Mitochondrial transfer technologies with molecular insights into clinical applications

Abstract

Mitochondria are essential cell signaling, survival, and bioenergetic organelles that uniquely harbor a maternally inherited, multicopy genome called mitochondrial DNA (mtDNA). The occurrence or accumulation of mtDNA mutations underlies a spectrum of inherited and acquired mitochondrial syndromes and diseases and is increasingly recognized as a source of metabolic plasticity, clonal fitness, and therapy tolerance in cancer. Recent studies have revealed mitochondrial transfer as a potential mode of intercellular communication that could compensate for mtDNA mutation–associated mitochondrial dysfunction. Transfer of mitochondria can restore homeostasis in stressed recipient cells by rebuilding respiratory capacity, rebalancing redox state, and reshaping cell fate. Reported mechanisms of transfer include tunneling nanotubes, extracellular vesicles, cell fusion, and others, such as macropinocytosis. Here, we review and evaluate emerging technologies developed for mitochondrial transfer studies and define the impact of transfer on cell physiology and pathology. We discuss translational opportunities for mitochondrial transfer–based interventions, as well as how mitochondrial exchange may represent a new framework for understanding tumor heterogeneity, adaptation, and aggressiveness.

Graphical Abstract

Graphical Abstract.

Significance statement.

The transfer of mitochondria containing mitochondrial DNA (mtDNA) between cells facilitates intercellular communication and may enable stressed or damaged recipient cells to restore bioenergetic and metabolic homeostasis. This process supports tissue repair but can also be exploited by cancer cells to enhance metabolic flexibility, promote aggressiveness, impede immune responses, and facilitate therapy resistance. Here, we discuss basic mitochondrial biology and mtDNA genetics, evidence for mitochondrial transfer in physiological and pathological settings, and highlight transfer technologies for modeling and measuring transfer in the laboratory. Defining the mechanisms and consequences of mitochondrial acquisition will help translate mitochondrial transfer into new clinical strategies for tissue repair and will clarify how transfer influences tumor heterogeneity and disease progression.

Introduction

Mitochondria are vital organelles in mammalian cells that generate adenosine triphosphate (ATP), produce intermediate metabolites, sustain cell viability, and function as signaling hubs to coordinate cell physiology programs.^1,2 They impact cell lineage fates and stability by controlling the abundance of multiple metabolites, such as acetyl-CoA and succinate that fine-tune chromatin regulating enzymes, histone modifications, and gene accessibility, influencing transcription.^3,4 Mitochondria are unique among mammalian organelles because they contain multiple, often hundreds of copies of a maternally inherited genome, mitochondrial DNA (mtDNA). Each copy of human and mouse mtDNA contains 16 569 and 16 299 base pairs, respectively, in a circular configuration that encode 13 essential subunits of the electron transport chain (ETC), along with 22 tRNAs and 2 rRNAs required for intramitochondrial translation.^5,6 Every cell except mature red blood cells and terminally differentiated central lens fiber cells⁷ contains mitochondria and mtDNA. Environmental damage, inefficient DNA repair, and low-fidelity mtDNA polymerases yield a high mutation rate compared with genomic DNA in the nucleus, which can result in the expression of genetically diverse mtDNA molecules within a single cell.^8,9 Many to most cells harbor a mixture of wild-type and mutant mtDNA molecules, a phenomenon termed heteroplasmy.⁸ Heteroplasmy can arise from point mutations and small or large mtDNA sequence insertions or deletions that affect only a subset of mtDNA molecules within a cell.⁸ These mtDNA variants can lead to hereditary syndromes and disease, with the first definitive linkage between a mtDNA point mutation and a human disease, Leber hereditary optic neuropathy (LHON), providing foundational evidence that mtDNA variants directly cause pathology.¹⁰ Subsequent work established additional mutant mtDNA caused syndromes, including mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS),^11,12 myoclonic epilepsy with ragged-red fibers (MERRF),^13,14 and disorders caused by large mtDNA deletions, such as Kearns–Sayre syndrome (KSS),¹⁵ collectively cementing mitochondrial genetics as a central component of human disease biology.

Importantly, mtDNA variation is not confined to rare inherited disorders. Sequencing approaches demonstrate that mtDNA variants are widespread throughout normal human somatic cells, with many variants present at low heteroplasmy fractions within individual cells.⁹ However, mtDNA mutations can accumulate over time and, under some conditions, undergo clonal expansion, gradually reshaping the respiratory capacity of specific tissues,^16,17 suggesting that mtDNA mutations can contribute to aging phenotypes, an area of active investigation. For example, mice engineered to accumulate elevated levels of somatic mtDNA mutations develop premature aging-like phenotypes with a reduced lifespan.¹⁸ Notably, age-associated mtDNA mutagenesis can occur in a variety of tissues, with dopaminergic neurons of the substantia nigra containing high burdens of mtDNA deletions linked to ETC impairment,¹⁹ intestinal colonic crypts displaying clonally expanded mtDNA mutations in stem-cell-derived lineages,²⁰ and skeletal muscle fibers accumulating high levels of mtDNA deletions.²¹ In skin, chronic ultraviolet light exposure provides an additional organ-specific driver of mtDNA mutagenesis, with sun-exposed sites having increased photoaging-associated deletions including a large, canonical ∼4,977 bp deletion.²² In these varied settings, the phenotypic consequences of mtDNA sequence alterations are frequently dose-dependent because cellular dysfunction often emerges based on the level of cellular energy demand and only after mutant genomes reach a heteroplasmy threshold.⁸

