Metabolic reprogramming drives immune cell fate
Immune activation is accompanied by metabolic reprogramming, with mitochondria providing energy and regulating redox status during this process18. Taking macrophages as an example, M1 pro-inflammatory macrophages rely on glycolysis for rapid energy supply while interrupting the tricarboxylic acid cycle, leading to accumulation of succinate, a phenomenon known as the Warburg effect (also termed aerobic glycolysis)19. In contrast, M2 anti-inflammatory macrophages depend on intact mitochondrial oxidative phosphorylation (OXPHOS)20. Research indicates that succinate stabilizes hypoxia-inducible factor 1-α (HIF-1α) by inhibiting prolyl hydroxylase (PHD), thereby promoting the transcription of proinflammatory factors such as interleukin (IL)-1β19. Pyruvate kinase M2 (PKM2) serves as a key regulatory factor in this pathway, which forms a complex with HIF-1α, enhancing its binding to the IL-1β promoter while simultaneously promoting IL-10 production, orchestrating macrophage activation, and counteracting inflammatory responses21. Additionally, Rab32-mediated mitochondrial microautophagy directly degrades mitochondria, promoting the shift of macrophages toward glycolytic metabolism, which in turn supports and amplifies inflammatory responses. This process is independent of macroautophagy or the endosomal sorting complexes required for transport (ESCRT) machinery, instead being jointly mediated by the Rab32 GTPase, phosphatidylinositol 3,5-bisphosphate, ubiquitination, and p62/SQSTM122.
Mitochondrial dynamics in immune regulation
The dynamic balance between mitochondrial fission and fusion critically regulates immune cell metabolism, inflammatory signaling, and antigen presentation, thereby shaping innate immunity. Mitochondrial fission is driven by the recruitment of dynamin-related protein 1 (Drp1), a process orchestrated by a coordinated interplay between receptors. Specifically, mitochondrial elongation factor (MIEF)1/2 (MiD51/49) acts as an essential adapters that link Drp1 and mitochondrial fission factor (Mff) into a trimeric complex. MIEFs serve as regulators, where low-to-moderate MIEF levels promote fission while high levels may sequester Drp1 and lead to elongation23. This precise molecular balance directly impacts immune signaling. In macrophages, the recruitment of Drp1, facilitated by adapters such as phosphoglycerate mutase family member 5 (PGAM5), or deubiquitination by ubiquitin-specific peptidase 16 (USP16), is crucial for the subsequent release of IL-1β and TNF-α or NLRP3 inflammasome activation24,25.
Conversely, the fusion machinery, particularly mitochondrial fusion protein mitofusin-2 (MFN2) and optic atrophy 1 (OPA1), is indispensable for maintaining mitochondrial integrity by sustaining TCA cycle flux and OXPHOS with metabolic plasticity26,27. Upon infectious or inflammatory stimulation, MFN2 undergoes a functional switch from a homeostatic stabilizer to a pro-inflammatory mediator, supporting HIF-1α-driven glycolysis26 and is regulated by mitochondrial ubiquitin ligase 1 (Mul1) to maintain essential endoplasmic reticulum (ER)-mitochondrial contacts28. Dynamic fusion–fission cycle of mitochondria, orchestrated by OPA1, is critical for maintaining cristae architecture and efficient OXPHOS. Its deficiency impairs nuclear respiratory factor 1 (NRF1)-dependent OXPHOS, thereby disrupting electron transport chain complex assembly and ATP synthesis. These mitochondrial defects ultimately culminate in lysosomal degradation of major histocompatibility complex class I (MHC-I) and compromised antigen cross-presentation29. Thus, the finely tuned equilibrium of mitochondrial dynamics serves as a central hub, dictating the outcome of innate immune responses.
