Mitochondrial DNA‐Mediated Immune Activation After Resuscitation From Cardiac Arrest

Abstract

Background

Postcardiac arrest syndrome is characterized by systemic inflammation that contributes to poor outcomes after resuscitation from sudden cardiac arrest. Mitochondrial DNA (mtDNA) has been implicated as a proinflammatory stimulus in other contexts, but its role in postcardiac arrest syndrome is unclear. We determined if postcardiac arrest syndrome is characterized by a rise in circulating mtDNA, how mtDNA activates immune cells, and if targeting mtDNA‐sensing pathways attenuates leukocyte activation.

Methods

Plasma mtDNA and nuclear DNA levels were measured ~4‐hours after return of spontaneous circulation following sudden cardiac arrest in swine (n=8) and humans (n=57). Additionally, porcine peripheral blood mononuclear cells were treated with mtDNA or extracellular vesicles (EVs) isolated from porcine plasma collected after return of spontaneous circulation. Pharmacological agents were used to inhibit TLR9 (toll‐like receptor 9)‐ and cGAS (cyclic GMP–AMP synthase)‐mediated mtDNA sensing.

Results

A ~250‐fold elevation in circulating mtDNA was observed after return of spontaneous circulation in swine despite negligible changes in circulating nuclear DNA, a finding that was corroborated in humans. Circulating mtDNA was largely encapsulated within EVs in both species, suggesting a conserved mechanism of release. In vitro studies demonstrated that peripheral blood mononuclear cell internalization of mtDNA‐containing‐EVs was required for leukocyte activation. This response was attenuated by EV disruption, DNA degradation, and blockade of TLR9 or cGAS pathways, identifying novel targets to modulate inflammation in postcardiac arrest syndrome.

Conclusions

Brief whole‐body ischemia and reperfusion in the context of resuscitation from sudden cardiac arrest elicits mtDNA release, primarily within EVs, that triggers leukocyte activation. Targeting mtDNA release or its downstream sensors may offer a new therapeutic strategy to improve outcomes after sudden cardiac arrest.

Nonstandard Abbreviations and Acronyms

BI

brief ischemia

dsDNA

double stranded DNA

EV

extracellular vesicle

mtDNA

mitochondrial DNA

nucDNA

nuclear DNA

OHCA

out‐of‐hospital cardiac arrest

PBMCs

peripheral blood mononuclear cells

PCAS

post‐cardiac arrest syndrome

ROSC

return of spontaneous circulation

SCA

sudden cardiac arrest

sR‐OHCA

successfully resuscitated out‐of‐hospital cardiac arrest

TR

transfection reagent

Clinical Perspective.

What Is New?

• Our results demonstrate that circulating mitochondrial DNA, primarily encapsulated in extracellular vesicles, is released into the bloodstream after resuscitation from sudden cardiac arrest.

• Extracellular vesicle‐encapsulated mitochondrial DNA triggers immune cell activation, evidenced by phenotypic shifts toward inflammatory dendritic cells and macrophages, as well as increased proinflammatory cytokine secretion that can be attenuated by pharmacological inhibition of TLR9 (toll‐like receptor 9) and cGAS (cyclic GMP–AMP synthase) pathways.

What Are the Clinical Implications?

• Identification of mitochondrial DNA as a key driver of sterile inflammation in postcardiac arrest syndrome reveals a potential target for interventions aimed at reducing multiorgan damage and improving neurological outcomes, as therapeutic strategies to block mitochondrial DNA release or downstream signaling (eg, TLR9/cGAS inhibition) may limit harmful proinflammatory cascades and bolster long‐term survival following resuscitation from cardiac arrest.

Each year in the United States, nearly 600 000 adults experience sudden cardiac arrest (SCA), often with little warning and frequently outside the hospital. ¹ Although improvements in cardiopulmonary resuscitation have increased initial survival rates, <10% of these patients regain normal brain function and achieve long‐term survival. ² , ³ A primary reason for these dismal postresuscitation outcomes is post‐cardiac arrest syndrome (PCAS), a condition characterized by systemic inflammation, cardiac dysfunction, and neurological injury following return of spontaneous circulation (ROSC). ⁴ , ⁵ Higher levels of inflammatory cytokines (eg, TNFα [tumor necrosis factor alpha], IL‐1β [interleukin 1beta], IL‐6 [interleukin‐6]) are linked to poor outcomes in survivors of cardiac arrest, emphasizing the possible clinical impact of reducing unnecessary immune activation. ⁶ , ⁷ Unfortunately, the fundamental mechanisms underlying the robust postresuscitation inflammatory response remain poorly understood, which has hindered the development of effective treatments for PCAS.

Recent findings indicate that innate immune activation triggered by danger‐associated molecular patterns is central to PCAS severity, but specific pro‐inflammatory stimuli have yet to be identified. ⁸ One intriguing candidate is mitochondrial DNA (mtDNA). Because mitochondria descended from α‐proteobacteria, ⁹ mtDNA retains bacteria‐like features, such as unmethylated cytosine‐phosphate‐guanine repeats, that can provoke an immune response. ¹⁰ , ¹¹ Clinical and preclinical data suggest a link between elevated mtDNA levels and inflammation in a variety of contexts including heart failure, ¹² atherogenesis, ¹³ and ischemic heart disease. For example, in experimental myocardial infarction ¹⁴ , ¹⁵ and clinical cardiopulmonary bypass, ¹⁶ elevated circulating mtDNA levels correlate with increased circulating and tissue levels of inflammatory cytokines. Preliminary work also suggests that mtDNA can accumulate in the bloodstream after oxidative stress without overt necrosis, ⁶ , ¹⁷ potentially acting as a powerful trigger for sterile inflammation. However, whether mtDNA is released after brief whole‐body ischemia in the context of SCA has not been directly studied. ¹⁸ Moreover, how circulating mtDNA might be detected by leukocytes (via TLR9 [toll‐like receptor 9]) ¹⁹ or the cyclic GMP‐AMP synthase–stimulator of interferon genes [cGAS‐STING] pathway ²⁰ ) and whether blocking these pathways can dampen postresuscitation immune activation is unknown. Accordingly, we aimed to delineate the role of mtDNA in activating immune cells after SCA and determine whether interfering with mtDNA‐mediated inflammatory signaling pathways could mitigate proinflammatory leukocyte activation. By integrating clinical data from survivors of out‐of‐hospital cardiac arrest (OHCA) with mechanistic experiments in porcine models of systemic (SCA) and regional brief ischemia (BI), we tested the hypothesis that release of mtDNA, rather than nuclear DNA (nucDNA), triggers robust post‐ROSC immune activation via TLR9 and cGAS signaling pathways. As detailed in the results below, our results show that extracellular vesicle (EV)‐encapsulated mtDNA serves as a potent inflammatory stimulus in PCAS, offering a novel therapeutic avenue to potentially improve outcomes among patients resuscitated from SCA.

Methods

Data Availability

The data supporting the study’s findings are available from the corresponding author upon reasonable request. Please see Data S1 for complete methodological details.