In cancer, large-scale sequencing studies have revealed that somatic mtDNA mutations are relatively common in almost all tumor types and are shaped by mutational biases, selection pressures, and tumor evolution.^23,24 These analyses also support the existence of recurrently mutated sites within the mitochondrial genome, including within regions of mtDNA-encoded rRNAs, suggesting that particular alterations may be favored in specific malignancy contexts.^23,24 Functionally, mtDNA mutations can reprogram tumor metabolism, reshape redox balance, and alter signaling states that influence tumor–microenvironment interactions, thereby affecting therapeutic responses.^2,23 Distinct cancers also display characteristic mitochondrial genetic patterns. For example, Hürthle cell thyroid carcinomas frequently harbor recurrent, often homoplasmic (or all mtDNA molecules in a cell containing the same mutation) mtDNA complex I mutations that arise as an early clonal event coupled to widespread chromosomal losses, suggesting strong selection for specific mitochondrial states.^25–27 Likewise, renal cell carcinoma and renal oncocytoma commonly exhibit disruptive mtDNA mutations affecting respiratory chain genes, particularly complex I, associating mitochondrial genetic remodeling with the oncocytic phenotype.²⁸ In colorectal tumorigenesis, mtDNA mutations that cause defective oxidative phosphorylation (OXPHOS) can accumulate along the adenoma–carcinoma sequence, and experimental systems demonstrate that age-associated mtDNA-caused OXPHOS deficiency can actively accelerate intestinal tumor formation through metabolic remodeling programs.²⁹ Finally, in hematologic malignancies, mitochondrial genetic variation has been systematically catalogued in acute myeloid leukemia and related pre-malignant states, supporting roles for mtDNA mutational burden and heteroplasmy in disease biology and risk stratification.^30,31

In addition to heteroplasmy induced by mutations during mtDNA replication and repair, a rapidly growing body of work is showing that mtDNA heteroplasmy can also be influenced by intercellular mitochondrial transfer, in which intact mitochondria, or mtDNA-containing material, moves between cells to buffer mitochondrial dysfunction or rewire recipient-cell fitness. Mechanistically, mitochondrial transfer is thought to occur through different routes that vary by tissue and context (Figure 1). Tunneling nanotubes (TNTs), thin, actin-rich membranous bridges, enable direct organelle transfer between connected cells and have become a leading model for contact-dependent mitochondrial exchange.³² Studies have reported on the important role of MIRO1 Rho-GTPase, an intracellular mitochondrial transport protein, in facilitating TNT-mediated mitochondrial transfer in various biological contexts.^33,34 Mitochondria or mitochondrial components can also be packaged into extracellular vesicles, providing a cell contact-independent mechanism for mitochondrial cargo exchange.³⁵ In addition, macropinocytosis, a process where cells engulf extracellular contents, is also a mechanism by which cells internalize mitochondria.³⁶ Together, these pathways illustrate that mitochondrial genomes are not strictly cell-autonomous but can be shared and are dynamically influenced by intercellular exchange.

Figure 1.

Pathways of intercellular mitochondrial transfer. Mitochondria can be transferred between cells through: (1) tunneling nanotubes, a direct cell-to-cell contact pathway, (2) macropinocytosis, which involves engulfing free mitochondria from the extracellular space, and (3) extracellular vesicles, where the donor cell delivers mitochondria packaged into vesicles to recipient cells.

Mitochondrial transfer has been shown in the context of regeneration and inflammation, where transfer can restore bioenergetic capacity and reprogram stress responses in damaged recipient cells. In mouse lung injury models, bone-marrow-derived stromal cells transfer mitochondria to injured alveolar epithelium to protect against acute lung injury.³⁷ In addition to direct epithelial rescue, mitochondrial transfer also reshapes innate immunity during tissue injury. In this context, mesenchymal stromal cells can donate mitochondria to pulmonary macrophages, enhancing OXPHOS and phagocytosis, thereby improving antimicrobial defenses in clinically relevant acute respiratory distress syndrome models.^38,39 Mechanistic studies further suggest that mitochondrial transfer is a regulated process rather than passive leakage. For example, MIRO1 promotes mitochondrial movement from mesenchymal stem cells (MSCs) to stressed epithelial cells and improves rescue efficacy in models of airway injury and inflammation.³³ Similar reparative transfer has been observed in the nervous system, where astrocyte-derived mitochondria can be imported by neurons following ischemic stroke, supporting neuronal survival and recovery.⁴⁰ Furthermore, MSCs can donate mitochondria to oxidant-damaged corneal epithelial cells via TNT-like structures, improving mitochondrial function and accelerating repair.⁴¹ In the kidney, systemically delivered bone marrow MSCs can transfer mitochondria to damaged proximal tubular epithelial cells in vivo, restoring mitochondrial structure and function and improving features of diabetic nephropathy–associated injury.⁴² Collectively, these studies support mitochondrial transfer as a broadly conserved injury-response program, positioning mitochondrial transfer as both a physiological repair mechanism and a potential therapeutic target.⁴³ However, while these findings suggest that mitochondrial transfer represents a conserved stress-response and repair program that supports tissue recovery, this process can be exploited by tumors to enhance metabolic plasticity and therapeutic resistance within the tumor microenvironment.^43,44

