Mitochondrial fission is defined as a complex condition that divides single mitochondrion into two separate mitochondria (Ni et al., 2015). Previous work has established that excessive activation of mitochondrial fission is an underlying factor in after-stroke brain injury (Table 1). Excessive fission subsequently destroyed mtDNA integrity and decreased respiratory complexes (Wu et al., 2019). Mice with transient middle cerebral artery occlusion (tMCAO) that lacked A-kinase anchoring protein 1 (AKAP1, protein kinase A (PKA) scaffold) impaired the phosphorylation of Drp1 (fission-associated protein) at Ser637. Then, this led to overmuch fission, oxidative respiratory chain damage, Ca²⁺ dysregulation, and consequently deteriorative injury in ischemic brain issues (Flippo et al., 2018). AMP-activated protein kinase (AMPK) mediating the phosphorylation of AKAP1 caused the disassembly of PKA/AKAP1, which ultimately facilitates mitochondrial fragments (Zhang et al., 2017). In agreement, a few articles have reported that the maintenance of the phosphorylated state of Drp1 Ser637 via either the administration of Kaempferol (Wu et al., 2017) or the knockout of a neuron-specific Drp1 promotor [protein phosphatase 2A regulatory domain Bβ2 (Flippo et al., 2020)] can elongate the mitochondria and alleviate the damage in mice brain infarct area and primary neurons exposed to oxygen-glucose deprivation (OGD).

Multiple evidence have advocated that the disruption of mitochondrial activities is prone to generating mitochondria-related damage-associated molecular patterns (DAMPs) and redundant ROS into the cytoplasm and an extracellular matrix, which are the major culprits in eliciting neuroinflammation (van Horssen et al., 2019; Liu et al., 2020). An experimental study using two-photon imaging in the neocortex of mice confirmed that an increment of mitochondrial fragmentation (fission) is obvious and reversible in cerebral ischemia injury (Kislin et al., 2017). Moreover, the expression of Drp1 was validated to increase in rat’s ipsilateral infarct area and BV-2 microglial cells subjected to OGD/reperfusion (OGD/R) (Rutkai et al., 2019; Hu et al., 2020). The inhibition of Drp1 phosphorylation at Ser616 and the translocation with Atractylenolide III prevent extreme mitochondrial fission, which subsequently moderates the CNS immune response by decreasing pro-inflammatory cytokines and increasing anti-inflammatory molecules (IL-10, CD206, and Arg-1, for instance) in the microglial cells of tMCAO rats (Zhou et al., 2019). In addition, a ketogenic diet was reported to prevent tMCAO-induced endoplasmic reticulum stress and mitochondrial cleavage. This restriction of fission restrained the activation of NACHT, LRR, and PYD domains-containing protein 3 (NLRP3) inflammasome, thereby reducing pro-inflammatory factors like IL-1β, IL-18, and caspase 1 in mice ischemic brain region and OGD/R SH-SY5Y cells (Guo et al., 2018; Toldo and Abbate, 2018).

Excessive mitochondrial fission was demonstrated to activate an amplification of the inflammation cascade between microglia and astrocytes. Park et al. (2013) have observed the phenomenon of increased mitochondrial fission in cultured microglia stimulated by lipopolysaccharides (LPSs). These mitochondrial fragments then regulated the microglia to be polarized into a pro-inflammatory phenotype (M1) for potentiating the expression of pro-inflammatory mediators (Park et al., 2013). The fragmented mitochondria and inflammatory factors released from microglia triggered reactive astrocytes to be transformed into a pro-inflammatory A1 phenotype that upregulated the expression of classic complements, i.e., C1r, C1s, C3, and C4 in vitro (Joshi et al., 2019; Liu et al., 2020). By highly expressing C3 that could bind with C3 receptors on microglia, A1 astrocytes exaggerated microglial inflammation and impaired microglial phagocytosis of myelin debris (Zheng et al., 2021). These results imply that the maintenance of mitochondrial homeostasis is beneficial to restore cerebral blood flow, dwindle infarct brain regions, alleviate cerebral edema, reduce cell apoptosis, and improve neurological outcomes.

