Ancestral allele of DNA polymerase gamma modifies antiviral tolerance

Abstract

Mitochondria are critical modulators of antiviral tolerance through the release of mitochondrial RNA and DNA (mtDNA and mtRNA) fragments into the cytoplasm after infection, activating virus sensors and type-I interferon (IFN-I) response^1-4. The relevance of these mechanisms for mitochondrial diseases remains understudied. Here we investigated mitochondrial recessive ataxia syndrome (MIRAS), which is caused by a common European founder mutation in DNA polymerase gamma (POLG1)⁵. Patients homozygous for the MIRAS variant p.W748S show exceptionally variable ages of onset and symptoms⁵, indicating that unknown modifying factors contribute to disease manifestation. We report that the mtDNA replicase POLG1 has a role in antiviral defence mechanisms to double-stranded DNA and positive-strand RNA virus infections (HSV-1, TBEV and SARS-CoV-2), and its p.W748S variant dampens innate immune responses. Our patient and knock-in mouse data show that p.W748S compromises mtDNA replisome stability, causing mtDNA depletion, aggravated by virus infection. Low mtDNA and mtRNA release into the cytoplasm and a slow IFN response in MIRAS offer viruses an early replicative advantage, leading to an augmented pro-inflammatory response, a subacute loss of GABAergic neurons and liver inflammation and necrosis. A population databank of around 300,000 Finnish individuals⁶ demonstrates enrichment of immunodeficient traits in carriers of the POLG1 p.W748S mutation. Our evidence suggests that POLG1 defects compromise antiviral tolerance, triggering epilepsy and liver disease. The finding has important implications for the mitochondrial disease spectrum, including epilepsy, ataxia and parkinsonism.

Fig. 1. Dysregulated immune signalling in fibroblasts of patients with MIRAS to viral PAMP mimetics.

a, The genotype–phenotype association of MIRAS POLG1 variant (rs113994097). Significance (P values) and disease categories are shown. The triangles indicate diseases or traits: upward-pointing triangles show a positive association, and vice versa. The dotted line shows the cut-off for significance. Analysis was performed using SAIGE mixed model logistic regression. Data are from ref. . b, POLG1 protein levels in patients with MIRAS (patient) and control fibroblasts. Western blot and quantification. The loading control was HSP60. Fibroblasts are from six patients and six control individuals, all female. c, Schematic of antiviral innate immune signalling responses to viral PAMPs. d, IFN-I signalling pathway genes induced by viral PAMP mimetics (dsRNA/poly(I:C) or dsDNA) in patient and control fibroblasts (as in b). Quantitative PCR (qPCR) analysis of cDNA. The reference gene was ACTB. Top, box plot. Bottom, heat map showing the average gene expression per condition. e, IFN-I signalling pathway protein induction by viral PAMP mimetic (dsRNA (poly(I:C)) or dsDNA) in patient and control (C) fibroblasts. Representative western blot analysis of four female control individuals and patients. The loading control was HSP60. Quantification is shown in Extended Data Fig. 2b. f, Paracrine immune signalling of fibroblasts in response to treatment with viral PAMP mimetic. Representative western blot of four female control individuals and patients (Pt). The loading control was HSP60. Quantification is shown in Extended Data Fig. 2e. g, mtDNA and mtRNA release into cytosolic extracts of fibroblasts (as in b; Extended Data Fig. 3e,f) after viral PAMP mimetic exposure for 7 h. Cytosolic versus whole-cell MT-CYB and MT-CO1 DNA or cDNA was analysed using qPCR. h,i, Immune signalling (h) and necroptosis activation (i) in fibroblasts (as in b) after prolonged viral PAMP mimetic treatment. Quantification of the western blot is shown for the indicated treatment times (Extended Data Fig. 3g,i). The loading control was β-actin. For b, d, g, h and i, the box plots show minimum to maximum values (whiskers), 25th to 75th percentiles (box limits) and median (centre line). Statistical analysis was performed using two-tailed unpaired Student’s t-tests. See also Extended Data Figs. 1–3 and Supplementary Table 1. Source Data

