Neutrophil extracellular traps promote tPA-induced brain hemorrhage via cGAS in mice with stroke

Abstract

Intracerebral hemorrhage associated with thrombolytic therapy with tissue plasminogen activator (tPA) in acute ischemic stroke continues to present a major clinical problem. Here, we report that infusion of tPA resulted in a significant increase in markers of neutrophil extracellular traps (NETs) in the ischemic cortex and plasma of mice subjected to photothrombotic middle cerebral artery occlusion. Peptidylarginine deiminase 4 (PAD4), a critical enzyme for NET formation, is also significantly upregulated in the ischemic brains of tPA-treated mice. Blood-brain barrier (BBB) disruption after ischemic challenge in an in vitro model of BBB was exacerbated after exposure to NETs. Importantly, disruption of NETs by DNase I or inhibition of NET production by PAD4 deficiency restored tPA-induced loss of BBB integrity and consequently decreased tPA-associated brain hemorrhage after ischemic stroke. Furthermore, either DNase I or PAD4 deficiency reversed tPA-mediated upregulation of the DNA sensor cyclic GMP-AMP (cGAMP) synthase (cGAS). Administration of cGAMP after stroke abolished DNase I-mediated downregulation of the STING pathway and type 1 interferon production and blocked the antihemorrhagic effect of DNase I in tPA-treated mice. We also show that tPA-associated brain hemorrhage after ischemic stroke was significantly reduced in cGas-/- mice. Collectively, these findings demonstrate that NETs significantly contribute to tPA-induced BBB breakdown in the ischemic brain and suggest that targeting NETs or cGAS may ameliorate thrombolytic therapy for ischemic stroke by reducing tPA-associated hemorrhage.

Graphical abstract

Figure 1.

tPA activates neutrophils to release NETs after ischemic stroke. (A-B) Representative immunoblots and quantitative determinations of the amount of neutrophils in the ischemic cortex at 24 hours after stroke in mice treated with vehicle or tPA compared with mice undergoing sham surgery (n = 5). (C) Quantification of the numbers of infiltrating neutrophils in the ischemic cortex at 24 hours after stroke (n = 6). (D) Quantification of MPO activity in the ischemic brain (n = 6). (E-F) Representative immunoblots and quantitative determinations of H3Cit levels in the ischemic cortex (n = 5). (G) Representative confocal images of H3Cit⁺ neutrophils in the peri-infarct cortex. Scale bar, 50 μm. (H) Quantification of plasma DNA (n = 6). (I) Representative images of isolated peripheral blood neutrophils from mice undergoing sham surgery and ischemic mice treated with vehicle or tPA. Neutrophils were incubated with LPS for 2.5 hours. DNA was stained with Hoechst 33342 (blue), and neutrophils were stained with H3Cit (green). Arrows indicate NETs. Scale bar, 50 μm. (J-K) Quantification of the percentage of H3Cit⁺ neutrophils and NETs in unstimulated (US) and LPS-stimulated peripheral blood neutrophils (n = 8). (L-M) Quantification of the percentage of H3Cit⁺ neutrophils and NETs in peripheral blood neutrophils isolated from naïve mice, mice undergoing sham surgery, or ischemic mice (n = 5-8). Neutrophils from naive mice or mice undergoing sham surgery were treated with vehicle or 100 μg/mL of tPA. Neutrophils from ischemic mice were treated with vehicle or 25, 50, or 100 μg/mL of tPA. (N) Representative immunoblots of PAD4 expression in neutrophils isolated from ischemic mice. Neutrophils were treated with vehicle or 100 μg/mL of tPA. (O-P) Quantification of the percentage of H3Cit⁺ neutrophils and NETs in neutrophils isolated from ischemic mice (n = 8). Neutrophils were treated with vehicle, 100 μg/mL of tPA, or tPA in combination with the PAD inhibitor Cl-amidine. (Q) Representative immunoblots of lipoprotein receptor–related protein 1 (LRP-1) expression in neutrophils isolated from ischemic mice. Neutrophils were treated with vehicle or 100 μg/mL of tPA. (R) Representative immunoblots of PAD4 expression in neutrophils isolated from ischemic mice. Neutrophils were treated with 100 μg/mL of tPA or tPA in combination with the LRP antagonist RAP. (S-T) Quantification of the percentage of H3Cit⁺ neutrophils and NETs in neutrophils isolated from ischemic mice (n = 8). Neutrophils were treated with vehicle, 100 μg/mL of tPA, or tPA in combination with the LRP antagonist RAP. Values are means ± standard deviation. *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 2.

