Targeting mitophagy for neuroprotection: mechanisms and therapeutic opportunities Authors: Jiahui Yang, Jianfeng Li, Xintong Hou, Yifei Zheng, Zeyu Zhao, Ting Zhou, Tongwei Jing, Jiming Kong, Guohui Zhang, Ying Guo Published: 2026 ABSTRACT AbstractMitochondria are essential for neuronal energy production, cellular homeostasis, and overall neuronal function. Due to their high metabolic demands and limited regenerative capacity, neurons are particularly vulnerable to mitochondrial dysfunction, which leads to ATP depletion, excessive reactive oxygen species (ROS) production, and calcium imbalance—ultimately causing oxidative stress, metabolic disruption, and neuronal death. Mitophagy is a selective process that removes damaged mitochondria through the autophagy-lysosome pathway. As a key mechanism of mitochondrial quality control, mitophagy preserves energy production, limits oxidative damage, and maintains mitochondrial network integrity. This process is regulated by pathways such as PINK1-Parkin and receptor-mediated mechanisms involving BNIP3 and FUNDC1, all of which help sustain cellular health by preventing mitochondrial dysfunction. Impaired mitophagy is a common feature of several neurodegenerative diseases, including Alzheimer’s, Parkinson’s, amyotrophic lateral sclerosis (ALS), and Huntington’s disease, exacerbating mitochondrial damage and neuronal stress. Emerging therapeutic strategies that target mitophagy—ranging from pharmacological agents and gene therapies to dietary interventions—show promise in restoring mitochondrial quality and protecting neurons from degeneration. Nevertheless, challenges remain in translating these findings into effective clinical treatments. Mitophagy represents a critical mechanism for preserving neuronal integrity and offers a compelling target for innovative therapies against neurodegenerative disorders. Subject terms: Cell biology, Neuroscience Introduction Mitochondria are crucial organelles that regulate energy homeostasis, cellular signaling, and cell death. They also serve as the cellular energy centers of all eukaryotic cells1. Through beta oxidation, the tricarboxylic acid cycle (TCA cycle), and oxidative phosphorylation of the electron transport chain (ETC), they supply the necessary “fuel” for maintaining the fundamental functions of cells2. Mitochondrial quality control (MQC) sustains the homeostasis and functionality of mitochondria3. Among these, mitophagy—the selective autophagic degradation of mitochondria—eliminates dysfunctional or damaged mitochondria. It also coordinates with mitochondrial biogenesis and dynamics to regulate mitochondrial quality and quantity, thereby determining cell fate4. Neurons are highly metabolic cells that demand substantial energy to maintain processes such as membrane potential, neurotransmitter release, and reuptake. Hence, neuronal mitochondria must generate ATP efficiently to fulfill this high-energy requirement. Neurons possess complex structures, including long axons and dendrites. Mitochondria within neurons must undergo dynamic distribution through mitochondrial dynamic processes, such as mitochondrial fission, fusion, and transportation, to ensure the timely provision of energy, particularly in synapses and other regions with high energy demands5. Mitochondria in neurons are not only responsible for energy generation but also play a key role in regulating intracellular calcium (Ca2+) concentration. By absorbing and releasing calcium ions, mitochondria act as buffers in the regulation of neuronal excitability and signal transduction6. Compared with other cell types, neurons typically have longer lifespans; consequently, their mitochondria must also maintain long-term functionality. This makes MQC particularly significant in neurons to prevent the accumulation of damaged mitochondria, which can lead to neuronal dysfunction and neurodegenerative diseases (NDDs)7,8. These characteristics render mitochondria vitally important in the normal function and health of neurons. In damaged mature neurons, mitochondria can also be injured, resulting in local mitochondrial depolarization and ATP loss, causing local energy deficits in axons and thereby restricting the regeneration of damaged neurons9. Not only that, both neural development and the maintenance of long-term neuronal health require the effective removal of aggregated proteins and aged or defective organelles. Selective mitophagy can clear aged and dysfunctional mitochondria, thereby maintaining mitochondrial function and cellular homeostasis3. Excessive mitophagy or insufficient mitophagy can lead to mitochondrial dysfunction and disruption of cellular homeostasis8. Not only can it fail to provide sufficient energy for the normal physiological activities of neurons, but it can also cause the excessive accumulation of dysfunctional mitochondria within neurons, thereby inducing neuronal injury processes. This is also an important cause for the occurrence and development of neurodegenerative disease3. Currently, functional impairments resulting from mitophagy disorders have become an important factor in the progression and development of NDDs. However, the precise mechanism by which mitophagy specifically regulates neuronal functions and thereby affects NDDs still requires in-depth research. Mitophagy Mitophagy, a specialized form of selective autophagy, is an evolutionarily conserved process dedicated to the degradation of damaged or dysfunctional mitochondria, thereby ensuring mitochondrial quality control and cellular homeostasis3,10,11. Under physiological or pathological stress, such as excessive ROS, nutrient deficiency, or cellular aging, mitochondria undergo depolarization damage. To maintain a stable mitochondrial network, cells utilize the mitophagy machinery to selectively encapsulate and eliminate these abnormal organelles through four sequential steps: 1) Cells accurately recognize damaged mitochondria targeted for removal and activate the corresponding signaling pathways12. 2) The initial damage signal is amplified, leading to the recruitment of adaptor proteins or receptors that link mitochondria to the autophagy machinery. This step involves both PINK1/Parkin-mediated and receptor-mediated pathways, ultimately enabling engagement of the autophagic system13,14. 3) The autophagy machinery is recruited to the vicinity of the targeted mitochondria, where an isolation membrane forms and progressively engulfs the organelle15. 4) The mature autophagosome subsequently fuses with a lysosome, resulting in the final degradation of the mitochondria3. By integrating these pathways, mitophagy effectively prevents the accumulation of dysfunctional mitochondria and maintains the integrity of the intracellular environment. Ubiquitin-dependent mitophagy The PINK1/Parkin-mediated ubiquitin-dependent pathway is currently the most extensively studied mitophagy pathway. PINK1 is a kinase encoded by the PARK6 gene, and Parkin is an E3 ubiquitin ligase encoded by the PARK2 gene16–18. This pathway consists of three key components (Fig. 1): PINK1 as a mitochondrial damage sensor, Parkin as a signal amplifier, and ubiquitin chains as a signal effector19. They work together to activate mitophagy of damaged mitochondria. Multiple regulatory pathways of mitophagy and their pathophysiological significance. PINK1-Parkin pathway: When mitochondria are damaged and lose their membrane potential, the PINK1 protein accumulates and becomes activated on the outer mitochondrial membrane. It then recruits and activates Parkin. Parkin tags proteins on the outer mitochondrial membrane with ubiquitin chains. These chains recruit autophagy adaptor proteins, such as p62 and NBR1, which bind to LC3 on the forming autophagosome. This ultimately leads to the engulfment and degradation of the damaged mitochondria. Receptor-Mediated Pathways: BNIP3 and NIX are receptors located on the outer mitochondrial membrane. Their expression increases upon mitochondrial damage, allowing them to bind directly to LC3 and initiate mitophagy. FUNDC1 is activated under hypoxic conditions, which exposes its LC3-binding domain to trigger mitophagy. In addition, under mild oxidative stress, mitochondria generate mitochondria-derived vesicles (MDVs), which selectively sequester oxidized proteins, lipids and damaged mitochondrial DNA and deliver them to lysosomes for degradation. Under physiological conditions, mitophagy functions as a sustained, low-level quality control process that maintains cellular homeostasis. When mitophagy is deficient—as seen in ALS (where SOD1 or TDP-43 mutations disrupt the pathway), PD (caused by PINK1 or Parkin mutations), and AD (where Aβ or tau inhibits FUNDC1)—damaged mitochondria accumulate, ultimately triggering neuronal apoptosis. Conversely, excessive mitophagy can lead to the erroneous removal of healthy mitochondria. Figure created with BioRender.com. The process begins with the loss of mitochondrial membrane potential (ΔΨm). This signal prevents PINK1 from entering the inner membrane for degradation. Instead, PINK1 stabilizes and accumulates on the outer mitochondrial membrane (OMM). It then recruits the cytosolic E3 ubiquitin ligase, Parkin, to the surface of the damaged mitochondrion3. Crucially, this recruitment