Mitochondria in antiviral immunity
MAVS-mediated signaling in mitochondria is a core pathway regulating viral immunity17,30. During viral infection, viral nucleic acids, as key components of PAMPs, trigger host resistance by being recognized through cytoplasmic retinoic acid-inducible gene I (RIG-I)-like receptors (RLR) signaling pathways (including RIG-I and melanoma differentiation-associated protein 5 (MDA5) receptors)31. As a result, MAVS on the mitochondrial outer membrane oligomerizes and recruits downstream factors to activate the TANK-binding kinase 1 (TBK1)/inhibitor of nuclear factor kappa-B kinase subunit epsilon (IKKε) complex, leading to interferon regulatory factor 3 (IRF3) phosphorylation and induction of type I interferons (IFN-I)32. However, recent studies reveal a non-canonical pathway: cellular RNA may directly regulate the MAVS signaling complex without relying on RIG-I or MDA5. MAVS interacts directly with the 3’ untranslated regions (3’ UTRs) of cellular mRNAs through its central disordered domain (amino acids 103–467), thereby activating downstream pathways and inducing antiviral activity33. Multiple studies have explored mechanisms regulating viral replication and antiviral immune responses: as orchestrators in cholesterol homeostasis, miRNA-33/33* inhibits mitochondrial autophagy (mitophagy) via an AMP-activated protein kinase (AMPK)-dependent mechanism, thereby suppressing MAVS activation and ultimately reducing RIG-I signaling, which impacts antiviral immunity17. Beyond its role in mitochondrial fission, Mff independently serves as a MAVS activator to initiate antiviral responses; these two functions are molecularly distinct and subject to discrete regulatory mechanisms. Activated AMPK phosphorylates Mff, thereby diminishing its capacity to promote MAVS aggregation and weakening antiviral signaling34. Although antiviral immunity is often discussed in the context of MAVS, DNA viruses can also directly activate cGAS. Recent work has shown that upon herpes simplex virus 1 (HSV-1) infection, nuclear cGAS is released from chromatin into the nuclear soluble fraction, where it directly senses viral DNA and produces cGAMP to restrict viral replication35. Similarly, the HSV-1 ubiquitin ligase ICP0 and the cytomegalovirus IE1 protein have been found to trigger nuclear cGAS activation by perturbing centromere integrity and inducing centromeric DNA amplification36. Thus, direct cGAS activation by viral DNA represents a key, MAVS-independent pathway that complements mitochondrial-centric immune sensing. However, viruses have evolved multiple strategies to disrupt mitochondria-mediated immune pathways, as summarized in Table 1. Disabling this central hub allows diverse viruses to paralyze the host’s innate defense network.
Table 1 Viral evasion mechanisms targeting mitochondrial immunityMitochondria fuel inflammasome activation
Mitochondria precisely regulate inflammasome activation through integrated structural–functional alterations, metabolic reprogramming, and quality control mechanisms. As the most studied inflammasome, the NLRP3 relies on mitochondrial disorders linking structural disintegration and metabolic failure to trigger activation, acting as a prerequisite for sensing cellular stress. Multiple NLRP3 activators (such as nigericin) disrupt mitochondrial cristae architecture and reduce ATP production by inhibiting OXPHOS37. However, contrary to the conventional view that cristae disruption promotes cytochrome c release, a recent study reveals that nigericin does not simply disrupt cristae but induces closure of crista junctions, trapping cytochrome c within the cristae lumen and inhibiting apoptosis. This ATP suppression provides a “mitochondrial signal” that requires a second signal (e.g., NLRP3 aggregation to Endosomes/Golgi induced by imidazoline drugs, or potassium efflux induced by the Piezo-1 channel activator Yoda-1) to activate the inflammasome38. This dual signaling mechanism ensures precise cellular choice between apoptosis and pyroptosis. Notably, mitochondrial damage, as a universal event, also activates NLRP10 and AIM2. NLRP10 directly senses damage and recruits apoptosis-associated speck-like protein containing a CARD (ASC) to form specks that colocalize with mitochondria39, while released mtDNA is recognized by AIM240 (Fig. 1A). Thus, whether a single mtDNA molecule engages distinct sensors simultaneously, or selectivity is determined by cell type, stimulus, and subcellular localization, remains unresolved.
Fig. 1: Mitochondria are the central hub for innate immune regulation.
The alternative text for this image may have been generated using AI.
