Mitochondrial Cleanup and a DNA Sensor May Team Up to Drive Stroke Inflammation

A new review published in Frontiers in Molecular Neuroscience argues that two cellular systems long studied separately after ischemic stroke—mitophagy, the selective disposal of damaged mitochondria, and the cGAS–STING innate immune pathway—may work in tandem to shape the neuroinflammation that worsens brain injury after a clot cuts off blood flow. Written by a team led by Yu Xiao and Yuanyuan Xiang of Jiujiang University, the article does not report new experiments. Instead, it assembles existing animal and cell-model findings into a single mechanistic framework, mapping how leaking mitochondrial DNA connects mitochondrial quality control to inflammatory signaling across the different cell types of the brain’s neurovascular unit.

The starting point is a well-established clinical problem. Stroke is the second leading cause of death worldwide, and ischemic stroke accounts for more than 80 percent of all cases. When blood flow returns after an ischemic episode, cellular energy failure and disrupted ion balance damage mitochondria, prompting them to release damage-associated molecular patterns that ignite inflammatory cascades in the central nervous system. Attenuating this neuroinflammation is considered an important step toward reducing tissue injury and improving functional outcomes, yet the review’s authors stress that every mechanism they discuss has been established in animal models such as middle cerebral artery occlusion and in vitro oxygen–glucose deprivation systems, not in patients.

Mitophagy is the cell’s quality-control answer to mitochondrial damage. Under normal conditions, the kinase PINK1 is imported into mitochondria through the TOM/TIM23 translocase complexes and quickly degraded by proteases including MPP and PARL. When mitochondrial membrane potential collapses—a hallmark of ischemia–reperfusion injury—PINK1 can no longer enter the inner membrane and accumulates on the outer membrane, where it dimerizes, autophosphorylates, and phosphorylates ubiquitin attached to outer-membrane proteins. This phosphorylated ubiquitin recruits the E3 ligase Parkin, which tags outer-membrane proteins for capture by autophagy receptors that bind LC3 on growing autophagosomes. A parallel receptor-dependent route runs through BNIP3 and its homolog NIX, whose transcription is driven by HIF-1α during hypoxia, and through FUNDC1, whose inhibitory phosphorylation at Ser13 and Tyr18 is removed by the phosphatase PGAM5 during ischemia, freeing it to bind LC3 directly.

Experimental models of cerebral ischemia–reperfusion show increased expression of PINK1, Parkin, and LC3-related markers, indicating that these pathways are activated under stroke-like conditions. Their consequences, however, appear to be double-edged. On the protective side, enhanced mitophagy clears damaged mitochondria before they spill their DNA, and studies in ischemia–reperfusion injury have shown that activating the HIF-1α/BNIP3 pathway or boosting Parkin-dependent mitophagy downregulates TLR9 and NLRP3 inflammasome activity, while PINK1/Parkin-mediated mitophagy reduces microglial pyroptosis and neuronal injury. On the harmful side, one in vitro study found that microglia engulfing apoptotic neurons upregulated Drp1, triggering excessive mitochondrial fission and mitophagy that drained the cells’ energy and impaired their phagocytic capacity; blocking mitophagy eased the inflammation. Another study showed that the chemokine-like factor CKLF1 increased autophagosome numbers while impairing lysosomal function, so incomplete mitophagic flux left damaged mitochondria piled up in microglia and drove excessive activation.

The review’s authors conclude that no clear signaling threshold currently separates beneficial from detrimental mitophagy. What seems to matter instead is the extent of mitophagy, cellular energy demands, and the integrity of the full mitophagic flux from autophagosome formation through lysosomal fusion, biogenesis, and acidification. Therapeutically, they argue, the goal should not be simply to promote or inhibit the process but to restore effective flux: enhancing mitophagy when damaged mitochondria accumulate faster than they can be cleared, and restraining it when sustained activity depletes cellular energy.

The second player, the cGAS–STING pathway, is a cytosolic DNA-sensing system. After ischemic stroke, the double-stranded DNA that activates cGAS comes chiefly from leaked mitochondrial DNA and oxidized mitochondrial DNA, which retain unmethylated CpG motifs that make them strongly immunogenic. Once cGAS binds cytosolic DNA, it synthesizes the second messenger 2′3′-cGAMP, which binds STING on the endoplasmic reticulum. STING, a roughly 42-kilodalton four-pass transmembrane protein, then dimerizes and oligomerizes, exits the ER in COPII-coated vesicles, and travels to the ER–Golgi intermediate compartment and Golgi apparatus. There it recruits TBK1, which phosphorylates IRF3, driving type I interferon transcription, while STING also recruits MAP3K14/NIK and the IKK complex to activate NF-κB and induce inflammatory mediators such as IL-6 and TNF-α. Palmitoylation of a membrane-proximal cysteine stabilizes the signaling complex, and blocking this modification dampens the interferon response. Neutrophil extracellular traps offer a second source of activating DNA, and inhibiting NET release has been shown to ease cGAS–STING-driven inflammation and promote neuronal survival in models.