Apart from inflammation, mitochondrial fission is also accompanied by apoptosis, indicating a cross-talk between apoptosis and mitochondrial dynamics. Mitochondrial fission is generally considered as an observed index of cellular apoptosis and is a crucial factor in apoptotic signaling pathways (Jeong and Seol, 2008; Maes et al., 2019). Drp1 was found to be overactivated in astrocytes surrounding hemotoma, which inevitably contributed to secondary brain injury after intracranial hemorrhage (ICH) in rodents (Wu et al., 2020; Zhang et al., 2021). Drp1 interference in ICH rats was shown to exert neuroprotection. While some articles suggest that these neuroprotective effects are attributed to the alleviation of inflammation, the downregulation of Drp1 can also suppress neuronal apoptosis following ICH and ischemic stroke. It was proven by the raised level of anti-apoptosis-associated protein B-cell lymphoma-2 (Bcl-2) and the reduced apoptosis-associated proteins (apoptosis-inducing factor, cytochrome c, caspase-3, etc.) (Xu et al., 2017; Zhang et al., 2021). Activated Bcl-2-associated-x protein (Bax) could be inserted into the outer mitochondrial membrane (OMM) to conduct oligomerization and finally increase OMM permeabilization (OMMP) (Zhao et al., 2014). Similarly, Mul1 was found to be upregulated in the ischemic tissues of rats in the tMCAO model, which intensified the fission by sumoylating Drp1 and ubiquitinating Mfn2 (Ren et al., 2019). The knockdown of Mul1 decreased a mitochondrial fragment while increasing energy generation, thus weakening caspase-3 activity, delaying apoptosis, and ameliorating the neurological deficit score (Ren et al., 2019). The application of photobiomodulation could reduce overwhelmed Drp1 GTPase activity and therefore restrain mitochondria-dependent caspase-3/9-apoptosome activities in the hippocampal CA1 regions of cardiac-arrest-induced global cerebral ischemic rats (Wang R. et al., 2019). Being independent of Drp1, mitochondrial division dynamin can promote Bax-associated cytochrome c release and apoptosis in the cell strain RDY84, which can be attenuated by the administration of a mitochondrial division inhibitor (mdivi-1) (Cassidy-Stone et al., 2008). Interactionally, the pro-apoptotic Bcl-2 subfamily, including Bax and BH3 homolog agonist killer (Bak), was capable of recruiting Drp1 to the fission site and actuating its fission activity (D’Orsi et al., 2017).

Although lots of research as mentioned above considers mitochondrial fission as a harmful process in stroke-induced brain injury, some evidence suggests its neural protective effect by eliminating the wasted mitochondrial segments, promoting signal transduction, boosting ATP production, and stabilizing mtDNA. Research on HEK293 and HeLa cells has found that the inhibition of Drp1-induced fission impaired the selectivity of mitophagy, thus enhancing its rate of elimination of even healthy mitochondria via the unchecked PINK1/Parkin pathway (Burman et al., 2017). After selective overexpression of uncoupling protein-2 (UCP2) in the corpus striatum of stroke-prone spontaneously hypertensive (SHRSP) rats, Busceti et al. (2020) found the upregulation of OPA1 and Fis1, which respectively amplified mitochondrial fusion and fission in the striatum tissue. Although the causality was not explicit, this experiment implied that UCP2 exerted neuroprotection by driving mitochondrial fusion and fission. A growing body of literature has investigated that asymmetric division of the mitochondrial body could separate the functional segment from the dysfunctional one. The dysfunctional one is then eliminated by mitophagy (Zorov et al., 2019). These physiological processes can confine the flawed mitochondria to precise clearance and maintain the integrity of the mitochondrial structure. Fission and mitophagy damages may give rise to the accumulation of abnormal mitochondrial elements that initiate various pathologies, jeopardize the whole mitochondrial network, and even kill the cell (Ma et al., 2020). Though mitochondrial fission is conventionally thought to be detrimental to cellular energy production and homeostasis, Horn et al. (2020) have identified that mitochondrial fragments enable mouse embryonic fibroblasts (MEFs) and HeLa cells to prime protective signal pathways on focal cell membrane damage. This process is indispensable to repair and regenerate the cells damaged by laser assay (Horn et al., 2020). Furthermore, increasing mitochondrial fragments seemed to amplify the energy production in primary astrocytes during hypoxia in vitro (Quintana et al., 2019). Mice that postnatally experienced Drp1 knockdown were shown to suffer from a deficit in bioenergetics and synaptic transmission (Shields et al., 2015). In addition, fission and other MQC mechanisms are involved in maintaining mtDNA integrity through the removal of mutational and dysfunctional mtDNA (Mishra and Chan, 2014). Irregularity in these processes could result in mitochondrial disorder and neuronal damage.