A Mitochondrial signals (cristae disruption, reduced ATP, or mtDNA) integrate with second signals (K⁺ efflux, NLRP3 aggregation) to activate the NLRP3 inflammasome. B Mitochondrial reactive oxygen species (mtROS) and ox-mtDNA form an amplification loop, linking metabolic stress to sustained inflammation. C To prevent excessive inflammation, quality control mechanisms such as mitochondria autophagy (mitophagy) and the adapter protein containing PH domain, PTB domain, and leucine zipper motif 1 (APPL1)–Rab5 axis restrict mitochondrial stress by clearing damaged mitochondria, thereby negatively regulating inflammasome signaling. ASC apoptosis-associated speck-like protein containing a CARD, OXPHOS Oxidative phosphorylation, mtDNA mitochondrial DNA, DAMP damage-associated molecular pattern, LPS Lipopolysaccharide, RET reverse electron transport, mtROS mitochondrial reactive oxygen species, ox-mtDNA oxidized mitochondrial DNA, mitophagy mitochondrial autophagy, APPL1 adapter protein containing PH domain, PTB domain, and leucine zipper motif 1. Created in BioRender. Xue, Y. (2026) https://BioRender.com/f6i291u.
Unlike AIM2’s direct double-stranded DNA (dsDNA) binding, the precise role of mtDNA in NLRP3 activation remains debated. A recent study shows that wild‑type NLRP3 binds non‑oxidized mtDNA with high affinity (IC₅₀ ~ 4.8 nM), but oxidized mtDNA (ox-mtDNA) is much weaker (IC₅₀ ~ 247 nM), challenging the view that only ox-mtDNA acts as an activating ligand41. Thus, the current debate centers on whether oxidation acts as a critical switch or merely modulates an intrinsic DNA‑binding capacity of NLRP3. Moreover, whether NLRP3 functions redundantly in mtDNA sensing depends on undefined contexts such as adapters, spatial compartmentalization, or signaling thresholds.
DAMPs from mitochondrial dysfunction amplify inflammation. Mitochondrial reactive oxygen species (mtROS) and ox-mtDNA couple metabolic reprogramming to inflammatory responses (Fig. 1B). In proinflammatory macrophages, lipopolysaccharide (LPS)-induced succinate accumulation drives reverse electron transport (RET) at electron transport chain complex I, efficiently generating superoxide that directly regulates IL-1β release during NLRP3 activation42. Notably, NLRP3 activators trigger mitochondrial dysfunction and ROS bursts, leading to mitochondrial membrane damage, inhibition of OXPHOS37, and the generation and release of ox-mtDNA43, thereby creating conditions for inflammatory amplification. Thus, mtROS and ox-mtDNA form a molecular bridge connecting upstream metabolic signals to downstream inflammatory outputs. However, mtROS function exhibits duality: superoxide generated by Complex III is also essential for macrophage anti-inflammatory IL-10 secretion following TLR stimulation44.
To prevent excessive inflammation, mitophagy plays a key negative regulatory role (Fig. 1C). Early endosome-dependent mitophagy mediated by the adapter protein containing PH domain, PTB domain, and leucine zipper motif 1 (APPL1)-Rab5 axis suppresses NLRP3 overactivation by clearing damaged mitochondria, thereby inhibiting ROS and ox-mtDNA accumulation. Hematopoietic cell-specific APPL1 deficiency exacerbates sepsis, obesity-associated inflammation, and glucose metabolism disorders, accompanied by elevated systemic IL-1β levels45. The transcription factor MafB sustains p62 expression to promote mitophagy, specifically inhibiting NLRP3 without affecting NLRP1, NLRC4, or AIM2. Preclinical evidence suggests MafB induction attenuates NLRP3 activity, pointing to upstream transcriptional regulation as a potential strategy46.
An interesting recent finding reveals that during pyroptosis, activated Gasdermin D N-terminal (GSDMD-NT) penetrates both mitochondrial membranes via cardiolipin externalization on the outer mitochondrial membrane, disrupting respiration, triggering ROS bursts, and releasing mtDNA37. This dual-membrane disruption refers to GSDMD-NT forming pores in both the outer and inner mitochondrial membranes, causing release of intermembrane space proteins (e.g., cytochrome c) and matrix contents (e.g., aconitase 2 (ACO2) and mtDNA), distinguishing pyroptosis from apoptosis. Mitochondrial damage serves as an IL-1β secretion checkpoint: blocking it (e.g., via Crls1 or Plscr3 knockout) abolishes IL-1β release even when GSDMD is cleaved, establishing a positive feedback loop37. Nevertheless, why identical mitochondrial signals activate multiple inflammasomes while mtDNA or other DAMPs sometimes show specificity for AIM2 or NLRP3 remains unclear; resolving this will help illuminate innate immune regulatory mechanisms underlying mitochondrial inflammation.
