Article Text

Extended report
Mitochondrial mutagenesis correlates with the local inflammatory environment in arthritis
  1. Leonard C Harty1,
  2. Monika Biniecka1,
  3. Jacintha O'Sullivan2,
  4. Edward Fox3,
  5. Kevin Mulhall4,
  6. Douglas J Veale1,
  7. Ursula Fearon1
  1. 1Translation Rheumatology Research Group, Dublin Academic Medical Centre, St Vincent's University Hospital, Elm Park, Dublin, Ireland
  2. 2Department of Surgery, Institute of Molecular Medicine, Trinity College Dublin, St James's Hospital, Dublin, Ireland
  3. 3Department of Pathology, University of Washington, Seattle, Washington, USA
  4. 4Department of Orthopaedics, Dublin Academic Medical Centre, Mater Misericordiae University Hospital, Dublin, Ireland
  1. Correspondence to Dr Ursula Fearon, Translational Rheumatology Group, Dublin Academic Medical Centre, St Vincent's University Hospital and The Conway Institute of Biomolecular and Biomedical Research, Dublin 4, Ireland; ursula.fearon{at}


Background To examine the association between mitochondrial mutagenesis and the proinflammatory microenvironment in patients with inflammatory arthritis.

Methods Fifty patients with inflammatory arthritis underwent arthroscopy and synovial tissue biopsies, synovial fluid and clinical assessment were obtained. Fifteen patients pre/post-TNFi therapy were also recruited. Normal synovial biopsies were obtained from 10 subjects undergoing interventional arthroscopy. Macroscopic synovitis/vascularity was measured by visual analogue scale. Cell-specific markers CD3 (T cells) and CD68 (macrophages) were quantified by immunohistology. TNFα, IL-6, IFNγ and IL-1β were measured in synovial fluids by MSD multiplex assays. Synovial tissue mitochondrial mutagenesis was quantified using a mitochondrial random mutation capture assay (RMCA). The direct effect of TNFα on oxidative stress and mitochondrial function was assessed in primary cultures of rheumatoid arthritis synovial fibroblast cells (RASFCs). Mitochondrial mutagenesis, reactive oxygen species (ROS), mitochondrial membrane potential (MMP) and mitochondrial mass (MM) were quantified using the RMCA and specific cell fluorescent probes.

Results A significant increase in mtDNA mutation frequency was demonstrated in inflamed synovial tissue compared with control (p<0.05), an effect that was independent of age. mtDNA mutations positively correlated with macroscopic synovitis (r=0.52, p<0.016), vascularity (r=0.54, p<0.01) and with synovial fluid cytokine levels of TNFα (r=0.74, p<0.024) and IFNγ (r=0.72, p<0.039). mtDNA mutation frequency post-TNFi therapy was significantly lower in patients with a DAS<3.2 (p<0.05) and associated with clinical and microscopic measures of disease (p<0.05). In vitro TNFα significantly induced mtDNA mutations, ROS, MM and MMP in RASFCs (all p<0.05).

Conclusion High mitochondrial mutations are strongly associated with synovial inflammation showing a direct link between mitochondrial mutations and key proinflammatory pathways.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Rheumatoid arthritis (RA) and psoriatic arthritis (PsA) are two of the most common forms of inflammatory arthritis characterised by synovial inflammation, pain and progressive damage.1 The mechanisms involved in synovial inflammation and invasion are not fully understood; however, an early event in inflammation is angiogenesis, which facilitates the persistent infiltration of immune cells into the joint resulting in destruction of adjacent articular cartilage and bone.2