Fig. 2. Aberrant immune response of fibroblasts of patients with MIRAS to bona fide viral infection.

a, Schematic of HSV-1 infection of human primary fibroblasts. b,c, Viral load and host inflammatory protein level modulation in fibroblasts during HSV-1 infection. Western blot (b) and quantification (c) of protein amounts over time during infection in a control and patient fibroblast line. Protein loading stain was used as the loading control. Exp., exposure. d, Quantification of viral and host cellular protein amounts at the specific timepoint after HSV-1 infection. n = 6 female controls and patients. The western blot is shown in Extended Data Fig. 4a–c,e. The loading control was HSP60. e, mtDNA levels in fibroblasts (as in d) during HSV-1 infection. qPCR analysis of mtDNA (MT-12S (MT-RNR1)) relative to a nuclear gene (B2M). f, The viral load in fibroblasts after 48 h of TBEV infection. Immunofluorescence analysis of TBEV antigen (green) and dsRNA (red; detects viral RNA) with DAPI co-staining (blue) is shown. Each channel is shown in greyscale. Scale bars, 100 μm. n = 2 female control individuals and patients. g, Quantification of the viral load in fibroblasts after 48 h of TBEV or SARS-CoV-2 infection. Top, the percentage of TBEV-positive cells. n = 4 control and 3 MIRAS images of 2 female control individuals and patients shown in f. Bottom, quantification of the western blot analysis of SARS-CoV-2 nucleocapsid protein (Extended Data Fig. 5h). The loading control was HSP60. n = 6 female control individuals and patients. h, POLG1 and inflammatory protein level in fibroblasts after 48 h of TBEV or SARS-CoV-2 infection. Quantification of the western blot is shown (Extended Data Fig. 5e–h). The loading control was HSP60. n = 6 female control individuals and patients. i, Schematic of the response of cells from patients with MIRAS to virus infection. For d, e, g (bottom) and h, the box plots show minimum to maximum values (whiskers), 25th to 75th percentiles (box limits) and median (centre line). For g (top), data are mean ± s.e.m. Statistical analysis was performed using two-tailed unpaired Student’s t-tests. See also Extended Data Figs. 4 and 5. Source Data

Fig. 3. Compromised in vivo activation of antiviral IFN-I signalling in MIRAS mice.

a, mtDNA replisome protein amount in mitochondria isolated from the mouse brain (cerebral cortex), liver and spleen. Western blot and quantification is shown. The loading control was HSP60. n = 4 female mice (aged 3 months) per genotype. b, mtDNA maintenance in the MIRAS mouse cerebral cortex and liver. mtDNA replication was analysed using south-western blotting for BrdU incorporation into mtDNA (the arrowhead indicates the band of interest for replicating mtDNA detected using anti-BrdU) relative to total mtDNA (Southern blot, mtDNA hybridization); full-length mtDNA is around 16 kb. Bottom left, quantification of BrdU-labelled mtDNA/total mtDNA. Bottom right, the mtDNA levels were assessed using qPCR analysis of mtDNA (mt-Co1) relative to nuclear gene (Actb). n = 5 female mice (aged 3 months) per genotype. c, The experimental design of TBEV infection. B, baseline uninfected. d, Circulatory IFN-I levels at day 4 after TBEV infection compared with uninfected mice. n = 5 female mice (aged 12 months) per condition. e, The expression of IFN-I-response components in the mouse cerebral cortex and spleen at day 4 after TBEV infection (as in d). cDNA was analysed using qPCR. The reference gene was Actb. f, The transcriptome profile of MIRAS and control mouse cerebral cortex on day 4 after TBEV infection compared with the baseline uninfected state (as in d). The volcano plot shows significance (adjusted P (P_adj), Wald test with Benjamini–Hochberg adjustment) and fold change (FC). Immune/antiviral-response-related genes are highlighted in blue. g, Disease and function enrichment analysis (Ingenuity pathway analysis) of the cerebral cortex transcriptome of MIRAS mice compared with parallel infected control mice on day 4 after TBEV infection (as in d) on transcripts with adjusted P < 0.05. Annotations with P < 0.05 (Fisher’s exact test) with activation z ≥ 1 are shown; and those with z ≥ 2, indicating predicted significant activation, are highlighted with a black border. For a, b, d and e, the box plots show minimum to maximum values (whiskers), 25th to 75th percentiles (box limits) and median (centre line). Statistical analysis was performed using two-tailed unpaired Student’s t-tests. See also Extended Data Figs. 6 and 7a,b and Supplementary Table 2. Source Data