mtDNA leakage and cGAS-STING activation
As a key DAMP, mtDNA leakage translates mitochondrial dysfunction into systemic inflammatory responses by activating the cytosolic DNA-sensing cGAS-STING pathway47. mtDNA leakage is mediated by a stress network composed of multiple mechanisms. The opening of the mitochondrial permeability transition pore (mPTP) represents a direct pathway, disrupting the membrane barrier under pathological conditions such as oxidative stress and facilitating direct mtDNA release48. In radiation-induced brain injury, excessive ROS leads to sustained mPTP opening, triggering massive mtDNA release accompanied by cristae disruption49. In addition to mPTP, the BCL2-associated X protein (BAX) mediates a distinct pathway: under LPS stimulation, BAX translocates to the outer mitochondrial membrane and increases its permeability to promote mtDNA leakage. The mtDNA released into the cytoplasm further activates the cGAS–STING pathway and amplifies inflammatory signaling50.
Crosstalk between PRR pathways via mitochondria
Mitochondrial damage signals, specifically mtDNA and mtROS, act as central coordinators that simultaneously regulate AIM2, NLRP3, cGAS–STING, and Z-DNA binding protein 1 (ZBP1) pathways48,51,52,53,54. Their interconnected crosstalk forms a self-amplifying inflammatory circuit, dramatically escalating the immune response (Fig. 2). TLR9 recognition of mtDNA CpG motifs further amplifies this process by promoting neutrophil extracellular traps (NETs) formation via p38/extracellular signal-regulated kinase (ERK) phosphorylation55. Following dsDNA internalization from NETs, AIM2 binds to pyrin and ZBP1, forming the AIM2-PANoptosome. This simultaneously activates GSDMD, caspase-3, and phosphorylated mixed lineage kinase domain-like protein (p-MLKL), triggering PANoptosis55.
Fig. 2: Mitochondrial damage drives a self-amplifying inflammatory loop.
The alternative text for this image may have been generated using AI.
Green (mtROS/ox-mtDNA Pathway): Damaged mitochondria generate reactive oxygen species (mtROS) and release oxidized mitochondrial DNA (ox-mtDNA) into the cytosol, serving as the core initiation signal for downstream cascades. Yellow (TLR9/p38-NETs Pathway): Endosomal toll-like receptor 9 (TLR9) recognizes mtDNA CpG motifs, triggering p38/extracellular signal-regulated kinase (ERK) phosphorylation to promote the formation and release of neutrophil extracellular traps (NETs). Red (AIM2/PANoptosis Pathway): Cytosolic ox-mtDNA or internalized DNA from NETs activates AIM2, recruiting ASC, Caspase-1, Pyrin, and Z-DNA binding protein 1 (ZBP1) to form the AIM2-PANoptosome, which executes PANoptosis (pyroptosis, apoptosis, and necroptosis). Pink (TXNIP/NLRP3 Pathway): mtROS facilitates the dissociation of thioredoxin-interacting protein (TXNIP) from thioredoxin, enabling its binding to NLRP3, driving inflammasome assembly. Purple (NF-κB Pathway): mtROS induces disulfide bond formation in NF-κB essential modulator (NEMO), activating the inhibitor of nuclear factor kappa-B kinase (IKK) complex and extracellular signal-regulated kinase (ERK)1/2 signaling to drive the nuclear translocation of NF-κB for pro-inflammatory gene transcription. Blue (cGAS-ZBP1/STING Pathway): Cytosolic DNA is sensed by cGAS to produce cGAMP, triggering STING translocation, or the recruitment of ZBP1/receptor-interacting protein kinase (RIPK) complexes to enhance signal transducer and activator of transcription 1 (STAT1) phosphorylation and sustain the type I interferons (IFN-I) response. mtROS mitochondrial reactive oxygen species, ox-mtDNA oxidized mitochondrial DNA, TLR9 toll-like receptor 9, dsDNA double-stranded DNA, mtDNA mitochondrial DNA, NETs neutrophil extracellular traps, ERK extracellular signal-regulated kinase, ZBP1 Z-DNA binding protein 1, GSDMD Gasdermin D, p-MLKL phosphorylated-mixed lineage kinase domain-like protein, TXNIP thioredoxin-interacting protein, NEMO NF-κB essential modulator, IKK inhibitor of nuclear factor kappa-B kinase, ERK extracellular signal-regulated kinase, RIPK receptor-interacting protein kinase, STAT1 signal transducer and activator of transcription 1, IFN-I type I interferons. Created in BioRender. Xue, Y. (2026) https://BioRender.com/80btzqp.
