Mitochondrial Diseases in the Era of Precision Medicine: Advances in Molecular Pathogenesis, Diagnosis, and Therapeutic Intervention

Abstract

Pathogenic mutations in mitochondrial DNA or nuclear genes encoding mitochondrial proteins essential for mitochondrial function give rise to mitochondrial diseases, a highly heterogeneous group of inherited multisystemic disorders. These genetic defects impair the respiratory chain and disrupt oxidative phosphorylation, resulting in critical adenosine triphosphate depletion and tissue-specific energy failure, predominantly affecting high-energy organs such as the brain, heart, and skeletal muscle. Disease pathogenesis is shaped by heteroplasmy thresholds and is further exacerbated by elevated reactive oxygen species, aberrant calcium signaling, and defective mitophagy. Although next-generation sequencing and biochemical profiling have significantly transformed diagnostic frameworks, these progressive metabolic disorders have historically been managed primarily with supportive or palliative care. Driven by precision medicine, the therapeutic paradigm has recently shifted toward genetic and organelle-targeted interventions. Advanced mitochondrial genome-editing platforms, including zinc-finger nucleases, transcription activator-like effector nucleases, and DddA-derived cytosine base editors, enable precise elimination or correction of mutant mitochondrial DNA variants, thereby modulating heteroplasmy. Concurrently, metabolic regulatory agents, mitochondria-targeted peptides, and organelle transfer techniques, encompassing germline mitochondrial replacement therapy to prevent maternal transmission and mitochondrial transplantation for tissue regeneration, are being actively investigated. Despite promising preclinical outcomes, clinical translation remains constrained by challenges in safe delivery, off-target effects, and ethical concerns associated with germline modification. This review bridges the gap between fundamental molecular pathology and cutting-edge therapeutic strategies, highlighting multidisciplinary approaches poised to transform mitochondrial disease management from symptomatic care toward prevention and disease-modifying intervention.

INTRODUCTION

Serving as the primary energy-producing structures within eukaryotic cells, mitochondria are essential organelles. Their evolutionary origin is explained by the endosymbiotic theory, which proposes that mitochondria arose from a symbiotic relationship between a primitive eukaryotic host and a bacterial ancestor that subsequently evolved into a permanently integrated organelle.¹ Beyond energy production, mitochondria participate in a wide range of cellular processes, and disruption of mitochondrial function can impair cellular homeostasis and contribute to diverse human diseases.²

Mitochondrial diseases represent a heterogeneous group of genetic disorders stemming from defects in mitochondrial function, affecting diverse organs due to the organelle’s universal cellular presence.³ These conditions primarily arise from inherited mutations in mitochondrial DNA (mtDNA), including point mutations, deletions, and duplications, that compromise oxidative phosphorylation (OXPHOS) and reduce adenosine triphosphate (ATP) production.⁴ Crucially, clinical expression is dictated by heteroplasmy, where the ratio of mutant to normal mtDNA must reach a specific pathogenic threshold to trigger symptoms, particularly in high-energy tissues like the brain, skeletal muscle, and heart.⁵ Beyond genetic mutations, mitochondrial dysfunction is amplified by elevated reactive oxygen species (ROS), defective mitophagy, and Ca²⁺ homeostasis perturbations, creating a self-reinforcing cycle of cellular damage.⁴ Ultimately, the interplay between nuclear DNA (nDNA) and mtDNA mutations, tissue specificity, and mutation burden determines disease severity. Understanding these complex molecular mechanisms remains the central focus for elucidating pathogenesis and developing targeted interventions for these progressive and potentially fatal metabolic disorders.

Despite major advances, the molecular mechanisms underlying mitochondrial diseases remain incompletely defined, while extensive genetic and phenotypic heterogeneity continues to complicate accurate diagnosis and contribute to diagnostic uncertainty.⁴ Precise identification therefore requires an integrated framework incorporating clinical evaluation, biochemical and histopathological analyses, molecular genetic testing, and functional assessment. Improved understanding of pathogenic mechanisms, including ROS overproduction, impaired mtDNA repair, and disrupted mitochondrial regulatory pathways, remains essential for elucidating disease progression and developing targeted therapies. Current therapeutic strategies integrate gene-based interventions, mitochondrial replacement, metabolic modulation, and pharmacological support to address diverse mitochondrial defects within a multidisciplinary treatment framework.

This review highlights recent progress in mitochondrial diseases, focusing on organelle structure, function, and diagnostic overviews. Additionally, it discusses existing treatment strategies and potential future therapeutic advances, bridging the gap between fundamental pathology and clinical management.

MITOCHONDRIAL BASICS: STRUCTURE AND FUNCTION

Mitochondrial abundance strongly reflects cellular metabolic demand, with high-energy tissues containing more mitochondria. In eukaryotic cells, mitochondria may occupy up to 25% of the cytoplasmic volume, although this proportion varies by cell type. Low-demand cells contain fewer mitochondria, whereas cardiac, skeletal muscle, and liver cells exhibit higher mitochondrial content.⁶ This disparity is most evident in gametes, where oocytes contain 100 000 to 600 000 mitochondria and sperm approximately 100, thereby ensuring maternal mitochondrial inheritance through selective paternal mitochondrial elimination during zygote formation.^6,7

Widely acknowledged as the “powerhouses” of the cell, mitochondria are semiautonomous double-membrane organelles that primarily generate ATP through OXPHOS. Their specialized structure includes an outer membrane, intermembrane space, inner membrane, and matrix (Figure 1).⁸ The outer membrane contains porin channels that facilitate ion and small-molecule transport, whereas the highly selective inner membrane forms cristae that expand the surface area for electron transport chain complexes involved in ATP production. The matrix contains metabolic enzymes involved in the Krebs cycle and fatty acid oxidation, together with mtDNA, ribosomes, and transcriptional machinery.⁸

Figure 1. Mitochondrial Structure, Genome Organization, and Disease-Associated Mutations.

(Top) The schematic illustrates the double-membrane architecture of a mitochondrion, including the outer membrane, inner membrane, intermembrane space, and matrix. The inner membrane forms cristae and contains the electron transport chain: complexes (I to IV), cytochrome c (Cyt c), and ATP synthase, which collectively mediate OXPHOS and ATP generation. The matrix houses mitochondrial ribosomes and multiple copies of circular mtDNA. (Bottom) The circular, double-stranded human mitochondrial genome is shown with common pathogenic mtDNA mutations, highlighting genotype-phenotype correlations associated with diverse mitochondrial diseases. Mutation mapping and nomenclature were based on the MITOMAP database (https://www.mitomap.org/MITOMAP). CPEO, chronic progressive external ophthalmoplegia; KSS, Kearns-Sayre syndrome; LHON, Leber hereditary optic neuropathy; LS, Leigh syndrome; MELAS, mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes; MERRF, myoclonic epilepsy with ragged-red fibers; NARP, neuropathy, ataxia, and retinitis pigmentosa; PS, Pearson syndrome.

