Oxidized DNA fragments exit mitochondria via mPTP- and VDAC-dependent channels to activate NLRP3 inflammasome and interferon signaling

Abstract

Mitochondrial DNA (mtDNA) escaping stressed mitochondria provokes inflammation via cGAS-STING pathway activation and, when oxidized (Ox-mtDNA), it binds cytosolic NLRP3, thereby triggering inflammasome activation. However, it is unknown how and in which form Ox-mtDNA exits stressed mitochondria in non-apoptotic macrophages. We found that diverse NLRP3 inflammasome activators rapidly stimulated uniporter-mediated calcium uptake to open mitochondrial permeability transition pores (mPTP) and trigger VDAC oligomerization. This occurred independently of mtDNA or reactive oxygen species, which induce Ox-mtDNA generation. Within mitochondria, Ox-mtDNA was either repaired by DNA glycosylase OGG1 or cleaved by the endonuclease FEN1 to 500-650 bp fragments that exited mitochondria via mPTP- and VDAC-dependent channels to initiate cytosolic NLRP3 inflammasome activation. Ox-mtDNA fragments also activated cGAS-STING signaling and gave rise to pro-inflammatory extracellular DNA. Understanding this process will advance the development of potential treatments for chronic inflammatory diseases, exemplified by FEN1 inhibitors that suppressed interleukin-1β (IL-1β) production and mtDNA release in mice.

Keywords: FEN1; NLRP3 inflammasome; OGG1; Ox-mtDNA; VDAC; cGAS-STING; mPTP; mitochondria; mtDNA.

Figure 1.. Base excision repair inhibits Ox-mtDNA production and attenuates NLRP3 inflammasome activation

(A) H&E, Sirius red, F4/80 and myeloperoxidase (MPO) antibody staining of lung tissue from mice challenged with 5 mg/kg LPS 24 h prior to tissue collection. n=3 for mock treatment and n=6–7 mice for LPS treatment. 10–12 images per mouse were analyzed. Scale bar, 100 μm. (B) Area (in %) occupied by F4/80 or MPO positive cells in lung sections from (A). Data are means ± s.d. (C) IL-1β and TNF concentrations in BALF from (A) measured by ELISA. Data are means ± s.d. (D and E) 8-OH-dG content of mtDNA from cytosol (left) or mitochondria (right) of LPS (200 ng/mL, 4 h)-primed WT and mt-Ogg^Tg (D) or WT and Ogg1^−/− (E) BMDM stimulated −/+ ATP (4 mM, 1 h). (F) Immunoblot (IB) analysis of lysates of WT and Ogg1^−/− BMDM. (G and H) Relative cytosolic mtDNA amounts in LPS-primed WT and mt-Ogg^Tg (G) or WT and Ogg1^−/− (H) BMDM stimulated −/+ ATP. The relative ratios of D-loop mtDNA, Cox1 mtDNA, or non-NUMT mtDNA are shown. (I) IB analysis of Casp1 p20 and mature IL-1β in supernatants and NLRP3, Pro-IL-1β, Pro-Casp1, and ASC in lysates of LPS-primed WT or mt-Ogg1^Tg BMDM stimulated −/+ ATP (4 mM, 1 h) or nigericin (Nig) (10 μM, 1h) (left). Mitochondrial OGG1 IB in WT and mt-Ogg1^Tg BMDM (right) is shown. (J) IL-1β and TNF secretion by LPS-primed WT or mt-Ogg1^Tg BMDM challenged with different NLRP3 activators (ATP, 4 mM for 1 h, nigericin, 10 μM for 1h, MSU, 600 μg/mL for 6 h and alum, 500 μg/mL for 6 h). (K) IB of Casp1 p20 and mature IL-1β in supernatants of LPS-primed WT and Ogg1^−/− BMDM stimulated −/+ ATP (4 mM) or nigericin (10 μM) for 1h. (L) Representative fluorescent microscopy images of WT or mt-Ogg1^Tg BMDM co-stained for Atp5b and ASC before or after LPS priming followed by ATP (4 mM) or nigericin (10 μM) stimulation for 1h. DAPI stains nuclei. Arrows indicate ASC specks. Scale bar, 5 μm. (M) Percentages of cells shown in (L) with ASC specks. n=150 cells per group from 3 independent experiments. IBs are one representative out of 3 independent experiments. Results in (D, E, G, H, J and M) are mean ± s.d. (n=3). *p<0.05; **p<0.01; ***p<0.001. ns, not significant. Two-sided unpaired t-test. See also Figure S1.