Ethics Statement

Investigators were blinded to each patient and animal’s experimental group (eg, baseline versus 1‐hour post‐ROSC versus 4‐hour post‐ROSC) while performing blood sampling, in vitro experimentation, and data analysis. All experimental procedures and protocols conformed to institutional guidelines for the care and use of animals in research and were approved by the State University of New York at Buffalo Institutional Animal Care and Use Committee (ID: MED15083Y). All clinical procedures and protocols conformed to institutional guidelines for human subjects research and were approved by the Institutional Review Board at the University at Buffalo (ID: STUDY00000284).

Porcine SCA Model

A total of 8 Yorkshire‐cross farm‐bred pigs (2 male/6 female) were studied in the closed‐chest state using a protocol summarized in Figure S1, as described previously. ²¹ Briefly, animals were subjected to 10 minutes of cardiac arrest, followed by up to 20 minutes of cardiopulmonary resuscitation with manual chest compressions, mechanical ventilation, and defibrillation until achieving ROSC (unassisted arterial systolic blood pressure ≥80 mmHg for at least 1 minute). If ROSC was achieved, blood plasma samples were collected at 1 and 4 hours post ROSC.

Porcine Brief Myocardial Ischemia Model

A total of 5 male Yorkshire‐cross farm‐bred pigs were studied in the closed‐chest state to induce a transient 10‐minute occlusion of the left anterior descending coronary artery, as described previously. ²² In short, animals were subjected to 10 minutes of BI localized to the left anterior descending coronary artery perfusion territory via inflation of a percutaneous angioplasty balloon. After 10 minutes of ischemia, the balloon was deflated, blood flow was restored, and subjects survived for 1 hour post reperfusion (post BI). Plasma samples from the coronary sinus were collected before the localized ischemic event (pre‐BI) and 1 hour post BI (full methodological details provided in Data S1).

Human Out‐of‐Hospital Cardiac Arrest Plasma Sample Collection

Blood plasma samples were collected from 57 successfully resuscitated patients of nontraumatic OHCA from 4 tertiary‐level hospitals in the Buffalo‐Niagara metropolitan area and followed for 30 days following ROSC. Patients with active cancer, infection, cardiac pathology, or surgery in the 3 months before presentation were excluded. Cardiac arrest characteristics, demographics, comorbidities, and therapeutic interventions were recorded (Table 1), and plasma samples were collected between 2 and 6 hours following ROSC and frozen at −80 °C to be analyzed via quantitative polymerase chain reaction (qPCR). OHCA samples were compared against healthy control subjects (44±3 years old, 6 male/3 female). Likewise, blood plasma was also collected from a validation cohort of 32 additional successfully resuscitated patients of nontraumatic OHCA, between 2 and 6 hours following ROSC.

Table 1.

Patient Demographics Stratified by 30‐Day Survivorship

Patient demographics	All sR‐OHCA	30 days survivors	30 days nonsurvivors	P value (survivors vs nonsurvivors)
Age, y	63.1±2.0	61.9±2.7	64.2±3.0	0.555
Sex, male/female	34/17	14/8	20/9	0.77
Cardiopulmonary resuscitation time to return of spontaneous circulation, min	14.8±2.0	11.4±2.9	17.3±2.8	0.162
No. of 1 mg epinephrine doses	1.9±0.3	1.3±0.4	2.8±0.4*	0.009
Comatose (%)	78±6	61±9	96±4*	0.002
Intubated (%)	85±5	71±9	100±0*	0.003
Targeted temperature management (%)	37±7	32±9	42±10	0.530
Length of hospital stay, h	238.9±29.5	288.8±43.2	194.1±39.2	0.109
24‐m survival (%)	45	86	0	‐‐‐
Admission white blood cell count, 103/μL	13.3±0.9	12.0±0.6	14.6±1.7	0.165

Categorical variables are presented as frequencies (%) and continuous data are presented as mean±SEM. Between‐group differences were assessed by 2‐tailed Student’s t test, except sex distribution, which was assessed by Fisher’s exact test.

sR‐OHCA indicates successfully resuscitated out‐of‐hospital cardiac arrest.

^* P<0.05 vs survivors, represents 2‐tailed Student’s t test.

Plasma mtDNA and nucDNA Digestion and EV Lysis

To digest circular mtDNA and nucDNA from plasma and plasma‐EVs, samples were treated with 50 U/mL DNase I (Thermo, Ref: 89836) at 37 °C for 1 hour. Subsequently, DNase I was inactivated with 50 mM EDTA for 10 minutes at 65 °C, per the manufacturer’s protocol. To lyse open EVs, either from whole plasma or EV isolate, 1× RIPA buffer was added to samples and incubated on ice for 30 minutes and then were diluted with PBS.

Quantification of Plasma mtDNA, nucDNA, and Total Double‐Stranded DNA Via qPCR

Whole plasma from patients and swine was clarified by centrifugation of 500g for 5 minutes before transferring the supernatant to a DNeasy Blood and Tissue Kit (Qiagen, Ref: 69504) to isolate plasma DNA, per the manufacturer’s protocol. To identify EV‐enclosed DNA, isolated plasma EVs were additionally treated with DNase I (described previously) to remove free‐floating and nonencapsulated DNA before isolation via a DNeasy kit. Following DNA isolation, human and porcine samples were quantified by qPCR using SsoAdvanced Universal SYBR Green Supermix (BIO‐RAD, Ref: 1725270) against a standard curve of mtDNA and nucDNA in a known concentration using primers to detect mitochondrial (Cyt‐B, D‐Loop) or nuclear (β2M) genes (Table S1), as previously described. ²³ Additionally, nonspecific quantification of double‐stranded DNA (dsDNA) was conducted using Quanti‐iT PicoGreen dsDNA Assay Kit (Thermo, Ref P7589), per the manufacturer’s protocol.

qPCR of Inflammatory Transcripts

RNA was extracted from thawed cell lysate via RNeasy Kit (Qiagen, Ref: 74104), per the manufacturer’s protocol. A NanoDrop spectrophotometer (Thermo Scientific, Ref: ND2000) was used to determine RNA purity and concentration. Total RNA was reverse transcribed and analyzed by qPCR as previously described. ²⁴ SsoAdvanced Universal SYBR Green Supermix (BIO‐RAD, Ref: 1725270) was used in combination with primers for IFNα (interferon alpha), IL‐1α, IL‐1β, IL‐6, IL‐8, TNFα, and β2M (beta‐2‐microglobulin) (Table S1). The comparative cycle threshold method was used to determine the mRNA expression for each target gene using the gene for β2M as the reference.

Quantification of Inflammatory Cytokine Release, Reactive Oxygen Species/Reactive Nitrogen Species, and Intracellular 2′3′‐Cyclic Guanosine Monophosphate–Adenosine Monophosphate

Conditioned media was collected 24 hours post activation from the supernatant of each cell culture dish well. Conditioned media samples were clarified by centrifugation at 500g for 5 minutes to remove cellular debris. Conditioned media was thawed and quantified in duplicate with porcine‐specific ELISA kits according to manufacturer’s instructions (Table S2). Additionally, conditioned media levels of reactive oxygen and nitrogen species were measured using a DCF‐based kit (Abcam, Ref: ab238535), and intracellular levels of 2′3′‐cyclic guanosine monophosphate–adenosine monophosphate (cGAMP; Invitrogen, Ref: EIAGAMP) after a 30‐minute incubation with 1× RIPA buffer on ice, following the manufacturer’s recommendations.