process is not autonomous but is facilitated by the F-box protein FBXO7 (encoded by the PARK15 gene). FBXO7 functions as a scaffolding protein that interacts with both PINK1 and Parkin, thereby markedly accelerating Parkin translocation from the cytosol to damaged mitochondria20. In addition, FBXO7 promotes the aggregation of damaged mitochondria, clustering them to ensure efficient downstream clearance21. These ubiquitin chains serve as “eat-me” signals, recruiting autophagy receptors such as p62 and NBR122,23. Biochemical and genetic analyses confirm that PINK1 acts upstream of Parkin. PINK1’s kinase activity and mitochondrial localization are prerequisites for recruiting Parkin and initiating the autophagy cascade24. Once recruited, Parkin “tags” the damaged mitochondrion by ubiquitinating key proteins on the OMM. Research shows that Parkin mediates the ubiquitination and degradation of the mitochondrial fusion proteins Mitofusin 1 and 2 (Mfn1/2). This not only prevents the damaged mitochondrion from re-fusing with the healthy mitochondrial network but also promotes its fragmentation for clearance25. Furthermore, Parkin ubiquitinates the mitochondrial motor adaptor protein Miro, leading to its rapid degradation. This severs the mitochondrion’s connection to microtubules, causing motility arrest. This ensures the damaged organelle is anchored in place, ready for engulfment by the autophagosome26. In PINK1/Parkin-mediated mitophagy, evidence suggests that NDP52 and OPTN serve as the primary autophagy receptors. They are recruited to the ubiquitin-coated mitochondrial surface. Through their LC3-interacting region (LIR) motifs, they bind to LC3 proteins on the expanding phagophore, effectively guiding the isolation membrane to enwrap the mitochondrion15,27. The presence of ubiquitin chains alone is not sufficient. Downstream receptors must efficiently recognize these chains, a process aided by the kinase TBK1. After Parkin ubiquitinates the mitochondrion, TBK1 is recruited and activated. TBK1 phosphorylates the autophagy receptors OPTN and NDP52, significantly enhancing their binding affinity for ubiquitin chains. This creates a positive feedback loop that accelerates mitochondrial engulfment28. PINK1/Parkin-mediated mitophagy is closely linked to NDDs. Amyloid-beta (Aβ) can directly inhibit the activation of PINK1 and the recruitment of Parkin to mitochondria, resulting in the obstruction of mitophagy29. At the same time, ROS produced by damaged mitochondria will promote the abnormal cleavage of Aβ precursor protein, accelerating the generation of Aβ, forming a vicious cycle29–31. PINK1 and Parkin play a crucial role in immune cells, such as dendritic cells. They promote mitophagy, which prevents mitochondrial antigens from being presented on the cell surface via MHC class I molecules. When the PINK1/Parkin pathway fails, mitochondrial antigens are mistakenly presented to cytotoxic T cells. This can lead the immune system to attack the body’s own neurons32. This mechanism may explain why gut infections—which can trigger mitochondrial antigen presentation—could be a potential trigger for Parkinson’s disease. In addition, there are ubiquitin-dependent mitophagy pathways that do not rely on Parkin. These pathways complement the Parkin pathway, working together to maintain mitochondrial homeostasis. MUL1 is an E3 ubiquitin ligase located on the outer mitochondrial membrane. Unlike Parkin, it does not need to be “recruited”; it continuously monitors mitochondrial quality by ubiquitinating mitofusin 2 (Mfn2), thereby functioning as a parallel and compensatory pathway to Parkin-mediated mitophagy. Studies in both Drosophila and mammalian cell models have demonstrated that MUL1 expression rescues mitochondrial accumulation as well as muscle and neurodegenerative phenotypes caused by Parkin deficiency. When PINK1 or Parkin function is impaired, MUL1-mediated ubiquitination and degradation of Mfn2 suppress mitochondrial fusion while simultaneously tagging the mitochondria with a ubiquitin label, promoting their clearance via autophagy29. After RNF185 ubiquitinates BNIP1, p62 recognizes the ubiquitin chain through its ubiquitin-binding domain (UBA) and binds to the autophagosome through the LIR30. Furthermore, research has found that in the absence of Parkin, the autophagy receptor AMBRA1 can recruit another large E3 ubiquitin ligase, HUWE1, to the outer mitochondrial membrane. HUWE1 extensively ubiquitinates OMM proteins, such as Mfn2. The formation of these Parkin-independent ubiquitin chains is sufficient to recruit downstream autophagy receptors like p62 and OPTN, thereby initiating mitophagy. This study also confirmed that this pathway is regulated by phosphorylation from IKKα and is crucial for preventing cell death under oxidative stress31. In contrast, under conditions of excessive mitochondrial fission in the absence of Parkin, excessive activation of the fission protein Drp1 (dynamin-related protein 1) can cause physical rupture of the OMM, exposing inner mitochondrial membrane (IMM) proteins. These exposed IMM proteins, such as Prohibitin 2, can then be ubiquitinated by non-Parkin E3 ligases, generating signals that recruit the autophagy machinery to complete mitochondrial clearance. This mechanism provides an explanation for the persistence of basal mitophagy observed in Parkin-deficient neurons or cardiomyocytes33. Ubiquitin-independent pathway In mitophagy, there also exists a ubiquitin-independent pathway in which mitochondrial outer membrane proteins are directly recognized by autophagy receptors. This receptor-mediated mitophagy primarily involves FUNDC1, BNIP3, NIX (BNIP3L) (Table 1 and Fig. 1). These pathways play a crucial regulatory role in NDDs, and their abnormalities are closely related to the imbalance of mitochondrial homeostasis in neurons. Abnormalities in Mitochondria-Associated Endoplasmic Reticulum Membranes are an early pathological feature of Alzheimer’s disease (AD)34. FUNDC1 is also localized to MAMs and interacts with the IP3R2 receptor on the endoplasmic reticulum. The presence of FUNDC1 helps maintain the structural stability of MAMs. When FUNDC1 is absent, the MAM structure is disrupted, leading to impaired mitochondrial fission and abnormal calcium signaling. Given that MAMs are often over-strengthened or dysfunctional in AD, FUNDC1’s fine-tuning of MAMs provides a new mechanism for understanding calcium homeostasis imbalance in NDDs35. Other research shows that when neurons suffer from ischemia/hypoxia-reperfusion injury, the mitochondrial kinase CK2 is inhibited. This leads to the dephosphorylation of FUNDC1 at the Ser13 site. This dephosphorylation activates FUNDC1, enabling it to recruit LC3 and initiate mitophagy. Data indicate that activating FUNDC1-mediated autophagy can clear damaged mitochondria, significantly reducing neuronal apoptosis and infarct volume. Conversely, knocking out FUNDC1 significantly worsens neural damage36. BNIP3 (BCL2 Interacting Protein 3) is a single-pass transmembrane protein located on the outer mitochondrial membrane, belonging to the BH3-only subfamily of the BCL-2 family. It is induced by HIF-1α under hypoxic stress and acts as an important receptor-mediated autophagy receptor. BNIP3 directly interacts with the autophagy-related protein LC3 (via its LIR motif), triggering the specific clearance of mitochondria37. Mitophagy dysfunction has been reported in neurons derived from patients with genetic forms of Parkinson’s disease, particularly those carrying mutations in PINK1 (PARK6) or PRKN (PARK2), which disrupt the canonical PINK1/Parkin pathway. Interestingly, overexpression of the mitochondrial outer membrane receptor NIX (BNIP3L) has been shown to restore mitophagy in these neurons by activating a receptor-mediated pathway independent of PINK1/Parkin signaling38. However, whether this compensatory mechanism operates in other genetic forms of PD or in idiopathic PD remains to be determined. NIX does not require phosphorylation by PINK1 nor the ubiquitination activity of Parkin. It can directly bind to LC3 to pull damaged mitochondria into the autophagosome38. This discovery provides theoretical support for the idea that the NIX-mediated pathway can compensate for the loss of the Parkin pathway, protecting neurons from degenerative death. Table 1Mitophagy pathwayPathway / Key ComponentCore mechanismRelevance to NDDsKey ReferencesUbiquitin-Dependent MitophagyPINK1/Parkin PathwayMitochondrial membrane potential (ΔΨm) loss stabilizes PINK1 on the OMM. PINK1 recruits cytosolic Parkin, a process accelerated by the scaffolding protein FBXO7. Parkin ubiquitinates OMM proteins (e.g., Mfn1/2, Miro), tagging mitochondria for clearance. Autophagy receptors (NDP52, OPTN), phosphorylated by TBK1, bridge ubiquitin chains to LC3 on the phagophore.In AD, Aβ inhibits PINK1 activation and Parkin recruitment, blocking mitophagy and creating a vicious cycle with ROS. PINK1/Parkin loss in immune cells can lead to inappropriate mitochondrial antigen presentation, potentially triggering autoimmunity in PD.3,16–1821,25,32Non-Parkin E3 LigasesMUL1: An OMM-resident E3 ligase that ubiquitinates Mfn2, promoting mitophagy and compensating for Parkin loss.RNF185: Ubiquitinates its substrate BNIP1 with