An important early study in cancer showed that A549 human lung carcinoma cells lacking mtDNA grown in culture recovered respiration by co-culture with either adult non-hematopoietic stem/progenitor cells from human bone marrow stromal cells or with skin fibroblasts containing wild-type mitochondria, suggesting that mitochondria uptake could occur and rescue bioenergetic collapse.⁴⁵ Building on this work, a more recent cancer study showed that a mouse melanoma cell line lacking mtDNA regained respiratory capacity and tumorigenic potential when implanted into a syngeneic mouse by acquiring functional mtDNA from host mouse cells in vivo.⁴⁴ Follow-up work strengthened this model by showing that host-derived rescue can occur by horizontal transfer of whole mitochondria, restoring mitochondrial function and supporting tumor formation.⁴⁶ Additional recent studies have shown that mitochondrial transfer is functionally important in different cancer contexts. Bone marrow stromal cells can donate mitochondria to acute myeloid leukemia cells to enhance bioenergetic capacity and promote chemotherapy resistance.⁴⁷ Mitochondrial transfer can also promote increased tumorigenesis, metabolic adaptation, and drug resistance in glioblastoma cells,^48,49 and carcinoma-associated mesenchymal stromal cells support ovarian cancer heterogeneity and metastatic fitness through mitochondrial donation.⁵⁰ Mitochondrial transfer also promotes metastasis with neuron-to-cancer mitochondrial transfer reported to enhance tumor metastasis.⁵¹ In parallel, mitochondrial transfer has emerged as a direct regulator of anti-tumor immunity. Two recent studies report that tumor–immune cell mitochondrial exchange reshapes T cell functions and therapeutic responses,^52,53 and another study showed that tumor cells can hijack mitochondria from immune cells, reducing the immunogenic capacity of these cells, which can support lymph nodes metastasis.⁵⁴

Altogether, these findings suggest that mitochondrial transfer is a conserved stress-response and repair program that supports tissue recovery, but one that can be used by tumors to enhance metabolic plasticity and therapeutic resistance within the tumor microenvironment.

Mitochondrial transfer technologies

Multiple in vitro methods have been developed to study mitochondrial transfer, which will be discussed in detail in the following section. Methods for mitochondrial transfer, with briefly described advantages and limitations of each technique, are also provided in tabular format (Table 1). Historically, mtDNA could not be manipulated easily or independently of DNA in the cell nucleus.^55,56 Although mtDNA status was suspected to contribute to disease pathogenesis and to have a central role in cellular metabolism, direct studies of mtDNA were limited by the inability to experimentally isolate, manipulate, or control mitochondrial genomes within cells.⁵⁷ As a result, early work was largely correlative, motivating the development of new technologies that physically transfer intact, functional mitochondria to examine mtDNA in a controlled context for both basic cell biology and disease modeling.

Table 1.

Mitochondrial transfer technologies: methods of delivery, advantages, and limitations of each method.

