Targeting mitochondrial oxidative stress with MitoQ reduces NET formation and kidney disease in lupus-prone MRL-lpr mice

Objectives Recent investigations in humans and mouse models with lupus have revealed evidence of mitochondrial dysfunction and production of mitochondrial reactive oxygen species (mROS) in T cells and neutrophils. This can provoke numerous cellular changes including oxidation of nucleic acids, proteins, lipids and even induction of cell death. We have previously observed that in T cells from patients with lupus, the increased mROS is capable of provoking oligomerisation of mitochondrial antiviral stimulator (MAVS) and production of type I interferon (IFN-I). mROS in SLE neutrophils also promotes the formation of neutrophil extracellular traps (NETs), which are increased in lupus and implicated in renal damage. As a result, in addition to traditional immunosuppression, more comprehensive treatments for lupus may also include non-immune therapy, such as antioxidants. Methods Lupus-prone MRL-lpr mice were treated from weaning for 11 weeks with the mitochondria-targeted antioxidant, MitoQ (200 µM) in drinking water. Mice were then assessed for ROS production in neutrophils, NET formation, MAVS oligomerisation, serum IFN-I, autoantibody production and renal function. Results MitoQ-treated mice manifested reduced neutrophil ROS and NET formation, decreased MAVS oligomerisation and serum IFN-I, and reduced immune complex formation in kidneys, despite no change in serum autoantibody. Conclusions These findings reveal the potential utility of targeting mROS in addition to traditional immunosuppressive therapy for lupus.


AbstrAct
Objectives Recent investigations in humans and mouse models with lupus have revealed evidence of mitochondrial dysfunction and production of mitochondrial reactive oxygen species (mROS) in T cells and neutrophils. This can provoke numerous cellular changes including oxidation of nucleic acids, proteins, lipids and even induction of cell death. We have previously observed that in T cells from patients with lupus, the increased mROS is capable of provoking oligomerisation of mitochondrial antiviral stimulator (MAVS) and production of type I interferon (IFN-I). mROS in SLE neutrophils also promotes the formation of neutrophil extracellular traps (NETs), which are increased in lupus and implicated in renal damage. As a result, in addition to traditional immunosuppression, more comprehensive treatments for lupus may also include non-immune therapy, such as antioxidants.
Methods Lupus-prone MRL-lpr mice were treated from weaning for 11 weeks with the mitochondria-targeted antioxidant, MitoQ (200 µM) in drinking water. Mice were then assessed for ROS production in neutrophils, NET formation, MAVS oligomerisation, serum IFN-I, autoantibody production and renal function. Results MitoQ-treated mice manifested reduced neutrophil ROS and NET formation, decreased MAVS oligomerisation and serum IFN-I, and reduced immune complex formation in kidneys, despite no change in serum autoantibody . Conclusions These findings reveal the potential utility of targeting mROS in addition to traditional immunosuppressive therapy for lupus.

