An inherited mitochondrial DNA mutation remodels inflammatory cytokine responses in macrophages and in vivo in mice

Abstract

Impaired mitochondrial bioenergetics in macrophages promotes hyperinflammatory cytokine responses, but whether inherited mtDNA mutations drive similar phenotypes is unknown. Here, we profiled macrophages harbouring a heteroplasmic mitochondrial tRNA^Ala mutation (m.5019A>G) to address this question. These macrophages exhibit combined respiratory chain defects, reduced oxidative phosphorylation, disrupted cristae architecture, and compensatory metabolic adaptations in central carbon metabolism. Upon inflammatory activation, m.5019A>G macrophages produce elevated type I interferon (IFN), while exhibiting reduced pro-inflammatory cytokines and oxylipins. Mechanistically, suppression of pro-IL-1β and COX2 requires autocrine IFN-β signalling. IFN-β induction is biphasic: an early TLR4-IRF3 driven phase, and a later response involving mitochondrial nucleic acids and the cGAS-STING pathway. In vivo, lipopolysaccharide (LPS) challenge of m.5019A>G mice results in elevated type I IFN signalling and exacerbated sickness behaviour. These findings reveal that a pathogenic mtDNA mutation promotes an imbalanced innate immune response, which has potential implications for the progression of pathology in mtDNA disease patients.

Fig. 1. Reduced mitochondrial respiratory chain complexes and oxidative phosphorylation in m.5019A>G macrophages.

a Schematic of m.5019A>G mt-Ta mutation model. b Pyrosequencing results of pre-differentiation bone marrow and post-differentiation bone marrow-derived macrophages (BMDMs) from m.5019A>G mice (n = 6). Heteroplasmy range between 70% and 87%. Colours indicate matched bone marrow and BMDMs. c Mitochondrial DNA (mtDNA) copy number in non-stimulated (non-stim) and lipopolysaccharide (LPS)-stimulated wildtype (WT) and m.5019A>G BMDMs (n = 3; LPS 6 h). d ³⁵ S-methionine labelling and quantification of mitochondrial proteins in non-stim WT and m.5019A>G BMDMs (n = 3). e Seahorse XFe24 oxygen consumption rate (OCR) trace in non-stim WT (n = 3) and m.5019A>G BMDMs (n = 4). f Coenzyme Q (CoQ) redox measurements with or without antimycin A (Ant A) in non-stim WT and m.5019A>G BMDMs (n = 3). g Heatmap of all identified complex I (CI), CIII and CIV subunits and assembly factors in non-stim WT (n = 6) and m.5019A>G (n = 7) BMDMs. h CV-ATP5A, CIII-UQCRC2, CIV-MT-COI, CII-SDHB and CI-NDUFB8 protein levels in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h & 24 h). Representative blot shown. i Comparison of log₂FC values of CI structural subunits from proteomics (n = 6; WT and n = 7; m.5019A>G) and RNA sequencing (n = 3) data with Pearson r correlation and two-tailed statistical analysis applied. j Mitochondrial mass (P = 0.0000000322) and (k) normalised mitochondrial membrane potential (MMP) measurements in non-stim m.5019A>G vs WT BMDMs using MitoTracker Green (MTG) and tetramethyl rhodamine methyl ester (TMRM) (n = 8). Data are scaled log₂ intensities, log₂FC or mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of two independent experiments. P-values calculated using two-tailed Student’s t test for two group comparisons or multiple two-tailed unpaired t tests corrected for multiple comparisons using Benjamini, Krieger and Yekutieli method. a Created in BioRender. Dwane, L. (2025) https://BioRender.com/to6x4hj.

Fig. 2. Increased aerobic glycolysis in resting and inflammatory m.5019A>G macrophages.