K63-linked chains, recruiting p62 for mitophagy.AMBRA1/HUWE1: In Parkin’s absence, AMBRA1 recruits HUWE1 to ubiquitinate OMM proteins, recruiting p62/OPTN.OMM Rupture: Excessive fission causes OMM rupture, exposing IMM proteins (e.g., PHB2) for ubiquitination by other E3 ligases.These pathways provide essential backup mechanisms to maintain mitochondrial homeostasis when the primary PINK1/Parkin pathway is impaired, as in familial PD.29–31,33Ubiquitin-Independent MitophagyFUNDC1An OMM receptor regulated by phosphorylation. Dephosphorylation (e.g., via CK2 inhibition during hypoxia) activates it to recruit LC3. Also stabilizes Mitochondria-Associated ER Membranes (MAMs).MAM dysfunction is an early AD feature. FUNDC1 deficiency disrupts MAMs, impairing mitochondrial fission and calcium signaling. Its activation protects against ischemic neuronal injury.35,36BNIP3 & NIXHypoxia-induced OMM proteins that act as receptors, using their LIR motif to bind LC3 directly, bypassing ubiquitination.In PD patient neurons, impaired PINK1/Parkin mitophagy can be rescued by overexpressing NIX, highlighting its compensatory role.37,38Other Receptors (BCL2L13, FKBP8, PHB2)BCL2L13: OMM receptor inducing fragmentation and LC3 recruitment.FKBP8: Uses an “escape mechanism” to avoid degradation and preferentially binds LC3A.PHB2: IMM protein exposed upon OMM damage; its LIR then binds LC3.These receptors reveal complex, parallel mechanisms for mitochondrial quality control, offering alternative pathways when primary systems fail in diseases like PD.39–41 Mitophagy pathway 3,16–18 21,25,32 MUL1: An OMM-resident E3 ligase that ubiquitinates Mfn2, promoting mitophagy and compensating for Parkin loss. RNF185: Ubiquitinates its substrate BNIP1 with K63-linked chains, recruiting p62 for mitophagy. AMBRA1/HUWE1: In Parkin’s absence, AMBRA1 recruits HUWE1 to ubiquitinate OMM proteins, recruiting p62/OPTN. OMM Rupture: Excessive fission causes OMM rupture, exposing IMM proteins (e.g., PHB2) for ubiquitination by other E3 ligases. BCL2L13: OMM receptor inducing fragmentation and LC3 recruitment. FKBP8: Uses an “escape mechanism” to avoid degradation and preferentially binds LC3A. PHB2: IMM protein exposed upon OMM damage; its LIR then binds LC3. Additionally, recent studies have identified several autophagy receptors involved in mitophagy, including BCL2L13, FKBP8, and PHB2 (Table 1). BCL2L13 is the functional homolog of the yeast mitophagy receptor Atg32 in mammals and operates independently of Parkin. Located on the outer mitochondrial membrane, it contains two key domains: a LIR motif (W-x-x-I at positions 270–273) for binding LC3, and a BH domain. BCL2L13 can simultaneously induce mitochondrial fragmentation and recruit autophagosomes39. Unlike other receptors, FKBP8 possesses a unique “escape mechanism” and preferentially binds to the LC3A isoform. While most outer mitochondrial membrane proteins are ubiquitinated and degraded by the proteasome during autophagy, FKBP8 avoids this process. It strongly recruits LC3A via its N-terminal LIR motif (F-E-V-I at positions 22–25) and directly mediates Parkin-independent mitophagy40. PHB2 is normally located on the inner mitochondrial membrane and remains “invisible” to cytosolic LC3. When mitochondria are damaged and the outer membrane ruptures or is degraded by the proteasome, PHB2 becomes exposed. Its LIR motif (Y-V-L-P at positions 121–124) then binds LC3 to initiate autophagy41. These findings not only reveal the complexity and precision of the autophagy mechanism, but also offer insights into how cells can still maintain a certain degree of mitochondrial homeostasis even when Parkin function is impaired, as seen in conditions like Parkinson’s disease. The bidirectional regulatory effect of mitophagy on neurons Mitophagy acts as a “double-edged sword.” Due to their extremely high energy demands and complex morphology, neurons are highly susceptible to dysfunctional mitochondria42. While mitophagy plays a protective role by removing damaged mitochondria, excessive mitophagy may lead to mitochondrial depletion and subsequent bioenergetic failure, ultimately contributing to neuronal death43,44. Therefore, striking a balance in mitophagy is of crucial importance to ensure the timely clearance of damaged mitochondria while preventing the adverse consequences of excessive activation. Appropriate mitophagy maintains the stability of neuronal function Moderate to mild autophagy serves as a survival-promoting mechanism for neurons by removing protein aggregates and damaged organelles, as well as regulating apoptosis and inflammation. Consequently, this process maintains central nervous system homeostasis. Mitophagy maintains neuronal bioenergetic homeostasis Neurons require a continuous supply of ATP to support electrical signaling and synaptic plasticity45. The majority of this ATP is generated by mitochondria through oxidative phosphorylation (OXPHOS)46,47. Mitophagy is essential for maintaining bioenergetic homeostasis because it selectively removes dysfunctional mitochondria. Damaged mitochondria are inefficient and may reverse ATP synthase activity, consuming rather than producing ATP—a process termed “substrate wasting“8. When the mitochondrial membrane potential (ΔΨm) decreases, the PINK1/Parkin pathway initiates the degradation of these low-efficiency organelles. In PD models, activation of this pathway stabilizes energy supply and reduces dopaminergic neuron loss48,49. Specific mitophagy receptors regulate mitochondrial clearance under different metabolic stresses (Fig. 1). For instance, FUNDC1 mediates mitophagy during hypoxia. Under normal conditions, Src kinase inhibits FUNDC1; however, during hypoxia, FUNDC1 is dephosphorylated, allowing it to bind LC3 and reduce mitochondrial fragmentation50. This selective removal ensures that only functional mitochondria contribute to the cellular metabolic pool. The balance between mitochondrial clearance and energy demand is particularly critical in AD. In HsAPP Tg mice, stimulating OXPHOS without restoring mitophagy leads to the accumulation of damaged mitochondria, creating a metabolic “bottleneck” that exacerbates synaptic loss51. Enhancing lysosomal function can clear these damaged organelles and restore metabolic function52. Furthermore, mitophagy works synergistically with mitochondrial biogenesis via the SIRT3/AMPK/PGC-1α axis. The “cleavage-renewal” cycle ensures that degraded mitochondria are replaced by healthy ones, thereby maintaining ATP efficiency and long-term bioenergetic stability53,54. Mitophagy suppresses neuroinflammation In the pathological landscape of neurodegenerative diseases, mitochondria serve not only as metabolic hubs but also as critical rheostats of the innate immune response55. The accumulation of dysfunctional mitochondria is a primary trigger for chronic neuroinflammation. When mitochondrial quality control fails, compromised mitochondria undergo outer membrane permeabilization or rupture, releasing a barrage of Damage-Associated Molecular Patterns—most notably ROS and mitochondrial DNA (mtDNA)—into the cytosol56. Specifically, excessive mitochondrial ROS facilitates activation of the NLRP3 inflammasome, leading to proteolytic cleavage of pro-caspase-1 and subsequent maturation and secretion of pro-inflammatory cytokines, such as IL-1β57. Simultaneously, the leakage of mtDNA into the cytoplasm triggers the cGAS-STING pathway, further driving the expression of Type I interferons and other inflammatory mediators, which exacerbate the neuroinflammatory microenvironment58. Mitophagy serves as a vital “quality control” checkpoint by selectively sequestering and degrading these damaged organelles via the PINK1/Parkin or receptor-mediated pathways. By removing dysfunctional mitochondria before they can release DAMPs, mitophagy effectively extinguishes the inflammatory fire at its source59. Consequently, defects in mitophagic flux are increasingly recognized as central drivers of the neuroinflammatory “storm” observed in AD and PD, suggesting that pharmacological enhancement of mitophagy may offer a robust therapeutic strategy to suppress chronic brain inflammation52. Mitophagy maintains synaptic function and plasticity Mitophagy is indispensable for preserving mitochondrial quality at distal neuronal processes, thereby directly supporting synaptic architecture and plasticity. The importance of this regulation is underscored by the specialized role of PINK1 in mitochondrial dynamics and dendritic spine development. In Pink1 knockout neurons, mitochondria in dendrites and axons are significantly elongated and reduced in number due to defective fission60. This morphological shift results in a “supply deficit” at synapses, where fewer presynaptic boutons contain mitochondria, and those present are often abnormally large. Consequently, Pink1 knockout neurons exhibit a lower density of mature dendritic spines and excitatory synapses, characterized by reduced Homer1 expression and impaired long-term potentiation60. Mechanistically, PINK1 maintains synaptic integrity by phosphorylating Drp1 at the S616 site, which promotes the mitochondrial fission necessary for their recruitment into dendritic spines. Whether alternative pathways compensate for this fission