Fig. 4. Infection-induced metabolome alteration and acute GABAergic neuronal loss in the mouse brain.

a–d, The metabolome of MIRAS and control cerebral cortex on day 4 after TBEV infection. n = 5 female mice (aged 12 months). a, Pearson r correlation of metabolites with P_adj < 0.05. P values were calculated using two-sample Student’s t-tests with Benjamini–Hochberg multiple-testing correction. b, Volcano plot showing significance (q, two-sample Student’s t-test with Storey–Tibshirani multiple-testing correction) and metabolite fold change. The dashed line indicates q = 0.05. c,d, Canonical pathway (c) and brain-related disease and function (d) enrichment analyses (Ingenuity pathway analysis) of metabolites (q < 0.05 as in b). Statistical analysis was performed using Fisher’s exact tests. e, mtDNA levels in the mouse cerebral cortex on day 4 after infection (as in a). qPCR analysis of mtDNA (mt-Nd4, mt-12s and mt-Co1) relative to nuclear gene (Actb) is shown. Statistical analysis was performed using two-tailed unpaired Student’s t-tests. f, RNA-seq analysis of Slc25a33 levels in the mouse cerebral cortex on day 4 after infection (as in a). Statistical analysis was performed using Wald tests. g, RNA-seq analysis of GABAergic-related gene expression in the mouse cerebral cortex at the baseline (uninfected) and on day 4 after TBEV infection. n = 5 female mice (aged 12 months) per condition. Statistical analysis was performed using Wald tests; *P < 0.05. h, GABAergic marker (GABRB2 and GAD67) staining in the mouse neocortical region on day 5 after TBEV infection. The region between the dotted lines shows interneurons in mid-cortical laminar layer 4. Representative image (left) and semiquantitative scoring (right); n = 6 female mice (aged 12 months) per condition. Statistical analysis was performed using two-tailed unpaired Student’s t-tests. Scale bars, 100 μm (top) and 200 μm (bottom). For e, f and h, the box plots show minimum to maximum values (whiskers), 25th to 75th percentiles (box limits) and median (centre line). See also Extended Data Fig. 7c–e and Supplementary Tables 2 and 3. Source Data

Fig. 5. Infection triggered exacerbated liver inflammation and pro-inflammatory circulatory cytokines in MIRAS and compromised mtDNA replisome and antiviral responses in patient brains.