The regulations of mitochondrial fission that contain various molecules, organelles, and signaling pathways can diversely influence several cellular functions, which require more research to explore the mechanism of its interaction with stroke progression. It remains controversial whether restraining fission is beneficial to protect cell viability after stroke. As mitochondrial fission appears to be a double-edged sword, there is a possible delicate balance between the physical compensation and the pathological initiation. Further clinical research must be conducted to validate the therapeutic effect of mitochondrial fission administration because of the complex MQC network and a wide heterogeneity of individual stroke pathology (for instance, location, duration, severity, and fundamental conditions).

Several studies have suggested that mitophagy is suppressed after stroke (Table 3; Di et al., 2015; Wang H. et al., 2019). Agents promoting mitophagy could be potential therapeutic candidates to protect mitochondrial function and ameliorate neural injury after stroke. Deng et al. (2021) showed that decreased PINK1 expression and mislocated Parkin existed both in the peripheral-hematoma brain tissues of ICH rats and in the oxygen hemoglobin-treated primary neurons. The elevation of mitochondrial membrane potential using methylene blue or electroacupuncture augmented mitophagy and subsequently attenuated neural injury in the ischemic area of tMCAO rats and OGD neuron-like PC12 cells (Di et al., 2015; Wang H. et al., 2019). Nicotinamide phosphoribosyltransferase (Nampt) knockdown inhibited autophagy, contributing to aggravated apoptosis and necrosis in the cultured cortical neurons subjected to OGD and the ischemic area of tMCAO rats (Wang et al., 2012). The block of Nampt can enhance the phosphorylation of mTOR and S6K1, which abolished the expression of autophagic proteins LC3 and Atg6 (Wang et al., 2012).

While reduced mitophagy has been intensively reported, some articles suggest that mitophagy is activated in stroke models with evidence of typical morphology, mitophagy makers, and mitochondrial mass. This phenomenon was found in MCAO models. The levels of LC3II/LC3I, PINK1, Parkin, Beclin1, LC3-COX4 copositive puncta were significantly elevated in ischemic hemisphere brain tissues (Li Q. et al., 2014; Li et al., 2018; Zuo et al., 2019; He et al., 2021). Research using long time-lapse imaging has discovered that Parkin was selectively recruited to the damaged mitochondria in the somatodendritic regions of neurons during CCCP exposure (Cai et al., 2012). It subsequently enhanced the translocation of the mitochondria to lysosomes for autophagy. Through a mitochondria-targeting two-photon ratiometric probe, enhanced mitophagy was demonstrated at 0, 3, 6, and 12 h after ODG/R in PC12 cells (Cheng F. et al., 2021). Increased mitophagy may be related to oxidative stress. Anti-inflammatory treatment can partially inhibit mitophagy (Cheng F. et al., 2021). Under physiological conditions, the autophagy adaptors BNIP3 and Beclin1 bind with Bcl-2 at the BH3 domain, which decreases mitophagy activity. In the cortex, striatum, and hippocampus of tMCAO mice and OGD-disposed hippocampal neuronal HT22 cells, uncoupling of Beclin-1/Bcl-2 complexes induced mitophagy, which led to reduced infarct size and brain edema (Chen et al., 2020). Nur77 is an orphan nuclear receptor that is dupregulated during inflammation and then in turn restricts inflammation development (Lith and de Vries, 2021). In MCAO rats and ODG/R human SH-SY5Y cells, Ping et al. (2021) have reported that increased mitophagy was induced by the translocation of Nur77 from the nucleus to mitochondria. Heat shock protein B8 (HSPB8) upregulated in tMCAO was shown to reinforce mitophagy in CA1 sector of the hippocampus (Li et al., 2018). The overexpression of HSPB8 further promoted mitophagy and therefore attenuated cell apoptosis after ischemic stroke. Meanwhile, this neuroprotective effect of HSPB8 could be compromised by inhibiting mitophagy (Li et al., 2018). In conclusion, it is likely that mitophagy is transiently induced after stroke and serves as a spontaneously neuroprotective response.