Beyond intrinsic mitochondrial signaling, the integration of mitochondrial distress with the ER stress response represents a critical regulatory layer. The cGAS–STING pathway not only provides the “licensing signal” for inflammation but also acts as a rheostat for ER homeostasis. Upon activation, STING directly binds and activates protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK), leading to the α-subunit of eukaryotic initiation factor 2 (eIF2α) phosphorylation56. This process is independent of STING translocation from the ER to the Golgi and is distinct from the canonical unfolded protein response (UPR), as it does not engage the inositol-requiring enzyme 1 alpha (IRE1α)—the spliced form of X-box-binding protein 1 (XBP1s) or activating transcription factor 6 (ATF6) branches57. Specifically, IRE1α operates as a structural scaffold at mitochondria-associated membranes (MAMs) to govern ER-to-mitochondria Ca²⁺ transfer7, and the activation of its RNase domain promotes pro-inflammatory cytokine production58. This mitochondria–ER axis is further fortified by ZBP1, which forms a DNA-dependent complex with cGAS, recruiting receptor-interacting protein kinase (RIPK)1/3 through ZBP1’s RIP homotypic interaction motif (RHIM) domain to enhance signal transducer and activator of transcription 1 (STAT1) Ser727 phosphorylation and to sustain the IFN-I response54,59. In addition, the mitochondrial–lysosomal axis reveals an indirect mtDNA release mechanism. The mitochondrial E3 enzyme mitochondrial-anchored protein ligase (MAPL) mediates the targeting of mitochondrial-derived vesicles (MDVs) containing mtDNA to lysosomes, where it activates caspase-3/7 to cleave GSDME. The N-terminal domain of GSDME forms pores in the lysosomal membrane, allowing mtDNA to enter the cytoplasm60. This activates cGAS–STING and further promotes NLRP3 inflammasome assembly, ultimately triggering pyroptosis via caspase-1 cleavage of GSDMD60.
Protein-specific modifications by mtROS precisely modulate signaling networks. mtROS induces intermolecular covalent linkage of NF-κB essential modulator (NEMO) through disulfide bonds formed by Cys⁵⁴ and Cys³⁴⁷, which is essential for inhibitor of nuclear factor kappa-B kinase (IKK) complex activation and subsequent signaling through the ERK1/2 and NF-κB pathways that lead to proinflammatory cytokine secretion61. Simultaneously, mtROS promotes the dissociation of thioredoxin-interacting protein (TXNIP) from thioredoxin, enabling TXNIP to bind and activate NLRP3, thereby driving inflammasome assembly and pyroptosis62.
Notably, emerging evidence indicates that mitochondria can initiate cell death through non-canonical, caspase-independent mechanisms. One such pathway is a recently identified form of lytic cell death that is distinct from pyroptosis, PANoptosis, and necroptosis. Under BAX/BCL2 antagonist/killer 1 (BAK1)/BH3 interacting domain death agonist (BID) dependence, mitochondria undergo oxidative stress and maintain prolonged contact with the plasma membrane, triggering localized oxidative damage, a process termed “mitoxyperiosis”, which in turn leads to membrane lysis and cell death, referred to as “mitoxyperilysis”. The cell death pathway is regulated by mechanistic target of rapamycin complex 2 (mTORC2), while inhibition of mTOR restores cytoskeletal activity, allowing mitochondria to detach from the plasma membrane and thereby preserve membrane integrity. In vivo experiments indicate that activating this pathway promotes tumor regression in an mTORC2-dependent manner63.