a,b, Liver histopathology after TBEV infection. a, Representative haematoxylin and eosin staining. The arrows indicate immune cell infiltration. N, necrotic cells. Scale bars, 200 μm. b, Liver inflammation. Top, semiquantitative scoring of the overall severity. Bottom, the total number and size of immune cell infiltrates. n = 3 views per mouse, 5 female mice (aged 12 months) per condition. c, Liver necroptotic activation on day 4 after TBEV infection (as in a). Western blot analysis and quantification is shown. The loading control was protein loading stain. n = 4 mice. d, Quantification of immune cell marker staining in livers (as in a). Representative immunohistochemical staining is shown in Extended Data Fig. 7g. e, IL-6 cytokine levels in mouse sera (as in a). f, IL-6 and TNF cytokines in patient sera. n = 15 patients (7 male, 8 female) and 23 controls (9 male, 14 female) (IL-6); and n = 8 patients (4 male, 4 female) and 8 controls (3 male, 5 female) (TNF). g, The mtDNA replisome components and native complex levels (mitochondrial fractions), their respective quantifications and mtDNA levels (from whole tissue) were analysed in autopsy-derived samples from cerebral cortex. Left, western blot and quantification. The loading control was ATP5A. Middle, POLG and HSP60 complex were analysed using native complex analyses. Right, mtDNA levels were analysed using qPCR (MT-CO1 relative to nuclear gene B2M). n = 3 patients and 6 controls (5 for native complex analysis), all female. h, RNA-seq analysis of the transcriptome of patient and control cerebral cortex. The volcano plot shows P values and the fold change (patient/control) of protein-coding transcripts. Statistical analysis was performed using Wald tests. Immune/antiviral response genes (P < 0.05) are shown in blue. n = 3 patients and 5 controls, all female. i, InnateDB pathway analysis of transcripts with P < 0.05 in the patient or control cerebral cortex transcriptome (as in h). Statistical analysis was performed using hypergeometric tests with Benjamini–Hochberg multiple-testing correction. Pathways predicted downregulated in cerebral cortex (pathway P < 0.05) are tabulated. The darker blue nodes indicate pathway P_adj < 0.05. j, Working model of MIRAS disease pathology. For b–f, the box plots show minimum to maximum values (whiskers), 25th to 75th percentiles (box limits) and median (centre line). For g, data are mean ± s.e.m. Statistical analysis for b–g was performed using two-tailed unpaired Student’s t-tests. See also Extended Data Figs. 7f–j and 8–10 and Supplementary Tables 1, 4 and 5. Source Data

Extended Data Fig. 1. FinnGen data and characterization of MIRAS patient fibroblasts.

(a) Genotype-phenotype association; FinnGen database. Clinical phenotype enriched in carriers of mitochondrial and related disease gene variants: rs80356540, c.1523A>G, p.Tyr508Cys in TWNK (Twinkle helicase) causing IOSCA (infantile onset spinocerebellar ataxia); rs28937590, c.232A>G, p.Ser78Gly in BCS1L (assembly factor for complex III) causing GRACILE syndrome (complex III deficiency); and rs148890852, c.92C>T, p.Ser31Leu in ZNHIT3 (Zinc finger HIT-type containing 3) causing PEHO (Progressive encephalopathy with oedema, hypsarrhythmia, and optic atrophy). Significance (p-value) and disease categories tabulated. Triangles: diseases or traits. Upward-pointing triangle: positive association and vice versa. Dotted line: cut-off for significance; Mixed model logistic regression method SAIGE. Data are from ref. . (b) Primary fibroblast characteristics: Age-of-sampling, MIRAS patients and control individuals. N = 6 for both patients (all homozygous for p.W748S + E1143G, females) and age- and gender-matched healthy control. Mean age of control group: 33 years, MIRAS patients: 30 years. Refer to Supplementary Table 1 for patient sample details. (c) Cell viability of control and MIRAS patient fibroblasts. N = 6 female controls and patients. (d) mtDNA replisome protein and gene expression in control and MIRAS patient fibroblasts. Western blot and quantification, loading control HSP60 (See Fig. 1b for POLG1 and HSP60 data); Gene expression, qPCR of cDNA, reference gene ACTB. N = 6 female controls and patients. (e) mtDNA and mtRNA amount in control and MIRAS patient fibroblasts. qPCR of DNA or cDNA (MT-12S (or MT-RNR1), MT-CYB and MT-CO1 relative to nuclear/reference gene ACTB). N = 6 female controls and patients. In b, c, d and e, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test. Abbreviations: Pt, patient. Source Data

Extended Data Fig. 2. Intracellular and paracrine IFN-I signalling in MIRAS fibroblasts exposed to viral PAMP mimetics.