PBMC Isolation and Culture

Naïve peripheral blood mononuclear cells (PBMCs) were isolated from healthy White Yorkshire×Landrace swine (Oak Hill Genetics, Ewing, IL). Approximately 8 mL of whole blood was collected and added to Ficoll CPT Vacutainers (BD, Ref: 362761), followed by centrifugation at 1500g for 28 minutes at 4 °C. As per the manufacturer’s protocol, the PBMC layer was collected, filtered through a 40 μm filter into a 50 mL Falcon tube, and washed with 20 mL of PBS. The filtered PBMCs were then centrifuged at 500g for 5 minutes and washed with PBS. Before plating, isolated PBMCs were treated with red blood cell lysis buffer for 5 minutes at room temperature, followed by centrifugation and PBS washing. Purified PBMCs were resuspended in RPMI 1640 supplemented with 1× penicillin/streptomycin, 1× sodium pyruvate, and 10% FBS. The cells were cultured at 37 °C in a humidified atmosphere of 5% CO₂ and ambient oxygen for 1 week, with bidaily media changes. After 1 week, the cells were subjected to activation assays. After aspiration of the culture medium and washing with PBS, naïve PBMCs were activated for 24 hours with either 1× lipopolysaccharide (5 μM; Invitrogen, Ref: 00‐4976‐93) or 2 μg/mL mtDNA with Lipofectamine 2000 Transfection Reagent (TR; Invitrogen, Ref: 11668027), as previously described. ²⁵ During this activation period, pre‐ and post‐ROSC plasma EV samples were labeled with 5 μM Vybrant DiD Lipophillic Dye (Thermo, Ref: V22887) and were added in place of FBS. At the end of the 24‐hour stimulation period, the conditioned medium was harvested and stored at −20 °C for subsequent ELISA and reactive oxygen species/reactive nitrogen species assays, and the cells were collected for flow cytometric analysis or lysed in Qiazol Reagent for downstream qPCR. Each figure represents the collection of biologically distinct PBMC populations, therefore leading to minute differences between experimental results.

Small Molecule Inhibition of mtDNA‐Induced PBMC Inflammation

To mimic the conditions under which a therapeutic agent would be delivered to successfully resuscitated patients, small molecular inhibitions were added in suspension to basal media at the same time as activation by mtDNA+TR. Inhibitors were used in a concentration‐response assay against TLR9 (2–8 μM ODN 2088; InvivoGen, Ref: tlrl‐2088 or 2–8 μM ODN inh‐18; InvivoGen, Ref: tlrl‐inh18), cGAS (5–20 μM G140; InvivoGen, Ref inh‐g140), STING (1–4 μM H151; InvivoGen, Ref: inh‐h151), and PKG1 (protein kinase G type 1; 1–4 μM Rp‐8‐Br‐PET‐cGMP; Tocris, Ref: 3028), as previously described. ²⁶ , ²⁷ , ²⁸ All small molecular inhibitors were prepared and frozen at −20 °C until being thawed immediately before use.

Plasma EV Isolation

Plasma from humans or swine was centrifuged at 500g for 5 minutes to remove large particulates. The resulting supernatant was filtered through a 0.2 μm filter to eliminate residual debris. Further purification was achieved using an Amicon Ultra Centrifugal Filter (100 kDa MWCO), with centrifugation at 4000g for 60 minutes at 4 °C. The isolated plasma‐EV were then diluted in ratios of 1:1000 and characterized using a ZetaView ×20 Series (Particle Metrix) particle tracking analysis. Additionally, the plasma‐EV preparations were evaluated using the Exo‐Check Exosome Antibody Array (SBI, Ref: EXORAY200), according to the manufacturer’s instructions.

Leukocyte Isolation and Flow Cytometry

Following a 24‐hour activation period, cultured leukocytes were labeled for porcine‐specific inflammatory cell‐surface markers ²⁹ , ³⁰ and viability was assessed by 7AAD staining. Flow cytometric assessment was performed on a BD LSRFortessa Cell Analyzer and analyzed via De Novo Software FCS Express 7 Plus (version 7.16.0035) following consistent gating strategy implementation (Figure S2) to quantify subpopulations of inflammatory granulocytes (CD172+CD163−), inflammatory dendritic cells (CD172+CD16+), and inflammatory macrophages (CD14+CD163+) ²⁹ , ³⁰ (Table S3). Additional staining of plasma‐EVs with 5 μM Vybrant DiD Lipophillic Dye allowed the detection of plasma‐EV uptake by PBMC populations.

Statistical Analysis

Data are reported as mean±SEM, and sex differences were assessed by Fisher’s exact test. Between‐group differences in end points measured across multiple time points or conditions were assessed by repeated measures ANOVA and the post hoc Tukey honestly significant difference) test, and between‐group differences in end points of 2 samples were assessed by 2‐tailed unpaired Student’s t test. For concentration‐dependent inhibition curves, simple linear regressions were used to determine the 50% inhibitory concentration (IC50) and the coefficient of determination (R2). Significance was set to an alpha level of 0.05. To ensure independence of sampling, all statistical analyses were performed on values representing individual biological subjects (eg, animals or human donors). Where technical replicates were collected (eg, qPCR wells or tissue sections), they were used for quality control only, and a single representative value per subject per condition was included in comparative analyses, as noted in figure legends for transparency. All data were compiled with Microsoft Excel (version 2402), and analysis and plotting were performed with GraphPad Prism Software (version 10.2.0).

Results

Elevation of Circulating mtDNA in Pigs and Humans With PCAS

To investigate the relevance of mtDNA release in clinical PCAS, we measured circulating mtDNA levels in patients with successfully resuscitated OHCA (sR‐OHCA) (n=57; Table 1) within 6 hours of ROSC. Relative to healthy controls (n=9, 44±3 years old, 6 male/3 female), patients with sR‐OHCA exhibited significantly elevated plasma mtDNA (~14‐fold) and nucDNA (~10%; both P<0.001) that did not differ between 30‐day survivors and nonsurvivors (Figure 1A and 1B). This finding was further supported in a validation cohort, where compared with healthy controls, patients with sR‐OHCA exhibited significantly elevated plasma mtDNA (~18‐fold; P=0.007; Figure S3); however, these levels did not differ between 30‐day survivors and nonsurvivors.