Mitophagy eliminates the damaged mitochondria and mitigates inflammation after stroke. It was shown that mitochondrial stress was booming after exhaustive exercise of PINK1^–/– mice or mtDNA mutation of Parkin^–/– mice. These mitophagy-deficient mice release extensive DAMPs, which remarkably activate a strong inflammatory response in the heart tissue of mice subjected to exhaustive exercise (Sliter et al., 2018). A study on SHRSP rats has provided evidence that autophagy inhibition in the whole brain may lead to increased morbidity from stroke (Forte et al., 2020). This may be associated with the deficient removal of the damaged mitochondria and the accumulation of misfolded proteins (Forte et al., 2020). Mechanically, the impaired mitophagy was likely the consequence of Ndufc2 gene downregulation in the SHRSP brain, contributing to NAD⁺ reduction (Forte et al., 2020). The reduction of NAD⁺ weakens cellular antioxidation and energy metabolism, leading to apoptosis and necroptosis in OGD/R-treated primary cultured neurons and in the ipsilateral areas of MCAO/R mice (Huang et al., 2018).

Recently, the central roles of mitophagy in mitigating neuronal apoptosis and increasing neural function have also been reported. In the ischemic cerebral cortex and the OGD hippocampal neuron cell line HT22, ischemia-reperfusion injury-induced oxidative stress and increased the production of caspase3 and Bax while decreasing the expression of the antiapoptotic molecule Bcl-2. These effects can be reversed by tissue-type plasminogen activator (tPA) treatment (Cai et al., 2021). tPA inhibited apoptosis and maintained the mitochondrial function through upregulating FUNDC1. FUNDC1 is an OMM protein that recruits LC3 to trigger mitophagy in mammals (Chen et al., 2016). The knockdown of FUNDC1 destroyed LC3/mitochondrial colocalization, hence lessening the neuroprotective effects of tPA-mediated mitophagy (Cai et al., 2021). A similar conclusion was acquired from a study on subarachnoid hemorrhage (SAH) in rats. Mitophagy promoted by Mitoquinone was able to mitigate mitochondrial oxidative stress, preserve mitochondrial integrity, and ultimately counteract apoptosis via Nrf2/Keap1 and PINK1/Parkin/LC3II pathways in the nucleus and mitochondria, respectively (Zhang et al., 2019). Moreover, enhanced mitophagy was suggested to raise neural resistance to ischemic stroke and lengthen the time window of thrombolytic therapy in MCAO-treated mice (Shen et al., 2017). This dynamic process increases the efficiency of eliminating the depolarized mitochondria in mature cortical neurons, which makes sense in attenuating neural damage.