Oxidative stress arises from an imbalance between reactive oxygen species (ROS) production and the capability of opposing antioxidant forces. Mitochondria provide cellular energy through oxidative phosphorylation in the mitochondrial electron transport chain with resultant ATP production and ROS generation as a by-product.3 Antioxidant defence systems generally mop up leaked ROS production; however, under inflammatory and pathological conditions, ROS can exceed the capacity of antioxidant defence systems and damage occurs to lipids, proteins and DNA. DNA adducts such as 8-oxo-7,8-dihydro-2′-deoxyguanine and lipid peroxidation-inducing agents such as 4-hydroxy-2-nonenal are highly elevated in synovial fluid and tissue of RA and PsA patients and correlate with angiogenic growth factor expression.2 4 Oxidative damage can induce cyclooxygenase-2, angiogenesis, MMP9/13 expression, can regulate antiapoptotic pathways and can alter nuclear factor kappa B (NF-κB) signalling, key processes involved in the pathogenesis of inflammatory arthritis.5,,8 Furthermore antioxidant treatments can improve disease progression in animal models of inflammatory arthritis.9 10

Mitochondrial DNA (mtDNA) lacks the intrinsic repair mechanisms of nuclear DNA and is thought to be more sensitive to increased oxidative damage and thus more vulnerable to high mutation rates.11 Mutation of the mitochondrial genome in genes encoding proteins for subunits of mitochondrial respiratory chain complexes IV, ribosomal RNA and transfer RNA have been associated with neurodegenerative diseases and cancer.11,,13 Inherited mitochondrial disease is considered a distinct entity predisposing individuals to heritable conditions such as chronic progressive external ophthalmoplegia, lebers hereditary optic neuropathy and mitochondrial myopathy, encephalopathy, lactic acidosis and stroke syndrome.14 Furthermore, type-1 tumour necrosis factor receptor (TNFR1) mutant cells show evidence of altered mitochondrial function paralleled by increased oxidative capacity and mitochondrial ROS generation, blockade of which resulted in reduced proinflammatory cytokine expression. This suggests a potential role for mitochondrial ROS in the regulation of cytokine pathways in inflammatory diseases.15

In this study, we demonstrate a significant increase in mtDNA mutation frequency in the inflamed synovium compared with normal synovium. We demonstrate for the first time a significant relationship between mtDNA mutation frequency with macroscopic/microscopic measures of inflammation and with key proinflammatory mediators: TNFα and interferon γ (IFNγ). Finally, we demonstrate in vivo and in vitro that TNFα significantly alters mtDNA mutation frequency and mitochondrial function.

Materials and methods

Patient selection and arthroscopy

Fifty inflammatory arthritis patients with active disease (13 men, 37 women) were recruited prior to commencing biologic therapy (RA, n=39; PsA, n=11) fulfilling the criteria of the American College of Rheumatology16 and classification criteria for PsA,17 respectively, from the Rheumatology Clinic. Ten patients undergoing interventional arthroscopy for cruciate ligament tears were also recruited (eight men; two women; mean age 38 years (IQR 28–43)). All patients provided fully informed consent. The mean age of the inflammatory arthritis cohort was 52 years (IQR 41–60) and the clinical characteristics for these patients included tender joint count 13 (3–21), swollen joint count 11 (7–18), erythrocyte sedimentation rate 37 mm/h (27–65), C reactive protein 29 mg/l (9–57), disease activity score (DAS) 28 5.76 (3.9–6.32) and disease duration 19 months (5–57). Sixty-five per cent of patients were rheumatoid factor positive and 77% anticitrullinated protein antibody positive. A subgroup of 15 patients (RA, n=8; PsA, n=7) were assessed at baseline and 3 months post-TNF inhibitor (TNFi) therapy; their median age was 52 years (IQR 55–49), 33% were rheumatoid factor and anticyclic citrullinated protein positive. Median baseline DAS28 was 4.76 (5.2–3.4). Seven patients achieved a low disease activity state (DAS28<3.2) after 3 months of treatment according to van Gestel et al.18

Arthroscopy, macroscopic assessment and synovial biopsy

Under local anaesthetic, arthroscopy of the inflamed knee was performed using a Wolf 2.7 mm needle arthroscope (Richard Wolf, Vernon Hills, Illinois, USA) as previously described.2 Macroscopic synovitis and vascularity were scored on a visual analogue scale (0–100 mm).19 Synovial tissue biopsies were obtained by 2 mm grasping forceps, under direct visualisation, embedded in mounting media for immunohistochemical analysis or snap frozen in liquid nitrogen for mitochondrial mutagenesis analyses. Paired synovial fluid was obtained and stored at −70°C for cytokine analysis. Control synovial biopsies were also obtained from 10 patients undergoing interventional arthroscopy for cruciate ligament tears.