Figure 1. Elevation of circulating mtDNA following SCA in humans and swine.

qPCR analysis of 6‐h post‐ROSC whole plasma levels of (A) mtDNA (Cyt‐b) and (B) nucDNA (β2M) within patients with sR‐OHCA stratified by 30‐d survivorship and compared with healthy controls. n = 9, 57, 28, 29, respectively. qPCR analysis representing detected porcine circulating mtDNA and nucDNA before 10‐min untreated VF (pre‐SCA), 1‐h post‐ROSC, and 4‐h post‐ROSC, compared with a standard curve. C, mtDNA (Cyt‐B) detection was normalized to isolated plasma volume, and fold‐change was calculated compared with pre‐SCA. D, nucDNA (β2M) detection was normalized to isolated plasma volume and fold‐change was calculated compared with pre‐SCA. Data represent biological n=8 and technical n=4. qPCR analysis representing detected porcine coronary sinus levels of mtDNA and nucDNA before (pre‐BI) and 1‐h after reperfusion from regional brief ischemia (10‐min) of the LAD perfusion territory, compared with a standard curve. E, mtDNA (Cyt‐F) detection was normalized to isolated plasma volume, and fold‐change was calculated compared with pre‐BI. F, nucDNA (β2M) detection was normalized to isolated plasma volume and fold‐change was calculated compared with pre‐BI. Data represent biological n=5 and technical n=4. Data represent mean±SEM. BI indicates brief ischemia; LAD, left anterior descending coronary artery; mtDNA, mitochondrial DNA; nucDNA, nuclear DNA; ROSC, return of spontaneous circulation; SCA, sudden cardiac arrest; sR‐OHCA, successfully resuscitated out‐of‐hospital cardiac arrest; and VF, ventricular fibrillation.

To corroborate these findings, we measured plasma levels of both mtDNA and nucDNA at multiple time points up to 4 hours post ROSC in swine. Despite a brief ischemic period of insufficient duration to induce necrosis, ²² circulating mtDNA was markedly elevated (~250‐fold increase) at 1 hour post ROSC (P=0.029) and remained ~175‐fold higher than baseline at 4 hours post ROSC (P=0.015; Figure 1C). These increases in circulating mtDNA occurred despite a negligible rise in circulating nucDNA (Figure 1D), although total dsDNA was also elevated at 1 and 4 hours post ROSC (Figure S4).

Based on speculation that ischemic myocardium may be a primary source of mtDNA release, mtDNA and nucDNA levels were measured in the blood samples collected from the coronary sinus before and after brief myocardial ischemia induced by a 10‐minute coronary occlusion. We have previously shown that this duration of regional myocardial ischemia causes a transient increase in myocyte TUNEL‐positivity without pathologic evidence of myocardial necrosis, ²² whereas left anterior descending coronary artery occlusion of prolonged duration (>15 minutes) is known to produce necrosis and infarction. ³¹ BI led to a rise in coronary sinus levels of mtDNA (4.6±1.4‐fold versus pre‐BI; P=0.054) but not nucDNA (P=0.354) in the coronary sinus plasma 1 hour after reperfusion (Figure 1E and 1F). These data demonstrate that brief myocardial ischemia is associated with mtDNA release, even in the absence of necrosis. However, because the magnitude of mtDNA release after brief myocardial ischemia was dramatically lower than that observed after brief whole‐body ischemia in swine with PCAS, extracardiac sources are likely a primary source of mtDNA after resuscitation from SCA.

Collectively, these clinical and preclinical data suggest that mtDNA is released following brief ischemia, both systemically (SCA) and regionally (BI), which positions mtDNA as a potential proinflammatory stimulus to trigger leukocyte activation in PCAS.

Intracellular mtDNA Increases Leukocyte Inflammatory Activation

We next explored whether elevated mtDNA (or nucDNA) drives innate immune cell responses in vitro. PBMCs were stimulated with either 2 μg/mL of mtDNA or nucDNA, delivered in suspension or via TR. Lipopolysaccharide served as a positive control, and a TR alone condition served as a negative control. Only transfected mtDNA (mtDNA+TR) induced surface marker changes comparable to lipopolysaccharide stimulation, including increased frequency of inflammatory dendritic cells (CD172+CD16+), granulocytes (CD172+CD163−), and macrophages (CD14+CD163+; Figure 2A through 2C). In contrast, nucDNA, whether transfected or in suspension, did not provoke a similar inflammatory shift. Consistent with these phenotypic changes, mtDNA+TR stimulated significant secretion of TNFα (9.5±9.5 versus 402.0±38.9 pg/mL; P<0.001 Figure 2D), IL‐1β (5.6±3.6 versus 46.8±6.9 pg/mL; P=0.003; Figure 2E), and IL‐6 (0.0±0.0 versus 34.1±0.4 pg/mL; P<0.001; Figure 2F), as well as higher expression of inflammatory genes (NFκB [nuclear factor kappa B], TNFα, IL‐6, IFNα; Figure S5). Additionally, only mtDNA+TR stimulated a significant enrichment of intracellular levels of cGAMP, the endogenous second messenger of cGAS‐STING (0.0±0.0 versus 2.7±0.4 pmol/mL lysate; P<0.001; Figure 2G). These findings indicate that leukocyte uptake of mtDNA is essential for triggering proinflammatory pathways, whereas nucDNA alone does not induce the same effect.

Figure 2. Intracellular mtDNA increases leukocyte inflammatory activation.

Porcine PBMCs (CD45+7AAD−) were cultured for 24 h in the presence of 2 μg/mL nucDNA or 2 μg/mL mtDNA in suspension (+media) or with lipofectamine 2000 (+TR). Flow cytometry was used to determine population proportion of (A) inflammatory dendritic cells (CD172+CD16+), (B) inflammatory granulocytes (CD172+163−), and (C) inflammatory macrophages (CD14+CD163+). Following clarification, conditioned media levels of inflammatory cytokines (D) TNFα, (E) IL‐1β, (F) IL‐6, and (G) intracellular cGAMP levels were detected via ELISA. Data represent biological n=3 and technical n=3. Values are mean±SEM. 7AAD indicates 7‐aminoactinomycin D; cGAMP, 2′3′‐cyclic guanosine monophosphate–adenosine monophosphate; IL‐1β, interleukin‐1β; IL‐6, interleukin‐6; LPS, lipopolysaccharide; mtDNA, mitochondrial DNA; nucDNA, nuclear DNA; PBMC, peripheral blood mononuclear cell; TNFα, tumor necrosis factor‐α; and TR, transfection reagent.

Post‐ROSC Plasma and mtDNA Stimulate Proinflammatory Cell Expansion and Cytokine Release From Cultured PBMCs

To further assess whether mtDNA in post‐ROSC plasma can shift leukocytes toward a proinflammatory state, PBMCs were cultured with plasma collected from swine 4‐hour post‐ROSC, mtDNA+TR, or lipopolysaccharide, compared with pre‐SCA plasma, TR (−), or basal media, respectively. PBMCs exposed to mtDNA+TR displayed a surface marker profile that paralleled the effect of 4‐hour post‐ROSC plasma (Figure 3A through 3C), including expansion of inflammatory cell subsets (dendritic cells, granulocytes, and macrophages). Like the effects seen following 4‐hour post‐ROSC plasma activation, mtDNA+TR stimulated a significant increase in secretion of TNFα (39.6±2.6 versus 310.1±9.9 pg/mL; P<0.001; Figure 3D) and IL‐1β (65.3±23.2 versus 236.0±34.0 pg/mL; P=0.014; Figure 3E), compared with the TR (−) control. Together, these data suggest that mtDNA within post‐ROSC plasma is capable of shifting leukocytes toward a proinflammatory phenotype and inducing inflammatory cytokine release.