LRP-1 mediates the effects of tPA on neutrophil recruitment and NET formation after ischemic stroke. (A-B) Representative immunoblots and quantitative determinations of LRP-1 expression in the ischemic cortex at 24 hours after stroke in mice treated with vehicle or tPA compared with mice undergoing sham surgery (n = 5). (C) Representative confocal images of LRP-1 immunostaining in the ischemic cortex. DNA was stained with 4′,6-diamidino-2-phenylindole (blue). Independent experiments are repeated ≥3 times. Scale bar, 40 μm. (D) Quantification of MPO activity in the ischemic brain (n = 6). (E) Representative immunoblots of H3Cit in the ischemic cortex. (F) Quantification of MPO activity in the ischemic brain (n = 6). (G) Representative immunoblots of H3Cit in the ischemic cortex. (H) Representative confocal images of fibrin intravascular deposits (green) and CD31⁺ microvessels (red) in the infarct areas at 24 hours. Scale bar, 10 μm. (I) Quantification of fibrin intravascular deposits for each group (n = 6). Values are means ± standard deviation. *P < .05. NS, not significant; TXA, tranexamic acid.

Figure 3.

DNase I treatment reduces tPA-mediated BBB breakdown and cerebral hemorrhage after ischemic stroke. (A) Permeability coefficient (P) of human brain endothelial monolayers to 40 KDa of FITC-dextran at 0 to 6 hours after normoxia or OGD with or without NETs (1.5 μg/mL). (B) Representative immunoblots of H3Cit levels in the ischemic cortex at 24 hours after stroke in mice treated with vehicle, tPA, or tPA in combination with DNase I. (C) Quantification of the numbers of H3Cit⁺ neutrophils in the ischemic cortex (n = 6). (D) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show Evans blue extravasation 24 hours after stroke in mice treated with vehicle, tPA, tPA in combination with DNase I, or DNase I alone. (E) Quantification of Evans blue fluorescence intensity for each group (n = 8). (F-G) Representative immunoblots and quantification of immunoglobulin G (IgG) levels in capillary-depleted brain tissue at 24 hours after stroke (n = 5). (H) Representative immunoblots of the tight junction proteins zonula occludens-1 (ZO-1), claudin-5, and occludin and the adherens junction protein vascular endothelial–cadherin (VE-cadherin) in isolated brain microvessels. (I) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show cerebral hemorrhage 24 hours after stroke in mice treated with vehicle, tPA, tPA in combination with DNase I, or DNase I alone. (J) Quantification of cerebral hemorrhage by spectrophotometric hemoglobin assay (n = 8). (K-M) DNase I treatment improved neurological functions in forelimb force test and beam walking test (n = 10). Values are means ± standard deviation. *P < .05 vs control group (A), ^#P < .05 vs OGD group (A), *P < .05 (C,E,G,J-M).

Figure 4.

NETs and PAD4 contribute to tPA-mediated BBB disruption and cerebral hemorrhage after ischemic stroke. (A-B) Representative immunoblots and quantification of PAD4 levels in the ischemic cortex 24 hours after stroke in mice treated with vehicle or tPA compared with mice undergoing sham surgery (n = 5). (C) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show Evans blue extravasation 24 hours after stroke in WT and Pad4^−/− mice treated with tPA. (D) Quantification of Evans blue fluorescence intensity for each group (n = 8). (E-F) Representative immunoblots and quantification of IgG levels in capillary-depleted brain tissue at 24 hours (n = 5). (G-K) Representative immunoblots and quantification of zonula occludens-1 (ZO-1), occludin, claudin-5, and vascular endothelial–cadherin (VE-cadherin) in isolated brain microvessels (n = 5). (L) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show cerebral hemorrhage 24 hours after stroke in WT and Pad4^−/− mice treated with tPA. (M) Quantification of cerebral hemorrhage by spectrophotometric hemoglobin assay (n = 8). (N-P) Effects of PAD deficiency on forelimb force test and beam walking test 24 hours after stroke (n = 10). Values are means ± standard deviation. *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 5.