a, b Overrepresentation analysis (ORA) using KEGG terms of all differentially expressed genes from RNA sequencing (n = 3) and differentially abundant proteins (n = 6; WT and n = 7; m.5019A>G) increased in non-stimulated (non-stim) m.5019A>G vs wildtype (WT) BMDMs. c, d Proton efflux rate (PER) measurements in non-stim WT and m.5019A>G BMDMs (n = 3; WT and n = 4; m.5019A>G)(P = 0.000036). e Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) measurements in non-stim and lipopolysaccharide (LPS)-stimulated WT and m.5019A>G BMDMs (n = 6; WT and n = 7; m.5019A>G). f Lactate/pyruvate ratio in cell culture medium (CCM) from metabolomics in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h & 24 h). g Heatmap of hypoxia-inducible factor 1-α (HIF-1α) targets and glycolytic enzymes from proteomics in LPS-stimulated WT and m.5019A>G BMDMs (n = 4; WT and n = 5; m.5019A>G; LPS 6 h). h Schematic of U-¹³C-glucose tracing into lactate and the tricarboxylic acid (TCA) cycle, indicating the first round labelling pattern. i m + 3 labelling in lactate and m + 2 labelling in citrate from U-¹³C-glucose in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h). Data are scaled log₂ intensities or mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of two independent experiments. P-values calculated using two-tailed Student’s t test for two group comparisons or multiple two-tailed unpaired t tests corrected for multiple comparisons using Benjamini, Krieger and Yekutieli method.

Fig. 3. Tricarboxylic acid (TCA) cycle remodelling in m.5019 A>G macrophages.

a Heatmap comparing metabolite levels in non-stimulated (non-stim) (α-KG, P = 0.001812; Fumarate, P = 0.006172; Malate, P = 0.000451) and lipopolysaccharide (LPS)-stimulated (α-KG, P = 0.005704; Succinyl-CoA, P = 0.025485; Aspartate, P = 0.047568) wildtype (WT) and m.5019A>G BMDMs (n = 5; LPS 6 h; left), and comparing metabolite levels in non-stim WT and m.5019A>G BMDMs in the presence or absence of glutamine (Gln) (n = 3; LPS 4 h; right) (Citrate, P = 0.008855; Isocitrate, P = 0.000054; α-KG, P = 0.0000001; Succinate, P = 0.0128505; Fumarate, P = 0.0000541; Malate, P = 0.0000116; Aspartate, P = 0.0297496). b α-Ketoglutarate (α-KG)/succinyl-CoA and α-KG/succinate ratio in non-stim WT and m.5019A>G BMDMs (n = 5). c Oxoglutarate dehydrogenase complex (OGDHC) and pyruvate dehydrogenase complex (PDHC) E3 subunit (DLD) levels from proteomics in non-stim WT (n = 4), non-stim m.5019A>G (n = 5), LPS-stimulated WT (n = 6; 6 h) and LPS-stimulated m.5019A>G BMDMs (n = 7; 6 h) (P = 0.000047). d Schematic of U-¹³C-glutamine tracing into the tricarboxylic acid (TCA) cycle, indicating oxidative versus reductive labelling patterns. e m + 5 labelling from U-¹³C-glutamine in isocitrate and citrate in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 5; LPS 6 h). f m + 5 labelling in L-2-hydroxyglutarate (L-2-HG) from U-¹³C-glutamine in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h). g Schematic of U-¹³C-glucose tracing into the TCA cycle, indicating oxidative versus reductive labelling patterns. h m + 3 labelling from U-¹³C-glucose in citrate, malate (P = 0.000405; P = 0.000405) and succinate in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 5; LPS 24 h). Data are mean or mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of three independent experiments. P-values calculated using two-tailed Student’s t test for two group comparisons or multiple two-tailed unpaired t tests corrected for multiple comparisons using Benjamini, Krieger and Yekutieli method. *** P < 0.001 ** P < 0.01 * P < 0.05.

Fig. 4. Reductive glutamine and pyruvate carboxylation fuel the AAS and NO in m.5019 A>G macrophages.

a Schematic of U-¹³C-glutamine tracing into the aspartate-argininosuccinate shunt (AAS), indicating oxidative versus reductive labelling patterns and NO production. b m + 3 labelling from U-¹³C-glutamine in aspartate, argininosuccinate (P = 0.000026) and fumarate (P = 0.000005) in non-stimulated (non-stim) and lipopolysaccharide (LPS)-stimulated wildtype (WT) and m.5019A>G BMDMs (n = 5; LPS 6 h). c Schematic of U-¹³C-glucose tracing into the AAS indicating oxidative versus reductive labelling patterns and nitric oxide (NO) production. d m + 3 labelling from U-¹³C-glucose in aspartate (P = 0.000070; P = 0.000002), argininosuccinate (P = 0.000015) and fumarate (P = 0.000039) in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 24 h). e Nitrite levels in cell culture medium (CCM) in non-stim and LPS-stimulated WT and m.5019A>G BMDMs in the presence or absence of glutamine (Gln) (n = 3; LPS 24 h) (P = 0.00000007). f, g Nos2 expression and inducible nitric oxide synthase (iNOS) protein levels from LPS time course analysis in WT and m.5019A>G BMDMs (n = 3; LPS 0, 1, 2, 6 & 24 h). Representative blot shown. Data are mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of three independent experiments. P-values calculated using multiple two-tailed unpaired t tests corrected for multiple comparisons using the Benjamini, Krieger and Yekutieli method or one-way ANOVA corrected for multiple comparisons using the Tukey method.

