disease progression and joint destruction.
Oxidative stress may also contribute to somatic mtDNA mutation. mtDNA mutations were known to have a key role in ageing-related diseases and carcinogenesis. Currently, there is a growing body of evidence suggesting the role of mitochondrial alterations in rheumatoid disorders [
46]. Recent studies showed higher accumulation of mtDNA damage in chondrocytes from OA patients compared with those from normal donors [
47]. Higher incidence of mtDNA somatic mutations has also been detected in synoviocytes and synovial tissue of RA than OA controls [
48]; however, the frequency of mitochondrial mutations has not been examined. Recently, using synovial tissue of baseline IA patients, we have screened a large number of mtDNA molecules for the presence of unexpanded random mutations, which may be crucial in driving disease progression. We demonstrated, for the first time that greater levels of mtDNA point mutations were significantly associated with higher hypoxia
in vivo, oxidative stress and disease activity [
49].
TNF-α was demonstrated to induce
in vitro mitochondrial ROS release and DNA damage in human chondrocytes and overexpression of the DNA repair enzyme prevents mtDNA alterations following TNF-α exposure [
50]. In this study, we determined whether TNF therapy affect the levels of mtDNA mutations. We observed that the increase in tpO
2 after treatment was associated with significant decrease in the levels of mtDNA mutations and reduction of disease activity scores DAS28-CRP. Contrary, no significant improvements in the levels of mtDNA mutations and DAS28-CRP were found in patients who had more hypoxic synovium after receiving TNF blocking treatment. Our findings strongly support the hypothesis that an increase in mutation frequency is a consequence of elevated hypoxia and oxidative damage to the mitochondrial genome. Furthermore, our results are in agreement with another report indicating the role of oxidative stress and diminished mtDNA integrity in the progression of OA, where high levels of mutagenesis following exposure to ROS were associated with reduced mtDNA capacity and cell viability [
47]. In addition, our study is the first to show that successful anti-TNF-α therapy reduces the frequency of mitochondrial synovial mutagenesis in IA. It may suggest a central role of mitochondrial mutagenesis in the cellular mechanism of anti-TNF-α response or resistance to the treatment
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MB conducted most of the experiments and analysis of data. AK, CTN, TCC, EB, EF and UF performed some of the experiments. JNO, UF, DV and MB participated in the data analysis and manuscript preparation and final approval of the version to be published. JNO, UF and DV participated in the study design and supervised the research. DV and CTN recruited all patients, performed the arthroscopies and oxygen measurements and provided all clinical information. All authors read and approved the final manuscript.