Figure 3. Post‐ROSC plasma and mtDNA transfection lead to cultured PBMC proinflammatory cell differentiation and cytokine release.

Porcine PBMCs (CD45+7AAD−) were cultured for 24‐h in the presence of LPS, 4‐h post‐ROSC plasma, or 2 μg/mL mtDNA with lipofectamine 2000 (TR+mtDNA), and compared with basal media, pre‐SCA plasma, and TR—controls, respectively. Flow cytometry was used to determine population proportion of (A) inflammatory dendritic cells (CD172+CD16+), (B) inflammatory granulocytes (CD172+163−), and (C) inflammatory macrophages (CD14+CD163+). Data represent biological n=4 and technical n=3. Following clarification, conditioned media levels of inflammatory cytokines (D) TNFα and (E) IL‐1β were detected via ELISA. Data represent biological n=3 and technical n=3. Values are mean±SEM. 7AAD indicates 7‐aminoactinomycin D; IL‐1β, interleukin‐1β; LPS, lipopolysaccharide; mtDNA, mitochondrial DNA; PBMC, peripheral blood mononuclear cell; ROSC, return of spontaneous circulation; SCA, sudden cardiac arrest; TNFα, tumor necrosis factor‐α; and TR, transfection reagent.

dsDNA Is Required for Post‐ROSC Plasma‐Induced Immune Cell Activation

To determine whether mtDNA drives innate immune cell activation elicited by exposure to post‐ROSC plasma, we selectively removed DNA and vesicles from 4‐hour post‐ROSC plasma. DNase I degraded free‐floating DNA, whereas lysis buffer disrupted vesicles, and the combination of lysis and DNase I removed both vesicle integrity and dsDNA. As expected, 4‐hour post‐ROSC plasma induced expansion of inflammatory leukocyte subpopulations (dendritic cells, granulocytes, macrophages) that was significantly reduced following lysis or lysis+DNase I and rescued by reintroduction of mtDNA+TR (Figure 4A through 4C). A similar pattern emerged for the secretion of TNFα and IL‐1β (Figure 4D and 4E) and for proinflammatory transcripts (NFκB, TNFα, IL‐1α, IL‐1β, IL‐6, IL‐8, IFNα, IFNβ; Figure S6). These results highlight the proinflammatory role of EV‐encapsulated mtDNA and establish that mtDNA‐containing EVs are required for post‐ROSC plasma‐induced inflammatory leukocyte activation.

Figure 4. dsDNA is required for Post‐ROSC plasma‐induced immune cell activation.

Porcine PBMCs (CD45+7AAD−) were cultured for 24‐h in the presence of 4‐h post‐ROSC plasma, either in its native state or following treatment with lysis buffer to disrupt vesicles, DNase I to degrade dsDNA, or both. Rescue was achieved by the addition of 856 ng mtDNA with lipofectamine 2000 (TR+mtDNA), as determined by 4‐h post‐ROSC plasma mtDNA levels, in comparison to pre‐SCA plasma (BL), and TR and mtDNA+TR controls. Flow cytometry was used to determine population proportion of (A) inflammatory dendritic cells (CD172+CD16+), (B) inflammatory granulocytes (CD172+163−), and (C) inflammatory macrophages (CD14+CD163+). Following clarification, conditioned media levels of inflammatory cytokines (D) TNFα and (E) IL‐1β were detected via ELISA. Data represent biological n=3 and technical n=3. Values are mean±SEM. 7AAD indicates 7‐aminoactinomycin D; dsDNA, double‐stranded DNA; IL‐1β, interleukin‐1β; mtDNA, mitochondrial DNA; PBMC, peripheral blood mononuclear cell; SCA, sudden cardiac arrest; and TNFα, tumor necrosis factor‐α.

Post‐ROSC Plasma‐EVs Containing mtDNA Induce Inflammatory Leukocyte Activation

Given that EV disruption abrogated the immune response to post‐ROSC plasma, we investigated whether plasma‐EVs harboring mtDNA might be a primary driver of leukocyte activation. EVs were isolated from 4‐hour post‐ROSC plasma, with their identity confirmed by nanoparticle tracking (Figure 5A and 5B) and EV marker analysis (CD63, ANXA5 [annexin A5], TSG101 [tumor susceptibility gene 101], Flot1 [flotillin‐1], ICAM [intercellular adhesion molecule], ALIX [ALG‐2 interacting protein X], CD81; Figure S7A and S7B). Compared with pre‐SCA EVs, 4‐hour post‐ROSC EVs contained significantly higher levels of D‐Loop (~300‐fold; Figure 5C) and Cyt‐B (~300‐fold; Figure 5D), with undetectable nucDNA levels (Figure 5E). In agreement with these preclinical data, EV‐encapsulated mtDNA accounted for 79±3% of total circulating mtDNA in patients with sR‐OHCA compared with 55±7% in healthy controls (P=0.002; Figure S8). These results suggest that EV‐mediated transfer of mtDNA to leukocytes may occur during the first several hours post ROSC and implicate EV‐encased mtDNA as a potential therapeutic target to attenuate post‐ROSC leukocyte activation in PCAS.

Figure 5. Post‐ROSC plasma‐EVs containing mtDNA induce inflammatory leukocyte activation.

A, Representative ZetaView image illustrating the circular morphology of the porcine post‐ROSC plasma EVs. B, Representative histogram depicting the diameter distribution of EVs in plasma following ZetaView analysis indicate particle size of ~100 nanometers. Porcine post‐ROSC plasma EVs were subjected to qPCR for detection of levels of mitochondrial genes (C) D‐Loop, (D) Cyt‐B, and (E) nuclear β2M and expressed relative to their pre‐SCA plasma EV values. Data represent biologic n=3. Porcine PBMCs (CD45+7AAD−) were cultured for 24 h in the presence of 4‐h post‐ROSC EVs (DiD stained) and were compared with pre‐SCA plasma EVs (DiD stained). Basal media (neg −) and LPS were used as controls. Flow cytometry was used to determine subpopulation proportion of (F) inflammatory dendritic cells (CD172+CD16+), (G) inflammatory granulocytes (CD172+163−), and (H) inflammatory macrophages (CD14+CD163+), while determining levels of all cells (blue), EV uptake via DiD+ cells (pink), and no EV uptake via DiD− cells (purple). Following clarification, conditioned media levels of inflammatory cytokines (I) TNFα and (J) IL‐1β were detected via ELISA and (K) levels of ROS/RNS were detected via DCF. L, Levels of intracellular cGAMP were detected via competitive ELISA. Data represent biological n=3 and technical n=2. Values are mean±SEM. 7AAD indicates 7‐aminoactinomycin D; cGAMP, 2′3′‐cyclic guanosine monophosphate–adenosine monophosphate; EV, extracellular vesicles; IL‐1β, interleukin‐1β; LPS, lipopolysaccharide; mtDNA, mitochondrial DNA; NO, nitric oxide; PBMC, peripheral blood mononuclear cell; qPCR, quantitative polymerase chain reaction; ROS/RNS, reactive oxygen species/reactive nitrogen species; ROSC, return of spontaneous circulation; SCA, sudden cardiac arrest; and TNFα, tumor necrosis factor‐α.