defect remains an area for future investigation. Beyond local dynamics, synaptic health requires the efficient clearance of damaged organelles, a process heavily dependent on retrograde transport. Han et al. demonstrated that synaptic mitochondrial integrity is maintained through the coordination of Parkin-mediated mitophagy and SNAPIN-driven retrograde transport61. In AD models, an imbalance occurs where increased mitophagy initiation is coupled with defective retrograde transport, leading to the accumulation of partially degraded mitochondria within synapses. Notably, enhancing SNAPIN-mediated transport facilitates the removal of these damaged organelles from the synapse to the soma, effectively alleviating mitophagy stress and rescuing synaptic deficits in AD mice61. Complementing this transport mechanism, the pro-apoptotic protein BOK has emerged as a key regulator of mitophagy flux in AD. BOK facilitates Parkin translocation to damaged mitochondria by competing with the MCL1/Parkin complex, thereby promoting the initiation of clearance. In APPswe/PS1dE9 models, the loss of BOK is closely linked to mitochondrial accumulation; however, hippocampal overexpression of BOK can restore mitophagy flux and improve cognitive function by alleviating mitochondrial dysfunction62. In summary, the integration of PINK1-Drp1 mediated fission, BOK-facilitated Parkin recruitment, and SNAPIN-dependent retrograde transport constitutes a multi-layered quality control system. These pathways ensure that synapses are populated only by healthy, functional mitochondria, offering promising therapeutic targets for restoring synaptic plasticity in NDDs. Excessive mitophagy leads to neuronal dysfunction In recent years, evidence has emerged suggesting that excessive activation of mitophagy may drive neuronal degeneration and neuroinflammation by compromising mitochondrial network integrity, exacerbating bioenergetic dysfunction, or triggering cytotoxicity. The following section elaborates on the mechanisms by which aberrant mitophagy affects neurons through distinct molecular targets, while addressing the current limitations in the field. Normal mitochondria are mistakenly cleared The primary function of mitophagy is to selectively clear damaged mitochondria to maintain cellular energy homeostasis. However, excessive activation or malfunction of mitophagic recognition mechanisms can lead to the inappropriate elimination of healthy mitochondria, contributing to neurodegenerative pathology (Fig. 1). This pathological overactivation is primarily mediated by dysregulation of the PINK1/Parkin pathway. Under normal conditions, PINK1 recruits E3 ubiquitin ligase Parkin to the outer mitochondrial membrane (OMM), where Parkin catalyzes the ubiquitination of OMM proteins, such as Mfn2 and VDAC1. This ubiquitination promotes the recruitment of autophagy receptors (e.g., p62/SQSTM1) and initiates clearance of damaged mitochondria3,63. Under pathological conditions, however, excessive ubiquitination can target even mildly damaged or healthy mitochondria for degradation. In PD models, aberrant Mfn2 ubiquitination blocks mitochondrial fusion and exacerbates fragmentation, contributing to neuronal damage. For instance, in MPTP-induced PD mice, Mfn2 ubiquitination increased by 50%, correlating with a 30% loss of dopaminergic neurons in the substantia nigra64. Importantly, functional studies demonstrate that restoring Mfn2 levels via viral-mediated overexpression significantly attenuates MPTP-induced dopaminergic neurons loss and motor deficits65,66, providing direct evidence that Mfn2 degradation is a causal contributor to neuronal degeneration in this model. Beyond downstream ubiquitin targets, recent studies have revealed that upstream regulators also play a crucial role in pathological mitophagy. For example, glucose-6-phosphate dehydrogenase (G6PD) regulates mitophagy by maintaining the stability of PINK1. In the SH-SY5Y neuronal model, overexpression of G6PD prolonged the half-life of PINK1 protein by two-fold and increased the aggregation of PINK1 on the mitochondrial membrane by 30%, thereby activating the Parkin-mediated ubiquitination cascade reaction, resulting in the erroneous clearance of normal mitochondria (a 40% decrease in mitochondrial DNA copies). Conversely, in neurons with inhibited G6PD, the level of PINK1 protein decreased by 60%, the mitophagy flux was significantly inhibited (the ratio of LC3-II/LC3-I decreased by 50%), and accumulated damaged mitochondria were observed. Interestingly, while it is traditionally believed that G6PD generates NADPH through the pentose-phosphate pathway (PPP) to counteract oxidative stress, the study revealed that the PPP was not involved in the mitophagic inhibition caused by G6PD deletion67. Targeting G6PD can directly affect ubiquitin-dependent mitophagy by regulating the stability of PINK1. This provides an intervention target for the interaction between metabolism and autophagy in the treatment of NDDs. However, it is necessary to be cautious of the bidirectional risks brought about by excessive inhibition or activation. Current research indicates that the excessive activation of the PINK1/Parkin pathway drives abnormal mitochondrial clearance through the ubiquitination of membrane proteins such as Mfn2, representing a key pathogenic mechanism in NDDs. Despite these insights, critical scientific controversies persist regarding the precise regulation of this process. One major gap is the lack of a clearly defined damage threshold for ubiquitination. Although it is established that factors like Parkin mutations or Aβ oligomers trigger excessive Mfn2 ubiquitination, the specific physiological criteria required for the pathway to recognize mildly damaged mitochondria remain unclear. Key parameters, including the precise amplitude of membrane potential decline and the specificity of ubiquitination sites, have yet to be fully characterized. For instance, further analysis is needed to determine whether G6PD overexpression exacerbates excessive autophagy by influencing PINK1 phosphorylation, rather than merely affecting its stability. Additionally, the significance of disease-specific ubiquitination targets has often been underestimated. While Mfn2 ubiquitination in PD leads to mitochondrial fragmentation, VDAC1 ubiquitination in AD appears to preferentially impair metabolic functions such as ATP production. The distinct contributions of these target preferences to neuronal damage have not yet been systematically compared. Aggravation of energy metabolism disruption Mitophagy is a key mechanism for cells to selectively eliminate damaged mitochondria. However, its excessive activation can lead to excessive consumption of functional mitochondria, thereby exacerbating energy metabolism disruption. NIX and BNIP3 act as receptors for mitophagy, mediating selective clearance of mitochondria by directly binding to the autophagosome membrane proteins LC3/GABARAP37,68. Excessive or prolonged activation of NIX/BNIP3-mediated mitophagy promotes aberrant mitochondrial clearance, which in turn disrupts neuronal energy metabolism and triggers apoptosis, contributing to the development of NDDs69 (Fig. 1). Research indicates that following severe metabolic stress (such as hypoxia/ischemia), HIF-1α within neurons transcriptionally activates BNIP370. BNIP3 then inserts into the mitochondrial membrane, directly causing the loss of mitochondrial membrane potential and the opening of the mitochondrial permeability transition pore. This BNIP3-mediated cell death is delayed neuronal death. Knocking out the BNIP3 gene can significantly improve neuronal survival in the damaged brain regions69,71,72. This suggests that inhibiting BNIP3 could be a target for preventing neuronal degenerative death. NIX is also a key responder to severe injury stress. Studies have found that under conditions of spinal cord injury and hypoxic stress, NIX in spinal cord neurons is highly upregulated73. NIX subsequently mediates a burst of mitophagy by binding to LC3. Other research has shown that in PD models, NIX can activate the JNK pathway via the POSH protein, inducing apoptosis74. Furthermore, stimulation with high concentrations of glutamate leads to Drp1-dependent excessive mitochondrial fission, which subsequently triggers large-scale mitophagy. This excessive autophagic clearance directly causes a catastrophic drop in ATP levels (ATP depletion) within neurons. Inhibiting mitochondrial fission and subsequent autophagy—whether pharmacologically or genetically—can maintain ATP levels and rescue neurons75. Therefore, following severe metabolic stress (such as hypoxia/ischemia), the excessive activation of BNIP3 and NIX may be a significant trigger for neuronal energy metabolism collapse. Inhibiting these pathways holds promise as a potential therapeutic strategy for treating NDDs. Aggravation of mitochondrial fragmentation Mitochondrial dynamics, characterized by the balance between fusion and fission, are tightly regulated by a specific set of proteins. Key players include the fission protein Drp1 and the fusion proteins Mfn1/2 and OPA1. The kinase PINK1 exerts context-dependent effects on this balance. While PINK1 is widely reported to promote mitochondrial fission by