(a) Inflammatory cytokine gene induction by viral PAMP mimetics (dsRNA/poly(I:C) or dsDNA) in fibroblasts. qPCR of cDNA, reference gene ACTB. N = 6 female controls and patients. (b) IFN-I signalling pathway proteins; fibroblasts exposed to viral PAMP mimetics (dsRNA/poly(I:C) or dsDNA). Western blot of two controls and patients in addition to the two controls and patients shown in Fig. 1e. Quantification of 24 h treatment protein level relative to basal condition of each cell line (n = 4 female controls and patients), loading control HSP60. (c) IFN-I signalling pathway proteins in fibroblasts at 24 h post dsRNA/poly(I:C) or dsDNA treatment. Western blot, quantification, loading control HSP60. N = 6 female controls and patients. (d) Experimental setup; transfer of cell culture media from 24 h dsRNA/poly(I:C)-treated control or patient cells to naïve cells for another 24 h incubation to investigate paracrine immune signalling. (e) Paracrine immune signalling of fibroblasts as illustrated in d. Quantification of western blot, loading control HSP60. N = 4 female controls and patients. Refers to representative western blot in Fig. 1f. In a, b, c and e, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test in a, b and c, 2-way ANOVA with Tukey’s multiple comparisons test in e. Abbreviations: C, control; Pt, patient; IFN, interferon; PAMP, pathogen-associated molecular pattern. Source Data

Extended Data Fig. 3. Immune signalling in MIRAS patient fibroblasts in response to viral PAMP mimetics.

(a) Induction of IFN-I signalling by exogenous RIG-I and MAVS expression. MIRAS patient fibroblasts transfected with empty vector (EV) or RIG-I (N; constitutive active) and MAVS vectors. Western blot and quantification, loading control HSP60; N = 3 female patients; protein level of each MIRAS patient fibroblast line quantified. (b) IFN-I signalling in response to dsRNA/poly(I:C) treatment upon exogenous expression of RIG-I and MAVS. MIRAS fibroblasts as described in a, further treated with 7 h of dsRNA/poly(I:C). qPCR of cDNA, reference gene ACTB; RIG-I protein level detected in each of three patient fibroblast lines before dsRNA/poly(I:C) treatment (as quantified in a is overlaid onto the chart, showing positive association between RIG-I protein amount and the subsequent dsRNA/poly(I:C)-induced expression of IFN-β. N = 3 female patients; 3 technical replicates per patient; mean ± SEM. (c,d) RIG-I and MAVS exogenous expressions improve IFN-I signalling in MIRAS fibroblasts in response to viral PAMP mimetics. c: Western blot representative, quantification, loading control HSP60; d: qPCR of cDNA, reference gene ACTB, in control and MIRAS fibroblasts with or without exogenous expression of RIG-I(N) and MAVS and subjected to 7 h dsRNA/poly(I:C) treatment. N = 3 female controls and patients; mean ± SEM; two-tailed unpaired student’s t test. (e) Cytosolic extract purification from fibroblasts. Western blotting for protein markers of mitochondria, nucleus and cytosol. T, total (whole-cells); C, cytosol. Representative of 3 independent experiments. (f) mtDNA and mtRNA amount in fibroblasts post 7 h dsDNA treatment. qPCR of DNA or cDNA (MT-CYB and MT-CO1 relative to reference gene B2M (mtDNA) or ACTB (mtRNA)). N = 6 female controls and patients per condition. (g) Immune signalling and necroptosis activation in fibroblasts after prolonged viral PAMP mimetics exposure. Western blot at 32 h of dsRNA/poly(I:C) or dsDNA treatment, loading control beta-actin (Quantification in Fig. 1h, i). N = 6 female controls and patients per condition. Arrowhead: protein band of interest. (h) IFN-β (IFNB1) expression in fibroblasts after prolonged viral PAMP mimetics exposure. qPCR of cDNA, reference gene ACTB, at 32 h of treatment. N = 6 female controls and patients per condition. (i) Necroptosis activation in fibroblasts after prolonged (32 h or 72 h) dsDNA treatment. Western blot, loading control beta-actin (Quantification in Fig. 1i). N = 6 female controls and patients per condition. Arrowhead: protein band of interest. In f and h, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test. Abbreviations: C, control; Pt, patient; IFN, interferon; PAMP, pathogen-associated molecular pattern. Source Data

Extended Data Fig. 4. HSV-1 infection of MIRAS patient and control fibroblasts.