Mitochondrial-immune interactions in disease and therapy
Abnormal mitochondrial function plays a significant role in various autoimmune and chronic inflammatory diseases. In systemic lupus erythematosus (SLE) (Fig. 3A), mtDNA release is considered a key driver of immune system overactivation64. A study on SLE demonstrated that ox-mtDNA directly interacts with GSDMD-N, promoting its oligomerization and membrane pore formation, thereby amplifying pyroptosis and mtDNA leakage. Specific knockout or inhibition of GSDMD in neutrophils alleviates disease phenotypes, suggesting the ox-mtDNA–GSDMD axis as a critical component of a pathological positive feedback loop65. However, the precise receptor(s) or sensor(s) (e.g., NLRP3, cGAS) through which ox-mtDNA triggers GSDMD activation and downstream signaling are not fully defined. In addition, how ox-mtDNA is generated and released in large quantities in the initial event of SLE is not sufficiently addressed. The primary drivers of the “first wave” of significant mitochondrial oxidative damage and subsequent membrane permeabilization are also not identified. Some interventions targeting mitochondria have been explored in preclinical models. For instance, the cytoskeletal protein nestin was reported to enhance mitophagy and modulate mtROS production, thereby promoting nephrin expression and phosphorylation to protect podocytes from injury, which ameliorated proteinuria in lupus nephritis mouse models66. Whether such strategies can intervene in the ox-mtDNA–GSDMD loop in SLE remains to be tested. Beyond acute pyroptosis, chronic mitochondrial dysfunction may persistently alter immune memory cell differentiation and function, potentially contributing to the relapsing-remitting disease course and long-term autoantibody persistence in autoimmune conditions67,68. Investigating how sustained mitochondrial damage affects the maintenance and recall responses of memory T and B cells could thus yield critical insights into the pathophysiology of immune dysfunction.
Fig. 3: Mitochondria-targeted therapeutic strategies across disease contexts.
The alternative text for this image may have been generated using AI.
A In Systemic lupus erythematosus (SLE), damaged mitochondria release oxidized mtDNA (ox-mtDNA), which binds Gasdermin D N-terminal (GSDMD-N) to drive pyroptosis and further mitochondrial DNA (mtDNA) leakage. Nestin enhances mitochondrial autophagy (mitophagy) and reduces mitochondrial reactive oxygen species (mtROS). B In cerebellar neurodegenerative diseases (CBND), dynamin‑related protein 1 (Drp1)-KO induces cerebellar degeneration; stereotactic transplantation of healthy liver mitochondria restores ATP, membrane potential, and respiration, while reducing excessive PTEN-induced putative kinase 1 (PINK1)/Parkin mitophagy and caspase-3 apoptosis, alleviating ataxia. C In idiopathic inflammatory myopathies (IIM), PN-101 (umbilical cord mesenchymal stem cells (MSC)-derived mitochondria) restores bioenergetics and suppresses inflammation in damaged muscle. D For aging and health, urolithin A (UA), a mitophagy activator, is shown to increase the proportion of natural killer (NK) cells and non‑classical monocytes, enhance phagocytic capacity, and improve muscle strength and aerobic endurance in middle‑aged and elderly individuals. UA acts by boosting mitophagy, thereby counteracting age‑related immune decline and chronic inflammation. E In cancer immunotherapy, multiple mitochondria‑targeting strategies are depicted: mitoNIDs induce autophagic mitochondrial degradation; nanoadjuvants trigger mtDNA leakage to enhance dendritic cells (DC)/CD8⁺ responses; engineered mitochondria activate TLR2 on DC for anti-PD-1 synergy; inhibiting isocitrate dehydrogenase 1 (IDH1) inhibitors sensitize pancreatic cancer. Together, these five panels highlight that context‑specific modulation of mitochondrial damage, mitophagy, and mitochondrial transfer can rewire immune outcomes and disease progression. SLE systemic lupus erythematosus, GSDMD-N Gasdermin D N-terminal, ox‑mtDNA oxidized mitochondrial DNA, mtDNA mitochondrial DNA, mtROS mitochondrial reactive oxygen species, mitophagy mitochondrial autophagy, CBND cerebellar neurodegenerative diseases, Drp1 dynamin‑related protein 1, IIM idiopathic inflammatory myopathies, MSC mesenchymal stem cells, TBK1 TANK-binding kinase 1, IRF3 interferon regulatory factor 3, IFN-I type I interferons, UA urolithin A, NK Natural Killer cells, GNPs gold nanoparticles, DC dendritic cells, CTL cytotoxic T lymphocytes, IDH1 inhibiting isocitrate dehydrogenase 1. Created in BioRender. Xue, Y. (2026) https://BioRender.com/tx9hosu.