phosphorylating Drp1 to facilitate mitophagy76, other studies highlight its role in supporting mitochondrial fusion by stabilizing OPA1 and modulating Mfn2 activity77,78. Normal mitophagy (such as the PINK1/Parkin pathway) selectively eliminates damaged mitochondria, maintaining mitochondrial network homeostasis; but excessive autophagy breaks this balance, leading to excessive fission and fragmentation of mitochondria. Drp1 is the core regulatory factor for mitochondrial fission. Its excessive activation leads to mitochondrial fragmentation and excessive autophagy, causing neuronal damage in NDDs. For example, the PINK1 kinase directly phosphorylates the S616 site of Drp1, enhancing its mitochondrial recruitment and fission activity76. Experimental evidence shows that mutant Huntingtin (mHtt) protein not only aggregates in the nucleus but also binds to Drp1 on mitochondria. mHtt acts as a potent activator, significantly increasing the GTPase activity of Drp1. This abnormal elevation in enzymatic activity causes the mitochondrial network to fragment. The resulting mitochondrial fragments are too small or too impaired to be transported efficiently, leading to defects in axonal transport and subsequent neuronal death79. Pathologically high intraocular pressure can also lead to the phosphorylation of Drp1 at the S616 site by activating ERK1/2, triggering mitochondrial fragmentation and extensive apoptosis in retinal ganglion cells, resulting in visual field defects in glaucoma patients80. In the PD model, oxidative stress also leads to enhanced acetylation at the K711 site of Drp1, promoting oligomerization and mitochondrial division. The SIRT3 deacetylase can rescue mitochondrial dysfunction by inhibiting K711 acetylation and reducing dopaminergic neuron death81. Similar to the phosphorylated S616 site of Drp1 in PD, in the AD model, Aβ triggers Drp1 phosphorylation via CDK5, increasing the mitochondrial division rate by 60% and reducing synaptic mitochondrial density by 45%. In PINK1-deficient models, excessive mitochondrial fission and impaired fusion contribute to mitochondrial fragmentation and dysfunction. For instance, Lutz et al. and Sandebring et al. reported that blocking Drp1-dependent fission—either by genetic or pharmacological means—can partially rescue mitochondrial morphology, membrane potential, and function. Additionally, dominant-negative Drp1 mutants or pharmacological inhibitors like Mdivi-1 restore mitochondrial network integrity and have protective effects on neuronal survival and function in PINK1-deficient and PD models60,82. These findings underscore that PINK1 acts within a complex regulatory network influencing both fission and fusion, and that therapeutic strategies targeting excessive fission can alleviate mitochondrial dysfunction caused by PINK1 deficiency. Furthermore, the above-mentioned PINK1/Parkin mechanism, which excessively activates the ubiquitination of Mfn2, also works in synergy with Drp1, thereby exacerbating mitochondrial fragmentation. This is because the mitochondrial fragments produced by the division of Drp1 are more easily recognized by PINK1/Parkin, triggering ubiquitination tags and autophagic clearance83. However, the therapeutic use of Drp1 inhibitors like Mdivi-1 raises critical questions: does the suppression of pathological fission also interfere with essential physiological division, such as that required for axonal transport? Prolonged Drp1 inhibition might disrupt the positive regulation of mitochondrial quality control, potentially introducing new pathological risks. Abnormal mitophagy in NDDs NDDs such as Alzheimer’s disease, Parkinson’s disease, and Huntington’s disease (HD) are closely related to the decline and progressive death of neurons. Abnormal mitophagy plays a crucial role in this pathological process. As shown in Fig. 2, genetic variations, oxidative stress, protein aggregation, and other factors can interfere with the key signaling pathways of mitophagy (such as PINK1/Parkin, mTOR, NIX/BNIP3, etc.), resulting in decreased mitochondrial clearance efficiency or abnormal activation. This leads to the accumulation of damaged mitochondria, excessive production of reactive oxygen species, and a cascade of apoptosis. This dynamic imbalance of mitophagy not only exacerbates the metabolic dysfunction of neurons but also forms a vicious cycle with the abnormal deposition of disease-specific pathological proteins (such as α-Syn, amyloid-beta, Tau protein), jointly promoting the progression of NDDs. Elucidating the molecular mechanisms of aberrant mitophagy may thus offer a novel perspective for developing therapeutic strategies that target mitophagy regulation to treat these disorders. Mitophagy and NDDs. Figure created with BioRender.com. Alzheimer’s disease The pathological features of the brain in AD patients are the presence of Aβ plaques and neurofibrillary tangles composed of hyperphosphorylated tau (p-tau) protein in specific brain regions, leading to extensive neuronal death and brain atrophy84. Synaptic damage and mitochondrial dysfunction are considered as the early cellular changes in AD85. Specifically, tau accumulation impairs mitophagy, resulting in the retention of dysfunctional mitochondria and increased neuronal toxicity86. Furthermore, N-terminal tau fragments can cause aberrant recruitment of Parkin and UCHL-1 to mitochondria, triggering excessive organelle removal and synaptic loss, thereby accelerating AD progression87. Beyond tau-mediated damage, Aβ accumulation induces excessive mitochondrial ROS production, which damages the mitochondrial network and promotes further tau aggregation88. Aβ and tau also interfere with mitochondrial transport, leading to further synaptic dysfunction89. Notably, Aβ also inhibits the initiation of mitophagy by the PINK1/Parkin pathway (Fig. 1). Du et al. found that Aβ downregulates the expression or activity of PINK1 and Parkin, hindering the ubiquitination marking of damaged mitochondria. For example, in Aβ-treated SH-SY5Y cells, the protein level of PINK1 decreased by 40%, and the recruitment of Parkin to mitochondria decreased, resulting in the collapse of mitochondrial membrane potential (ΔΨm) and mitochondrial fragmentation90. Moreover, Aβ also downregulates the expression of mitophagy receptor proteins. In the brain tissues of AD patients, the expression levels of OPTN, BNIP3, and NIX, which are mitophagy-related proteins, are significantly lower than those in normal tissues91. This abnormality is closely related to mitochondrial clearance disorders, oxidative stress, and synaptic function deterioration in AD. In the hippocampus and prefrontal cortex, the mRNA and protein levels of OPTN decreased by 30%-50% compared to normal controls, and were negatively correlated with the degree of Aβ deposition92. This may be related to the activation of GSK-3β phosphorylation of OPTN (Ser473 site) by Aβ oligomers, promoting its ubiquitin-mediated degradation, resulting in reduced protein stability92. In AD, several studies have demonstrated downregulation of the fusion proteins Mfn1 and Mfn2 in patient neurons, leading to impaired mitochondrial fusion and abnormal morphology. For example, Chen et al. (2003) reported that Mfn1/2 double-knockout mice showed mitochondrial fragmentation and reduced energy metabolism, which resemble AD-like pathological features. Loss of Mfn2 can also disrupt ER–mitochondria contact sites, disturbing calcium signaling and lipid metabolism. Additionally, Aβ accumulation reduces OPA1 expression, promoting mitochondrial fragmentation and impairing autophagy, which contributes to tau hyperphosphorylation and neurofibrillary tangle formation93. Recent studies have shown that tau acetylation can exacerbate mitochondrial fragmentation by suppressing mitochondrial biogenesis and fusion proteins (Mfn1/2, OPA1). Importantly, treatment with BGP-15 has been reported to counteract this by activating the PINK1/Parkin pathway, enhancing mitophagy, and promoting cognitive recovery in AD models94. The deposition of Aβ plaques and pathological tau protein are the core pathological features of AD. Impaired mitophagy triggers the accumulation of Aβ and Tau by increasing oxidative damage and cellular energy deficiency, leading to synaptic dysfunction and cognitive impairment95. This dysfunction creates a vicious cascade that perpetuates neurodegeneration. Therefore, targeting tau acetylation or the Aβ-PINK1 axis represents a key direction for precision therapeutic interventions in AD. Parkinson’s disease The hallmark of PD is the degeneration of dopaminergic neurons within the substantia nigra96,97. The neurons are vulnerable in part because of DA oxidation, as the resulting oxidative metabolites directly attack mitochondria and damage mtDNA. Simultaneously, oxidized dopamine reduces the activity of the lysosomal enzyme glucocerebrosidase. This lysosomal dysfunction impedes the clearance of α-Syn, causing a synergistic “mitochondrial-lysosomal” dual failure98. Besides this internal toxicity, genetic factors also increase mitochondrial stress. α-Syn was first identified in Alzheimer’s plaques99. Early studies found mutations in the α-Syn gene in families with early-onset PD100. However, toxic α-Syn also damages mitochondria in