(a-c) Immune signalling pathway protein and viral protein amount in fibroblasts during HSV-1 infection. Western blot (top) and quantification (bottom) at a: 6 h, b: 24 h, and c: 48 h of HSV-1 infection, loading control HSP60; arrowhead: band of interest. N = 6 female controls and patients per condition. Quantification also in Fig. 2d. (d) Comparative fold expression of HSV-1 ICP-27 and host cellular pro-inflammatory protein, NF-κB-p65 in the parallelly infected MIRAS patient and control fibroblasts at the indicated infection time from panels a-c above; two-tailed unpaired student’s t test. (e) POLG1 protein level and necroptosis sensitivity of fibroblasts during TBEV infection. Western blot of POLG1 and necroptotic activating p-MLKL protein, loading control HSP60; arrowhead: band of interest. N = 6 female controls and patients per condition. Quantification in Fig. 2d. (f) Fibroblast cell viability at 48 h of HSV-1 infection. N = 6 female controls and patients per condition; In a, b, c and f, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test. Abbreviations: Pt, patient. Source Data

Extended Data Fig. 5. Response of MIRAS mutation-corrected fibroblast-like cells to HSV-1 infection, and control and patient fibroblasts response to TBEV or SARS-CoV-2 infection.

(a) Schematic for generation of induced fibroblast-like cells from iPSC cell lines. Created using BioRender.com. See method section for details. (b) POLG1 protein amount in controls, MIRAS patient and MIRAS mutation-corrected induced fibroblast-like cells. Western blot and quantification, loading control SDHA. Mean ± SEM. N = 4 controls (2 males, 2 females), 1 male patient, 2 independent MIRAS mutation-corrected clones of the male patient (Cr1, Cr2). (c) Viral protein, host inflammatory and necroptotic protein in cells (as in b) at 48 h post HSV-1 infection. Western blot and quantification, loading control SDHA. Mean ± SEM. (d) MtDNA amount at uninfected basal condition and 48 h post HSV-1 infection in cells (as in b). mtDNA, MT-12S to nuclear gene, B2M. Mean ± SEM. (e-h) Immune signalling (e,f) and necroptotic activation (g,h) of fibroblasts post 48 h of TBEV (e,g) or SARS-CoV-2 (f,h) infection. Western blot and quantification, loading control HSP60, arrowhead: band of interest. N = 6 female controls and patients per condition. Refer to Fig. 2g, h for other protein quantification. (i,j) Cellular viability of fibroblasts post 48 h of TBEV (i) and SARS-CoV-2 (j) infection. N = 6 female controls and patients per condition. In e, f, i and j, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test. Abbreviations: Pt, patient. Source Data

Extended Data Fig. 6. Generation and characterization of MIRAS knock-in mice.

(a) Generation of MIRAS knock-in mice. (b) DNA sequencing verification of wildtype (control), heterozygote and homozygote MIRAS mice. DNA sequence alignment indicates MIRAS variants c. 2177G>C in exon 13 and c. 3362A>G in exon 21 of Polg1. (c) Genotype verification of wildtype, heterozygotic or homozygotic MIRAS mice. Amplification of exon 13 and lox^P site containing intron 13, shows a 120 bp longer fragment in MIRAS than wildtype control. The genotyping procedure was performed on all mice analysed in the study. (d) Weight of mice at 2, 6 and 12 months of age. Age 2 months: N = 9 control, 10 MIRAS; age 6 months: N = 10 control, 10 MIRAS; age 12 months: 10 control, 11 MIRAS; all females; mean ± SEM. Source Data

Extended Data Fig. 7. Impact of TBEV infection on mouse tissues.