IntROduCtIOn
SLE is a complex and heterogeneous autoimmune syndrome characterised by numerous abnormalities, including various autoantibodies, the appearance of unusual CD4-CD8-TCR-αβ + cells, altered metabolism of lymphocytes, a type I interferon (IFN-I) gene signature in peripheral blood mononuclear cells (PBMC), increased formation of neutrophil extracellular traps (NETs), and deposition of immunoglobulins and complement at renal glomeruli. [1][2][3][4][5][6][7][8][9][10] Each of these immune abnormalities may provoke different aspects of the disease, and each also may be driven by different environmental and genetic aberrations. Recent work has also revealed the importance of non-immune factors, such as oxidative stress, in the development of endorgan damage in SLE, shifting the paradigm of SLE pathogenesis from that of a disease provoked solely by a disturbed immune system. 1 7 Thus, fully effectively treatment of SLE may require inhibiting pathways in addition to those targeted by traditional immunosuppressive medications.
Neutrophil activation in response to glomerular immune complexes contributes to lupus glomerulonephritis, in part through the production of mitochondrial reactive oxygen species (mROS) that can directly injure tissue through oxidation of lipids, proteins, DNA and induction of apoptosis. 11 12 We have shown previously that immune complex activation of neutrophils can also lead to the formation of NETs, which can release oxidised genomic and mitochondrial DNA and promote the production of IFN-I. 10 12 Furthermore, a subset of low-density granulocytes from patients with SLE has an enhanced capacity to form NETs in an mROS-dependent manner, and these structures have been observed in the kidneys of patients with lupus nephritis as well as in other lupus-affected tissues. 11 13 We have also previously observed that T lymphocytes from patients with SLE manifest mitochondrial dysfunction as evidenced by enlarged mitochondria and elevated mROS that can induce oxidative damage. 14 The mROS triggers the spontaneous oligomerisation of the mitochondrial antiviral stimulator (MAVS) protein and downstream IFN-I Lupus Science & Medicine production. 15 This was likely due in part to induction of cysteine disulfide bonding in the caspase recritment domain (CARD) of MAVS, since a C79F polymorphism of MAVS known to be associated with milder SLE 16 did not efficiently oligomerise nor induce IFN-I with mROS. 15 In addition, we observed that oligomerised MAVS is reversed in vitro in the presence of the mitochondrial antioxidant, MitoQ. 15 Consequently, mROS and resulting oxidative damage emerge as possible driving forces in certain aspects of SLE, and hence represent a target for therapy.
The process that promotes the mitochondrial abnormalities in human SLE T cells is unknown. Hence, we examined whether a similar phenomenon of enlarged mitochondria and spontaneous MAVS oligomerisation might occur in any of the T-cell subsets of lupus-prone MRL-lpr mice. T cells in these mice accumulate in large numbers in lymphoid organs through dysregulated homeostatic proliferation that is enhanced in the absence of the death receptor, Fas (CD95). 17 The dysregulation of T cells includes the emergence with age of an increasingly large proportion of polyclonal CD4 -CD8 -TCR-αβ + cells that derive from CD8 + precursors during homeostatic proliferation. 17 A CD4 -CD8 -TCR-αβ + subset also occurs in human SLE. 3 4 Similar to human SLE T cells, the lpr CD4 -CD8 -TCR-αβ + subset also manifested enlarged mitochondria and spontaneous MAVS oligomerisation. We thus investigated further the ability of the mitochondriatargeted antioxidant MitoQ in vivo to reverse mROS and NET formation, MAVS oligomerisation, as well as to test its therapeutic potential on lupus disease manifestations in MRL-lpr mice.

Mice
Mice were bred and housed in the Association for Assessment and Accreditation of Laboratory Animal Care International-approved animal facilities of The University of Vermont Larner College of Medicine. Original breeding pairs of MRL/MpJ-Fas lpr (MRL-lpr) mice were obtained from Jackson Laboratory (Bar Harbor, Maine, USA). All breeding and animal studies were conducted in accordance with the policies of The University of Vermont's Animal Care and Use Committee.

MitoQ treatment
Mice were weaned at 4 weeks and placed on either drinking water alone or 200 µM MitoQ (MitoQ, Auckland, NZ) in drinking water. Bottles were changed weekly. After 11 weeks of treatment, mice were euthanised and kidneys harvested for histological analysis. Brachial and axillary lymph nodes were assessed for cell number and lymphocyte subsets were determined by flow cytometry. Neutrophils were purified from bone marrow as previously described. 10 Serum was obtained for autoantibodies, creatinine and blood urea nitrogen (BUN). Urine was obtained for creatinine measurement.
For direct staining, single cell suspensions were washed with cold (4°C) PBS and then incubated with Live Dead Fixable Blue Stain (Life Technologies-Molecular Probes, Eugene, Oregon, USA) in PBS for 30 min at 4°C. The cells were washed with cold PBS containing 1% w/v bovine serum albumin (BSA) fraction V (Sigma) (PBS/1% BSA) and then incubated with a mixture of rat IgG and hamster IgG (50 µg/mL each) for 30 min at 4°C. After washing, the cells were incubated with the appropriate antibodies in PBS/1% BSA, washed and fixed with freshly made 1% v/v methanol-free formaldehyde (Ted Pella, Redding, California, USA) in PBS/1% BSA. Flow cytometry was performed on an LSR II (BD Biosciences) and the data were analysed using FloJo software (Tree Star, Ashland, Oregon, USA). Analysis of mitochondrial morphology by transmission electron microscopy Cells were fixed for 1 hour at 65°C in 2% paraformaldehyde and 2.5% glutaraldehyde (Polysciences, Warrington, Pennsylvania, USA) in 100 mM sodium cacodylate buffer (pH 7.2). Samples were washed in cacodylate buffer and then postfixed for 1 hour in 1% osmium tetroxide (Polysciences). Samples were then extensively rinsed in distilled H 2 O before undergoing en bloc staining for 1 hour with 1% aqueous uranyl acetate (Ted Pella). After several rinses in distilled H 2 O, the samples were dehydrated in a graded series of ethanol and then embedded in Eponate 12 resin (Ted Pella). Sections (95 nm in thickness) were cut with an Ultracut UC7 ultramicrotome (Leica Microsystems, Wetzlar, Germany), stained with uranyl acetate and lead citrate, and viewed on a JEOL 1400 transmission electron microscope (JEOL USA) equipped with an XR611 high-resolution, 11-megapixel mid-mount charge-coupled device camera (Advanced Microscopy Techniques, Woburn, Massachusetts, USA).