A distinctive contribution of the review is its cell-by-cell mapping of this pathway within the neurovascular unit. In microglia, cGAS–STING activation has been associated with shifts toward a pro-inflammatory phenotype, exacerbation of NLRP3- and AIM2-inflammasome-mediated responses and pyroptosis, and microglial synaptic engulfment of neurons that the STING inhibitor H151 can reverse. In neurons, oxidative stress generates oxidized mtDNA that activates the pathway, and in a rat transient MCAO model STING appeared in neurons as early as four hours after reperfusion, rising in microglia only later and peaking at 12 to 24 hours—suggesting a temporal sequence that begins in neurons. In astrocytes, ischemia downregulates SIRT3, releasing its brake on the pathway; restoring SIRT3 suppresses signaling, cuts pro-inflammatory cytokine release, and improves neurological function in models. In brain microvascular endothelial cells, ischemia activates the mechanotransduction regulators YAP/TAZ, and their conditional deletion suppresses CXCL10 release from the pathway, reshapes myeloid-cell recruitment, restrains microglial activation, and reduces chronic infarct volume. The pathway may also spread between cells: damaged cells can transfer cGAMP to neighbors through gap junctions, and gap-junction blockade suppressed cGAS/STING activation in neuron–microglia co-culture systems.

The molecular bridge between the two systems is mitochondrial DNA itself. mtDNA escapes into the cytosol mainly through BAX/BAK pore permeabilization of the outer membrane and opening of the mitochondrial permeability transition pore, and its presence activates cGAS–STING, driving type I interferons, IL-1β, TNF-α, IL-6, and chemokines such as CXCL10. By clearing damaged mitochondria before this leakage occurs, mitophagy can suppress the axis at its source. The influence also runs the other way, though the evidence is more tentative. STING trafficking to the ERGIC promotes LC3 lipidation and noncanonical autophagy through WIPI2 and ATG5; in MCAO and oxygen–glucose deprivation–reperfusion models, STING directly bound LC3 via its transmembrane domain, promoting a pro-inflammatory microglial phenotype, and cGAS–STING activation has been linked to excessive autophagy in neurons and to NCOA4-mediated ferritinophagy. Inhibition of cGAS–STING promoted STX17–SNAP29–VAMP8-mediated autophagosome–lysosome fusion and restored autophagic flux, while STING activation promotes TFEB nuclear translocation and lysosomal biogenesis through GABARAP lipidation. Most intriguingly, recent mechanistic studies show that STING recruited to damaged mitochondria activates TBK1, which phosphorylates the autophagy receptor OPTN at Ser177, promoting PINK1–Parkin-dependent mitophagy—a potential feedback loop in which the inflammatory sensor also flags damaged mitochondria for destruction.

The authors are careful about the limits of this crosstalk story. The STING–TBK1–OPTN axis has been demonstrated mainly in mechanistic cellular models and has not yet been validated in ischemic stroke models, so they treat it as a potential mechanism rather than an established stroke-specific one. Findings linking STING to autophagy concern general autophagy rather than mitophagy specifically, and in astrocytes and endothelial cells direct evidence connecting STING activation to mitophagy remains very limited. Overall, they describe the relationship as functional synergy and potential molecular crosstalk rather than a fully established bidirectional regulatory loop, and they note that studies of mitophagy modulating cGAS–STING exist mainly for ischemia–reperfusion injury in other organs, with cerebral data still sparse.

The therapeutic section surveys pharmacological and genetic interventions tested in stroke models. PINK1/Parkin activators including glycosides of Buyang Huanwu decoction, ginsenoside Rg1, and esculin reduced infarct size, neuronal damage, and NLRP3-associated inflammation in MCAO/R rodents. Cyclosporine A, by inhibiting mitochondrial permeability transition, reduced oxidized mtDNA release and downstream cGAS–STING activation. STING and cGAS inhibitors—C-176, H-151, A151, RU.521, melatonin, edaravone dexborneol, and others—reduced infarction, edema, microglial activation, and pyroptosis across mouse and rat models, while DNase I degraded neutrophil extracellular traps and reduced blood–brain barrier disruption. Gene-based approaches, including HIF-1α overexpression plasmids, AAV-mediated ATF4 overexpression, lncRNA Tug1 knockdown, and cGAS knockdown or m6A-demethylation via FTO overexpression, similarly improved outcomes in animals. Early nanomedicine platforms—carbon dots with antioxidant enzyme mimicry, strontium-substituted Prussian blue analogs, DNase-mimicking artificial enzymes loaded with C-176, and biomimetic mitochondria-targeted vesicles—showed preclinical promise but carry unresolved questions about off-target effects, biosafety, membrane reproducibility, and degradation products.

Translational hurdles remain substantial. Most small-molecule cGAS–STING inhibitors cross the blood–brain barrier poorly, and because the pathway also serves critical antiviral and antibacterial defense, systemic suppression must be approached cautiously. The review closes by emphasizing that its synthesis rests almost entirely on animal and cellular models: the therapeutic window of mitophagy is unclear and likely context dependent, cell-type-specific mechanisms within the neurovascular unit require further study, and clinical validation of any intervention targeting the mitophagy–cGAS–STING axis in stroke patients has yet to begin. What the article offers, its authors say, is a mechanistic perspective on immune intervention and a theoretical foundation for future translational work—not a treatment ready for the clinic.

Subject of Research: The interplay between mitophagy and the cGAS–STING pathway in neuroinflammation after ischemic stroke

Article Title: Synergistic regulation of neuroinflammation by mitophagy and the cGAS–STING pathway following ischemic stroke

Article References: Xiao, Y., Fu, P., Zhao, C., Song, W., Lu, Y., Wang, H., Yin, X., Chen, Z., & Xiang, Y. (2026). Synergistic regulation of neuroinflammation by mitophagy and the cGAS–STING pathway following ischemic stroke. Frontiers in Molecular Neuroscience, 19, Article 1990236. https://doi.org/10.3389/fnmol.2026.1990236

Image Credits: AI Generated

DOI: 10.3389/fnmol.2026.1990236

Keywords: ischemic stroke, mitophagy, cGAS-STING, neuroinflammation, mitochondrial DNA, microglia, neurovascular unit, PINK1-Parkin, TBK1, nanomedicine, inflammasome, review