Technology	Method of delivery	Advantages	Limitations
Cell fusion/cybrids11,13,14,58–63	Polyethylene glycol /electrofusion of enucleated cytoplasts with ρ0 recipients	Stable mtDNA replacement; nuclear background control; selectable via uridine deprivation; high heteroplasmy stability after clonal selection	Requires artificial mtDNA depletion; cytoplasmic mixing transfers RNAs/proteins; clonal bottleneck selection; nuclear adaptation over time; low physiological relevance; immortalized backgrounds
Direct co-culture49,50,69,70	Stress-induced tunneling nanotubes; contact-dependent transfer	Physiologically relevant; no artificial manipulation; heteroplasmy stability not well known	Stochastic; no dose control; confounded by paracrine signaling; donor selection bias; unpredictable directionality; difficult quantification of mitochondrial internalization
Co-incubation33,72–76	Endocytosis/macropinocytosis uptake of isolated mitochondria by recipient cells	Simple and scalable; low technical barrier	Predominantly lysosomal routing; limited cytosolic escape; transient metabolic rescue; mtDNA dilution with cell division; ex vivo mitochondrial degradation; contamination misinterpreted as integration; delivery efficiency is lower than other methods; integration of mitochondria/mtDNA not well known
Biolistics87,88	Microprojectile bombardment of mtDNA constructs into yeast	Direct mtDNA genetic manipulation; homologous recombination in yeast	Proof of concept studies performed in yeast; inefficient in mammals; organelle damage risk; limited to DNA (not organelle replacement); low throughput; species restricted
Microinjection89	Direct needle injection into cytoplasm of individual cells	Precise dosing; bypasses endocytosis	Extremely low throughput; mechanical trauma; operator variability; not scalable; limited to large cells; no inherent selection mechanism unless using ρ0 recipients
MitoCeption77	Centrifugation-enhanced uptake	Increased efficiency compared to passive co-incubation; simple and scalable; stable integration of mtDNA in immortalized cells	Still endosomal; centrifugal stress; limited stable heteroplasmy shift; utility not demonstrated in primary cells
Photothermal nanoblade84	Laser-induced transient membrane disruption	Direct cytosolic delivery; reduced lysosomal degradation and bypasses endocytosis; demonstrated stable rescue and remodeling of metabolome	Low throughput; specialized optics; adherent cell restriction; localized membrane stress; high technical barrier; performance is user dependent
BLAST 90	Laser-assisted microfluidic photothermal delivery, operates on similar principal as Photothermal nanoblade but with larger scale delivery	Higher throughput than nanoblade; delivers large cargo	Laser-induced stress; limited long-term heteroplasmy data; instrumentation complexity including specialized optics and device fabrication
MitoPunch80–83	Pressure pulse drives mitochondria through membrane	Scalable; nuclear background control; tunable force to control dosage of delivered mitochondria; stable mtDNA replacement demonstrated; supports clonal isolation; demonstrated in primary cells	Requires artificial mtDNA depletion; Mechanical stress responses; adherent-cell bias; device standardization needed; potential selection for stress-resistant clones; moderate technical barrier
Magnetomitotransfer79	Delivery of TOM22 antibody-coated magnetic beads conjugated with mitochondria	High mitochondria uptake efficiency; compatible with primary cells	Surface modification alters mitochondrial membrane; bead persistence; unclear long-term integration; artificial trafficking pathways
Microfluidics91,92	Mechanical membrane disruption via constriction channels or co-encapsulation of recipient cells with isolated mitochondria	High throughput; works for suspension cells	Shear or surfactant stress resulting in cell damage risk; dosing variability; viability tradeoffs; long-term integration of mtDNA unclear
FluidFM85,86	Atomic force microscope cantilever injects cargo into single cells	Precise single-cell control; direct cytosolic delivery	Extremely low throughput; costly; operator dependent; not scalable; membrane repair variability

Abbreviation: mtDNA, mitochondrial DNA.

Cell fusion

A major advance was stimulated by mitochondrial disorder research in which establishing genetic causality was challenging. Patients can harbor both genomic and mtDNA sequence variants, making it difficult to determine whether mtDNA variants were primary drivers of observed pathologies.⁵⁸ This issue was initially addressed by cell fusion technology and cytoplasmic hybrid (cybrid) cell generation, which manipulated mtDNA in cells by either replacing or mixing entire mitochondrial genomes rather than by direct editing in situ.^55,56 Cell fusion is the basis of hybridoma generation for producing antibodies and typically uses polyethylene glycol (PEG) to cause cell membranes to fuse.⁵⁹ This method is highly reproducible and compatible with multiple cell types, although fusion and mitochondria delivery is stochastic and long-term PEG exposure is cytotoxic.⁵⁷ Additional cell fusion approaches utilized Sendai virus⁶⁰ or electrical impulses,⁶¹ whereas more recent methods employed physical fusion with microfluidic devices.⁶² These fusion approaches circumvented the need for mtDNA delivery or RNA import into mitochondria to engineer the expressed mitochondrial genome.^55–57 Cell fusion experiments in the 1970s produced systems from chloramphenicol resistant and sensitive HeLa carcinoma cells, with mixed nuclear genomes and cytoplasmic components, including mitochondria from both parent cells.⁶³ While these systems enabled the manipulation of mtDNA heteroplasmy in the resulting fused cells, the presence of multiple nuclear genomes and endogenous mitochondria made it difficult to study mtDNA variation in isolation.

This limitation led to the development of transmitochondrial cybrid cells, which allowed mitochondrial genomes to be inserted into cells with isogenic nuclear backgrounds.⁵⁹ In 1989, investigators eliminated mtDNA from a recipient cell to make so-called rho-zero (ρ0), also called rho-null, cells from 143B.TK^- osteosarcoma cell lines which they designated as ρ0101 and ρ0206. They then repopulated ρ0101 and ρ0206 cells by fusion with cytoplasts from HT1080-6TG human cells and selected for fusion products with functional mitochondria so that the mitochondrial genotype could be studied as an independent variable.⁶⁴ These transmitochondrial cybrid cells showed that mtDNA alone could impact respiration and metabolic phenotype. A number of insights were produced from this technological advance by studying the pathogenic mutations underlying MELAS (m.3243A>G),¹² LHON (m.11778G>A, m.3460G>A, or m.14484T>C),⁶⁵ MERFF (m.8344A>G),¹⁴ and KSS (large, spontaneous deletions, mostly occurring between m.8469 and m.13447) disorders.¹⁵ For example, the effect of mtDNA heteroplasmy depended on nuclear background as MELAS cybrids required a ∼90% mutant mtDNA load to impair cytochrome oxidase activity in 143B.TK^- osteosarcoma cells, whereas a ∼55% mutant mtDNA load in A549 lung carcinoma cells led to greater impairment.¹² Overall, cybrid cells revealed that mtDNA sequence variations could affect nuclear gene expression and epigenetics, signaling, cell fate, redox imbalance, heteroplasmy, and phenotypic penetrance, and laid the foundation for studies of mitochondria-nuclear communications.⁵⁹ However, while cybrids established that whole-mitochondria replacement could reprogram a cell phenotype, studies were limited by physiological relevance.⁶⁶ The work relied on fusions with immortalized cell lines, chemical mtDNA depletion methods, and the mixing of cytoplasmic contents during cell fusion, which led researchers to pursue improved methods to manipulate and study mitochondria and mtDNA.⁶⁶