Animal models
semidenaturing detergent agarose gel electrophoresis (sdd-AGe) for the detection of MAVs oligomers SDD-AGE was performed according to a published protocol with minor modifications. 18 Briefly, mitochondria were resuspended in sample buffer (0.5× tris-borate EDTA (TBE), 10% glycerol, 2% sodium dodecyl sulfate (SDS) and 0.0025% bromophenol blue) and loaded onto a vertical 1.5% agarose gel. After electrophoresis in running buffer (1× TBE, 0.1% SDS) for 35 min with a constant voltage of 75 V at 4°C, proteins were transferred to polyvinylidene difluoride (PVDF) membranes with a Trans-Blot Turbo Transfer System in preparation for western blot analysis. PVDF membranes were blocked in Tris-buffered saline and 5% non-fat powdered milk and analysed with MAVS-specific antibody (Santa Cruz Biotechnology, Dallas, Texas, USA). Immunoreactive proteins were visualised with horseradish peroxidaselabelled conjugates (Jackson ImmunoResearch, West Grove, Pennsylvania, USA) and developed with Clarity Western ECL Substrate (Bio-Rad, Hercules, California, USA). Chemiluminescence was detected and recorded with a Bio-Rad Chemidoc instrument. Densitometric measurements were performed in Image Lab image acquisition and analysis software (Bio-Rad).

Quantification of nets and mROs in bone marrow neutrophils
The isolation of bone marrow-derived neutrophils and quantification of NETs and mROS were performed as previously described. 10 Briefly, hindlimb marrow neutrophils were purified by Percoll gradient. Cells were seeded in a 96-well plate (2 00 000 cells/100 µL/well) in triplicates for each dye and allowed to form NETs in the presence of SYTOX (externalised DNA, 1 µM final concentration), Quant-It Picogreen (total DNA stock solution diluted 1:250) and MitoSox (200 ng/mL) (all from Thermo Fisher, Waltham, Massachusetts, USA). Fluorescence was measured at different time points for each dye, at the earliest time point 485/520 (Picogreen), 1 hour 510/580 (MitoSox) and 2 hour 486/520 (SYTOX), using a FLUOstar Omega BMG Labtech (Cary, North Caroline, USA) plate reader. Picogreen measurement was used as the initial number of cells or total DNA.

Autoantibody quantification
Serum concentrations of autoantibodies were determined using commercially available ELISA kits (Alpha Diagnostic International, Texas, USA). Serum was diluted (1:125) in non-specific binding (NSB) buffer and the assay done following manufacturer's instructions.