A defining characteristic of mitochondria is their unique possession of their own genetic material, mtDNA, a compact circular genome of approximately 16.6 kb composed of heavy and light strands that encode 13 OXPHOS subunits, 22 transfer RNAs, and two ribosomal RNAs essential for mitochondrial protein synthesis (Figure 1).^9,10 Unlike nDNA, mtDNA lacks histone protection and is located near the electron transport chain, rendering it highly vulnerable to ROS, elevated mutation rates, and limited repair capacity.¹⁰ Nevertheless, mitochondrial function remains closely coordinated with the nucleus because most mitochondrial proteins are nuclear encoded and imported into mitochondria following cytosolic synthesis.⁸

Acting as principal regulators of cellular metabolism and signaling, mitochondria generate ATP through OXPHOS by establishing a proton gradient across the inner membrane that drives ATP synthase activity.⁸ Beyond energy production, they maintain intracellular Ca²⁺ homeostasis, regulate ROS generation, and control intrinsic apoptosis through cytochrome c release and caspase activation. Although physiological ROS support cellular signaling, excessive accumulation promotes oxidative stress, molecular damage, impaired ATP synthesis, and Ca²⁺ dysregulation, thereby contributing to disease progression.⁸ Mitochondria also participate in essential biosynthetic pathways, including cholesterol, steroid hormone, heme, and nucleotide metabolism, to preserve cellular homeostasis.

FUNDAMENTALS OF MITOCHONDRIAL DISEASES

Characterized by marked clinical heterogeneity, mitochondrial diseases include inherited multisystemic disorders caused by mutations in either nDNA or mtDNA that impair the respiratory chain, disrupt ATP synthesis, and predominantly affect high-energy tissues, particularly the nervous and muscular systems. The molecular era of mitochondrial medicine began with Luft’s first clinical report in 1962,¹¹ followed by the identification of mtDNA deletions in myopathies¹² and point mutations associated with Leber hereditary optic neuropathy (LHON).¹³ These discoveries led to the characterization of more than 250 pathogenic mtDNA mutations (Figure 1) and numerous nDNA variants. Although mitochondrial diseases represent some of the most common inherited metabolic disorders and can arise at any age, they remain largely incurable, with current treatment strategies primarily focused on symptomatic management and prevention of maternal transmission.

Major Mitochondrial Syndromes and Disorders

Leigh Syndrome

In 1951, Denis Leigh first delineated Leigh syndrome (LS) in a 7-month-old male infant whose death was ascribed to subacute necrotizing encephalomyelopathy, with neuropathology showing bilaterally symmetrical necrotic lesions in the thalamus, brainstem, and posterior spinal cord.¹⁴ LS is a rare, genetically heterogeneous neurodegenerative disorder characterized by ataxia, dystonia, developmental delay or regression, cognitive decline, hypotonia, nystagmus, dysphagia, and respiratory dysfunction. A systematic review reported elevated lactate in blood or cerebrospinal fluid in 72%, followed by developmental delay (57%), hypotonia (42%), respiratory impairment (34%), seizures (33%), feeding difficulties (29%), and muscle weakness (27%).¹⁵ Genetically, LS arises from nDNA (38%) and mtDNA (32%) mutations causing defective OXPHOS. Over 85 genes are implicated, mainly affecting respiratory-chain complexes (~80% of cases). Complex I is most frequent (~35%), with NDUFS4 as a common nuclear gene, while other defects involve complex II (SDHA, SDHB, SDHD, SDHAF1), complex III (UQCRQ, BCS1L) (<8%), and complexes IV and V (15% and 11%), with SURF1 predominant in complex IV deficiency.^16,17 Maternally inherited Leigh syndrome (10%-20% of cases) is associated with MT-ATP6 m.8993T>G, where high heteroplasmy (>90%) causes maternally inherited Leigh syndrome, and ~70% heteroplasmy leads to milder neuropathy, ataxia, and retinitis pigmentosa (NARP) phenotypes, reflecting marked genotype-phenotype variability.¹⁶

Mitochondrial Encephalomyopathy, Lactic Acidosis, and Stroke-Like Episodes

Recognized as a common maternally inherited mitochondrial disorder, mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) is a genetically and clinically heterogeneous disease in which phenotypic severity is strongly influenced by heteroplasmy levels across affected tissues. It was first defined in 1975 by Shapira and colleagues, who reported stroke-like episodes associated with lactic acidemia and ragged-red fibers on muscle biopsy, and the term MELAS was subsequently introduced in 1984 by Pavlakis and coworkers following further case characterization.^18,19 Clinically, MELAS presents with diverse manifestations including cerebellar ataxia, muscle weakness, migraine, epilepsy, diabetes mellitus, cardiomyopathy, visual impairment, and sensorineural hearing loss. More than 25 mitochondrial gene mutations have been associated with MELAS, as reported in MITOMAP (http://www.mitomap.org). The most prevalent cause is the m.3243A>G mutation in MT-TL1, which encodes tRNA^Leu(UUR), identified by Goto et al. (1990), accounting for over 80% of cases and impairing OXPHOS through disruption of transfer RNA (tRNA) structure, stability, methylation, aminoacylation, and codon recognition.^20,21 Additional MT-TL1 mutations (m.3271T>C, m.3252A>G, m.3291T>C) and MT-ND5 variant m.13513G>A have also been implicated in pathogenesis. Further mutations include mt-tRNA^Lys m.8316T>C, mt-tRNA^His m.12147G>A, mt-tRNA^Leu(CUN) m.12299A>C, mt-tRNA^Phe m.583G>A, and mt-tRNA^Val m.1642G>A and m.1640A>G.²¹ Protein-coding gene variants such as MT-CO3 (m.9396G>A, m.9957T>C), MT-CYB m.14864T>C, MT-ND1 (m.3697G>A, m.3946G>A, m.3949T>C), MT-ND5 m.13513G>A, and MT-ND6 m.14453G>A have also been recognized.²² Moreover, nuclear gene defects including POLG, DGUOK, SUCLG2, TRNT1, LOXHD1, KCNQ1, KCNQ2, NEUROD1, and MYH7 are associated with MELAS either independently or in combination with m.3243A>G.²³

Myoclonic Epilepsy With Ragged-Red Fibers

Maternally inherited myoclonic epilepsy with ragged-red fibers (MERRF), first described by Fukuhara et al. in 1980, is a genetically heterogeneous mitochondrial disorder affecting the nervous system and skeletal muscles, characterized by myoclonus, generalized epilepsy, cerebellar ataxia, and ragged-red fibers (RRFs) on muscle biopsy.²⁴ Neurological features include peripheral neuropathy, dementia, deafness, migraine, and optic atrophy, while systemic features include cardiomyopathy, Wolff-Parkinson-White syndrome, muscle weakness, wasting, and lipomatosis. RRFs are observed in over 92% of patients, with muscle biopsies demonstrating reduced cytochrome c oxidase activity and SDH-positive, COX-deficient muscle fibers, reflecting progressive mitochondrial dysfunction.²⁵ At the molecular level, the m.8344A>G mutation in the MT-TK gene accounts for approximately 80% of cases,²⁶ impairing mitochondrial protein synthesis, disrupting respiratory-chain complexes I and IV, and reducing ATP production. At the molecular level, the m.8344A>G mutation in the MT-TK gene, which encodes mt-tRNA^Lys, accounts for ~80% of cases,²⁶ impairing mitochondrial protein synthesis, disrupting respiratory-chain complexes I and IV, and reducing ATP production. Together, m.8344AG and the additional MT-TK variants m.8356T>C, m.8361G>A, and m.8363G>A account for ~90% of cases. Other rare MT-TK variants, such as m.8315A>C, have also been reported. In addition, rare variants in other mitochondrial tRNA genes, including MT-TC (tRNA^Cys, m.5820C>A), MT-TP (tRNA^Pro, m.15967G>A), MT-TI (tRNA^Ile, m.4284A>G), and MT-TT (tRNA^Thr, m.15923A>G), may further contribute to phenotypic variability.²¹