Figure 2.. mPTP opening and VDAC oligomerization enable mtDNA cytosolic release and NLRP3 inflammasome activation

(A) VDAC oligomerization in BMDM that were pretreated −/+ VBIT-4 (10 μM, 16 h), primed −/+ LPS (200 ng/mL, 4 h) and stimulated −/+ ATP (4 mM, 1 h) was determined by IB analysis of cell lysates incubated with the cross-linking agent EGS (200 μM, 40 min, 30°C) to stabilize VDAC oligomers. Asterisk indicates nonspecific band. (B) Relative calcein fluorescence in BMDM pretreated −/+ CsA (1 μM, 16 h) or VBIT-4 (10 μM, 16 h), followed by priming −/+ LPS (200 ng/mL, 4 h) and stimulation −/+ ATP (4 mM, 1 h). The calcein-quenching assay measures mPTP opening. (C and D) Relative amounts of Ox-mtDNA (C) or total mtDNA (D) in cytosols of LPS-primed shCtrl- or shOgg1-transduced BMDM, pretreated −/+ CsA and VBIT-4 and stimulated with ATP as in (B). (E) VDAC oligomerization in BMDM treated −/+ EtBr (450 ng/mL, 4 days) to deplete mtDNA, and subsequently incubated −/+ ATP (4 mM) for 1, 2 or 3 h was analyzed as in (A). Asterisk indicates nonspecific band. (F) Experimental scheme for measuring Ox-mtDNA in the intermembrane space (IMS) based on OMM rupture with proteinase K. (G and H) Relative amounts of Ox-mtDNA (G) or total mtDNA (H) in supernatants (IMS fraction) of proteinase K digested mitochondria isolated from BMDM treated as indicated. (I) IB analysis of Casp1 p20 and mature IL-1β in supernatants of LPS-primed BMDM, pretreated −/+ CsA (1 μM, 16 h) and VBIT-4 (10 μM, 16 h) and challenged with ATP (4 mM, 1h). (J) Relative Casp1 p20 and IL-1β in supernatants of BMDM treated as in (I). (K) IL-1β and TNF secretion by BMDM primed −/+ LPS (200 ng/mL, 4 h), pretreated−/+ CsA (1 μM, 16 h) and VBIT-4 (10 μM, 16 h) and challenged with different NLRP3 activators. (L) Representative fluorescent microscopy images of BMDM, pretreated −/+ VBIT-4 (10 μM, 16 h), primed −/+ LPS (200 ng/mL, 4 h), stimulated −/+ ATP (4 mM, 1 h) or nigericin (10 μM, 1 h) and co-stained for Atp5b and ASC. Arrows indicate ASC specks. Scale bar, 5 μm. (M) Percentages of cells from (L) with ASC specks. n=150 cells per group from 3 independent experiments. IBs show one representative out of 3 independent experiments. Results in (B-D, G, H, J, K, and M) are mean ± s.d. (n=3). *p<0.05; **p<0.01; ***p<0.001. ns, not significant. Two-sided unpaired t-test. See also Figure S2.

Figure 3.. [Ca²⁺]_m uptake initiates mPTP opening and VDAC oligomerization to release Ox-mtDNA

(A) VDAC oligomerization in BMDM stimulated with different NLRP3 activators was analyzed as in Figure 2A. Asterisk indicates nonspecific band. (B) Relative VDAC oligomerization in cells from (A) (n=3). (C) Relative calcein fluorescence in BMDM −/+ LPS (200 ng/mL, 4 h) priming and challenged with different NLRP3 agonists, using the calcein-quenching assay to measure mPTP opening as in Figure 2B (n=4). (D) Representative fluorescent microscopy images of BMDM double-labeled with MitoTracker Green and Rhod-2 to detect [Ca²⁺]_m, followed by ATP, nigericin, MSU or alum stimulation. Scale bar, 10 μm. (E) Normalized Rhod-2 fluorescence traces of BMDM challenged with ATP (blue), nigericin (red), MSU (green) or alum (purple) at 40 seconds (n=8). (F) VDAC oligomerization in BMDM pretreated −/+ CsA (1 μM, 16 h), followed by priming −/+ LPS and stimulation −/+ ATP was analyzed as in Figure 2A. IBs are one representative out of 3. Results in (B, C, and E) are mean ± s.d.. *p<0.05; **p<0.01; ***p<0.001. Two-sided unpaired t-test. See also Figure S3.