Co-culture and co-incubation techniques: cell-to-cell mitochondria exchange and uptake of isolated mitochondria

In 1982 it was reported that mammalian cells could import and internalize purified mitochondria, transferring encoded antibiotic resistance without the need for transfection reagents or specialized delivery vehicles.⁶⁷ Subsequent work used this study as a starting point for co-incubation-based mitochondrial uptake, although later reports observed that import was cell type and condition dependent.⁶⁸ In 2004, researchers observed that mitochondria could move between mammalian cells through TNTs, and in 2006, investigators showed that mitochondria transfer could occur through physical contact between mammalian cells without cell fusion to restore aerobic respiration in mtDNA-depleted recipient cells.^32,45 Collectively, these studies showed that mitochondrial transfer, either by direct cell-cell contact or through cell uptake mechanisms, is a biological process that could be exploited for therapy or for engineering mtDNA study systems. This led to approaches that utilized endogenous modes of mitochondrial exchange without active mechanical or engineered delivery, and co-culture and co-incubation strategies depending on native cell uptake and exchange mechanisms.

Co-culture techniques leverage the ability of living cells to exchange mitochondria through direct physical contact, often under stress-responsive conditions. Transfer is generally tracked with fluorescently tagged mitochondrial donors and exogenous mitochondria labels, and functional readouts use bioenergetic characterizations of oxygen consumption, ATP generation, membrane potential, and growth measurements.⁶⁹ Co-culture approaches have been used to generate disease models for studying mitochondrial transfer in cancer, wound healing, tissue damage, neurodegeneration, and cardiovascular disease.⁷⁰ In cancer, acquisition of mtDNA by 4T1 mouse breast cancer cells from nerve cells increased aggression and metastatic potential, providing insight into lethal phenotypes observed for innervated tumors.⁵¹ In lung injury and lung disease, epithelial cell mitochondrial dysfunction is a pathological feature. The transfer of mitochondria from manually juxtaposed MSCs into alveolar epithelial cells reduced alveolar destruction and improved lung function in a cigarette smoke-induced lung damage rat model.⁷¹ Contact-dependent transfer was observed between adult human endothelial progenitor cells and GFP-expressing neonatal rat cardiomyocytes, with mitochondria localizing within TNTs under stress conditions.⁷² More recently, co-culture techniques have been used for cell engineering therapies that boost T cell fitness for immunotherapy. T cell exhaustion has been linked to mitochondrial dysfunction and culturing bone marrow stromal cells with CD8⁺ T cells increased the metabolic fitness of T cells, which was demonstrated by improved tumor killing efficiency.⁵²

Co-incubation relies on endocytic mechanisms that enable cells to internalize extracellular mitochondria.⁷³ Co-incubation with isolated mitochondria has been used to study transient rescue of bioenergetic defects in compromised cells, and to interrogate the mechanisms of mitochondrial uptake and transfer.⁷³ Using this technique, investigators showed successful integration of isolated mitochondria into human induced pluripotent stem cell (iPSC)-derived cardiomyocytes and primary cardiac fibroblasts within minutes, with transferred mitochondria often escaping degradation.⁷⁴ Others showed that uptake of isolated mitochondria can occur across species and can restore respiration in human cells lacking mtDNA, with uptake by macropinocytosis demonstrated in MDA-MB-435, COLO 205, and MCF-7 human cancer cell lines.⁷⁵ Moreover, DsRed2-labelled mitochondria isolated from human uterine endometrial stromal cells were shown to be internalized by rat H9c2 cardiomyoblasts, and live-cell microscopy was used to dynamically track intracellular movement and fate.³⁶ Recent co-incubation technology advances focus on increasing mitochondria uptake efficiency with cell-penetrating peptide Pep-1 and coating isolated mitochondria with dextran to mask a negative mitochondrial surface charge.^76,77  In vitro studies ultimately showed that co-incubation can be a physiologically gentle and accessible way to model disease, which has inspired in vivo mitochondria transplantation therapies for ischemic heart injury.⁷⁸

Biophysical mitochondrial transfer technologies

Mitochondrial transfer employing biophysical methods bypasses barriers to mtDNA manipulation by physically introducing whole mitochondria containing mtDNA into recipient cells without the issues associated with cell fusion. Biophysical transfer methods range from uptake-based approaches, such as MitoCeption^79,80 and Magnetomitotransfer,⁸¹ to direct cytosolic deliveries that include MitoPunch,^82–85 photothermal nanoblade,⁸⁶ and FluidFM,^87,88 offering progressively greater control over the amount of cargo delivered and genetic stability. These methods collectively rely on the handling of whole mitochondria and therefore mtDNA populations rather than editing specific nucleotide sequences, providing technological strategies for generating disease-relevant cell systems to study mitochondrial genetics. However, these techniques thus far require pre-existing mtDNA sequences, which could limit applications of biophysical methods of mitochondrial transfer to specific models.