Mitochondrial random mutation capture assay (RMCA)

Levels of mitochondrial point mutations in snap frozen synovial biopsies were analysed in a blinded fashion using mitochondrial RMCA as described previously.20 Biopsies were homogenised (Precellys 24, Stretton Scientific Ltd, UK) in 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 20 mM EDTA, 0.5% sodium dodecyl sulphate buffer and digested with proteinase K (Sigma–Aldrich) at a final concentration of 0.2 mg/ml and incubated overnight at 56°C. The mtDNA was extracted using phenol–chloroform–isoamyl alcohol (25:24:1 by volume; Sigma) added in a 1:1 ratio with the lysed tissue, mixed thoroughly by shaking and centrifuged at >400 000×g for 10 min. The aqueous phase was gently removed from the top of the solution without disturbing the interphase. The aqueous solution was again mixed with phenol–chloroform–isoamyl alcohol in a 1:1 ratio and re-extracted. One-tenth volume of 3 M sodium acetate was added and the samples were precipitated with 2–2.5 volumes of ethanol. The DNA samples were resuspended in 50 μl 10 mM Tris-HCl. Ten micrograms of mtDNA were digested with 100 units of TaqαI restriction enzyme (New England BioLabs), 1×bovine serum albumin and a TaqαI-specific digestion buffer (10 mM Tris-HCl, 10 mM MgCl2, 100 mM NaCl, pH 8.4) for 10 h; 100 units of TaqαI being added to the reaction mixture every hour. PCR amplification was performed in 25 μl reactions, containing 12.5 μl 2×SYBR Green Brilliant Mastermix (Stratagene), 0.1 μl uracil DNA glycosylase (New England Biosciences), 0.7 μl of 10 pM/μl forward and reverse primers (Integrated DNA Technologies) and 6.7 μl H2O. The samples were amplified using a Roche Lightcycler 480 using the following protocol: 37°C for 10 min and 95°C for 10 min followed by 45 cycles of 95°C for 15 s, 60°C for 1 min. Samples were held at 72°C for 7 min and, following melt curve analysis, immediately stored at −80°C. The primer sequences used for identifying random mtDNA mutations at the 12S RNA site (bp 1215–1218) were 5′-cctcaacagttaaatcaacaaaactgc-3′ (forward) and 5′-gcgcttactttgtagccttca-3′ (reverse); the primer sequences used for mtDNA copy number quantification were 5′-acagtttatgtagcttacctcc-3′ (forward) and 5′-ttgctgcgtgcttgatgcttgt-3′ (reverse). All PCR products were sequenced to identify the mutation at the TaqαI recognition site (UW High Throughput sequencing facility, Seattle, USA).


Seven-micrometre OCT (Optimal Cutting Temperature media) sections were allowed to reach room temperature, fixed in acetone for 10 min and air-dried. A routine three-stage immunoperoxidase labelling technique incorporating avidin-biotin immunoperoxidase complex (DAKO, Glostrup, Denmark) was used. Sections were incubated with primary mouse monoclonal anti-CD68 and anti-CD3 antibodies (DAKO, Glostrup, Denmark) at room temperature for 1 h. Sections were also incubated with an appropriate isotype matched mouse monoclonal antibody as a negative control. Colour was developed in solution containing diaminobenzadine tetrahydrochloride (Sigma-Aldrich, St Louis, Missouri, USA), 0.5% H2O2 in phosphate buffered saline (pH 7.6). Slides were counterstained with haematoxylin and mounted. Slides were analysed using a well-established semiquantitative scoring method ranging from 0 to 4 (0=no staining, 1=<25%, 2=25–50%, 3=50–75%, 4=>75% staining).1 4 21 22 Sections were scored separately for lining and sublining layers and results were expressed as the mean scores.1 4

MSD multi-array technology

The four-spot MSD human proinflammatory cytokine ultra sensitive kit, containing cytokines TNFα, IL-6, IFNγ and IL-1β (Meso Scale Discovery, Maryland, USA, Cat# K15009C-1) was used for the analysis of paired synovial fluid samples. Quantification was assessed using the Sector Imager 2400 (Meso Scale Delivery) instrument and software.