Figure 4.. MCU mediates mitochondrial Ca²⁺ uptake to trigger mPTP-VDAC channel opening and Ox-mtDNA cytosolic release

(A) Relative Rhod-2 fluorescence in LPS-primed BMDM pretreated −/+ the IP3R inhibitor xestospongin C (XeC) (5 μM, 16 h) or the MCU inhibitor ruthenium red (RuR) (10 μM, 16 h), followed by ATP (4 mM, 1 h) or nigericin (10 μM, 1 h) challenge (n=3). (B) Relative calcein fluorescence in BMDM treated as in (A) (n=3). (C) VDAC oligomerization in BMDM treated as in (A). (D and E) Amounts of Ox-mtDNA (D) or relative total mtDNA (E) in cytosols of LPS-primed BMDM pretreated −/+ XeC (5 μM, 16 h) or RuR (10 μM, 16 h) and stimulated with ATP (4 mM, 1 h) (n=3). (F) IB analysis of Casp1 p20 and IL-1β in supernatants of BMDM−/+ RuR pretreatment, primed −/+ LPS, and stimulated with ATP or nigericin as in (A). (G) IL-1β and TNF secretion by BMDM pretreated−/+ RuR (10 μM, 16 h), primed or not with LPS (200 ng/mL, 4 h) and challenged with different NLRP3 activators (n=3). (H) IB analysis of NLRP3, pro-IL-1β, pro-Casp1, ASC, and tubulin in lysates of BMDM pretreated −/+ RuR (10 μM, 16 h), before and after LPS priming (200 ng/mL, 4 h). IBs are one representative out of 3. Results in (A, B, D, E and G) are mean ± s.d.. *p<0.05; **p<0.01; ***p<0.001. ns, not significant. Two-sided unpaired t-test. See also Figure S4.

Figure 5.. FEN1-mediated mtDNA fragmentation is required for Ox-mtDNA cytosolic release and NLRP3 inflammasome activation

(A) Agarose gel electrophoresis of cytosolic (Cyto) and mitochondrial (Mito) DNA from BMDM −/+ LPS priming (200 ng/mL), −/+ nigericin (Nig) (10 μM) or ATP (4 mM) stimulation for 1 h. (B) Agarose gel electrophoresis of cytosolic and mitochondrial mtDNA PCR amplification products (591 bp and 5698 bp) form BMDM −/+ LPS priming that were challenged with different NLRP3 activators (n=3). (C) IB analysis of Casp1 p20 and IL-1β in culture supernatants of LPS (200 ng/mL, 4 h)-primed BMDM that were transduced with shCtrl, shMgme1, shMre11a, or shFen1 and stimulated −/+ ATP (4 mM, 1 h). (D) IB analysis of MGME1, MRE11A, FEN1 and Tom20 in mitochondria isolated from BMDM transduced with shCtrl, shMgme1, shMre11a, or shFen1. (E and F) Cytosolic Ox-mtDNA (n=10) (E) or relative mtDNA amount (n=3) (F) in BMDM that were transduced with shCtrl, shMgme1 and shFen1 treated as in (C). (G) Analysis of Cyto and Mito DNA from LPS-primed shCtrl- or shFen1-transduced BMDM stimulated −/+ ATP as in (C). (H) mtDNA PCR amplification products (entire genome, 16299 bp, or D-loop region, 591 bp) from LPS-primed BMDM stimulated −/+ ATP or nigericin as in (A). (I) Relative amounts of whole mtDNA genome in the BMDM analyzed in (H). (J) Analysis of Cyto or Mito DNA from BMDM that were pretreated with FEN1-IN-4 (10 μM, 16 h), LPS-primed and stimulated −/+ ATP. (K and L) Amounts of Ox-mtDNA (K) or relative total mtDNA (L) in the cytosol of BMDM treated as in (J) (n=3). (M) IB analysis of Casp1 p20 and mature IL-1β in culture supernatants of BMDM pretreated −/+ FEN1-IN-4, primed with LPS and challenged with ATP or nigericin. (N) Peritoneal IL-1β and TNF in mice treated with FEN1-IN-4 (40 mg/kg) 16 h prior to i.p. injections of alum (1 mg) or PBS, whose peritoneal exudates were analyzed 4 h later. Data are mean ± s.d. (n=6 PBS- and alum-injected groups; n=8 in alum+FEN1-IN-4-treated group). (O) Alum-induced peritoneal exudate cells (PEC), neutrophils (CD11b⁺Ly6G⁺F4/80⁻) and monocytes (CD11b⁺Ly6C⁺Ly6G⁻) in mice pretreated with FEN1-IN-4 at 40 mg/kg 16 h prior to i.p. injections of alum or PBS for 16 h. (n=6 PBS- and alum-treated groups; n=8 alum+FEN1-IN-4-treated group). The EtBr-stained agarose gels in (A, B, G, H and J) are representative of 3 independent experiments. IBs are one representative out of 3. Results in (E, F, I, K, L, N and O) are mean ± s.d.. *p<0.05; **p<0.01; ***p<0.001. ns, not significant. Two-sided unpaired t-test. See also Figure S5.