Because mtDNA is present within 4‐hour post‐ROSC plasma EVs and sR‐OHCA plasma EVs, we postulated that EVs isolated from post‐ROSC plasma could independently trigger immune cell activation. To test this hypothesis, we stimulated naïve PBMCs with DiD‐labeled EVs isolated from 4‐hour post‐ROSC plasma and assessed changes in surface marker expression (Figure S2, Bottom) and inflammatory cytokine secretion. Compared with coculture with EVs isolated from pre‐SCA plasma, PBMCs which had taken up 4‐hour post‐ROSC plasma EVs (DiD+) exhibited a significant increase in inflammatory surface marker expression (Figure 5F through 5H). Interestingly, post‐ROSC plasma EVs elicited an increase in inflammatory macrophages that was not dependent on EV uptake (ie, the expanded inflammatory macrophage population consisted of DiD+ and DiD− cells; Figure 5H). Compared with the effects of pre‐SCA plasma EVs, post‐ROSC plasma EVs induced a significant elevation of TNFα (144.7±22.6 versus 465.9±37.8 pg/mL; P<0.001; Figure 5I) and reactive oxygen species/reactive nitrogen species (0.51±0.12 versus 2.67±0.39 μM; P<0.001; Figure 5K) within the conditioned media, along with significantly elevated intracellular cGAMP levels (0.03±0.02 versus 0.36±0.02 pmol/mL lysate; P<0.001; Figure 5L). Together, these findings indicate that post‐ROSC EVs containing mtDNA can independently trigger leukocyte activation when isolated from blood plasma.

Post‐ROSC Plasma EV–Encapsulated mtDNA Is Necessary for Inflammatory Leukocyte Activation

To verify that mtDNA within post‐ROSC plasma EVs is essential for immune cell activation, we evaluated the effects of EV integrity disruption (lysis buffer) with and without simultaneous degradation of encapsulated DNA (lysis+DNase I) before PBMC stimulation. Following depletion and before PBMC stimulation, post‐ROSC plasma EV integrity (Figure S9A and S9B) and dsDNA levels (Figure S9C) were confirmed to show expected treatment effects. The inflammatory phenotypes induced by 4‐hour post‐ROSC EVs were diminished by lysis buffer alone or lysis+DNase I and were restored by reintroducing mtDNA+TR (Figure 6A through 6C). Similarly, TNFα and IL‐1β secretion (Figure 6D and 6E), intracellular cGAMP levels (Figure 6F), and inflammatory gene expression (NFκB, TNFα, IL‐1α, IL‐1β, IL‐6, IL‐8, IFNα, and IFNβ; Figure S10) were blunted after EV disruption but partially recovered with mtDNA+TR. These results add further support to the notion that EV‐encapsulated mtDNA within post‐ROSC plasma is a key driver of leukocyte activation in PCAS.

Figure 6. Post‐ROSC plasma EVs require mtDNA to induce inflammatory leukocyte activation.

Porcine PBMCs (CD45+7AAD−) were cultured for 24 h in the presence of 4‐h post‐ROSC plasma EVs, either in their native state, treated with lysis buffer to disrupt vesicles, or treated with a combination of lysis buffer and DNase I to disrupt vesicles and degrade dsDNA. Rescue was achieved by the addition of 804 ng mtDNA with lipofectamine 2000 (TR+mtDNA), as determined by 4‐h post‐ROSC plasma EV mtDNA levels, in comparison to pre‐SCA plasma EVs, and TR and mtDNA+TR controls. Flow cytometry was used to determine population proportion (A) inflammatory dendritic cells (CD172+CD16+), (B) inflammatory granulocytes (CD172+163−), and (C) inflammatory macrophages (CD14+CD163+). Following clarification, conditioned media levels of inflammatory cytokines (D) TNFα, (E) IL‐1β, and (F) intracellular cGAMP levels were detected via ELISA. Data represent biological n=3 and technical n=3. Values are mean±SEM. 7AAD indicates 7‐aminoactinomycin D; BL, baseline; cGAMP, 2′3′‐cyclic guanosine monophosphate–adenosine monophosphate; dsDNA, double‐stranded DNA; EV, extracellular vesicles; IL‐1β, interleukin‐1β; mtDNA, mitochondrial DNA; PBMC, peripheral blood mononuclear cell; ROSC, return of spontaneous circulation; SCA, sudden cardiac arrest; TNFα, tumor necrosis factor‐α; and TR, transfection reagent.

Molecular Inhibition of TLR9 and cGAS Attenuates mtDNA‐Induced Inflammation

Finally, to identify putative signaling pathways that could be targeted to attenuate mtDNA‐mediated immune activation, we tested inhibitors targeting known dsDNA/mtDNA‐sensing pathways: TLR9 ¹⁹ and cGAS, ²⁰ as well as STING ³² and PKG1 ³³ downstream of cGAS. Transfection of mtDNA in the absence of inhibitors caused marked increases in inflammatory surface marker expression (Figure 7A through 7C and Figure S11B and S11D), TNFα and IL‐1β secretion (Figure 7D and 7E and Figure S11E through S11G), and proinflammatory gene expression (Figure S12 and Figure S10H through S10K). In contrast, treatment with TLR9 antagonists (ODN 2088/inh‐18), a cGAS antagonist (G140), a STING antagonist (H151), or a PKG1 antagonist (Rp‐8‐Br‐PET‐cGMP) significantly attenuated these responses in a concentration‐dependent manner (Tables S4–S6). Likewise, intracellular levels of cGAMP were diminished following inhibition of TLR9 (ODN 2088), cGAS, STING, and PKG1 (Figure 7F), consistent with a mechanistic role of cGAMP within the signaling pathways driving mtDNA‐induced immune activation. These results confirm that blocking mtDNA‐sensing pathways effectively dampens the immune cell activation elicited by internalization of mtDNA.

Figure 7. Molecular inhibition of TLR9 or cGAS signaling pathways attenuate mtDNA‐induced leukocyte inflammatory activation.