The earliest form of biophysical mitochondrial transfer technology was biolistics, where microparticles coated with material derived from isolated mitochondria were propelled into yeast cells using ballistic force. Biolistics delivered DNA constructs containing mtDNA fragments coated on tungsten microparticles into yeast to repair a mitochondrial gene defect in oxi3, demonstrating that mtDNA could be manipulated, provide functional rescue, and achieve stable genetic transformation.⁸⁹ Others showed that a bacterial plasmid carrying the mitochondrial oxi1 gene could be introduced into ρ0 yeast cells with biolistics, creating synthetic cells that could grow in fermenting conditions.⁹⁰ Biolistics showed that forced delivery of mitochondrial components into cells is feasible, although the technique is rarely used today because of high rates of cellular damage, low efficiency, and the limitation of delivering mitochondrial material rather than whole, intact mitochondria.

Microinjection, an early method for transferring intact mitochondria, employed a hollow glass needle with a micron-sized open tip to directly deliver isolated mitochondria using light pressure into the cytosol of recipient cells.⁹¹ Isolated mitochondria, carrying a chloramphenicol-resistance gene, were injected into chloramphenicol-sensitive human cell lines (143B.TK^- and HT1080-6TG) that were partially mtDNA-depleted using ethidium bromide. Despite a low delivery efficiency, estimated at less than one mitochondrion per cell, delivered mitochondria were successfully retained in recipient cells. Analysis of recipient 143B.TK^- and HT1080-6TG cells with multiple mtDNA and genomic DNA polymorphisms and intracellular selection resulted in near-complete replacement of resident mtDNA by exogenous mtDNA within 6-10 weeks of microinjection.⁹¹ Although less damaging than biolistics, microinjection is low throughput with mitochondria transplanted into one cell at a time, with the persistence of delivered mitochondria in the recipient cells not examined. To minimize cellular damage, needle tip openings at ∼1-3 microns in size could easily clog, further reducing throughput.

MitoCeption^79,80 enhances co-incubation-based mitochondrial transfer and relies on increasing the passive uptake efficiency of isolated mitochondria with centrifugal force. Developed in 2015, MitoCeption is scalable and reproducible and enables the delivery of mitochondria into many cells with the rescue of recipient cell metabolism. This technique relies on the endocytic internalization of isolated mitochondria under isotonic conditions with a thermal shock. The applied centrifugal force aims to increase the frequency and strength of mitochondria-cell membrane contact and not to directly puncture the recipient cell membrane during co-incubation. Centrifugation at lower temperatures appeared to increase the efficiency of mitochondria uptake. MitoCeption was performed on ρ0 MDA-MB-231 human breast cancer, HeLa, HEK293, and fibroblast cell lines and showed metabolic rescue. However, with MitoCeption there is limited control over the amount of mitochondria taken up by cells, and like microinjection, the long-term persistence of donated mitochondria in non-immortalized cells is not fully known.⁸³

To retain the precision of microinjection while seeking to reduce mechanical invasiveness, a photothermal nanoblade (Figure 2A) was developed that uses laser-induced cavitation bubbles to transiently puncture the cell membrane, enabling cytosolic delivery of isolated mitochondria and other large cargo while bypassing endocytosis.⁸⁶ A 3-micron glass capillary microneedle tip is coated with a heat conducting metal, such as titanium, that is positioned to touch a cell membrane using a micromanipulator. Following microneedle positioning, a non-damaging 532 nm wavelength green YAG light laser pulse is applied to the cell membrane area in contact with the metal-coated tip, creating a small delivery portal, with microneedle contents pushed into the opened cell using light pressure. Similar to microinjection, cargo delivery is low throughput, occurring one cell at a time. A range of cargo including mitochondria, genomic DNA, quantum dots, and bacteria were delivered into MDA-MB-453 breast carcinoma and 143B.TK^- ρ0 osteosarcoma cells. Analysis of recipient 143B.TK^- ρ0 cells that received mitochondria showed restored metabolite profiles similar to the wild-type state, with long-term mitochondria retention occurring in ∼2%-4% of recipient cells.⁸⁶

Figure 2.

Schematic illustrations of mitochondria transfer technologies. (A) Photothermal nanoblade. A glass microcapillary pipette tip coated with titanium is placed in contact with the cell membrane. A laser pulse is applied, which triggers a cavitation bubble, resulting in a membrane opening, allowing delivery of mitochondria into the cell. (B) MitoPunch. Cells are plated on a porous PET filter. Donor mitochondria and placed on a PDMS reservoir, and mitochondria and delivered using a mechanically driven applied force.