Leber Hereditary Optic Neuropathy

German ophthalmologist Theodor Leber (1871) first clinically characterized LHON by reporting sudden vision loss in young males.²⁷ LHON is a maternally inherited mitochondrial disorder predominantly affecting young males, leading to bilateral subacute visual failure due to retinal ganglion cell degeneration and optic nerve atrophy. Wallace et al. (1988) established its association with MT-ND4 m.11778G>A,¹³ while MITOMAP identifies three primary mutations, namely m.11778G>A (MT-ND4), m.3460G>A (MT-ND1), and m.14484T>C (MT-ND6), accounting for ~95% of cases. Prognosis varies by mutation, with m.11778G>A showing the poorest visual recovery (11.3%).²⁸ LHON pathogenesis is driven by complex I dysfunction, resulting in impaired ATP synthesis, increased ROS, and oxidative stress-mediated retinal ganglion cell loss. Cytoplasmic hybrid (cybrid) studies show heterogeneous complex I impairment, including 79% reduction in m.3460G>A, 20% in m.11778G>A, and minimal dysfunction in m.14484T>C.²⁹ Additional modifiers, including mt-tRNA^Thr m.15950G>A and mt-tRNA^Glu m.14693A>G, further modulate penetrance by disrupting mitochondrial translation, enhancing oxidative stress, and promoting mitophagy-associated apoptosis.^30,31

Chronic Progressive External Ophthalmoplegia

A leading manifestation of mitochondrial myopathies, chronic progressive external ophthalmoplegia (CPEO), is characterized by progressive dysfunction of the extraocular and eyelid muscles, typically presenting in early adulthood with ophthalmoplegia and ptosis, and may be accompanied by dysphagia and sensorineural hearing loss.³² Muscle biopsy frequently reveals RRFs and COX-deficient fibers, reflecting impaired mitochondrial function. CPEO primarily results from defective OXPHOS caused by large-scale mtDNA deletions (~60%), duplications, or multiple deletions associated with nDNA defects involved in mtDNA maintenance.^32,33 Key implicated genes include POLG1, POLG2, TK2, PEO1 (TWNK), OPA1, RRM2B, SLC25A4, DNA2, TYMP, DGUOK, and RNASEH1, with both autosomal dominant and recessive inheritance patterns. Additionally, mt-tRNA variants involving tRNA^Leu, tRNA^Ile, tRNA^Leu(CUN), and tRNA^Lys contribute significantly to mitochondrial dysfunction,²¹ while the novel mt-tRNA^Glu m.14677T>C mutation further expands the genetic spectrum of CPEO.³⁴

Kearns-Sayre Syndrome

Originally delineated in 1958 by Thomas Kearns and George Sayre, Kearns-Sayre syndrome (KSS) is a rare multisystem mitochondrial cytopathy primarily caused by large-scale mtDNA deletions, with occasional point mutations.³⁵ KSS typically presents before 20 years of age with the classic triad of retinitis pigmentosa, external ophthalmoplegia, and cardiac conduction defects, often accompanied by cerebellar ataxia, hearing loss, short stature, elevated cerebrospinal fluid protein, and cognitive impairment. Disease severity is largely influenced by heteroplasmy levels.³⁶ The disorder mainly arises from spontaneous mtDNA rearrangements, including deletions and duplications, which impair OXPHOS by disrupting genes essential for mitochondrial protein synthesis. The common 4977 bp deletion accounts for over one-third of cases, while novel 7652 bp and 5888 bp deletions further expand the mutational spectrum.^36–38 Additionally, pathogenic tRNA mutations involving tRNA^Leu (m.3249G>A and m.3255G>A), tRNA^Leu(CUN) (m.12315G>A), and tRNA^Lys (m.8319A>G) also contribute to KSS phenotypes.²¹

Neuropathy, Ataxia, and Retinitis Pigmentosa

Initially reported by Holt and colleagues in a family of four, NARP is a rare maternally inherited mitochondrial disorder characterized by sensory neuropathy, ataxia, retinitis pigmentosa, seizures, and dementia, typically without RRFs.³⁹ NARP primarily results from heteroplasmic mtDNA mutations, most commonly m.8993T>G in MT-ATP6, which impairs the F_o subunit of ATP synthase complex V and reduces ATP production.^40,41 Disease severity strongly depends on heteroplasmy levels, with 70% to 90% mutant load associated with NARP, and levels exceeding 90% frequently causing maternally inherited Leigh syndrome.¹⁶ Additionally, the tRNA^Cys m.5789T>C mutation has been linked to a NARP-like phenotype through destabilization of the tRNA structure.⁴² Experimental low-frequency electromagnetic field exposure suggests mitohormetic restoration, enhancing complex V activity and ATP production in NARP cybrids.⁴¹

Pearson Syndrome

Pearson and colleagues in 1979 initially outlined Pearson syndrome, a rare, early-onset mitochondrial disorder marked by vacuolated marrow precursors and exocrine pancreatic dysfunction.⁴³ It manifests with lactic acidosis, failure to thrive, pancreatic insufficiency, hypotonia, and multisystem involvement affecting renal, hepatic, and endocrine systems.⁴⁴ With a prevalence of ~1 in 1 000 000 and over 200 reported cases, more than half of patients die before age 6 years, while survivors may later develop KSS-like features.⁴⁴ Pearson syndrome is primarily caused by single large-scale mtDNA deletions, notably the 4977 bp deletion and a 4734 bp variant, which impair OXPHOS and ATP production.^45,46 Clinical heterogeneity reflects uneven mutant mtDNA distribution and tissue-specific heteroplasmy thresholds.

DIAGNOSTIC FRAMEWORK FOR MITOCHONDRIAL DISEASES

Progress in clinical assessment has highlighted that mitochondrial genetic disorders are heterogeneous, complicating diagnosis. Early manifestations include fatigue, reduced exercise tolerance, and muscle weakness, progressing to multisystem neurological involvement such as developmental delay, epilepsy, hypotonia, and visual impairment, with epilepsy often appearing early and showing poor responsiveness to treatment and unfavorable outcomes. Diagnosis requires an integrated approach combining clinical evaluation, neuroimaging, biochemical profiling, and molecular DNA/RNA sequencing, while metabolic screening of blood, urine, and tissues, assessing lactate, pyruvate, amino acids, creatine kinase, glutathione, acylcarnitines, and organic acids, supports early detection.⁴⁷ Biochemical assays further strengthen diagnostic confirmation.