Assessment of kidney histology and function
Renal immune complex deposition was quantified as previously described 10   isothiocyanate (FITC)-antimurine C3 antibody (Immunology Consultants Laboratories, Portland, Oregon USA). Nuclei were stained with Hoechst (1:500; Life Technologies, Carlsbad, California, USA). For quantification, three random images were obtained from each stained frozen section. The images were analysed with Image J software selecting the glomerular compartment to quantify mean pixels for each fluorescence channel used.
To quantify serum creatinine and eliminate the influence of chromogens in mouse serum that interfere with the classic Jaffe method for creatinine detection, an high performance liquid chromatography assay was used as previously described. 19 Briefly, 5 µL serum were treated with 0.5 mL acetonitrile, centrifuged at 4°C at 13 000×g for 20 min, and supernatants were dried by SpeedVac and resuspended in mobile phase (5 mM sodium acetate, pH 5.1). Duplicates were run on a 100×4.1 mm PRP-X200 column (Hamilton, Reno, Nevada, USA) and isocratically eluted at 2 mL/min in an Agilent 1100 System, with ultraviolet detection at 234 nm. Absolute quantitation was determined with a standard curve of 2-50 ng creatinine (r 2 =0.999).

statistical analysis
Statistical analyses were performed using the graphing software Prism V.7 (GraphPad Software, La Jolla, California, USA). The following statistical tests were used: paired and unpaired t-test when comparing two conditions, one-way analysis of variance (ANOVA) with Tukey's test for correction for multiple comparisons when comparing multiple conditions and two-way ANOVA with Sidak test for correction for multiple comparisons when comparing multiple variables across multiple conditions. All data met the assumptions of the statistical tests used and variation among the compared groups was similar.

MRL-lpr Cd4 -Cd8 -tCR-αβ + cells have enlarged mitochondria, increased oxygen consumption and glycolysis
Our previous observations in human SLE T cells revealed that they manifest enlarged mitochondria, mROS production and spontaneous MAVS oligomerisation. 5 14 15 We thus examined lupus-prone MRL-lpr mouse T cells for similar features. Initial analysis revealed that lpr CD8 + T cells, the precursors of the CD4 -CD8 -TCR-αβ + T cells, 17 20 21 contained relatively low mitochondrial mass, Animal models  using MitoTracker and flow cytometry, whereas the CD4 -CD8 -TCR-αβ + T cells had markedly higher mitochondrial mass relative to the CD8 + T cells ( figure 1A). Further analysis by electron microscopy revealed that, similar to human SLE T cells 5 14 the lpr CD4 -CD8 -TCR-αβ + T cells contained very large and rounded mitochondria, in contrast to the more typical elongated mitochondria of the CD8 + subset ( figure 1B). This paralleled greater rates of oxygen consumption and glycolysis in the CD4 -CD8 -TCR-αβ + subset, as detected by Seahorse extracellular flux analysis ( figure 1C). The increased aerobic glycolysis of CD4 -CD8 -TCR-αβ + T cells is consistent with the known rapid proliferation by this subset in vivo. 22 This is paralleled by increased spontaneous cell death of the CD4 -CD8 -TCR-αβ + T cells compared with the CD4 + and CD8 + T cell subsets (figure 1D), consistent with previous observations that high levels of glycolysis in T cells, including CD4 -CD8 -TCR-αβ + T cells, drives high levels of active caspase-3, rendering them prone to cell death. 23 24 Such increased cell death could contribute to the inflammatory response in these mice.
Given these parallels between T cells from human SLE and the CD4 -CD8 -TCR-αβ + T cells of MRL-lpr mice, we examined whether they also manifested evidence of spontaneous MAVS oligomerisation and activation of IFN-I genes. Indeed, the CD4 -CD8 -TCR-αβ + (B220 + ) subset contained MAVS oligomers, which were diminished in the CD4 + and CD8 + (B220 − ) fraction and non-existent in wild-type mice ( figure 2A). This paralleled the upregulation of several IFN-I-stimulated genes in the CD4 -CD8 -TCR-αβ + T cells compared with the CD8 + precursors (figure 2B). Consistent with these findings, MRL-lpr mice had higher levels of serum IFNα compared with wild-type mice (figure 2C). Collectively, these findings suggest that mROS may drive MAVS oligomerisation in lpr CD4 -CD8 -TCR-αβ + T cells as it does in human SLE. We thus considered that mROS and possible oxidative damage might also drive some of the disease manifestations in MRLlpr mice, and elected to treat the mice in vivo with the mitochondria-targeted antioxidant MitoQ.
In vivo MitoQ treatment of MRL-lpr mice reduces ROs production and net formation by neutrophils Increased production of mROS and NET formation has been observed in neutrophils of patients with SLE. 10 11 As NET formation in human lupus is, at least in part, driven by mROS production, we initially examined whether the mitochondria-targeted antioxidant, MitoQ, would reduce neutrophil mROS and NET formation. MitoQ contains an antioxidant, ubiquinonol, coupled to a triphenylphosphonium moiety containing three phenyl groups to promote membrane permeabilisation, combined with a central positively charged phosphorous that draws the compound toward negative charges. As mitochondria are about 150-170 mV negative compared with the cytosol, which is itself a further 30-60 mV negative compared Lupus Science & Medicine  with the extracellular environment, Mito compounds are concentrated 500-1000× in mitochondria. [25][26][27] Mice were administered MitoQ (200 µM) in their drinking water beginning at 4 weeks of age and continued for 11 weeks. No adverse effects of MitoQ (eg, weight, development) were observed in mice at this dose, consistent with other studies. 27-30 mROS production in neutrophils was initially examined by MitoSox fluorescence plate assay. This revealed a reduction in mROS production in MitoQ-treated mice, both spontaneously and following mROS stimulation with the calcium ionophore, A23187 (figure 3A). These findings were observed for both male and female MRL-lpr mice. Consistent with these findings, NET formation was also reduced with MitoQ treatment. In male mice, this was statistically significant for spontaneous NET formation, and in females this was the case following A23187 stimulation ( figure 3B).
Further analysis of metabolism in the neutrophils revealed reduced oxygen consumption with MitoQ, both at basal levels and following stimulation of respiration with the mitochondrial uncoupler FCCP (respiratory capacity) (figure 4A). In contrast, glycolysis was largely unchanged, except in the case of female mice where the glycolytic capacity, following addition of ATP synthetase inhibitor oligomycin, was decreased with MitoQ ( figure 4B). These results indicate that oral MitoQ modulates lupus neutrophil immunometabolism and reduces their ability to form NETs.