Primary RA and PsA synovial fibroblast culture

RA and PsA synovial biopsies were digested with 1 mg/ml collagenase type 1 (Worthington Biochemical, Freehold, New Jersey, USA) in RPMI medium (Gibco-BRL, Paisley, UK) for 4 h at 37°C in humidified air with 5% CO2. Dissociated cells were grown to confluence in RPMI 1640, 10% FCS (Gibco-BRL), 10 ml of 1 mmol/l HEPES (Gibco-BRL), penicillin (100 units/ml; Biosciences), streptomycin (100 units/ml; Biosciences) and fungizone (0.25 ug/ml; Bioscience) before passaging. Cells were used between passages 4 and 8. K4IM cells, an immortalised normal human synoviocyte cell line (kind gift, Dr Evelyn Murphy), were cultured as above and used between passages 35 and 38. RA synovial fibroblast cells (RASFCs), PsA synovial fibroblast cells (PsASFCs) and K4IM were seeded onto 96-well plates and into T25 flasks and incubated with TNFα (10 ng/ml) for 24 h. mtDNA mutations, ROS, mitochondrial membrane potential (MMP) and mitochondrial mass (MM) were assessed as follows:

In vitro mitochondrial dysfunction

To determine the frequency of mtDNA point mutations, RASFC pellets were obtained following TNFα stimulation, were digested, purified mtDNA was extracted and RMCA was performed as outlined above.

To measure the level of ROS, following TNFα stimulation, cells were washed twice with a buffer (130 mM NaCl, 5 mM KCl, 1 mM Na2HPO4, 1 mM CaCl2, 1 mM MgCl2 and 25 mM Hepes, pH 7.4). Cells were loaded with 5 μM 2,7-dichlorofluorescein diacetate (DCFH-DA) (Invitrogen) for 40 min at 37°C. DCFH-DA is a non-fluorescent molecule which diffuses into the cells where it is deacetylated and rapidly oxidised to highly fluorescent 2,7-dichlorofluorescein in the presence of the generated ROS. DCF emits a fluorescent signal of the product which is linearly related to the intracellular hydrogen peroxide concentration. To measure MMP, following TNFα stimulation, cells were washed and loaded with 5 μM rhodamine-123 (Sigma) for 40 min at 37°C. Rhodamine-123 is taken up selectively by mitochondria which is dependent on MMP. Following 40 min incubation, ROS and MMP probes were removed, cells were washed and analysed using the Spectra Max Gemini System. DCFH-DA and rhodamine-123 were excited at 485 nm, and fluorescence emission at 538 nm was recorded. To measure MM, following TNFα stimulation, cells were incubated with Green-Fluorescent MitoTracker dye (Invitrogen). After incubation for 45 min, cells were visualised using a fluorescence microscope. The probe has absorption and emission peaks at 490 and 516 nm, respectively, and the fluorescence intensity is proportional to the MM. Mean fluorescence values from four wells for each condition were obtained. Mitochondrial dysfunction assays were normalised to cell number.

Statistical analysis

SPSS15 system for Windows was used for statistical analysis. Non-parametric Wilcoxon signed rank, Spearman correlations with Bonferroni corrections and Mann–Whitney U test were used for analysis. p<0.05 was determined as statistically significant.