Fig. 5. Inflammatory cytokine and oxylipin production is disrupted in m.5019 A>G macrophages.

a Volcano plot of differentially expressed genes from RNA sequencing in lipopolysaccharide (LPS)-stimulated m.5019A>G vs wildtype (WT) BMDMs (n = 3; LPS 1 h). b, c Ifnb1 expression and interferon-β (IFN-β) release from LPS time course analysis in WT and m.5019A>G BMDMs (n = 3; LPS 0, 1, 2, 6 & 24 h) (P = 0.000052). d Phospho-interferon regulatory factor 3 (IRF3) (ser396), IRF3, IRF7, and interferon-stimulated gene 15 (ISG15) protein levels from LPS time course analysis in WT and m.5019A>G BMDMs (n = 3; LPS 0, 1, 2, 6 & 24 h). Representative blot shown. e Oxylipin profiling of cell culture medium (CCM) in non-stimulated (non-stim) and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h) (P = 0.00001417). f Olink target T48 mouse cytokine and chemokine profiling of CCM in LPS-stimulated m.5019 A > G vs WT BMDMs (n = 3; LPS 6 h). g Cyclooxygenase 2 (COX2) and pro-interleukin-1β (pro-IL-1β) protein levels from LPS time course analysis in WT and m.5019A>G BMDMs (n = 3; LPS 0, 1, 2, 6 & 24 h). Representative blot shown. h COX2 and pro-IL-1β protein levels in non-stim and LPS-stimulated WT and m.5019A>G BMDMs treated with an anti-interferon-α/β receptor (IFNAR) monoclonal antibody (Ab) or isotype control Ab (n = 3; LPS 6 h). Representative blot shown. i COX2 and pro-IL-1β protein levels in non-stim and LPS-stimulated WT and m.5019A>G BMDMs treated with Ruxolitinib or vehicle control Ab (n = 3; LPS 6 h). Representative blot shown. j Inducible nitric oxide synthase (iNOS) protein levels in non-stim and LPS-stimulated WT and m.5019A>G BMDMs treated with Ruxolitinib or vehicle control Ab (n = 3; LPS 24 h). Representative blot shown. Data are log₂FC or mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of three independent experiments. P-values calculated using two-tailed Student’s t test for two group comparisons, multiple two-tailed unpaired t tests corrected for multiple comparisons using Benjamini, Krieger and Yekutieli method or one-way ANOVA corrected for multiple comparisons using Tukey method.

Fig. 6. Mitochondrial network remodelling in m.5019A>G macrophages.

a, b Representative immunofluorescence staining of cytochrome c (Cyt c) and translocase of the outer membrane 20 (TOM20) coupled to confocal microscopy in non-stimulated (non-stim) wildtype (WT) and m.5019A>G BMDMs (a) and mitochondrial morphology analysis in non-stim WT (n = 4), non-stim m.5019A>G (n = 6), lipopolysaccharide (LPS)-stimulated WT (n = 3) and LPS-stimulated m.5019A>G (n = 4) BMDMs (b) (LPS 6 h; minimum of 20 cells analysed per condition per biological replicate) (P = 0.000038). Scale bars: 5 μm. c, d Representative immunofluorescence staining of TOM20 and ATP synthase coupled to super-resolution microscopy in non-stim WT and m.5019A>G BMDMs (c) and mitochondrial morphology analysis of non-stim WT (n = 3), non-stim m.5019A>G (n = 3), LPS-stimulated WT (n = 2) and LPS-stimulated m.5019A>G (n = 3) BMDMs (d) (LPS 6 h; minimum of 33 cells analysed from independent biological replicates) (P = 0.000000183; P = 0.0000000003). Scale bars: 5 μm. e, f Representative immunofluorescence staining of dynamin-related protein 1 (DRP1) and TOM20 coupled to confocal microscopy (e) and Pearson r correlation analysis (f) in non-stim and LPS-stimulated WT and m.5019A>G BMDMs (n = 3; LPS 6 h; minimum of 20 cells analysed per condition per biological replicate) (P = 0.000059). Scale bars: 5 μm. Data are mean ± s.e.m or ± s.d. n number represents independent biological replicates (mice) from a minimum of two independent experiments. P-values calculated using multiple two-tailed unpaired tests corrected for multiple comparisons using the Holm-Sidak method or one-way ANOVA corrected for multiple comparisons using the Kruskal-Wallis method.