MitoQ in vivo reduces MAVs oligomerisation and serum IFn-I
CD4 -CD8 -TCR-αβ + T cells purified from MitoQ-treated mice revealed reduced levels of MAVS oligomerisation compared with mice receiving water only (figure 5A). This was paralleled by a reduction in serum IFNα (figure 5B). In contrast to the reduction of NET formation and MAVS oligomerisation, MitoQ treatment did not affect lymph node total cell numbers, the distribution of T-cell subsets, including CD4 -CD8 -TCR-αβ + T cells, or B cells (not shown), nor the titres of serum autoantibodies (figure 6) in male or female MRL-lpr mice. This was somewhat anticipated as MitoQ was not expected to alter events believed to be upstream of mROS production.
Reduced renal injury with MitoQ MRL-lpr mice develop immune complex-mediated kidney damage with age. 31 32 MitoQ treatment of female mice caused a reduction of the glomerular deposition of both C3 and IgG (figure 7). This was less apparent in male mice, but male control mice had less C3 and IgG deposition than female control mice, so there was a narrower window in male mice in which to observe a difference with MitoQ. These findings were paralleled by increased urine excretion of creatinine, although without a change in urine albumin (figure 8A) and was also accompanied by a reduction in serum BUN and creatinine in female mice ( figure 8B).