DNase I treatment or PAD4 deficiency inhibits tPA-induced upregulation of cGAS-STING and type 1 IFN signaling. (A-B) Representative immunoblots and quantification of cGAS expression in the ischemic cortex 24 hours after stroke in WT mice treated with vehicle, tPA, or tPA in combination with DNase I and in WT and Pad4^−/− mice treated with tPA compared with mice undergoing sham surgery (n = 5). (C) Representative immunoblots for STING, phosphorylated (pTBK1) and total TBK1, and pIRF3 and total IRF3 expression in the ischemic cortex. (D-F) Quantification of band intensities for each group (n = 5). (G-H) Quantification of IFN-β and IL-6 levels by enzyme-linked immunosorbent assay in the ischemic cortex (n = 6). (I-J) Double immunostaining of cGAS and STING with microglial cells (Iba1), macrophages (RM0029-11H3), and neutrophils (Ly6G) in mice subjected to MCAO and tPA treatment. Scale bar, 20 μm. (K-L) Quantification of activated microglia and macrophage infiltration in the ischemic cortex (n = 6). (M-O) Quantification of the numbers of pIRF3⁺, IFN-β⁺, and IL-6⁺ microglial cells (Iba1) in the ischemic cortex (n = 6). Values are means ± standard deviation. *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NS, not significant.

Figure 6.

DNase I–mediated cerebrovascular protection and antihemorrhagic effects after ischemic stroke are abolished by cGAMP. (A) Representative immunoblots for STING, pTBK1 and total TBK1, and pIRF3 and total IRF3 expression in the ischemic cortex of tPA-treated mice after treatment with DNase I or DNase I in combination with cGAMP. (B-D) Quantification of band intensities for each group (n = 5). (E-F) Quantification of IFN-β and IL-6 levels by enzyme-linked immunosorbent assay in the ischemic cortex (n = 6). (G-H) Representative immunoblots and quantification of IgG levels in capillary-depleted brain tissue. (I-J) Quantification of extravascular dextran fluorescence in tPA-treated mice after treatment with DNase I or DNase I in combination with cGAMP. Scale bar, 20 μm. At 24 hours after stroke, mice were administered an intravascular injection of 40 KDa of FITC-labeled dextran, and brain sections were stained with CD31. (K) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show cerebral hemorrhage 24 hours after stroke. (L) Quantification of cerebral hemorrhage (n = 8). (M-O) Effects of DNase I or combination of DNase I and cGAMP on forelimb force test and beam walking test 24 hours after stroke in tPA-treated mice (n = 10). Values are means ± standard deviation. *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 7.

tPA-associated BBB disruption and hemorrhage after ischemic stroke are rescued by loss of cGAS. (A) Representative immunoblots for STING, pTBK1 and total TBK1, and pIRF3 and total IRF3 expression in the ischemic cortex of WT and cGas^−/− mice treated with tPA. (B-D) Quantification of band intensities for each group (n = 5). (E-F) Quantification of IFN-β and IL-6 levels by enzyme-linked immunosorbent assay in the ischemic cortex (n = 6). (G-H) Representative in vivo multiphoton microscopic images of IV injected FITC-dextran (molecular weight, 40 KDa) leakage from cortical vessels and quantification of the permeability surface (P) product of FITC-dextran at 24 hours after stroke in WT and cGas^−/− mice treated with tPA (n = 6). Scale bar, 100 μm. (I-J) Representative immunoblots and quantification of IgG levels in capillary-depleted brain tissue (n = 5). (K) Representative images of the dorsal surface (upper panel) and a coronal section (bottom panel) show cerebral hemorrhage 24 hours after stroke. (L) Quantification of cerebral hemorrhage (n = 8). (M-O) Effects of cGAS deficiency on forelimb force test and beam walking test 24 hours after stroke in tPA-treated mice (n = 10). (P) Model figure showing how tPA stimulates NET formation and how NETs are sensed by microglia after ischemic stroke. Values are means ± standard deviation. *P < .05. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.