Fig. 7. Late phase type I IFN in m.5019A>G macrophages is dependent on mitochondrial nucleic acids.

a, b Representative transmission electron microscopy (TEM) images (a) and cristae and mitochondrial aspect ratio (length/width) analysis (b) of non-stimulated (non-stim) and lipopolysaccharide (LPS)-stimulated wildtype (WT) and m.5019A>G BMDMs (n = 3; LPS 6 h; mitochondria from a minimum of 9 cells were analysed per condition per biological replicate). Scale bars: 0.5 μm. Black arrows indicate mitochondria. c Mitochondrial (mt)DNA and mtRNA levels in cytosolic fraction of non-stim WT and m.5019A>G BMDMs (n = 6). d, e Representative immunofluorescence staining of DNA and translocase of the outer membrane 20 (TOM20) coupled to confocal microscopy in LPS-stimulated WT and m.5019A>G BMDMs (d) and cytosolic DNA foci quantification in non-stim and LPS-stimulated WT and m.5019 A > G BMDMs (e) (n = 3; LPS 24 h; minimum of 20 cells analysed per condition per biological replicate). White arrows indicate cytosolic DNA foci. f Ifnb1 expression (LPS 24 h) and interferon-β (IFN-β) release (LPS 6 h) in LPS-stimulated WT and m.5019 A > G macrophages pre-treated with cyclic GMP-AMP synthase (cGAS) inhibitor RU.521 or vehicle control (DMSO) for 1 h (n = 3) (P = 0.000262). g mt-Nd1 (P = 0.000000008; P = 0.000946773), mt-Co3 (P = 0.000000005; P = 0.000236469), Ifnb1 (P = 0.000402; P = 0.000402) expression and IFN-β release in LPS-stimulated WT and m.5019A>G macrophages pre-treated with inhibitor of mitochondrial transcription 1 (IMT1) or vehicle control (DMSO) for 24 h (n = 3; LPS 24 h). Data are mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of three independent experiments. P-values calculated using multiple two-tailed unpaired t tests corrected for multiple comparisons using the Benjamini, Krieger and Yekutieli method.

Fig. 8. Elevated type I IFN levels in m.5019 A>G mice following an in vivo LPS challenge.

a Schematic of in vivo lipopolysaccharide (LPS) endotoxemia model experiment. b, c Volcano plot of Olink target T48 mouse cytokine and chemokine profiling of m.5019A>G mice vs wildtype (WT) mice serum injected intraperitoneally (i.p.) with phosphate-buffered saline (PBS) (n = 4; 2 h) or LPS (n = 10; 2 h). d Interferon-β (IFN-β) levels in serum of WT and m.5019A>G mice injected i.p. with PBS (n = 4; 2 h) or LPS (n = 10; 2 h). e Ifnb1 expression in kidney tissue of m.5019A>G mice vs WT mice serum injected i.p. with PBS (n = 4; 2 h) or LPS (n = 10; 2 h) (P = 0.0002). f Volcano plot of kidney tissue proteomic analysis of m.5019A>G mice vs WT mice injected i.p. with LPS (n = 10; 2 h). g Sepsis score of WT and m.5019A>G mice injected i.p. with PBS (n = 4; 2 h) or LPS (n = 10; 2 h) (P = 0.00001699). Data are log₂FC or mean ± s.e.m. n number represents independent biological replicates (mice) from a minimum of two independent experiments. P-values calculated using multiple two-tailed unpaired t tests corrected for multiple comparisons using the Benjamini, Krieger and Yekutieli method or one-way ANOVA corrected for multiple comparisons using the Tukey method. a Created in BioRender. Dwane, L. (2025) https://BioRender.com/kk5tnv3.