dIsCussIOn
The current findings suggest that inhibiting mROS and oxidative stress can mitigate certain aspects of autoimmune disease and organ damage in lupus-prone MRL-lpr mice. It has been appreciated for some time that T cells and neutrophils in patients with SLE manifest increased levels of ROS, 5 6 but the consequences of this on other molecular abnormalities, immune dysregulation or disease manifestations remain largely unexplored. The The IFN-I gene signature in PBMC is one of the hallmarks of SLE, 8 9 yet the mechanism is unclear. The IFN-I signature may be driven in part through engulfment of dead cells by macrophages and dendritic cells. 33 The increased rate of cell death of neutrophils in patients with SLE, as well as the increased cell death of CD4 -CD8 -TCR-αβ + T cells in MRL-lpr mice and human SLE T cells, may both contribute to the source of dead cells. The released nucleic acids can trigger both ROS and IFN-I production. This is clear from genetic deficiencies in DNA clearance (eg, three prime repair exonuclease (TREX)) which lead to interferonopathies. 34 35 Another source of ROS in human SLE comes from T cells, some of which manifest enlarged mitochondria. 5 14 We have shown previously that ROS can induce MAVS oligomerisation and IFN-I production in different cell types. 15 We now observe that, similar to human SLE, mitochondrial enlargement and spontaneous MAVS oligomerisation arises in the MRL-lpr CD4 -CD8 -TCR-αβ + subset. This has been attributed to an imbalance of mitochondrial biogenesis 36 and turnover by mitophagy. 37 Specifically, our previous studies of SLE T cells have demonstrated human retrovirus endogenous sequence (HRES)-1/rasrelated protein 4 (RAB4)-mediated depletion of dynamin related protein 1 (Drp1), a mediator of mitochondrial fission. 36 The accumulation of mitochondria in lupus T cells is sensitive to mTOR blockade, with clinical benefit in patients 38 and mice 39 with SLE.

Animal models
The CD4 -CD8 -TCR-αβ + subset, which also occurs in human SLE, including inflamed kidneys, 40 arises from recurrent homeostatic proliferation of CD8 + precursor T cells. 17 41 CD4 -CD8 -TCR-αβ + T cells also derive from antigen-specific CD8 + T cells in wild-type mice only when the antigen is presented as a self-antigen. 42 Gene expression profiling of these wild-type derived CD4 -CD8 -TCR-αβ + T cells has revealed upregulation of genes for programmed cell death 1 (PD-1), interleukin 17, IFNγ, C-X-C Motif Chemokine Ligand 2 (CXCL2) and downregulation of CD127, 43 exactly the same pattern observed in lpr CD4 -CD8 -TCR-αβ + T cells. 22 Thus, these unusual T cells are not unique to Fas-deficient lpr mice. These findings, in addition to our previous observations that T-cell homeostatic proliferation also leads to the upregulation of genes involved with cytolysis and inflammation 22 Lupus Science & Medicine adds further importance why it is critical to regulate the homeostatic expansion of T cells with Fas. In addition to these gene expression changes, our findings also suggest that part of the programme of dysregulated T-cell homeostatic proliferation in lpr mice includes mitochondrial enlargement leading to mROS production and MAVS oligomerisation.
MitoQ has shown therapeutic potential in a variety of conditions in which mROS has been implicated. These include animal models of Alzheimer's disease, 29 liver fibrosis, 30 NLR Family Pyrin Domain Containing 3 (NLRP3) inflammasome-mediated colitis, 28 cardiac ischaemia-reperfusion 44 and the metabolic syndrome in Apolipoprotein E (ApoE) -/mice. 27 In human trials, MitoQ prevented inflammatory damage in a phase II study in hepatitis C, 45 showed no benefit in Parkinson's disease but was well tolerated 46 and improved vascular function in older adults. 47 Oxidative stress could be functioning at several levels in SLE, including induction of MAVS oligomerisation 15 and oxidised mitochondrial DNA release during NET formation, 10 each contributing to the activation of the IFN-I pathway. In addition, immune complex and complement deposition in renal glomeruli induces oxidative stress in renal epithelium. 10 Thus, MitoQ may function at several points in reducing oxidative damage in the pathogenesis of SLE.
From our current understanding that MitoQ reduces mROS and oxidative stress and hence events downstream of mROS generation, we would not have anticipated MitoQ to affect events upstream of mROS production. These might include the development of lymphadenopathy of MRL-lpr mice, the generation of CD4 -CD8 -TCR-αβ + T cells and the production of autoantibodies, which may result from the absence of cell death occurring during homeostatic proliferation of lymphocytes. 17 48 Consequently, the upregulation of genes during T-cell homeostatic proliferation involved with cytolysis and inflammation, such as granzyme B, perforin and Fasligand, 22 is likely to still contribute to inflammation during MitoQ therapy. Thus, fully effective therapy for SLE may require a combination of traditional immunosuppression with non-immune antioxidant therapy.