Mutation of mtDNA in the synovium of active RA and PsA patients

A significant increase in mtDNA mutations was demonstrated in RA and PsA synovial tissue (mean 4.8×10−4) compared with normal synovium (7.3×10−5) (p<0.013). When mtDNA mutations were analysed separately for RA and PsA synovium, significantly higher mutation frequency was demonstrated for both RA and PsA compared with normal synovium (p=0.017 and p=0.041) (figure 1A). No significant difference in mutation frequency was observed between RA and PsA (p=0.673) (figure 1A). The frequency of mtDNA mutations did not correlate with age in normal synovium (r=0.006, p=0.98) or in age-matched RA and PsA synovium (r=0.236, p=0.511) demonstrating that the increase in mtDNA mutation frequency was independent of age. The spectrum of mutations identified demonstrated substitution of purines for pyrimidines in 49% of mutations and substitution of pyrimidines for purines in 51% of mutations (figure 1B).

Figure 1

(A) Box plots representing significant increase in frequency of mitochondrial DNA mutations between control, inflamed rheumatoid arthritis (RA) and psoriatic arthritis (PsA) synovium. Boxes represent the 25th to 75th percentiles, lines within the boxes represent the median, and lines outside the boxes represent the 10th and 90th percentiles. * p< 0.05, statistically significant. (B) Bar chart representing distribution of mitochondrial DNA mutation types most frequently seen in active RA and PsA synovial tissue. Purines were substituted for pyrimidines in 49% of cases whereas the opposite was true in 51% of cases.

Mutations in mtDNA correlate with macroscopic measures of inflammation

Representative images of macroscopic synovitis and vascularity from a patient with low mtDNA mutations (figure 2A,D) versus high mtDNA mutations (figure 2B,E) are shown. Significant correlation was demonstrated between mtDNA mutation frequency and macroscopic synovitis (r=0.52, p=0.016) and vascularity (r=0.55, p=0.01) suggesting that mtDNA mutation frequency is associated with the degree of inflammation in the joint (figure 2C,F).

Figure 2

Representative images of macroscopic synovitis from a patient with low mitochondrial DNA (mtDNA) mutations (A) versus high mtDNA mutations (B). (C) The correlation between point mutation frequency and macroscopic synovitis. Representative images of macroscopic vascularity from a patient with low mtDNA mutations (D) versus high mtDNA mutations (E). (F) The correlation between point mutation frequency and macroscopic vascularity. r=Spearman's correlation coefficient, *p<0.05 is statistically significant.

Mutations in mtDNA correlate with mediators of inflammation

Synovial tissue mtDNA mutation frequency significantly correlated with synovial fluid expression of TNFα (r=0.74, p=0.024) and IFNγ (r=0.718, p=0.039) demonstrating a strong relationship between mitochondrial instability and proinflammatory pathways (table 1, see online supplementary figure 1). Associations with CD68 sublining (r=0.5, p=0.052) and CD3 sublining layer expression (r=0.46, p=0.08) were observed, although following Bonferroni correction significance was not reached (figure 3).

Figure 3

Representative image of CD68+ sublining layer (SL) and CD3+ SL expression in inflamed synovial tissue from a patient with high frequency of mitochondrial DNA (mtDNA) mutation versus a patient with low frequency of mtDNA mutation.

Table 1

Correlation of mtDNA mutation frequency with pro-inflammatory cytokines

The effect of TNFi therapy on synovial mtDNA mutations

Synovial mtDNA mutations were decreased in patients following TNFi biologic therapy (n=15); however significance was only achieved in patients with DAS28 score<3.2, where 6/7 patients showed a parallel decrease in clinical score and synovial mtDNA mutation frequency (p<0.05) (figure 4). A corresponding decrease in C reactive protein from 19 to 4 (p<0.05), swollen joint count from 3 to 0 (p<0.05), visual analogue scale from 60 to 20 mm (p<0.05) and DAS28 from 4.3 to 2.5 (p<0.05) was also demonstrated. Furthermore, significant correlations between mtDNA mutation frequency and DAS28 (r=0.653, p<0.05) and sublining CD3 expression (r=0.485, p<0.05) was demonstrated. No significant association was observed in patients whose DAS28 remained >3.2. No significant difference was observed in baseline DAS28 in patients who achieved a DAS28<3.2 versus DAS28>3.2 post-TNFi therapy (p=0.325).