Excessive mitophagy is accompanied by multitudinous mitochondrial damage, degradation, and metabolic dysfunction. Using a neonatal Ischemia/Hypoxia (I/H) mouse model, Shi et al. (2014) observed that the expression of BNIP3 increased in a delayed manner (increasingly elevated in 1, 3, and 7 days), contributing to deteriorative neural loss. The knockdown of BINP3 obviously mitigated mitophagy and neuronal apoptosis in ischemic and hypoxic insults (Shi et al., 2014). Consistently, it was found that in response to cerebral ischemia-reperfusion injury, small nucleolar RNA host gene 14 was upregulated in mouse hippocampal neurons, which excessively stimulated mitophagy and induced cell apoptosis (Deng et al., 2020). Research on similar tMCAO (occlusion for 2 h and reperfusion for 22 h) rats reported that PINK1, Parkin, and LC3-II were increased in ischemic brain tissues (Zhang Y. et al., 2020). Rehmapicroside treatment prevented the accumulation of mitophagy-associated proteins on the mitochondria and simultaneously improved neurological deficit scores (Zhang Y. et al., 2020). Increased mitochondrial dysfunction and mitophagy also occurred in H₂O₂-treated primary cortical neuronal cells and at 24 h after Rose Bengal photothrombosis in cortical infarct mice (Hwang et al., 2020). ROS alleviation partially avoided mitochondrial damages and autophagy, which protected neurons against photothrombotic cerebral infarction (Hwang et al., 2020). A similar phenomenon was obtained from the severe ischemic model (Wang R. et al., 2019) – global cerebral ischemia (GCI) with bilateral common carotid arteries and vertebral artery occlusion. Wang R. et al. (2019) described that hippocampus CA1 neurons displayed marked nuclear membrane distortion, hyperchromasia, and fragmented mitochondria with broken cristae and membrane depolarization at 3 days after GCI. Additionally, the impaired mitochondrial network erupted numerous mitophagy, followed by the deficiency of cytochrome c oxidase and ATP contents and the increase of CA1 neuron apoptosis. Based on the abovementioned data, it seems likely that excessive mitophagy is not only a cause leading to the impaired mitochondrial network but also a result induced by the severe stroke.

The activation of mitophagy after stroke is different in variant NVU cells like neurons, microglia, astrocytes, endothelia, etc. (Lan et al., 2018). Referring to their unique biological characteristics, the mitophagy-induced downstream pathways might play distinct roles in the prognosis of stroke. Evidence provided by Cai et al. (2012) has confirmed several specific features of Parkin-mediated mitophagy in cortical neurons. Firstly, Parkin was selectively recruited to dysfunctional mitochondria, which occurred more slowly than non-neurocytes. Secondly, Parkin translocation is restricted to somatodendritic regions of a small fraction of neurons where mature lysosomes are abundant. These spatial characteristics allow neurons to efficiently clear the depolarized mitochondria and maintain cellular vitality (Cai et al., 2012). Research on the mouse hippocampal neuron cell (HT22) OGD model has revealed that the activation of mitophagy by inducing FUNDC1 phosphorylation contributed to less neuronal apoptosis (Cai et al., 2021).

Mitophagy has been shown to play a similar protective role in endothelial cells (Forte et al., 2020). Endothelial progenitor cells (EPCs), isolated from the peripheral venous of human subjects with an NDUFC2 variant, exhibit mitophagy impairment and senescence aggravation when cultured at a high salt level (Zuo et al., 2019). Pharmacological activation of mitophagy restores EPCs’ mitophagy and reduces stroke occurrence in SHRSP rats.

As for microglia, the neuroprotective effect of mitophagy was achieved via alleviating the inflammatory responses that were induced by an ischemic stroke in mouse (Han et al., 2021). Though little research on stroke explored the function of microglial mitophagy, the anti-inflammatory role of microglial mitophagy has been reported in several diseases, including stress-induced hypertension (Zhang S. et al., 2020), glaucomatous neurodegeneration (Jassim et al., 2021), Alzheimer’s disease (Agrawal and Jha, 2020), Parkinson’s disease (Ahmed et al., 2021), etc. Meanwhile, the underlying mechanisms of microglial mitophagy in anti-inflammation are unclear, requiring further elucidation.

In astrocytes, mitophagy was shown to be upregulated after ischemic stroke in vivo and in vitro (Cao et al., 2021). Further activation of mitophagy limited astrocyte polarization to a pro-inflammatory subtype, thereby reducing the production of inflammatory cytokines and neurotoxins (Cao et al., 2021). In another hypoxia/reoxygenation model of primary astrocytes in vitro (Quintana et al., 2019), 24.9 and 43.7% of mitochondria, respectively, undergo mitophagy at 3 h after hypoxia and 10 h of reoxygenation, comparing to 26.4% of mitochondria under normoxic conditions. In addition, the authors have also observed a reduction of astrocytic ATP during hypoxia, but a restoration of ATP content after reoxygenation (Quintana et al., 2019). Enhanced mitophagy and the recovery of ATP production during reoxygenation contribute to decreased mitochondrial quantity and limited astrocytic extensions. This confined astrocytic extension might ameliorate glial scar formation, and consequently reverse its inhibition of axonal growth and regeneration (Liu and Chopp, 2016; Okada et al., 2018).