Extending photothermal nanoblade delivery to population-scale delivery, BLAST (Biophotonic Laser-Assisted Surgery Tool) was developed as a high throughput photothermal delivery platform to simultaneously deliver micron-sized cargo into ∼10 000-100 000 mammalian cells at a time, emphasizing scalability rather than single-cell precision.⁹² In BLAST, cells are cultured on top of a porous silicon chip where the pore walls are coated with titanium. The silicon chip is positioned above a microfluidic chamber containing cargo. A pulsed 532 nm wavelength green YAG laser is scanned across the entire chip to generate cavitation bubbles that induced transient pores in the cell membranes, with light pressure cargo delivery into the cells through the silicon chip pores below. A successful range of large cargo including mitochondria, bacteria, enzymes, antibodies, and nanoparticles were delivered into human peripheral blood monocyte–derived macrophages, primary normal human dermal fibroblasts, human primary renal proximal tubule epithelial cells, and HeLa cells. Cells maintained high viability and delivered cargo remained in the cytosol, bypassing endosomal entrapment. BLAST was not specifically designed for mitochondrial transfer but established as an agnostic technology for the delivery of large cargo using physical, optically driven membrane disruption.

MitoPunch is a pressure-based mitochondrial transfer platform designed for large-scale delivery of isolated mitochondria into numerous adherent recipient cells simultaneously (Figure 2B). In this system, isolated mitochondria suspensions are loaded on a flexible polydimethylsiloxane (PDMS) membrane positioned below a porous, low protein-binding polyethylene terephthalate (PET) filter membrane onto which ρ0 recipient cells are seeded.^82–85 A rapid, controlled pressure pulse administered by a solenoid-actuated plunger deflects the PDMS membrane upwards, forcing the mitochondrial suspension through the pores of the PET membrane and through the plasma membrane of the recipient cells, permitting cytosolic entry independent of endocytosis. The magnitude of the applied force is tunable and may be optimized for different cell types to balance delivery efficiency and cell viability. Following transfer, recipient cells undergo selection in uridine-starved galactose-containing media, enriching for cells that successfully integrate delivered mitochondria and respire. By delivering mitochondria to large numbers of adherent cells, MitoPunch provides a reproducible, population-scale method for creating stable, mitochondrial transfer clones containing wild-type or mutant mtDNA of choice for primary and immortalized cells.

Microfluidic devices enhance mitochondrial delivery by inducing transient pores in the cell membrane through controlled mechanical deformation or by co-encapsulating isolated mitochondria with recipient cells to promote internalization, offering a scalable approach for mitochondrial transfer. Cells are squeezed through constricted microchannels, deforming the plasma membrane beyond its elastic limit. Recipient cells and isolated mitochondria or other cargo are typically co-suspended and flowed through a microchannel as delivery is synchronized with deformation.⁹³ Droplet-based microfluidics systems have also been used to deliver specific amounts of mitochondria into individual cells at high efficiency and throughput by co-encapsulating recipient cells and isolated mitochondria within microdroplets to enable controlled transfer.⁹⁴ In this approach, C2C12 mouse myoblasts receiving mitochondria isolated from C2C12 myoblast donors showed dose-dependent improvement of bioenergetics when more than ∼30 mitochondria were delivered per cell. When these bioenergetically enhanced C2C12 cells were transplanted into C57BL/6 mice following muscle injury, muscle regeneration was significantly improved above baseline repair levels. Notably, this study focused on augmenting mitochondrial content to enhance cellular therapeutic performance rather than achieving mtDNA replacement for cell-engineering or rescuing a respiration-deficient phenotype.

Magnetomitotransfer uses superparamagnetic beads coated with anti-TOM22 (a translocase of the mitochondrial outer membrane complex protein) antibodies, to which isolated mitochondria are bound through TOM22.⁸¹ Application of an external magnetic field pulls the mitochondria-coated beads from a liquid suspension to within proximity or contact of recipient cells, increasing the probability and rate of internalization. Using fluorescently labeled mitochondria, transfer efficiencies were reported in the 78%-92% range, exceeding the 17% efficiency reported by co-incubation alone. Mitochondria-adhered beads were delivered into HeLa cells and primary cells including fibroblasts and cardiomyocytes. Recipient cells showed increased respiration and mitochondrial density, although it was unclear whether the delivered mitochondria were permanently retained and integrated as long-term heteroplasmy over multiple passages was not quantified. Mitochondrial delivery is rapid, and uptake efficiency increased compared to MitoCeption.

FluidFM uses a hollow atomic force microscope cantilever to extract individual mitochondria from a donor cell and inject the extracted mitochondria into a recipient cell using controlled pressure and force.^87,88 This approach is very precise and is compatible with live cell imaging but is single-cell delivery and is therefore costly and has low throughput. Using a hollow cantilever, mitochondria harvested from HeLa cells were transferred into U2OS human osteosarcoma recipient cells fluorescently labeled with su9-BFP, enabling tracking of donor mitochondria relative to the endogenous mitochondrial network in real time.⁸⁷ Further examination showed that transferred mitochondria can fuse with the host mitochondrial network, exchange components, and remain integrated rather than becoming isolated or degraded. This approach provided evidence at organelle resolution that transferred mitochondria can integrate and remodel host mitochondrial networks rather than acting as a transient metabolic additive.