Figure 4

Box plots representing a change in frequency of mitochondrial DNA mutation in patients with disease activity score (DAS) 28 >3.2 (left) or in patients with DAS 28 <3.2 (right) at baseline and after 3 months of therapy. Baseline mitochondrial DNA mutations (grey bars) and post-TNFi therapy (white bars). Boxes represent the 25th to 75th percentiles, lines within the boxes represent the median, and lines outside the boxes represent the 10th and 90th percentiles. *p<0.05 is statistically significant.

In vitro TNFα induces mitochondrial dysfunction in primary RASFC.

In RASFC, TNFα significantly increased mtDNA mutation frequency from 6.62×10−6 to 3.12×10−5 (p<0.05) (figure 5A). In addition TNFα induced ROS production (p=0.007), MMP (p=0.045) and mitochondrial mass (p=0.023) compared with unstimulated control cells (figure 5B–D). In PsASFC, mtDNA mutations increased from 6.7×10−6 to 9.9×10−6 in response to TNFα (n=3, NS). In K4IM cells, basal mtDNA mutations were lower than RASFC/PsASFC, however increased from 5.02×10−6 to 8.52×10−6 following TNFα stimulation (p<0.05). In PsASFC and K4IM TNFα significantly induced MMP (p<0.05), ROS production (p<0.05) and mitochondrial mass (p<0.05) (see online supplementary figures 2, 3).

Figure 5

(A–D) Box plots representing mitochondrial DNA (mtDNA) mutation frequency, reactive oxygen species (ROS), mitochondrial mass (MM) and mitochondrial membrane potential (MMP) quantification in response to TNFα stimulation in rheumatoid arthritis synovial fibroblast cells (n=5). Boxes represent the 25th to 75th percentiles, lines within the boxes represent the median and lines outside the boxes represent the 10th and 90th percentiles. *p<0.05 is statistically significant.


In this study, we examine mitochondrial mutagenesis in patients with inflammatory arthritis and assess its relationship to inflammatory mechanisms in vivo and in vitro. We demonstrate that mitochondrial random mutations are significantly increased in inflammatory arthritis compared with normal synovial tissue. mtDNA mutations significantly correlate with level of inflammation at the macroscopic level. In particular, we demonstrate a significant association between mtDNA mutation frequency and macroscopic/microscopic scores of inflammation and pro-inflammatory cytokines: TNFα, IFNγ. Following TNFi biologic therapy, we demonstrate a significant decrease in synovial mtDNA mutations. Finally in vitro, we show that TNFα stimulation drives mitochondrial dysfunction in primary RASFC and PsASFC along with mtDNA mutations.

In this study, we used a newly validated mitochondrial RMCA which relies on single-molecule amplification to detect rare mutations among millions of wild-type bases and analyses mitochondrial mutagenesis at single base pair resolution and detects one mutation among 10 million wild-type bases.20 Many mitochondrial point mutations associated with mitochondrial dysfunction have been characterised.23 By assaying the random mutation frequency of a phenotypically neutral locus, 12S RNA, we estimate the likelihood at which any specific mitochondrial mutation and associated dysfunction, might occur. By using a TaqI site that does not span a protein coding region of the mitochondrial genome, we are able to extrapolate the extent of mutagenesis throughout the mitochondrial genome (ie, mutagenesis unaffected by phenotypic selection).