sporadic PD cases. Di Maio et al. found that toxic α-Syn binds to the TOM20 receptor. This binding blocks protein import into mitochondria by 50%–60%. This blockage triggers ROS production and the accumulation of damaged mitochondria101,102. Furthermore, α-Syn inserts into the inner mitochondrial membrane. This interferes with Complex I activity and further increases ROS production103. Oxidative stress and protein stress eventually overwhelm neuronal defenses. The failure of mitochondrial dynamics is a key driver. This involves fusion proteins Mfn2 and OPA1. First, Mfn2 connects mitochondria to the endoplasmic reticulum (MAMs). In PD, excessive loss of Mfn2 breaks this connection. This disrupts calcium transfer and causes energy failure104,105. Additionally, OPA1 maintains the structure of mitochondrial cristae. In PD, the loss of OPA1 causes cristae to collapse. This structural change releases cytochrome c to start apoptosis. It also reduces energy efficiency and causes a burst of ROS106,107. Neurons also fail to clear these damaged organelles. The protein Miro1 anchors mitochondria to microtubules. Miro1 must be degraded to stop mitochondrial movement. This allows autophagosomes to engulf them. However, Hsieh et al. showed that PD neurons fail to remove damaged mitochondria. This is due to Miro1 retention. They found this defect in neurons from 16 out of 17 sporadic PD patients. In healthy neurons, Miro1 levels drop to less than 20% after stress. In contrast, PD neurons keep Miro1 levels high at 80–100%. Consequently, the failure to degrade Miro1 halts mitophagy initiation, leading to the persistent retention of dysfunctional mitochondria108. The accumulation of damaged mitochondria triggers lethal signals. One major pathway is inflammation. Retained mitochondria release mtDNA. This mtDNA acts as a danger signal109. It triggers the cGAS-STING pathway110. This causes inflammation and kills dopaminergic neurons111. Another pathway links clearance defects to ferroptosis. Recent work shows that inhibiting mitophagy increases lipid ROS. This speeds up neuronal death. Research found that restoring mitophagy via USP30 knockdown reduced lipid ROS by about 40%. It also reduced neuronal death. This suggests a critical link between mitochondrial control and ferroptosis112. Understanding these mechanisms helps researchers find clinical biomarkers. Many focus on α-Syn. Tokuda et al. used ELISA to study α-Syn in cerebrospinal fluid (CSF). They found higher levels of α-Syn oligomers in PD patients. The ratio of oligomers to total α-Syn has high diagnostic value113. Subsequent studies extended these findings to plasma, identifying elevated oligomers as a robust blood-based biomarker114. However, total protein detection has limitations. It lacks specificity. Total protein includes many different forms. The monomeric form changes easily due to metabolism and kidney function115. Therefore, it is less stable than oligomers or phosphorylated forms116. These findings support the development of blood or CSF diagnostic tests. Huntington’s disease HD is a neurodegenerative disorder caused by the pathological expansion of a trinucleotide (CAG) repeat sequence within exon 1 of the huntingtin (HTT) gene117. The disease is characterized by the selective death of medium spiny neurons in the striatum, where a growing body of evidence implicates the dysregulation of mitophagy as a key contributor to neuronal damage in HD118,119. The role of mitophagy in HD appears complex and bidirectional. On one hand, studies indicate that the expanded polyQ tracts in mHtt can impair the formation of the autophagy initiation complex in striatal cells, thereby disrupting the autophagic process and contributing to mitophagic dysfunction119,120. Simultaneously, mHtt disrupts mitochondrial dynamics, leading to elevated reactive oxygen species (ROS) levels and systemic mitochondrial dysfunction79. These interconnected deficits in clearance and function likely create a permissive environment for neuronal toxicity. On the other hand, research also suggests that excessive or untimely mitophagy may play a detrimental role under specific conditions. For instance, the accumulation of Valosin-containing protein (VCP), which binds to mHtt, has been shown to induce excessive mitophagy, leading to neuronal death43. Furthermore, evidence points to the GTP hydrolase Rhes—highly enriched in the striatum—as another mediator; its upregulation may trigger mitophagy via the NIX receptor, contributing to striatal cell death121. This highlights that the net effect of modulating mitophagy is context-dependent, involving a balance between its insufficient and potentially excessive activation. In addition, mitophagy requires not only the formation of autophagosomes but also their transport to the cell body for fusion with lysosomes. The interaction between the mitophagy receptor Optineurin and its partner protein Rab8 is compromised by mHtt. This impairs the efficient retrograde transport of mitochondria-containing autophagosomes along microtubules to the cell body, where lysosomes are primarily located, thereby preventing the autophagic process from proceeding normally27. In summary, mHtt does not simply suppress or enhance mitophagy, but rather uncouples mitochondrial fission from degradation while simultaneously disrupting the autophagic flux. Simply using broad-spectrum autophagy activators (e.g., Rapamycin) might be insufficient or even detrimental if the downstream transport and lysosomal fusion machinery remain blocked by mHtt; conversely, inhibiting autophagy to prevent VCP-mediated death could exacerbate the accumulation of toxic aggregates. Consequently, future therapeutic strategies must move beyond global modulation. Targeting the regulation of mitophagy to restore its functional homeostasis, rather than merely inhibiting or activating it, represents a promising but complex research direction for treating HD. Amyotrophic lateral sclerosis ALS, also known as motor neuron disease, is characterized by abnormal aggregation of pathological proteins TDP-43 and SOD1122,123. The exact cause of ALS remains uncertain, but more and more studies have shown that ALS-related genes are involved in regulating mitophagy, including TDP-43, SOD1, Drp1 (Dynamin-related protein 1), and OPA127,124,125. Among them, increased expression of TDP-43 has been proven to induce mitochondrial dysfunction, including decreased mitochondrial membrane potential and increased ROS production. Conversely, inhibiting TDP-43 expression reduces mitochondrial Complex I activity and diminishes ATP synthesis126. Intriguingly, there are also studies that have found that inducing strong autophagy in the spinal cord exacerbated the phenotype of WTxQ331K mice with TDP-43 expression, manifested as shortened lifespan, accelerated loss of motor neurons and significant clearance of TDP-43 in the cell nucleus. This is likely related to the excessive clearance of mitochondria in motor neurons127. Additionally, Drp1 is dephosphorylated in several ALS models, including those with SOD1 and TDP-43 mutations, and this dephosphorylation is mediated by the pathological induction of protein phosphatase 1 (PP1) activity in these models. In primary neuron culture models, iPSC-derived human motor neurons and in vivo zebrafish models, inhibiting the PP1-Drp1 cascade reaction effectively prevented ALS-related symptoms, including mitochondrial fragmentation, damage to mitochondrial complex I, axonal degeneration and cell death125. Beyond these factors, OPTN mutations have been identified as pathogenic in ALS, as mitophagy critically depends on OPTN and its kinase TBK1128. Recent studies reveal that OPTN directly interacts with the microtubule and mitochondrial transport complex TRAK1/KIF5B in a C-terminal-dependent manner to stabilize them for proper anterograde axonal mitochondrial transport. Notably, the overexpression of OPTN/TRAK1/KIF5B not only prevents neurodegeneration induced by OPTN truncation but also protects against damage from high intraocular pressure while promoting optic nerve regeneration. These findings suggest that OPTN acts as a microtubule-dependent promoter of mitochondrial transport; its absence may constitute a primary molecular mechanism driving neurodegeneration129. Based on these findings, targeting mitochondrial quality control presents promising new directions for ALS treatment. One vital strategy involves mitigating TDP-43-induced mitochondrial toxicity by blocking its inhibition of Complex I and dampening oxidative stress. Additionally, maintaining Drp1 phosphorylation homeostasis is crucial to prevent excessive mitochondrial fragmentation and pathological autophagy. Furthermore, utilizing gene therapy or small-molecule agonists to restore the function of the OPTN-TRAK1-KIF5B complex offers a viable approach to rescuing anterograde axonal transport. Ultimately, these multifaceted interventions aim to restore the physiological distribution and clearance of mitochondria within axons, representing a critical frontier in ALS therapeutics. Targeting KEy molecular factors of mitophagy for the treatment of NDDs NDDs are closely related to mitochondrial dysfunction. The accumulation of damaged mitochondria leads to oxidative stress, bioenergetic