Biochemical signatures reflecting OXPHOS impairment have been used to identify mitochondrial disorders. Fibroblast growth factor 21 (FGF21) and growth differentiation factor 15 (GDF-15) serve as circulating biomarkers, showing elevated levels in mitochondrial dysfunction, particularly in MELAS and MERRF, supporting diagnostic sensitivity.⁴⁸ In parallel, urinary metabolomic profiling using proton nuclear magnetic resonance (¹H NMR) spectroscopy or mass spectrometry offers a noninvasive approach, revealing distinct biosignatures in pediatric cases, including increased tryptophan, homovanillic acid, and Krebs cycle intermediates, with reduced histidine, thereby improving diagnostic specificity.⁴⁹

Next-generation sequencing (NGS) has transformed mitochondrial disease diagnosis by shifting from biopsy-based methods to genome-wide analysis of blood and urine DNA. However, diagnosis remains complex due to the involvement of more than 350 nDNA and mtDNA genes, diverse inheritance patterns, and mtDNA heteroplasmy.⁵⁰ Rapid advances in NGS, particularly whole-exome sequencing (WES), have accelerated the discovery of novel genetic defects. Integrating WES with mitochondrial genome analysis in patients and their families further facilitates the identification of de novo variants and supports a comprehensive assessment of their pathogenicity. Mitochondrial genome NGS enhances sensitivity, heteroplasmy evaluation, and variant identification, thereby confirming primary mitochondrial disease.⁵⁰ Long-read NGS technologies are increasingly applied in mitochondrial disease diagnostics, enabling detection of single large-scale mtDNA deletions, complex rearrangements, and breakpoint structures that may be missed or incompletely characterized by short-read sequencing.^51,52 Frascarelli et al. demonstrated efficient detection of structural mtDNA alterations, supporting long-read sequencing as a unified approach for identifying mitochondrial disease-associated genetic defects.⁵¹ Fearn et al. developed NanoDel, a sensitive long-read pipeline for detecting known and novel large-scale mtDNA deletions, revealing recurrent associations with repeat sequences, G-quadruplex motifs, and the contact zone, suggesting sequence-dependent and structure-dependent mechanisms of mtDNA instability.⁵²

In cases where exome sequencing is inconclusive, RNA sequencing (RNA-seq) and transcriptomic profiling can provide complementary functional evidence by detecting abnormal gene expression, cryptic splicing, allelic imbalance, and other transcript-level defects that may remain undetected by DNA-based analyses.^53,54 These approaches can also aid in resolving variants of uncertain significance and selecting candidate genes. Integrating transcriptomic data with proteomic and metabolomic profiles may further improve variant interpretation and identify convergent molecular signatures, particularly in disease-relevant tissues.^55,56 Therefore, transcriptomic and multi-omics strategies offer valuable secondary diagnostic approaches for unresolved mitochondrial diseases, supporting molecular diagnosis and clarifying underlying disease mechanisms.⁵³

Clinically affected tissues harboring mtDNA mutations are essential for diagnosing mitochondrial diseases. Muscle and skin biopsies allow direct assessment of OXPHOS and electron transport chain activity through biochemical assays, flow cytometry, and respirometry. Noninvasive modalities, including positron emission tomography, magnetic resonance spectroscopy, and circulating mitochondrial proteins such as peroxiredoxin, provide complementary systemic indicators.⁵⁷ Skeletal muscle is particularly informative for evaluating electron transport chain complexes I through IV, especially when blood or buccal samples are inconclusive or mtDNA deletions are suspected.⁵⁸

Specimen selection is critical for accurate assessment of mtDNA heteroplasmy in the diagnosis of mitochondrial disease, particularly m.3243A>G-associated disease, as heteroplasmy levels can vary substantially among tissues. Genetic testing is essential for confirmation, with urinary epithelial cells generally preferred over blood because they exhibit higher heteroplasmy levels and better reflect affected postmitotic tissues.^59,60 Blood heteroplasmy is often lower than that in urine or muscle and declines with age, increasing the risk of false-negative results, particularly in mildly affected or asymptomatic carriers.⁶¹ Urinary sediment therefore provides a practical alternative, with heteroplasmy more closely associated with clinical features than blood measurements.^62,63 Digital polymerase chain reaction enables sensitive quantification in both specimens.⁶⁴

Several criteria-based scoring systems have been developed to standardize the diagnostic assessment of suspected mitochondrial diseases by integrating clinical, biochemical, histopathological, metabolic, and molecular findings. The Walker criteria were among the earliest approaches,⁶⁵ with distinct diagnostic parameters, and were subsequently modified by Bernier et al. for pediatric populations.⁶⁶ The modified Walker criteria place particular emphasis on histopathological and oxidative phosphorylation abnormalities alongside clinical and molecular evidence. Other established approaches include the Nijmegen criteria, originally developed by Wolf and Smeitink and subsequently simplified by Morava et al., and the mitochondrial disease criteria, which facilitate clinical assessment and diagnostic decision-making.^67,68 Although these criteria remain valuable, the increasing availability of genomic testing may reduce reliance on invasive investigations and reshape their application in contemporary mitochondrial disease diagnosis. Recent studies support the Nijmegen criteria as a practical tool for early clinical assessment, patient selection, and interpretation of NGS findings.^69,70 Their emphasis on clinical features facilitates application when biochemical or biopsy data are unavailable, particularly as genetic testing increasingly precedes invasive investigations. Hautakangas et al. further demonstrated that the Nijmegen and modified Walker criteria can identify primary mitochondrial disease despite secondary mitochondrial abnormalities.⁷¹ The modified Walker criteria may provide greater discriminatory value when molecular and biochemical data, including respiratory-chain enzyme deficiencies, are available. In pediatric patients, multisystem phenotypes may result in high Nijmegen scores despite limited biochemical abnormalities, potentially reducing sensitivity when these parameters are excluded.⁷¹

Following molecular diagnosis, reproductive counseling should be tailored to the underlying genetic mechanism and recurrence risk. Prenatal diagnosis and preimplantation genetic testing may be appropriate for Mendelian nuclear gene defects and selected heteroplasmic mtDNA variants, although the mitochondrial bottleneck can complicate the prediction of transmission. Oocyte and mitochondrial donation provide additional options for women at substantial risk of transmitting pathogenic mtDNA variants when conventional approaches are unsuitable.⁷²

THERAPEUTIC STRATEGIES TARGETING MITOCHONDRIAL DYSFUNCTION

Bioactive Compound-Based Approaches

No definitive cure for mitochondrial diseases has been established because of their extensive genetic and phenotypic heterogeneity. Current management primarily relies on symptomatic and supportive approaches, including mitochondrial cocktails containing coenzyme Q10 (CoQ10), vitamins, amino acids, antioxidants, and other cofactors, to enhance electron transport chain activity, improve substrate utilization, and reduce mitochondrial dysfunction. Concurrently, ongoing preclinical and clinical studies are evaluating repurposed agents that may modulate respiratory-chain function or stimulate mitochondrial biogenesis, thereby alleviating symptoms and slowing disease progression. Bioactive compound-based approaches, encompassing supplements, dietary interventions, and pharmacological therapies, are increasingly being explored as potential strategies for managing mitochondrial diseases (Table 1).