We demonstrated increased mtDNA mutation frequency in the inflamed synovium compared with normal controls, levels of which correlate with clinical measures of disease activity and with the local inflamed synovial tissue microenvironment. Our data is supported by previous studies showing increased levels of DNA damage and lipid peroxidation in patients with inflammatory arthritis.2 4 24 Significantly higher levels of lipid peroxidation and depolarised mitochondria in RA peripheral blood mononuclear cells have been shown to correlate with disease activity.24 Increased expression of 8-oxo-7,8-dihydro-2′-deoxyguanine and 4-hydroxy-2-nonenal have been demonstrated in RA synovial tissue and serum, and correlate with disease activity and angiogenic factors.2 4 Increased clonal mtDNA mutation frequency in the MT-ND1 gene for mitochondrially encoded NADH-dehydrogenase-1 has been detected in RASFC.15 Furthermore potential mutation sites in the major histocompatibility complex epitope in RA patients, but not osteoarthritis, have been identified suggesting mtDNA may become antigenic and drive immune mediated responses. In mouse fibroblasts, TNFR1 mutant cells show evidence of altered mitochondrial function, resulting in increased oxidative capacity, mitochondrial ROS generation and proinflammatory cytokines.15 In addition, in vivo synovial tissue oxygen levels correlate with oxidative damage and mitochondrial dysfunction.25

Mitochondrially derived ROS can induce TNFα cytotoxicity and may mediate the activation of transcriptional factor NF-κB which in turn can stimulate mitochondrial NADPH oxidase.26 27 TNFR1 mutant cells exhibit altered mitochondrial function, increased production of IL-6, TNFα, IL-8 and phosphorylation of mitogen-activated protein kinase pathways.15 Altered mitochondrial membrane permeability and Bcl-X overexpression leads to resistance to TNF cytoxicity.28 Combination of IL-1β and IFNγ can induce mitochondrial Bax translocation, cytochrome c release and caspase-3 cleavage.29 In addition, ROS can mediate IL-1- induced experimental arthritis,30 and cyclooxygenase-2 expression in RASFC.31 CD3+ T cells and CD68+ macrophages are potent producers of cytokines critical to synovial inflammation.32 33 In this study, reduction of synovial mtDNA mutation frequency and its association with measures of clinical disease activity and microscopic CD3+ cells following TNFi therapy, strengthens the hypothesis of a pathological link between mitochondrial dysfunction and pro-inflammatory pathways.

We have shown using RASFC that TNFα significantly increases mtDNA mutation frequency, coupled with increased ROS production, MMP and MM, thus recapitulating the in vivo responses to TNFi therapy. This is consistent with mice models showing interactions between oxidatively damaged mtDNA, increased NF-κB activity and TNFα production.30 TNFα increases mitochondrial ROS production and induces lipid-derived aldehyde formation.26 34,,38 High mutations were demonstrated at baseline in TNFi non-responders suggesting TNF independent pathways may be involved in mtDNA mutations. We and others previously showed strong associations between mitochondrial dysfunction and synovial tissue oxidative stress, angiogenic growth factors and cytokines.4 29,,31 Furthermore, hypoxia can increase mtDNA mutations in primary RA synoviocytes.25

Direct functional consequences of mitochondrial random mutations in inflammation are unclear; however, they may have an important role in the regulation of the innate immune response.39 Zhou et al, have demonstrated that the NLRP3 inflammasome can sense mitochondrial dysfunction, altering the inflammatory response.40 Furthermore, mitochondrial dysfunction regulates IL-1 and IL-18 secretion from macrophages in response to lipopolysaccharide.41 Additionally in an experimental arthritis model, manipulation of mitochondrial pathways reduces severity of cartilage lesions and synovial inflammation.42


In summary, we have shown, for the first time that mtDNA mutations are significantly higher in the inflamed synovium compared with the control. We have shown significant associations between mtDNA mutations and macroscopic/microscopic measures of synovial tissue inflammation and with local production of pro-inflammatory cytokines. Furthermore, we demonstrate both in vitro and in vivo, that DNA mutagenesis and mitochondrial function are altered by TNFα regulation. These data further support the concept of complex interplay between inflammatory pathways, oxidative damage and oxygen metabolism in the pathogenesis of inflammatory arthritis.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Funding This study was funded by the Centocor Newman Fellowship, the Health Research Board of Ireland and EU FP6 AutoCure.

  • Competing interest None.

  • Patient consent Informed consent obtained.

  • Ethics approval This study was conducted in accordance with the Declaration of Helsinki and approved by the St Vincent's University Hospital medical research and ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.