(a) IFN-I signalling, mouse cerebral cortex, TBEV 1 dpi. qPCR of cDNA, reference gene ACTB. (b) Inflammatory proteins in mouse cerebral cortex. TBEV 4 dpi. Western blot, quantification, loading control HSP60. (c) Schematic depicting brain cell metabolism, TBEV 4 dpi (MIRAS/Control). Alteration of metabolites (as in Fig. 4b, level indicated for metabolites of q-value ≤ 0.05) and transcripts encoding for pathway proteins (level indicated for transcripts of p-value < 0.05; RNA seq; Wald test). Pdh, pyruvate dehydrogenase; Pdk1, pyruvate dehydrogenase kinase 1; Glud1, glutamate dehydrogenase 1; Gls, glutaminase; Glul, glutamate-ammonia ligase (glutamine synthetase); Ggt7, gamma-glutamyl transpeptidase 7; Slc1a1, solute carrier family 1 member 1; GSSG/GSH, glutathione/glutathione-reduced; ROS, reactive oxygen species. (d) mtDNA amount; mouse spleen and liver; day 4 post-TBEV infection. qPCR of mtDNA (MT-ND4, MT-12S and MT-CO1) relative to nuclear gene (ACTB). (e) GABAergic neurons (GAD67 in uninfected mouse cerebral cortex at baseline. Semiquantitative scoring of histological staining. (f) Immune cell infiltration of mouse liver; TBEV 4 dpi. Representative Hematoxylin-eosin (H&E) staining. 5X; yellow dotted line: boundary of immune cell infiltrate. Right panel: 10X; yellow double-headed arrow: distance spanned by the immune cell infiltration from the portal tract. Refer to Fig. 5b for quantification. (g) Representative CD68+ (macrophages), CD3 + /4 + /8b+ (pan-T/helper-T/cytotoxic-T lymphocytes) staining in liver, TBEV 4 and 9 dpi. Quantification in Fig. 5d. (h) Liver fat content 4 and 9 dpi. Oil-red O staining. Red: lipid (on hematoxylin). Quantification: lipid as % of tissue. (i) TBEV RNA amount in mouse liver and cerebral cortex, 4 and 9 dpi. qPCR of cDNA (TBEV-NS5, Non-structural 5). Samples with no detectable TBEV-NS5 RNA: open circle at zero value. (j, k) Inflammatory response and viral protein expression in the mouse brain (frontal cortex; TBEV 9 dpi. H&E staining / immunohistochemistry, hematoxylin counterstaining. Bars = 50 µm. (j) Control mouse. A mild focal perivascular mononuclear infiltration in association with viral antigen expression in moderate numbers of neurons in the adjacent parenchyma (arrow). Inset: infected neurons with strong TBEV antigen expression in cell body and processes. MIRAS: mild focal perivascular mononuclear infiltration (arrows) and abundant infected neurons in the adjacent parenchyma. Inset: extensive viral antigen expression in neuronal cell bodies and cell processes (arrowheads). (k) Inflammatory cells, microglial response. Control; Perivascular infiltrates (arrows) contain several macrophages (Iba1 + ; inset: arrowheads), a few T cells (CD3 + ; arrowheads) and rare B cells (CD45R/B220 + ; arrowhead). Iba1 staining: disseminated activated microglial cells with their typical stellate shape (inset: arrow). MIRAS; perivascular infiltrates containing macrophages (Iba1 + ; arrowhead), T cells (CD3 + ; arrowheads) and several B cells CD45R/B220 + ; arrowheads). A few individual T cells (small arrows) in adjacent parenchyma. Iba1: mild focal microgliosis of the parenchyma adjacent to the affected vessel (asterisk). In a-d and f-k, N = 5 mice per condition; age 12 months, females; In e, 1 year-old mice (N = 5 control, 6 MIRAS, males) and 2+ year-old mice (N = 4 control, 4 MIRAS, females). In a, b, d, e, h and i, box plots show minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); two-tailed unpaired student’s t-test. Abbreviations: Dpi, day post infection. Source Data

Extended Data Fig. 8. Analyses of MIRAS patient brain autopsy samples.