dysfunction, and protein toxicity cascades. Mitophagy selectively eliminates defective mitochondria. How to achieve a precise balance between “removing dysfunctional” and “preserving healthy” mitochondria remains a core challenge in the development of drugs targeting the ubiquitination pathway. Therefore, regulators that target the regulation of mitophagy can play an important role in treating NDDs by activating the protective pathways of mitophagy or correcting excessive mitophagy. However, small molecule targeting is relatively scarce for the treatment of HD and ALS. Therefore, the following section will focus on the targeted regulation of mitophagy for the treatment of PD and AD, highlighting their mechanisms of action, therapeutic effects, and current research progress, and discussing their limitations and future research directions. Alzheimer’s disease Mitophagy regulators are considered promising therapeutic strategies for enhancing mitochondrial function in the elderly and improving AD-related symptoms as shown in Table 2. In this field, researchers have explored various synthetic compounds and natural products that regulate mitophagy to eliminate dysfunctional mitochondria. Table 2Mitophagy for the treatment of NDDsDisease TargetAgent / StrategyTypeMechanism of actionKey findings / Clinical statusReferenceAlzheimer’s DiseaseResveratrolNatural (Polyphenol)Activates proteasome and AMPK-SIRT1 pathway; modulates GSK-3β to reduce Tau phosphorylation.Preclinical: Promotes Aβ clearance (not inhibition of generation); reduces pTau. Clinical (Phase II): Safe profile; potential trend in stabilizing cognitive decline, though no significant drop in plasma Aβ.130–132,142Aloe-Emodin (AE)NaturalActivates AMPK/PGC-1α/SIRT3 pathway. Upregulation of mitochondrial SIRT3 is crucial.Reduces ROS, restores membrane potential, and regulates mitophagy proteins (Beclin1, LC3, etc.). Specificity confirmed via SIRT3 knockdown.133β-AsaroneNatural (Acorus tatarinowii)Upregulates PINK1 and Parkin.Positive: Reduces hippocampal Aβ, improves cognition. Caution: Excessive autophagy activation in late-stage AD may be detrimental (inhibiting Beclin-1 may be protective then).134,135Urolithin A (UA)Natural (Metabolite)Enhances lysosomal function (normalizes cathepsins like Cathepsin Z) to restore the “lysosomal sink.”Preclinical: Reduces Aβ Tau pathology, improves memory/synaptic plasticity. Clinical (Phase I): Safe, bioavailable, induces mitochondrial gene expression in elderly.136,137UMI-77SyntheticInhibits MCL-1/Bax interaction; releases MCL-1 to bind LC3A.Selectively eliminates damaged mitochondria. Reduces Aβ deposition and neuroinflammation; improves cognition.138MelatoninHormone/ SupplementN/A (Sustained-release formulation used).Clinical (Phase II): Significantly improved cognition (ADAS-Cog), sleep efficiency, and daily living ability in insomnia subgroups.141LifestyleInterventionDiet rich in antioxidants/polyphenols; regular exercise.Enhances mitochondrial health; may delay AD occurrence and progression.139,140Parkinson’s DiseaseLovastatinDrug (Statin)Agonist for phosphatase SHP2; promotes Parkin translocation and dephosphorylation (Tyr143) to unlock E3 ligase activity.Preclinical: Alleviates motor deficits in MPTP mice. Epidemiological: High cumulative dose associated with protective effect against PD onset in Japanese elderly.145,146Lycopene NanodotsNanomedicineTPP coupled with ferritin nanocages targets neuronal mitochondria.High biocompatibility. Promotes PINK1/Parkin mitophagy and accelerates α-Syn excretion. Restores motor function.147AMBRA1Co-regulatorInteracts with Parkin.Overexpression reduces apoptotic cell death induced by neurotoxins (rotenone/6-OHDA) in vitro.148,149USP30 Inhibitors (e.g., MTX115325)SyntheticInhibits USP30 (which antagonizes mitophagy via deubiquitylation).Prevents behavioral deficits, increases mitophagy, and reduces α-Syn induced neuron loss. MTX115325 is brain-penetrant.150,151GSK3357679ASyntheticInhibits pathogenic LRRK2 kinase activity.Restores mitophagy defects in LRRK2 G2019S mice; first demonstration of rescuing mitophagy damage in vivo.152SulforaphaneCompoundModulates Nrf2-mediated oxidative stress response.Reduces neuronal loss and inhibits microglial activation. Phase II trials ongoing for safety/efficacy.153 Mitophagy for the treatment of NDDs A major class of natural regulators involves the activation of the AMPK-Sirtuin signaling axis. Preclinical studies (in vitro and AD animal models) have shown that Resveratrol, a non-flavonoid polyphenolic compound, promotes the clearance of Aβ peptides. Rather than inhibiting Aβ generation, it primarily activates the proteasome130 and the AMPK-SIRT1-dependent autophagy pathway to degrade intracellular Aβ deposits131. Concurrently, it mitigates neurofibrillary tangles by reducing Tau phosphorylation at Ser396 via GSK-3β modulation and promoting ubiquitin-proteasome degradation of Tau through SIRT1-mediated deacetylation132. Similarly, Wang et al. found that Aloe-Emodin activates mitophagy via the AMPK/PGC-1α/SIRT3 pathway. In AD models, AE enhances mitochondrial function, reducing ROS and restoring membrane potential. Mechanistically, the upregulation of mitochondrial SIRT3 is crucial; this protein regulates mitochondrial proteins and promotes mitophagy, as evidenced by changes in Beclin1, LC3, P62, Parkin, and PINK1 levels. Notably, SIRT3 knockdown in HT22 cells abolished these benefits, confirming the pathway’s specificity133. Other natural compounds function by directly modulating mitophagy receptors or upstream kinases, though some exhibit stage-dependent effects. β-Asarone (from Acorus tatarinowii) has been confirmed to enhance mitophagy by up-regulating PINK1 and Parkin in neurons, reducing hippocampal Aβ load and improving cognition134. However, its role is complex and stage-dependent; in late-stage AD, β-Asarone may exert neuroprotection by inhibiting Beclin-1-dependent autophagy, suggesting that excessive autophagic activation in advanced pathology could be detrimental135. Additionally, Kaempferol (a flavonoid) and Rhapontigenin act as natural inducers that promote mitochondrial fission and increase mitophagy-related proteins, thereby reducing p-Tau and Aβ levels. Kaempferol is particularly notable for its ability to penetrate the blood-brain barrier, offering distinct clinical advantages. Emerging research has also highlighted the critical link between lysosomal function and mitophagy. Urolithin A (UA), a metabolite of ellagic acid, represents a new target for combined intervention. Preclinical studies in AD transgenic mice demonstrate that UA induces mitophagy not just by targeting mitochondria, but by enhancing lysosomal function—specifically by normalizing the activity of lysosomal cathepsins (such as cathepsin Z). This restoration of the “lysosomal sink” significantly reduces Aβ and Tau pathology, improves olfactory and cognitive functions, and restores synaptic plasticity136,137. Beyond natural products, synthetic strategies are being developed to target specific protein interactions. UMI-77, a MCL-1 inhibitor, targets the mitophagy receptor MCL-1. By inhibiting the interaction between MCL-1 and Bax, UMI-77 releases free MCL-1, which then binds to LC3A to selectively eliminate damaged mitochondria. In AD mouse models, two months of treatment significantly reduced Aβ deposition and neuroinflammation while improving cognition138. Furthermore, broad lifestyle interventions support these molecular findings; diets rich in antioxidants and polyphenols, alongside regular exercise, have been shown to enhance mitochondrial health and may delay AD progression139,140. Translating these preclinical findings into clinical applications is the next critical step. In a phase II randomized double-blind controlled trial, sustained-release Melatonin (2 mg daily) significantly improved cognitive performance, sleep efficiency, and, in subgroups with insomnia, daily living ability and MMSE scores141. Resveratrol has also been tested in a phase II trial for mild-to-moderate AD; while it showed an acceptable safety profile and a potential trend toward stabilizing cognitive decline, it did not significantly lower plasma Aβ levels142. Finally, Urolithin A has demonstrated safety, bioavailability, and the ability to induce mitochondrial gene expression in a first-in-human phase I trial in elderly individuals143. While not yet tested for efficacy in AD patients, these findings support further translational exploration. In conclusion, the stage-specific effects of compounds like β-Asarone underscore that mitophagy regulation requires a delicate balance; excessive activation in late-stage AD may trigger mitochondrial depletion. Moreover, the mechanism of Urolithin A suggests that inducing mitophagy is insufficient if downstream lysosomes cannot effectively process the sequestered waste. This mismatch creates “autophagic stress,” where cellular waste is “packaged” but not “incinerated.” Future drug screening should prioritize a “mito-lysosomal dual repair” mechanism to ensure that waste is thoroughly cleared rather than simply accumulated in autophagosomes. Parkinson’s disease Mitophagy is a crucial pathogenic mechanism underlying dopaminergic neuron loss in PD. Whether in familial or