Molecular insights into clinical applications

Intercellular mitochondrial transfer has emerged as a stress-responsive rescue program that cells use to restore homeostasis in neighboring recipients by rewiring bioenergetics and redox balance. Transfer is commonly triggered when recipient cells experience mitochondrial dysfunction, such as oxidative stress, impaired respiration, or mtDNA damage, creating selective pressure for the acquisition of viable mitochondria that can rebuild ATP-generating capacity. While this offers substantial therapeutic potential, this phenomenon is highly context dependent, as the same mechanisms that support tissue repair after injury can also be exploited in pathological settings, including cancer, where mitochondrial acquisition may enhance metabolic flexibility, promote tumor aggressiveness and therapy tolerance, and weaken immune interactions to drive tumor growth. As a result, mitochondrial transfer represents both a therapeutic opportunity and a potential biologic liability, depending on the cellular ecosystem in which it occurs.

Clinically, insights on mitochondrial transfer are converging into two major translational strategies: (1) direct mitochondrial transplantation, in which isolated mitochondria are delivered into damaged tissue and (2) cell-mediated transfer, most commonly using mesenchymal stromal/stem cells as mitochondrial donors. Early human feasibility efforts are underway, with a phase 1 clinical trial (NCT04998357) testing intra-arterial delivery of autologous mitochondria in acute ischemic stroke patients undergoing thrombectomy, examining real-world practicality of delivering mitochondria into ischemic tissue.⁹⁵ Yet critical translational questions remain unresolved. In humans, precise mitochondrial genome editing within injured tissues is not yet practical. So, what quantity of isolated mitochondria should be administered, how frequently, and for how long? An important uncertainty is the permanence of transferred mitochondria, as most studies have assessed the presence of physical mitochondria rather than sustained levels of transferred mtDNA as a source for improved cell fitness. Replacing deleterious mtDNA mutants requires the stable integration of functional mtDNA to propagate continued biosynthesis of functional mitochondria rather than just physical transfer of functional mitochondria. Because mitochondria go through several fusion and fission iterations, the benefit of a healthy mitochondrion may be lost over time. Thus, to gauge the potential for mitochondrial transfer as a clinical application depends on understanding the fate of transferred mitochondria. Transferred mitochondria and its native mtDNA could lead to improved cell fitness if the mtDNA is actively incorporated, replicated, and maintained as a stable component of the recipient’s mitochondrial genome. However, this issue is especially difficult to resolve because standard fluorescent imaging approaches cannot reliably determine whether transferred mtDNA becomes a persistent, functional contributor to the recipient mitochondrial pool, highlighting a current need for rigorous lineage tracing and quantitative genomic approaches to measure long-term mtDNA retention. Conversely, transferred mitochondria could also improve cell fitness through mtDNA-independent processes, such as mitochondrial transfer-driven acquisition of mitochondrial proteins and metabolites from the donor cell, which potentially can more readily affect bioenergetics, redox state, and signaling. However, because of the dynamic nature of mitochondria, the transfer of proteins and metabolites is most likely a short-lived bioenergetic rescue rather than a durable genetic replacement of impaired mtDNA.

In contrast, mitochondrial transfer in cancer may have a particularly profound impact by introducing an additional layer of tumor heterogeneity, one that is mitochondria- or mtDNA-driven rather than facilitated by genomic DNA expression, dynamics, or mutations. Even subtle changes in mtDNA heteroplasmy can shift respiratory capacity, redox handling, and metabolic node usage, thereby altering cellular physiology in specific tumor subclones. In turn, these metabolic changes can propagate into broader remodeling of cell identity through mitochondria-to-nucleus signaling, including effects on chromatin state and epigenetic regulation through metabolite-sensitive enzymes and cofactors. In this framework, mitochondrial transfer becomes more than a survival mechanism and may diversify tumor cell states, accelerate adaptation, and expand the range of therapy-permissive phenotypes within a tumor, improving tumor aggressiveness and facilitating severe sequelae including therapy resistance, tumor recurrence, and metastases.

While the therapeutic potential for exploiting mitochondrial transfer to treat disease is an opportunity that warrants further investigation, key challenges for clinical deployment remain. First, most of our mechanistic understanding of mitochondrial transfer comes from cell culture and animal models, where transfer has been demonstrated in diverse injury and cancer contexts. However, mitochondrial transfer in humans still lacks solid evidence, making human relevance a critical translational question to resolve. Additional considerations include determining whether transferred mtDNA becomes integrated or is transient in recipient cells, which could affect delivery schedules, establishing robust standardization guidelines, optimizing delivery routes and techniques, and managing risk to avoid unintended inflammatory consequences. Together, these considerations emphasize that the clinical future of mitochondrial transfer will depend not only on demonstrating efficacy but also on defining durability, retention or transience, and context-specific safety throughout regenerative, oncologic, and additional physiological and pathological settings.

Author contributions

Vahan Martirosian (Conceptualization, Writing—original draft), Berney Peng (Conceptualization, Writing—original draft), and Michael A. Teitell (Conceptualization, Writing—review & editing)

Funding

V.M. is supported by the California Institute for Regenerative Medicine UCLA Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research Training Program. B.P. is supported by the American Cancer Society Postdoctoral Fellowship PF-24-1318565-01-DMC. M.A.T. is supported by NIH grants P30CA016042, R01GM073981, and R01HL149687, by an Air Force Office of Scientific Research award FA9550-23-1-0737, and by the UCLA David Geffen School of Medicine and UCLA Health.

Conflicts of interest

None declared.