Porcine PBMCs (CD45+7AAD−) were cultured for 24 h in the presence of LPS or 2 μg/mL mtDNA with lipofectamine 2000 (TR+mtDNA) and compared with basal media (−) or TR—controls. At the time of stimulation, small molecular inhibitors were added against TLR9 (8 μM ODN 2088 or 8 μM ODN inh‐18), cGAS (20 μM G140), STING (4 μM H151), and PKG1 (4 μM Rp‐8‐Br‐PET‐cGMP). Flow cytometry was used to determine population proportion of (A) inflammatory dendritic cells (CD172+CD16+), (B) inflammatory granulocytes (CD172+163−), and (C) inflammatory macrophages (CD14+CD163+). Data represent biological n=3 and technical n=3. Following clarification, conditioned media levels of inflammatory cytokines (D) TNFα, (E) IL‐1β, and (F) intracellular cGAMP levels were detected via ELISA. Data represent biological n=3 and technical n=3. Values are mean±SEM. Dashed red line indicates levels of mtDNA+TR in the absence of inhibitors. 7AAD indicates 7‐aminoactinomycin D; cGAMP, 2′3′‐cyclic guanosine monophosphate–adenosine monophosphate; cGAS, cyclic GMP‐AMP synthase; IL‐1β, interleukin‐1β; LPS, lipopolysaccharide; mtDNA, mitochondrial DNA; ODN, oligodeoxynucleotide; PBMC, peripheral blood mononuclear cell; PKG1, protein kinase G type 1; STING, stimulator of interferon genes; TLR9, toll‐like receptor 9; TNFα, tumor necrosis factor‐α; and TR, transfection reagent.

Discussion

The present study provides several novel insights into how mtDNA drives postresuscitation immune activation in PCAS. First, in patients with sR‐OHCA, circulating mtDNA levels were significantly elevated within the first several hours after ROSC in the absence of marked changes in circulating nucDNA levels, consistent with release of mtDNA in clinical PCAS. Second, in a porcine model of SCA, we detected a robust (~250‐fold) increase in mtDNA and negligible changes in nucDNA 1‐hour post‐ROSC, reaffirming the nature of mtDNA release even after a brief ischemic insult and in the absence of comorbid conditions. Third, in vitro experiments revealed that intracellular uptake of mtDNA, particularly when packaged within EVs, was required for inducing pro‐inflammatory phenotypic shifts and cytokine secretion in cultured leukocytes, underscoring the functional significance of EV‐encapsulated mtDNA. Fourth, pharmacological inhibition of TLR9 and cGAS‐STING signaling notably dampened mtDNA‐mediated leukocyte activation, suggesting that targeting these DNA‐sensing pathways offers a viable strategy to curtail excessive inflammation. Collectively, these findings identify EV‐encapsulated mtDNA as a novel mechanism by which immune cells become activated in PCAS and provide signaling pathways that warrant investigation as new therapeutic targets for interventions aiming to improve outcomes in patients resuscitated from SCA.

Ischemia–Reperfusion Induced mtDNA Release and Immune System Activation

Previous studies have shown that cell‐free dsDNA accumulates in the bloodstream after cardiac arrest, with some reports highlighting its prognostic value for neurological outcomes or survival. ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ However, these studies did not distinguish whether the source of dsDNA was mitochondrial or nuclear. Our results refine these observations by establishing that the substantial increase in dsDNA after SCA primarily reflects elevations in mtDNA. This may be in indication that mtDNA is more susceptible to release during ischemic injury than nucDNA: nucDNA is tightly packed into chromatin with histones, which provide stability and protection against degradation, whereas mtDNA lacks histones, making it more susceptible to oxidative stress, fragmentation, and release into circulation upon mitochondrial damage. ⁴⁰ The nuclear envelope is also more resilient and less prone to rupture compared with mitochondrial membranes. ⁴¹ Additionally, DNases rapidly degrade circulating nucDNA, whereas mtDNA may be less efficiently cleared. ⁴²

Furthermore, we demonstrated that EVs enriched with mtDNA make up a significant fraction of the dsDNA released following ROSC and is a key mediator of immune cell activation. This not only clarifies the underlying mechanism by which dsDNA might drive inflammation but also highlights the significance of mtDNA‐loaded EVs in orchestrating the postresuscitation inflammatory response. The finding that mtDNA must be internalized by immune cells to incite a robust response aligns with prior work suggesting that mitochondrial components, particularly mtDNA, behave like bacterial DNA. ⁴³ , ⁴⁴ In our study, only transfected mtDNA (mtDNA+TR) elicited an inflammatory signature on par with lipopolysaccharide, whereas nucDNA did not. This underscores the unique ability of mtDNA to engage innate immune pathways, most notably TLR9 and cGAS‐STING, which, once activated, prompt the production of TNFα, IL‐1β, IL‐6, and related pro‐inflammatory mediators. ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹

EV Encapsulation of mtDNA and Inflammatory Pathway Activation

By selectively removing DNA and vesicles from post‐ROSC plasma, we found that EV integrity and encapsulated mtDNA are essential for its proinflammatory effects. This indicates that EVs are not merely passive carriers but actively determine the inflammatory potential of mtDNA by providing a mechanism for internalization by leukocytes. Similar conclusions emerge from prior studies linking mitochondria‐derived vesicles and EVs to systemic inflammation, particularly in the context of trauma or infection. ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁵⁰ , ⁵¹ Our data reinforce the concept that EV‐encapsulated mitochondrial components, rather than free‐floating mtDNA, drive leukocyte activation. Consistent with a “danger signal,” mtDNA in EVs can engage TLR9 and cGAS, eventually amplifying the innate immune cascade. ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷

The notion that post‐ROSC EVs can stimulate macrophages, dendritic cells, and granulocytes also aligns with research showing how mtDNA‐laden particles can propagate inflammatory signals beyond the cells initially stressed by ischemia/reperfusion injury. ³⁶ , ³⁹ Interestingly, inflammatory macrophage expansion involved both EV‐positive (DiD+) and EV‐negative (DiD–) cells, implying that once mtDNA‐stimulated signaling is initiated, it may spread to bystander immune cells, further enhancing systemic inflammation in PCAS.

Implications for Clinical Outcomes in PCAS

The “sterile sepsis”–like state that often arises after OHCA has been attributed to overactive immune responses. ⁴³ , ⁵² Our findings suggest that mtDNA release, specifically mtDNA encapsulated in EVs and released into plasma, may be a key contributor to this phenomenon. This perspective resonates with observational data linking high inflammatory biomarker levels to circulatory failure, organ dysfunction, and increased mortality after cardiac arrest. ⁵³ Moreover, although some studies have proposed that dsDNA or dsDNA/DNase ratios correlate with patient prognosis, ³⁴ , ³⁵ the presence of EV‐encapsulated mtDNA may complicate the predictive power of these measurements, since DNase has limited capacity to degrade mtDNA that is encased within vesicles.

Relative to healthy controls, patients with sR‐OHCA exhibited significantly elevated circulating levels of both mtDNA (~14‐fold) and nucDNA (~10%) that did not differ between 30‐day survivors and nonsurvivors. This is consistent with previous findings from Donnino et al., ⁵⁴ who observed elevated levels of cell‐free DNA and nuclear DNA in plasma of PCAS patients that were not different between survivors and nonsurvivors. Thus, our results do not support a prognostic role for early post‐ROSC circulating mtDNA concentrations, but future studies in larger cohorts and at additional post‐ROSC time points are needed to draw definitive conclusions. In our cohort of patients with sR‐OHCA, we also found that the portion of circulating mtDNA within EVs was significantly higher than in healthy controls. EV uptake can occur rapidly, with EVs being identified inside cells from as early as 15 minutes following exposure to the target cell. ⁵⁵ Further investigation of post‐ROSC EV kinetics, including timing, biodistribution, and cellular internalization, will be essential for deciphering how these differences may be related to patient outcomes and exploring whether combining measures of total dsDNA, nucDNA, and EV‐encapsulated mtDNA could refine risk stratification or treatment in PCAS.