sporadic cases, mitophagy dysfunction is a hallmark of the disease. As shown in Table 2, proteins encoded by the PINK1 and PRKN (Parkin) genes are central to the mitophagy process. While early-onset PD is often directly linked to defects in these genes, impaired mitophagy is also a consistent feature of sporadic PD144. Consequently, therapeutic strategies targeting the PINK1/Parkin pathway and its regulatory networks have emerged as a major area of drug discovery. One direct approach involves pharmacological agents that enhance Parkin activity. For instance, the lipid-lowering drug lovastatin functions as an agonist of the phosphatase SHP2. By binding to SHP2, lovastatin promotes its translocation to mitochondria, where it interacts with Parkin and dephosphorylates the Tyr143 residue. This specific dephosphorylation unlocks Parkin’s E3 ligase activity, thereby enhancing mitophagy and alleviating motor deficits in MPTP-induced PD mouse models145. Supporting the translational potential of this mechanism, a recent case-control study in an elderly Japanese population found that statin use was associated with a reduced risk of PD. Notably, a lower cumulative statin dose was associated with an increased risk of PD, whereas a higher cumulative dose showed a protective effect against its onset. These findings further underscore its clinical relevance, although optimal dosage control requires further investigation146. Beyond small molecules, novel delivery systems are being developed to overcome bioavailability challenges. Targeted nanomedicines, such as lycopene nanodots (TPP coupled with ferritin nanocages), have been designed to specifically target neuronal mitochondria. Due to their high biocompatibility, these nanodots can effectively protect mitochondrial function in MPTP models, promote PINK1/Parkin-mediated mitophagy, and accelerate α-Syn excretion, ultimately restoring motor function in mice147. In addition to direct PINK1/Parkin targeting, modulating co-regulators of mitophagy offers another therapeutic avenue. Studies indicate that AMBRA1 interacts with Parkin and is crucial for mitochondrial clearance. In in vitro PD models (SH-SY5Y cells), the overexpression of AMBRA1 was sufficient to promote mitophagy and synergistically reduce apoptotic cell death induced by neurotoxins such as rotenone and 6-OHDA. These findings support the concept that targeting AMBRA1-mediated mitophagy could be a promising therapeutic approach148,149. Conversely, inhibiting negative regulators of mitophagy, such as the deubiquitinase USP30, represents a parallel strategy. While PINK1 and Parkin orchestrate the clearance of dysfunctional mitochondria by inducing ubiquitylation, this process is antagonized by USP30-mediated deubiquitylation150. Recent studies demonstrate that the loss of USP30 in knockout mice prevents behavioral deficits, increases mitophagy, and reduces both phosphorylated S129 α-Syn levels and α-Syn-induced dopaminergic neuron loss. These findings were validated using MTX115325, a potent, selective, and brain-penetrant USP30 inhibitor151. These data strongly support further research into USP30 inhibition as a potential disease-modifying therapy. Additionally, targeting upstream kinases and oxidative stress pathways shows promise in restoring mitochondrial homeostasis. The kinase inhibitor GSK3357679A inhibits pathogenic LRRK2 activity and has been shown to restore mitophagy defects in LRRK2 G2019S mice. This study is the first to demonstrate that targeting LRRK2 can rescue mitophagy damage in vivo152. Furthermore, in MPTP-induced models, sulforaphane has been shown to reduce neuronal loss and inhibit microglial activation by modulating the Nrf2-mediated oxidative stress response. Clinical evaluation is currently underway, with phase II trials assessing the safety and efficacy of sulforaphane in improving movement symptoms, though full results are yet to be disclosed153. Current research suggests a shift in treatment strategies for PD. The focus is moving from non-specific “broad-spectrum autophagy activation” to “precise molecular regulation” based on disease mechanisms. While clearing damaged mitochondria is crucial, excessive or uncontrolled autophagy risks mitochondrial depletion or even synaptic loss. Current approaches, such as USP30 inhibition or AMBRA1 overexpression, mainly focus on accelerating the process. This is similar to “stepping on the gas”. Yet, determining the optimal “Therapeutic Window” remains a core challenge for pharmacological development. Researchers must find ways to precisely identify and clear “pathological” mitochondria without affecting healthy mitochondrial function. Furthermore, translating results from animal models to clinical applications requires further exploration. Conclusion Mitophagy plays a complex, dual role in NDDs, where both its impairment and overactivation can precipitate neuronal damage and subsequent neurological disorders. Despite the therapeutic potential of targeting this pathway, several critical limitations persist. First, the precise mechanisms distinguishing physiological from pathological mitophagy, along with the specific “damage threshold,” remain elusive. For example, in HD, the accumulation of Valosin-containing protein (VCP) induces excessive mitophagy, leading to neuronal death43. Similarly, in an ALS mouse model (WTxQ331K), inducing robust autophagy shortened lifespan and accelerated motor neuron loss127. In AD research, β-asarone may cause adverse effects in the late stages of the disease. This is due to the excessive activation of autophagy154. Moreover, effective mitophagy requires not only autophagosome formation but also their successful transport to lysosomes for degradation. In HD, mHtt blocks the retrograde transport of these autophagosomes. Consequently, utilizing broad-spectrum autophagy activators like Rapamycin could be detrimental if downstream transport remains obstructed, as this would exacerbate the accumulation of toxic aggregates155. This indicates a critical problem. We must clear damaged mitochondria. However, we must also avoid excessive clearance or toxic accumulation caused by transport blockages. Therefore, achieving precise bidirectional regulation of mitophagy is essential. It controls the quality and quantity of mitochondria in neurons. This ensures normal cell function. This challenge provides an important direction for future research. Furthermore, most studies rely on cell or animal models. We still need to verify their direct relevance to human disease. The link to human pathology is not yet fully confirmed. Similarly, we face challenges in clinical translation. Many drugs show significant efficacy in preclinical studies. However, it is often difficult to reproduce these results in human clinical trials. For example, Resveratrol showed significant effects in animal models. Researchers tested it in patients with mild to moderate Alzheimer’s disease. The drug had an acceptable safety profile. However, it did not significantly reduce plasma beta-amyloid levels142. Additionally, a major obstacle in treating NDDs is drug delivery. Drugs must effectively cross the blood-brain barrier and remain stable. Researchers are exploring nanocarrier technology to address this issue, and they have made some progress. For instance, in a mouse model of Parkinson’s disease, scientists constructed sequence-targeted H-ferritin nanocages. These nanocages were loaded with lycopene. This system successfully crossed the blood-brain barrier. It also specifically targeted mitochondria within neurons. This approach not only cleared reactive oxygen species (ROS) but also significantly restored the autophagic clearance of damaged mitochondria. Additionally, it accelerated the metabolism of α-syn147. Current research shows a shift in focus. The trend is moving from “broad-spectrum autophagy activation” to “precise molecular regulation.” Nano-targeted delivery technology offers new solutions for delivering drugs into the brain and mitochondria. This addresses a key bottleneck in treatment. It provides a very promising strategy for treating NDDs. Acknowledgements We gratefully acknowledge the financial support from the Natural Science Foundation of Hebei Province (C2024405023), Doctoral Start-up Research Grant of Hebei North University (BSJJ202512), and the Postdoctoral Science Foundation Project of the Chongqing Natural Science Foundation (CSTB2023NSCQ-BHX0119). Author contributions J.Y. conceived and drafted the original manuscript. J.L. and X.H. critically reviewed and edited the manuscript. Y.Z. and Z.Z. contributed to visualization and manuscript revision. T.Z. and T.J. participated in manuscript review and editing. J.K. and G.Z. contributed to manuscript revision. Y.G. supervised the project, acquired funding, and participated in manuscript review and editing. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work. Data availability No datasets were generated or analysed during the current study. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Jiahui Yang, Jianfeng Li. Contributor Information Jiming Kong, Email: Jiming.Kong@umanitoba.ca. Guohui Zhang, Email: 18931316008@163.com. Ying Guo, Email: guoying6106@gmail.com. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.