Recently, mitochondrial transplantation has shown benefits in certain disease models. Mitochondrial dysfunction has been shown to contribute to cerebellar neurodegenerative diseases (CBND) (Fig. 3B). In a Drp1 knockout-induced cerebellar degeneration model, stereotactic transplantation of healthy liver mitochondria into the cerebellum transiently (approximately 3 weeks) improved cellular mitochondrial membrane potential, ATP synthesis, and respiratory chain complex activity. Concurrently, it downregulated PTEN-induced putative kinase 1 (PINK1)-Parkin-mediated excessive mitophagy and caspase-3-dependent apoptosis, ultimately alleviating ataxia69. In idiopathic inflammatory myopathies (IIM), a group of autoimmune diseases primarily characterized by chronic muscle inflammation and progressive muscle weakness, ox-mtDNA activates the IFN-I pathway to drive disease pathogenesis (Fig. 3C). Mitochondria isolated from human umbilical cord mesenchymal stem cells (MSC) (PN-101) were reported to improve mitochondrial dysfunction and reduce myositis severity in a clinical study70.
Additionally, mitochondria have been suggested as potential targets for modulating age-related immune changes (Fig. 3D). In clinical trials involving middle-aged and elderly individuals, supplementation with urolithin A (UA), a mitophagy activator, was associated with increased proportion of natural killer (NK) cells and non-classical monocytes, as well as enhanced phagocytic capacity71. Furthermore, UA improves immune function by boosting mitophagy, potentially counteracting age-related immune decline, which is characterized by reduced naive T cells and chronic low-grade inflammation71. Another trial reported that long-term UA supplementation correlated with improved muscle strength and aerobic endurance by regulating mitochondrial efficiency and mitophagy, while reducing inflammation72. These findings hint at possible links between mitophagy and immune or muscle function during aging, but causal relationships remain to be established.
In tumor immunotherapy (Fig. 3E), tumor cells with high mtDNA levels exhibit resistance to CD8⁺ T cells and are prone to immune escape. The developed mitoNIDs, a dual-targeting nanosystem based on gold nanoparticles (GNPs), promote proximity between mitochondria and the autophagy-associated protein LC3, thereby inducing mitochondrial degradation via the autophagy pathway73. Nanoadjuvants such as synthetic protein aggregates enhance dendritic cells (DC) maturation and CD8⁺-dependent immunity by increasing mitochondrial membrane permeability to facilitate mtDNA leakage, thereby activating the cGAS–STING pathway, promoting IFN-I production, and ultimately inducing potent antigen-specific cytotoxic T lymphocyte (CTL) responses and tumor regression74. Additionally, engineered mitochondria have been explored as potential vaccine carriers, delivering tumor antigens. In mouse models, their rich content of phosphatidylserine activates the TLR2 pathway on DC, thereby promoting DC maturation, enhancing CTL responses, and synergizing with anti-PD-1 therapy75. Additionally, inhibiting isocitrate dehydrogenase 1 (IDH1) was reported to sensitize pancreatic cancer cells to chemotherapeutic agents by impairing mitochondrial function, as IDH1-derived α-ketoglutarate supports TCA cycle anaplerosis and OXPHOS76. Collectively, these mitochondria-centered strategies, ranging from inducing mitophagy and activating cGAS–STING to serving as vaccine carriers and disrupting metabolic support, highlight the broad therapeutic potential of targeting mitochondrial biology in cancer. While these strategies show promise in preclinical settings, their clinical utility and safety require further validation.