The rapidity of inflammatory activation following resuscitation presents a significant challenge for timely therapeutic intervention. Our data show that mtDNA levels rise sharply within 1 hour of ROSC and are already encapsulated in EVs during this early period, suggesting that innate immune signaling may be rapidly engaged. These findings underscore the importance of administering immunomodulatory therapies as early as possible—potentially during active resuscitation. Incorporating intra‐arrest drug delivery via intraosseous or endotracheal routes, ⁵⁶ may offer a feasible strategy for delivering agents that inhibit DNA‐sensing pathways, such as TLR9 or cGAS‐STING inhibitors, at a time point when immune activation is just beginning. Although our study did not directly assess the impact of targeted temperature management on mtDNA dynamics, prior evidence suggests that early hypothermia may preserve mitochondrial membrane integrity ⁵⁷ and limit danger‐associated molecular pattern release by stabilizing cellular membranes, ⁵⁸ potentially attenuating the proinflammatory consequences of mtDNA signaling. These protective effects may extend to limiting the immunogenic potential of mtDNA; however, data from our institution ⁵⁹ indicate that targeted temperature management does not significantly affect long‐term survival, highlighting that its clinical benefits may be restricted to early‐phase inflammation and confounded by other patient‐specific factors such as sex, duration of cardiac arrest, and baseline comorbidities. Integrating targeted temperature management with early pharmacologic targeting of mtDNA signaling may yield synergistic benefit, though this hypothesis warrants testing in prospective trials. Although the pharmacologic inhibition of mtDNA sensing has not yet been evaluated clinically in the context of cardiac arrest, emerging small‐molecule cGAS inhibitors such as VENT‐03 ⁶⁰ and IMSB301 ⁶¹ are currently under investigation for autoimmune and interferon‐driven diseases. Collectively, these insights support the rationale for early, mechanism‐based interventions to mitigate mtDNA‐driven sterile inflammation and improve clinical outcomes in patients with PCAS.

Targeting mtDNA‐Sensing Pathways

We observed that antagonizing TLR9 and cGAS‐STING significantly diminished intracellular levels of cGAMP and inflammatory responses induced by mtDNA. This bolsters evidence that these DNA‐sensing pathways are pivotal in orchestrating the post‐ROSC inflammatory milieu and suggests new therapeutic angles for modulating immune activation. ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ Although our study investigated TLR9 and cGAS‐STING, other cytoplasmic DNA sensors like Rad50 and Ku70 may also respond to mtDNA release, potentially contributing to the innate immune activation in PCAS. ⁴⁹ , ⁶² Broadening future investigation to include these sensors could uncover additional molecular targets to mitigate PCAS‐associated inflammation.

Future Directions

The precise mechanisms by which mtDNA is packaged into EVs or shuttled via mitochondria‐derived vesicles following ROSC remains incompletely understood, but it is plausible that oxidative stress, mitochondrial fission, and disrupted autophagic flux underlie this phenomenon. ⁵⁰ , ⁵¹ In parallel, the contributions of distinct immune cell subsets, diverse tissue sources, and the kinetic profile of EV‐mediated mtDNA release to systemic inflammation have yet to be systematically characterized. Pharmacologic approaches aimed at blocking mtDNA liberation, inactivating vesicle‐associated mtDNA, and inhibiting critical DNA‐sensing pathways hold promise for alleviating the inflammatory burden of PCAS. Validating these strategies in large animal models and subsequent clinical trials will be critical to determining their efficacy in mitigating mtDNA‐driven injury.

Additionally, it is still unclear why some individuals exhibit more pronounced mtDNA release or a more severe inflammatory response post ROSC. Heterogeneity in mitochondrial function, redox balance, or DNase activity may contribute to varying extents of mtDNA release. Studies also suggest that partial EV loading of mtDNA might occur under more subtle damage conditions, while severe cell death could release substantial amounts of free mtDNA. ³⁹ , ⁵⁰ Clarifying these distinct mechanisms could enable refined therapeutic interventions that target harmful mtDNA release without compromising vital metabolic and immunological functions.

Limitations

Although the porcine model of PCAS exhibits key features of the clinical syndrome, interspecies differences in physiology and disease course compared with human subjects should be considered when interpreting preclinical data. The partial translation of our findings in swine to human patients with sR‐OHCA provides initial clinical support; however, larger cohorts with extended follow‐up are needed to study a potential link between mtDNA‐EVs and clinical outcomes. Our speculation regarding accelerated EV uptake in nonsurvivors remains unproven, and mtDNA‐EVs may exert additional, uncharacterized roles—for instance, compensating for impaired mitochondrial quality control, warrants further investigation. We also did not explore other DNA‐sensing pathways beyond TLR9 and cGAS‐STING, leaving room to investigate additional innate immune danger‐associated molecular pattern receptors that might significantly influence the post‐ROSC inflammatory response. Finally, it is important to recognize that the availability of swine‐specific antibodies is limited and that immunophenotypic frameworks for characterization of porcine immune cell subsets are incompletely developed. Given these limitations, our approach focused on identifying broad shifts in innate immune cell populations using the most widely validated porcine markers available to demonstrate that mtDNA triggers a robust immunologic response consistent with that observed in PCAS, rather than define discrete subpopulations or polarization states. However, future studies leveraging multidimensional single‐cell technologies are necessary to delineate the functional heterogeneity and plasticity of porcine immune cell responses after resuscitation from SCA.

Conclusions

In summary, our findings identify a novel mechanism of mtDNA release following brief whole‐body ischemia and implicate EV‐encapsulated mtDNA as a central mediator of sterile inflammation in PCAS. By demonstrating that mtDNA release and vesicular packaging are indispensable for TLR9 and cGAS‐STING driven immune activation, we unveil a novel mechanism by which an excessive inflammatory response may exacerbate organ damage following SCA. These new insights fuel interest in evaluating the therapeutic potential of strategies targeting DNA‐sensing pathways to lessen the excessive inflammatory burden that accompanies resuscitation. Pharmacological disruption of mtDNA‐mediated immune activation, whether by inhibition of its release, vesicular encapsulation, or downstream receptor signaling, holds promise for improving outcomes in the large group of patients who develop PCAS following resuscitation from SCA.

Sources of Funding

This work was supported by the National Heart Lung and Blood Institute (1R01HL160538), the National Center for Advancing Translational Sciences (UL1TR001412), and the American Heart Association (24PRE1193924).

Disclosures

None.

Supporting information

Data S1–S3

Tables S1–S6

Figures S1–S12

References 63–66