As stroke develops, the activity of mitophagy keeps changing, which indicates its intricate impact on secondary brain injury. We proposed that mitophagy activated at the acute stage of stroke might be neuroprotective to eliminate the damaged mitochondria, but persistent or immoderate mitophagy in the subacute and chronic stages could be harmful due to enormous mitochondrial degradation.

Mitophagy upregulated by acidic postconditioning in the early stage of ischemic stroke (within 6 h) could render the brain tissue resistant to ischemia in mice (Shen et al., 2017). This enhanced mitophagy may facilitate the clearance of the damaged mitochondria, subsequently eliminating the ROS burden and retarding disease promotion. Intervening mitophagy at the different stages of stroke could result in contrasting effects. Lan et al. (2018) have observed that the levels of mitophagy markers on neurons increased at 6 h after tMCAO, and then reached their peak at 24 h. The severe cellular insult at 24 h excessively motivated mitophagy, leading to the deficit of mitochondrial quality and energy generation (Anzell et al., 2018; Lan et al., 2018).

However, in different stroke models, the concrete time point dividing acute and chronic phases is different. Research on SAH rats has shown that both autophagy proteins (LC3-II and Atg5) and mitophagy markers (Parkin and PINK1) were increased at 24 h after SAH in ipsilateral basal cortical samples nearing the hematoma (Cao et al., 2017). Further activation of mitophagy by melatonin at 24 h after SAH promoted the fragmented mitochondrial elimination, concomitant with a reduction of ROS generation and inflammatory cytokine release (Cao et al., 2017). Increased mitophagy at 48 h after SAH in rats’ cerebrums also played a significant role in attenuating neuronal apoptosis and necrosis (Li J. et al., 2014). In addition, Cheng Y. et al. (2021) have reported that Nrf2/OPTN-mediated mitophagy, at 3 days after ICH in mice, inhibited NLRP3 inflammasome and improved the mitochondrial function in perihematomal brain tissues.

Another possible cause might be the different severity of stroke derived from the varied protocol of animal model establishment. Excessive mitophagy could be attributed to prolonged brain hypoperfusion, massive cerebral infarction, cellular stress after reperfusion, etc. These severe brain injuries experienced nutrient deprivation, DNA destruction, oxidative stress, and mitochondrial depolarization, which played a role in motivating hyperactive and protracted mitophagy. In rat permanent focal ischemia, Baek et al. (2014) have reported that Drp1 and Parkin increasingly recruited to the mitochondria, which resulted in a release of cytotoxic substances, cytochrome C, and apoptosis-associated factors in the ischemic area of pMCAO rats and primary cortical neurons treated by NMDA. Carnosine treatment attenuated ischemic injury in the ipsilateral hemisphere by decreasing mitophagy (Baek et al., 2014). This phenomenon implies that enormous mitochondrial damage and degradation occur in severe stroke models, leading to the deficiency of the mitochondrial network and impaired energy production. A similar effect was reported in GCI following cardiac arrest. Excessive mitochondrial fission and mitophagy contribute to massive hippocampal CA1 neuron apoptosis as evidenced by increasing mitochondrial oxidation, fragmented structure, and disintegrated cristae (Wang R. et al., 2019). On the contrary, cellular mitophagy was considered to play a pro-survival role in preconditioned tMCAO rats that experienced 30 min of tMCAO before 100 min of tMCAO (Matic et al., 2016). The extent of brain injury in this preconditioned tMCAO model might be moderate, thus the optimal mitophagy as a defensive reaction maintains the MQC homeostasis and increases the mitochondrial resistance to reoxygenation and membrane depolarization. Recently, Youn et al. (2021) have compared the morphological changes and autophagy of mitochondria in cerebrospinal fluid cells between patients with and without delayed cerebral ischemia (DCI) SAH. They found a significant increase in mRNA expression of mitophagy markers (DAPK1, BNIP3L, and PINK1) in patients with DCI SAH. This enhanced mitophagy, concomitant with remarkable mitochondrial dysfunction, might be involved in DCI pathogenesis (Youn et al., 2021).