(a) mtDNA replisome and mtDNA amount in patient cerebellum. Western blot and quantification (loading control ATP5A) and native POLG replisome complex amount (loading control HSP60 complex) from isolated mitochondria; qPCR for mtDNA amount (mtDNA, MTCO1 relative to nuclear gene, B2M). N = 2 patient cerebellar samples and 6 (or 5 for native complex analysis) controls, all females. Mean ± SEM; two-tailed unpaired student’s t test. (b) Immune-related gene expression in MIRAS patients’ cerebral cortex. qPCR of cDNA, reference gene ACTB. N = 3 patients and 6 controls, all females; box plot: minimum-25^th-50^th(median)-75^th percentile-maximum (whiskers extend to the smallest and largest value); + symbol indicates the mean expression value; two-tailed unpaired student’s t test. (c-g) Pathway enrichment analyses of MIRAS patient cerebral cortex transcriptome. RNA-seq; N = 3 patients compared to 5 controls, all females. (c) Ingenuity pathway analysis using genes with p-value < 0.05 in MIRAS patient/control transcriptome. The most significantly affected (based on p-value; Fisher’s exact test) canonical pathways with their predicted activation or inactivation status are tabulated. Red: predicted activation (z-score≥2); blue: predicted inactivation (z-score ≤ −2); grey: pathway with no predicted significant activation/inhibition. (d) Inhibition of Interferon (IFN) signalling. Heatmap: fold change of IFN signalling gene expressions in MIRAS patients compared to the controls, and gene mapping onto the IFN-signalling pathway (generated by Ingenuity Pathway Analysis). (e) KEGG pathway enrichment analysis of downregulated genes (p-value < 0.05; Fold change > −1.5) in MIRAS patients compared to controls; generated using gProfiler server; Fisher’s one-tailed test with multiple testing correction using the server default algorithm g:SCS. (f, g) Integrated network summary of canonical pathways, upstream regulators, diseases and biological functions associated with patient/control cerebral cortex transcriptome (genes with p-value < 0.05). Generated using Ingenuity Pathway Analysis knowledgebase. The network in g shows also predicted relationships to transcripts with levels altered in the dataset. Source Data

Extended Data Fig. 9. MIRAS patient cerebral cortex proteome and transcriptome.

(a) Proteome profiling of MIRAS patient cerebral cortex. Volcano plot: significance (p-value) and fold change. 57 of the proteins with p-value < 0.05 (two-sample Student’s T-test) are associated with immune/antiviral pathways (highlighted in blue). N = 3 patients and 6 controls, all females. (b) Canonical pathway enrichment analysis of cerebral cortex proteome in MIRAS (as in a). Ingenuity pathways analyses on proteins significantly affected (p-value < 0.05) in MIRAS; pathway p-value (Fisher’s exact test) and ratio (number of molecules from the dataset that map to the pathway divided by the total number of molecules that map to the canonical pathway). Blue: pathways related to immune and/or viral process; orange: pathways related to mitochondrial functions. (c) Comparative pathway profiling of cerebral cortex transcriptome of patient/control (as in Fig. 5h) versus TBEV-infected/uninfected wildtype mice (as in Fig. 3f). Biological pathways (analysed using gProfiler server; Fisher’s one-tailed test with multiple testing correction using the server default algorithm g:SCS) enriched with downregulated genes (p-value < 0.05, fold change > −1.5) in patient/control or with upregulated genes (adj-p-value < 0.05, fold change>1.5) in TBEV 4 dpi/uninfected wildtype control mice. Left: Top common overlapping pathways of the two datasets (pathway adj-p-value < 0.05) are tabulated with the pathway-associated gene count. Right: the pathway-associated gene count and adj-p-value. 58 overlapping pathways between the two datasets are highlighted in green shade. Selected top pathways related to virus and immune response are annotated in blue text to depict the opposed regulation.

Extended Data Fig. 10. Overview of research data.

Created using BioRender.com.