Figure 6.. mt-OGG1 and FEN1-IN-4 restrain cell-free Ox-mtDNA secretion

(A) Relative ccf-mtDNA amounts in BALF from Figure 1A (n=3 for PBS-treated groups; n=6–7 for LPS-treated groups). (B) Relative ccf-mtDNA amounts in peritoneal exudate from Figure S1A (n=4 for PBS-treated groups; n=6 for alum-treated groups). (C) Relative ratio of Fpg-treated (+) to nontreated (−) ccf-mtDNA from (B), indicating the fraction of cell-free non-Ox-mtDNA (Fpg-resistant) relative to total secreted mtDNA. (D and E) Relative amounts of ccf-mtDNA (D) and cell-free non-Ox-mtDNA (Fpg-resistant) (E) in peritoneal exudates treated as in Figure 5N. (n=4–5 for PBS- and alum-treated groups; n=9 for alum+FEN1-IN-4-treated group). (F) Experimental scheme for measuring the signaling activity of secreted ccf-mtDNA. (G and H) Relative amounts of secreted mtDNA (G) or Ox-mtDNA (H) from LPS-primed BMDM challenged with different NLRP3 activators (n=3). (I) Q-PCR quantitation of Il1b, Il6, Tnf, Cxcl10, Isg15, Ifit3 and Irf7 mRNAs in recipient BMDM incubated for 24 h with DNA (−/+ DNase I digestion) secreted by LPS-primed donor BMDM challenged with ATP or nigericin (n=3). (J) Q-PCR quantitation of Il1b, Il6, Isg15, Ifit3 and Irf7 mRNAs in recipient BMDM incubated for 24 h with DNA secreted by donor BMDM treated as in Figure S6H (n=3). Results are mean ± s.d.. *p<0.05; **p<0.01; ***p<0.001. Two-sided unpaired t-test. See also Figure S6.

Figure 7.. Induction of mtDNA release by NLRP3 Inflammasome activators leads to STING Ser³⁶⁵ phosphorylation

(A) Representative fluorescent microscopy images of LPS (200 ng/mL, 4 h)-primed WT or mt-Ogg1^Tg BMDM stained for phospho-STING (Ser³⁶⁵) before or after ATP (4 mM) or nigericin (10 μM) stimulation for 1 h. DAPI stains nuclei. Scale bar, 5 μm. (B) Percentages of cells from (A) with p-STING (Ser³⁶⁵) puncta. n=150 cells per group from 3 independent experiments. (C) IB analysis of STING Ser³⁶⁵ phosphorylation in lysates and Casp1 p20 and mature IL-1β in supernatants of BMDM pretreated −/+ EtBr (450 ng/mL, 4 days) to deplete mtDNA, followed by LPS priming and −/+ ATP stimulation as in (A). (D) Relative p-STING (Ser³⁶⁵) to total STING amounts from cells in (C). (E) IB analysis of STING Ser³⁶⁵ phosphorylation and NLRP3 in lysates and Casp1 p20 and mature IL-1β in supernatants of LPS-primed WT and Nlrp3^−/− BMDM that were pretreated −/+ CsA (1 μM, 16 hrs), VBIT-4 (10 μM, 16 hrs) or FEN1-IN-4 (10 μM, 16 hrs) and stimulated −/+ ATP as in (A). (F) Relative p-STING (Ser³⁶⁵) to total STING amounts from cells in (E). IBs are one representative out of 3. Results in (B, D and F) are mean ± s.d.. *p<0.05; **p<0.01; ***p<0.001. Two-sided unpaired t-test. See also Figure S7.