Recent literature studies have also suggested that selective clearance of the damaged mitochondria can promote cellular survival, while non-selective mitophagy leads to neural injury. Physiologically, Lazarou et al. (2015) have described that ubiquitin chains would tag protein aggregates and impair the mitochondria to activate selective autophagy in HeLa cells. Then, PINK1 phosphorylates these ubiquitin chains to further modulate the autophagy signals in the mitochondria. The expression of PINK1 maintains a low level in healthy and polarized mitochondria, which is determined by voltage-dependent proteolysis. If the mitochondria sustained damage, PINK1 would rapidly accumulate to OMM and signal the depolarized mitochondria to Parkin, which amplifies the mitophagy cascade (Narendra et al., 2010; Lazarou et al., 2015). PINK1/Parkin-induced mitochondrial clearance provides protection against neuronal injury, especially the CA1 neurons in the hippocampus of cerebral ischemia in rats (Wang H. et al., 2019). The mislocalization of Parkin induced by oxygen hemoglobin in primary neurons resulted in impaired clearance of the damaged mitochondria and reduced neuronal viability (Deng et al., 2021). Interestingly, the highly selective PINK1/Parkin-mediated mitophagy has been confirmed to be monitored by mitochondrial fission (Burman et al., 2017). Fission-associated protein Drp1 KO HEK293 and HeLa cells developed a higher rate of mitophagy and increased the translocation of Parkin to misfolded protein aggregates of the mitochondria-localized mutant ornithine transcarbamylase (ΔOTC). However, with impaired fission, the selectivity of mitophagy to ΔOTC was destroyed, causing enhanced elimination of healthy mitochondria (Burman et al., 2017). In the early stage of ischemic stroke, Drp1-dependent fission was activated, which selectively triggered the autophagy of dysfunctional mitochondria (Zuo et al., 2014). The inhibition of Drp1 counteracted selective mitophagy and caused the accumulation of the damaged mitochondria (Zuo et al., 2014). It seems possible that mitophagy becomes detrimental partially due to its non-selective clearance of healthy mitochondria. Selective mitophagy targets the damaged mitochondria and reduces the release of DAMPs.

In addition to the abovementioned phenomenon and hypotheses, the controversial role of mitophagy may also be due to different upstream stimulatory factors, mitophagy markers, and intervention phases. To eliminate the damaged mitochondria while maintaining healthy mitochondria, future studies will be needed to uncover the mechanism of mitophagy regulation. A clear definition of the marginal value that identifies the appropriate mitophagy may promote the clinical translation of MQC.

In general, the intricate role of mitophagy seems to be associated with the different NVU cells, time courses of stroke, disease severity, mitophagy selectivity, etc. In addition, different protocols of mitophagy intervention, including drug species, routes, time points, and doses of administration, may be responsible for diverse results of studies. Mitophagy-induced downstream cascades could be various in neurons, glial cells, and endothelial cells. Activated mitophagy reduces cell apoptosis and senescence in neurons and endothelia while alleviating inflammatory responses in microglia and astrocytes. Different activities of mitophagy were reported during the development of stroke. Mitophagy can be activated by great mitochondrial damage at the acute stage of stroke, which might be neuroprotective by eliminating the damaged mitochondria. Due to the repair of damage in the subacute and chronic stages, mitophagy could be attenuated. However, extended or immoderate mitophagy in the chronic stage could be harmful due to massive mitochondrial degradation. The severity of stroke also determines the role of mitophagy. Under moderate stress with bits of mitochondrial damage, selective autophagy of mitochondria is protective to eliminate dysfunctional ones and maintain the homeostasis of the mitochondrial network. However, when suffering from prolonged ischemic insult or extensive hemorrhage, dysfunctional mitochondria explode and trigger excessive and unselective mitophagy, leading to severe neural injury.