Table 1. Supplements, Dietary Interventions, and Pharmacological Therapies for Mitochondrial Diseases.

Agent	Category	Mechanism/primary target	Target disease(s)	Highest phase reached	Regulatory status
Coenzyme Q10 (CoQ10)	Dietary supplement	Essential electron carrier in the mitochondrial respiratory chain; antioxidant that may reduce oxidative stress and support OXPHOS	Primary mitochondrial diseases (PMDs), particularly CoQ10 deficiency	Phase III	Not approved as a disease-specific therapy; widely used as a dietary supplement and commonly employed in clinical management of PMDs
Nicotinamide adenine dinucleotide (NAD+)/NAD+-boosting interventions	Dietary/metabolic intervention	Restores NAD+ balance, supports redox homeostasis, sirtuin signaling, mitochondrial biogenesis, and energy metabolism	PMDs, mitochondrial myopathy, MELAS, PEO	Phase I/II for NAD+-boosting approaches such as nicotinamide riboside	No disease-specific regulatory approval; clinical use remains investigational
Taurine	Dietary supplement	Restores taurine modification of mitochondrial tRNA, improving protein translation in m.3243A>G-associated MELAS	MELAS; prevention of stroke-like episodes	Phase III	Approved in Japan in 2019 to suppress and prevent stroke-like episodes in MELAS, the first mitochondrial disease-specific therapy
L-Arginine	Dietary/metabolic intervention	Precursor for nitric oxide; proposed to improve endothelial function and cerebral vasodilation during stroke-like episodes	MELAS	Phase III clinical studies	No disease-specific regulatory approval; clinically used for MELAS stroke-like episodes, supported mainly by open-label studies
Idebenone	Pharmacological therapy/CoQ10 analogue	Facilitates electron transfer and acts as an antioxidant, potentially bypassing impaired complex I electron transport	LHON; investigated in other mitochondrial disorders	Phase III/IV	Approved in the EU, UK, and several countries for LHON (Raxone); not approved by the FDA
Vatiquinone (EPI-743/PTC743)	Pharmacological therapy	Para-benzoquinone derivative; inhibits 15-lipoxygenase and modulates NQO1-associated redox pathways to reduce oxidative stress	Leigh syndrome, PMDs, FA	Phase II/III	Investigational; no current FDA approval for mitochondrial disease/FA
KL1333	Pharmacological therapy	Activates NQO1; increases the NAD+-to-NADH ratio and promotes mitochondrial biogenesis and oxidative metabolism	PMDs, including MELAS/m.3243A>G-related disease	Phase II	Investigational; no regulatory approval; the FALCON phase II study is ongoing
Elamipretide (SS-31)	Pharmacological therapy	Binds cardiolipin, stabilizes the inner mitochondrial membrane, and improves mitochondrial bioenergetics	Barth syndrome; mitochondrial myopathy; other mitochondrial disorders	Phase III	Approved by the FDA in 2025 (Forzinity) for Barth syndrome in adults and children weighing 30 kg or more; accelerated approval
Sonlicromanol (KH176)	Pharmacological therapy	ROS/redox modulator targeting cellular redox homeostasis	MELAS/m.3243A>G spectrum disorders and other PMDs	Phase III	Investigational; no marketing approval; a phase III registrational study (KHENERFIN) began recruitment in 2026
Omaveloxolone	Pharmacological therapy	Activates the Nrf2 pathway; reduces oxidative stress and inflammation	FA	Phase II/III development	Approved by the FDA in 2023 (Skyclarys) for Friedreich ataxia in patients aged 16 years and older, not PMDs broadly
Pyrimidine nucleosides (doxecitine + doxribtimine)	Pharmacological/metabolic therapy	Replenishes mitochondrial deoxynucleotide pools and supports mtDNA replication and maintenance	TK2 deficiency and other mtDNA depletion syndromes	Phase II; subsequent regulatory approval for TK2 deficiency	Approved by the FDA in 2025 (KYGEVVI) for TK2 deficiency in defined patient populations

Abbreviations: FA, Friedreich ataxia; FDA, US Food and Drug Administration; LHON, Leber hereditary optic neuropathy; MELAS, mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes; NQO1, NAD(P)H:quinone oxidoreductase 1; Nrf2, nuclear factor erythroid 2-related factor 2; PMDs, primary mitochondrial diseases; TK2, thymidine kinase 2.

CoQ10 remains the most widely used compound in the management of disorders linked to mitochondrial dysfunction. As an endogenously produced molecule and a dietary supplement regulated by the US Food and Drug Administration (FDA), CoQ10 is recognized for its potent intracellular antioxidant properties and its role as a mitochondrial electron carrier, although its ability to restore mitochondrial function and provide consistent clinical benefits remains unproven.⁷³ In Alzheimer disease models, oral CoQ10 reduced amyloid plaque accumulation in APP/PS1 transgenic mice,⁷⁴ while a systematic review reported potential cognitive benefits in aged rodents and patients with Alzheimer disease and other dementias, although further clinical validation is required.⁷⁵ In Parkinson disease, CoQ10 exerts neuroprotective effects by reducing ROS, preserving mitochondrial membrane potential, enhancing respiration, inhibiting mitochondria-dependent apoptosis, and promoting mitochondrial biogenesis, thereby improving neuronal function in experimental models.⁷⁴ However, although a phase II trial in Parkinson disease suggested slowed functional decline, a phase III trial failed to demonstrate mitochondrial restoration or clinical efficacy, indicating inconsistent translational outcomes.⁷⁶

A number of studies have advanced understanding of CoQ10 and its therapeutic limitations in mitochondrial disorders. Fuller et al. (2023) demonstrated through molecular dynamics and free energy perturbation simulations that the m.3460G>A mutation in ND1 associated with LHON disrupts the CoQ10 binding site, impairs channel passage, and reduces electron transfer from the N2 iron-sulfur cluster due to an alanine-to-threonine substitution.⁷⁷ Abi Nahed et al. (2025) showed that CoQ10 improves cellular bioenergetics in HaCaT keratinocytes under high-glucose stress, attenuating oxidative damage and implicating AMP-activated protein kinase signaling, while its depletion exacerbates energy stress and cell death.⁷⁸ Despite its established safety and widespread use as a dietary supplement in mitochondrial disease, CoQ10 has demonstrated limited clinical efficacy and is not approved by the FDA for therapeutic use. Consequently, idebenone, a water-soluble CoQ10 analogue, was developed as an alternative electron carrier and membrane antioxidant, demonstrating safety and clinical benefits, receiving European Medicines Agency approval as Raxone for LHON^79,80 and reducing respiratory complications in Duchenne muscular dystrophy.⁸¹

Studies demonstrate that idebenone bypasses complex I deficiency by transferring electrons from the cytoplasm to complex III, restoring ATP production and reducing lipid peroxidation, supporting therapeutic potential in LS, MELAS, and LHON.^82–84 It crosses the blood-brain barrier, reduces superoxide levels, preserves mitochondrial function, and modulates survival and synaptic plasticity pathways.^85,86 Clinically, it improves delayed memory in amnestic mild cognitive impairment, increases superoxide dismutase, and reduces inflammation.⁸⁶ In prion models, it reduces apoptosis and oxidative stress via SIRT3-dependent SOD2 regulation.⁸⁷

Vatiquinone (EPI-743/PTC743) is an orally active vitamin-E–derived compound developed to target oxidative stress and impaired cellular energy regulation.⁸⁸ Identified through glutathione depletion screening, it inhibits ferroptosis by targeting 15-lipoxygenase, thereby reducing lipid peroxidation and preserving neuronal integrity under excessive ROS conditions.^88,89 Vatiquinone also restores redox balance by increasing reduced glutathione, decreasing oxidized glutathione, and lowering ROS levels, which may alleviate symptoms and slow disease progression.^88,89 An open-label, subject-controlled study evaluated vatiquinone in 14 patients with genetically confirmed mitochondrial disease over 13 weeks.⁹⁰ Twelve patients survived, of whom 11 demonstrated clinical improvement, and 10 showed improved quality-of-life scores. Based on clinical, quality-of-life, and brain imaging measures, disease modification was reported in more than 90% of patients. However, these findings were from an uncontrolled study and require confirmation in randomized, placebo-controlled trials.⁹⁰ An LS case carrying the m.10158T>C ND3 mutation also showed improved motor, ocular, and bowel functions following treatment.⁹¹

As a central coenzyme in mitochondrial energy production, nicotinamide adenine dinucleotide (NAD⁺) is frequently depleted in mitochondrial disorders due to electron transport chain dysfunction. NAD⁺ precursors, particularly nicotinamide riboside (NR), have demonstrated the potential to restore mitochondrial function by improving redox balance, enhancing oxidative capacity, and promoting mitochondrial biogenesis.⁹² Acute NR supplementation reduced oxidative stress and improved isometric strength and fatigue resistance in older adults with low antioxidant capacity.⁹³ In animal studies, NR protected against obesity and insulin resistance while enhancing mitochondrial biogenesis and improving gut microbiota composition.^94,95 However, clinical trials in humans demonstrated enhanced NAD⁺ metabolism without significant improvements in adiposity or metabolic health.⁹⁶ Another NAD⁺ modulator, KL1333, functions as a substrate for NAD(P)H:quinone oxidoreductase 1, increasing intracellular NAD⁺ via NADH oxidation.⁹⁷ In MELAS fibroblasts, it activated SIRT1, AMPK, and PGC-1α, promoting mitochondrial biogenesis and ATP production while reducing lactate and ROS.⁹⁷ In phase 1a/1b trials, it showed safety and tolerability with dose-dependent gastrointestinal effects and improved fatigue, strength, and endurance.⁹⁸

Designed to target mitochondrial dysfunction, elamipretide (SS-31) is a mitochondria-penetrating tetrapeptide that selectively binds cardiolipin, thereby preserving cristae integrity, reducing oxidative stress, enhancing ATP production, and improving mitochondrial function.⁹⁹ Preclinical studies have demonstrated that elamipretide preserves respiratory complex activity in cardiac ischemia-reperfusion injury¹⁰⁰ and restores mitochondrial integrity, mitophagy, and respiratory efficiency in Barth syndrome models with tafazzin deficiency.¹⁰¹ Clinically, the TAZPOWER trial reported sustained improvements in walking distance, fatigue, cardiac function, and monolysocardiolipin-to-cardiolipin ratio in patients with Barth syndrome over 168 weeks.¹⁰² These findings supported the subsequent accelerated approval of elamipretide (Forzinity) by the FDA on September 19, 2025, to improve muscle strength in adult and pediatric patients with Barth syndrome weighing 30 kg or more, making it the first FDA-approved treatment for this rare X-linked mitochondrial disorder.¹⁰³ Although the ReCLAIM-2 trial in dry age-related macular degeneration failed to meet its primary end points, subgroup analyses revealed delayed ellipsoid zone loss and improved visual acuity.¹⁰⁴ Additionally, MMPOWER-3 analyses demonstrated improved walking ability in patients with primary mitochondrial myopathy, particularly those with CPEO.¹⁰⁵

Sonlicromanol (KH176), a chromanol piperidine oral candidate, improves cellular redox status by activating the thioredoxin/peroxiredoxin system, inhibiting microsomal prostaglandin E1 synthase, reducing ROS and lipid peroxidation, and protecting OXPHOS-deficient cells.¹⁰⁶ In neurons derived from MELAS-induced pluripotent stem cells (iPSCs) carrying the m.3243A>G mutation, it improved neuronal network activity and partially corrected transcriptomic defects,¹⁰⁷ while phase 2a/2b trials confirmed safety and improvements in cognition, mood, fatigue, and physical function.^106,108 Approved by the FDA in 2023 for Friedreich ataxia, omaveloxolone is an orally bioavailable triterpenoid that activates NRF2 signaling through KEAP1 inhibition, thereby reducing oxidative stress, suppressing inflammation, and enhancing mitochondrial biogenesis,¹⁰⁹ while clinical studies in patients aged 16 years and older have demonstrated sustained neurological improvement, good tolerability, and potential enhancement of exercise capacity via reduced heart rate and lactate accumulation, although multiagent approaches may still be required.^110,111

Taurine is a nonproteinogenic β-amino acid that serves as a mutation-mechanism–directed therapy for m.3243A>G-associated MELAS by restoring taurine modification at the wobble position of mt-tRNA^Leu(UUR). In a multicenter, open-label phase III trial, high-dose taurine significantly reduced the recurrence of stroke-like episodes,¹¹² contributing to its approval in Japan in February 2019 for the prevention of stroke-like episodes and subsequent reimbursement. L-Arginine is also well established in Japan for the acute and interictal management of stroke-like episodes, supporting the use of both therapies alongside antioxidant and supportive interventions.¹¹³

Among mitochondrial DNA depletion and deletion syndromes, pathogenic autosomal recessive mutations in the thymidine kinase 2 (TK2) gene, which encodes the mitochondrial matrix protein TK2, cause TK2 deficiency (TK2d), an ultrarare disorder that disrupts mitochondrial deoxynucleotide supply and mtDNA replication.¹¹⁴ Predominantly affecting skeletal muscle, TK2d causes progressive myopathy, respiratory failure, feeding difficulties, and premature death, with severity varying by age of onset.¹¹⁵ Pyrimidine nucleoside supplementation addresses the nucleotide imbalance associated with TK2d by restoring the availability of substrates required for mtDNA maintenance. Doxecitine and doxribtimine (KYGEVVI), a combination of two pyrimidine nucleosides, were approved by the FDA on November 3, 2025, for the treatment of TK2d in adult and pediatric patients with symptom onset at or before 12 years of age, representing the first approved pharmacological therapy for this disorder.¹¹⁶

Although drug treatments for mitochondrial diseases can help relieve symptoms and delay progression, their therapeutic impact is generally limited to targeting faulty cellular pathways rather than the genetic mutations that cause the condition. Consequently, therapeutic effectiveness can differ between patients, and sustained, lifelong treatment is typically necessary to preserve clinical benefits.

Gene Therapy and Genome-Editing Approaches

Conventional therapies have remained limited in mitochondrial diseases driven by multigene mutations and mtDNA-associated dysfunction, as they fail to address the primary genetic defects underlying pathogenesis. Gene therapy has therefore emerged as a promising alternative, directly targeting and correcting these genomic abnormalities. These strategies aim to deliver functional copies of mitochondrial-related genes into affected cells, shifting the focus from downstream signaling disturbances to primary genetic restoration, while genome-editing platforms enable precise targeting of mutant mtDNA, facilitating heteroplasmy shifting, reducing mutant load, and restoring wild-type mtDNA levels. The following sections provide an overview of gene therapy and genome-editing strategies developed to correct mtDNA mutations and regulate mitochondrial gene expression (Figure 2).

Figure 2. Molecular Tools Engineered to Target and Modify the Mitochondrial Genome.

Five major approaches (A through E) are designed to address mtDNA mutations and modulate mitochondrial gene expression. (A) AAV is employed to deliver a nuclear-encoded gene containing an MTS. Following transcription in the nucleus and translation in the cytosol, the allotopically expressed protein is recognized by the mitochondrial import machinery and subsequently translocated into the mitochondrial matrix via the TOM and TIM23 translocase complexes. (B) MitoTALENs are mitochondria-targeted nucleases that function as heterodimers, consisting of two TALEN monomers bound to closely spaced DNA sequences. This binding induces dimerization of the FokI nuclease domains, resulting in the cleavage of mutant mtDNA. (C) MitoZFNs are chimeric enzymes composed of a Cys2His2 ZFP fused to an MTS, the catalytic FokI domain, and NES peptides. These components ensure exclusive mitochondrial localization, allowing the selective cleavage of mutant mtDNA. (D) CRISPR-Cas9 nucleases are targeted to the mitochondria together with unmodified conserved sgRNA structural elements, including one tetraloop and three stem-loops, which are required for Cas9-sgRNA complex formation, thereby mediating mtDNA cleavage. (E) DdCBEs catalyze C•G-to-T•A base conversions in mtDNA with high specificity without inducing DSBs. They consist of split DddAtox halves fused to MTS-linked TALE proteins and a UGI. The UGI protects uracil from glycosylase-mediated removal during subsequent mtDNA replication, ultimately resulting in the replacement of uracil with thymine and the corresponding substitution of guanine on the complementary strand with adenine. AAV, adeno-associated virus; DdCBE, DddA-derived cytosine base editors; DddAtox, interbacterial deaminase-like toxin domain; MTS, mitochondrial-targeting sequence; NES, nuclear export signal; sgRNA, single-guide RNA; TALEN, transcription activator-like effector nucleases; UGI, uracil glycosylase inhibitor; WT, wild-type; ZFN, zinc-finger nucleases; ZFP, zinc-finger protein.

Adeno-Associated Virus-Mediated Allotopic Gene Therapy

Allotopic gene therapy has emerged as a promising strategy for mitochondrial disorders such as LHON, whereby recoded mtDNA transgenes are expressed from the nucleus and imported into mitochondria through mitochondrial-targeting sequences (MTSs) using adeno-associated virus (AAV) vectors with a ~4.7 kb carrying capacity.¹¹⁷ Preclinical studies using MTS-modified AAV vectors to deliver wild-type MT-ND1 and ND4 genes have restored complex I activity, enhanced ATP production, and preserved retinal ganglion cells and axonal integrity in mouse models carrying the m.3460G>A and m.11778G>A mutations.^118,119 Furthermore, the rAAV2/2-based therapy lenadogene nolparvovec has improved visual acuity in LHON, with evidence of contralateral ND4 RNA transfer mediated by connexin-43–dependent astrocyte networks.¹²⁰ A meta-analysis further showed a gradient of visual recovery, highest with GS010, intermediate with idebenone, and lowest in natural history cohorts.¹²¹ In the RESCUE trial, a single intravitreal injection failed to meet the primary end point at week 48 but showed delayed improvement at week 96 when administered early,¹²² whereas the REFLECT study demonstrated that bilateral administration significantly improved visual outcomes with a favorable safety profile, supporting the enhanced efficacy of bilateral gene delivery.¹²³

Mitochondria-Targeted Transcription Activator-Like Effector Nucleases

Originating from plant-pathogenic Xanthomonas bacteria, transcription activator-like effectors (TALEs) have been engineered into transcription activator-like effector nucleases (TALENs) by fusing TALE-derived DNA-binding domains with FokI nuclease domains to facilitate highly specific genome editing via targeted double-strand breaks.¹²⁴ Their modular 34-amino acid repeats, defined by repeat-variable diresidues, confer single-base recognition. Mitochondria-targeted TALENs (mitoTALENs), incorporating MTSs, support matrix import and selective cleavage of mutant mtDNA, promoting its degradation due to limited repair capacity and thereby reducing pathogenic mtDNA while restoring cellular function.¹²⁵

Hashimoto et al. (2015) developed mitoTALENs to selectively cleave pathogenic mtDNA variants, including m.8344A>G (MERRF) and m.13513G>A (MELAS/LS), reducing mutant load and restoring OXPHOS in cybrids.¹²⁶ Shortened mitoTALENs improved viral delivery efficiency, while mitoTev-TALEs incorporating I-TevI nuclease overcame size constraints and shifted heteroplasmy toward wild type in m.8344A>G models.¹²⁷ In vivo and germline studies demonstrated elimination of mutant mtDNA in mice and oocytes carrying m.14459G>A (LHON-dystonia) and m.9176T>C (NARP), thereby preventing transmission.¹²⁸ In m.3243A>G MELAS-iPSCs, editing removed mutations without off-target effects and restored mitochondrial function, while mRNA injection into porcine oocytes further reduced mutation burden, highlighting germline prevention potential.¹²⁹

Bacman et al. (2018) demonstrated that AAV9-delivered mitoTALENs reduced heteroplasmic m.5024C>T in tRNA^Ala mice via intramuscular, intravenous, or intraperitoneal administration, decreasing mutant mtDNA in heart and skeletal muscle, restoring tRNAAla expression, and reversing disease phenotypes.¹²⁵ They further showed that AAV-PHP.eB under a synapsin promoter facilitated neuronal targeting with widespread transduction and mutation load reduction.¹³⁰ In parallel, researchers developed optimized mtDNA-targeted platinum TALENs (mpTALENs) in patient-derived m.3243A>G-iPSCs, improving specificity via engineered ncRVD-LK/WK or NM variants and heterodimeric FokI domains, thereby allowing precise mutation load modulation without impairing differentiation.¹³¹

Mitochondria-Targeted Zinc-Finger Nucleases

Engineered as chimeric nucleases, mitochondrial zinc-finger nucleases (mtZFNs) consist of Cys2His2 zinc-finger DNA-binding domains fused to the FokI catalytic domain, along with MTSs and nuclear export signals to ensure mitochondrial localization.¹³² By binding opposite DNA strands at defined triplet sequences, they enable FokI dimerization and site-specific mtDNA cleavage. Early heterodimeric mtZFNs, however, lacked full specificity, often degrading both mutant and wild-type mtDNA and inducing mtDNA depletion. To improve selectivity, single-chain mtZFNs with dual FokI domains were developed, enhancing discrimination of pathogenic point mutations, although limited efficacy against large deletions and residual off-target activity prompted further redesign.¹³³

Gammage et al. (2014) demonstrated that mtZFNs efficiently eliminate pathogenic mtDNA variants, including the m.8993T>G mutation and the 4977-bp deletion, thereby reducing mutant load, promoting wild-type repopulation, and restoring respiratory function in cybrid models, supporting targeted heteroplasmy correction as a strategy to ameliorate mitochondrial disease phenotypes.¹³⁴ In mouse models, AAV-mediated delivery targeting the m.5024C>T mutation selectively removed mutant genomes in cardiac tissue, shifted heteroplasmy, and restored molecular and physiological function.¹³⁵ More recently, a tandem-linked mtZFN design packaged into a single AAV vector achieved efficient heteroplasmy modification at lower doses with improved efficacy and translational potential.¹³⁶

ZFNs and TALENs can modify mitochondrial heteroplasmy but are limited by large size, heterodimeric architecture, and reduced specificity. To address these constraints, the homing endonuclease I-CreI was engineered into compact monomeric ARCUS nucleases (1092 bp/40 kDa) that recognize palindromic 22-bp sequences and discriminate single-base variants.¹³⁷ AAV9-delivered mitoARCUS targeting m.5024C>T eliminated mutant mtDNA in tissues and restored mt-tRNA^Ala levels,¹³⁷ while systemic delivery in an m.3243A>G xenograft removed mutant genomes, preserved wild-type mtDNA, and improved respiratory function.¹³⁸

Clustered Regularly Interspaced Short Palindromic Repeats–Associated Protein 9

Although the clustered regularly interspaced short palindromic repeats (CRISPR)–associated protein 9 (Cas9) (CRISPR-Cas9) system has been widely used for nuclear genome editing, its application in mtDNA modification remains controversial due to major technical constraints. A mito-Cas9 platform incorporating single-guide RNA (sgRNA), MTSs, and regulatory elements from nuclear-encoded mitochondrial genes was developed to promote mitochondrial import of Cas9.¹³⁹ Proteinase and DNase protection assays confirmed mitochondrial localization and exogenous single-stranded DNA oligonucleotide (ssODN) integration into mtDNA. However, sgRNA delivery across the mitochondrial double membrane remains inefficient, limiting editing efficiency to 0.03% to 0.23%.¹⁴⁰

Researchers have advanced mito-CRISPR-Cas9 systems across multiple models, including human cells and zebrafish, which selectively target mtDNA, reduce copy number, enable precise knock-in, and produce heritable F1 modifications with increased nDNA repair factor expression.¹⁴¹ In yeast and Chlamydomonas, edit plasmids enable stable organelle genome integration via organelle-specific Cas9, guide RNAs, and donor DNA.¹⁴² CRISPR-Cas9 also corrected the c.1334G>A OPA1 mutation in autosomal dominant optic atrophy–iPSCs, restoring mitochondrial function,¹⁴³ and TYMP correction in mitochondrial neurogastrointestinal encephalomyopathy mouse models reestablished mitochondrial homeostasis and thymidine phosphorylase activity.¹⁴⁴ A novel mitochondrial gene-editing strategy has been established using an ND4-targeting guide RNA fused to an RNA transport-derived stem loop (RP-loop) and mitochondria-localized Cas9, enabling mtDNA targeting, reducing ND4 expression, and decreasing mtDNA levels in cells carrying the m.11205A>G variant.¹⁴⁵ This approach demonstrated sgRNA mitochondrial access for CRISPR-Cas9 cleavage. Subsequently, CRISPR-Cas9 ribonucleoprotein delivery via ribonucleoprotein-MITO-Porter crossed mitochondrial membranes and corrected the m.7778G>T mt-Atp8 mutation in mouse-derived cells.¹⁴⁶ More recently, RP11-46H11.3-derived 30-nt RNA MTSs facilitated sgRNA import, and RMTS-CRISPR achieved up to 26% heteroplasmic shifting,¹³⁹ supporting mtDNA correction.

DddA-Derived Cytosine Base Editors

Despite their therapeutic potential, mtDNA editing strategies, including mito-CRISPR-Cas9, mtZFNs, and mitoTALENs, have remained limited by technical challenges. To overcome these constraints, a CRISPR-independent Burkholderia cenocepacia mitochondrial base-editing platform derived from the bacterial cytidine deaminase toxin DddA of Burkholderia cenocepacia enabled cytidine modification in double-stranded DNA without inducing double-strand breaks.¹⁴⁷ This led to RNA-free DddA-derived cytosine base editors (DdCBEs), where split DddAtox fragments were fused with transcription activator-like effector arrays, mitochondrial-targeting sequences, and uracil glycosylase inhibitor (UGI) to achieve precise C•G-to-T•A conversion in human mtDNA. DdCBEs further demonstrated efficient mtDNA editing in embryos and animal models,^148–150 while engineered monomeric variants achieved up to 50% editing, near-homoplasmic conversion via AAV delivery, and reduced off-target activity in human cells.¹⁵¹

In mouse embryos, customized DdCBEs efficiently edited the mitochondrial MT-ND5 gene, introducing pathogenic m.12918G>A and nonsense m.12336C>T mutations, thereby generating mitochondrial disease models.¹⁵² To enhance precision, DdCBE nuclear export signals reduced nuclear off-target activity, while mitoTALENs selectively eliminated unmodified genomes, producing m.12918G>A mice exhibiting abnormal posture, mitochondrial dysfunction, hippocampal atrophy, and shortened lifespan.¹⁵³ Wei et al. (2024) improved DdCBE systems by fusing DddA11 with Xenopus laevis activation-induced cytidine deaminase to enhance C-to-T editing and expand targeting; DddA6 coupled with TALE-linked deoxyadenosine deaminase enabled A-to-G conversion, while fidelity mutations and nuclear export signals further reduced off-target effects in human cells and embryos.¹⁵⁴ Structure-guided engineering of DddAtox produced K1402D/E variants, achieving approximately 400-fold reduction in off-target edits while retaining robust on-target activity across mitochondrial and nuclear genomes.¹⁵⁵ In vivo, AAV9-delivered DdCBEs corrected m.5024C>T via compensatory m.5081G>A editing, restoring tRNA^Ala structure, increasing mt-tRNA^Ala levels to 78%, and reducing lactate accumulation, although high-dose AAV9 induced toxicity due to off-target mtDNA effects.¹⁵⁶ Separately, lipid nanoparticle-delivered DdCBE-modified RNA corrected m.4291T>C in patient fibroblasts, restoring mitochondrial membrane potential with limited recovery of ATP production and OXPHOS function.¹⁵⁷

Recent advances in DdCBE technology have significantly improved both precision and functional versatility for mitochondrial and nuclear genome editing. Zinc-finger–based editors (ZF-DdCBEs) and zinc-finger deaminases, which integrate zinc-finger arrays with split DddA and UGI, enable precise C-to-T substitutions without insertions or deletions in human cells.¹⁵⁸ Enhanced ZF-DdCBEs with optimized scaffolds and activity-modifying mutations have further increased editing efficiency and specificity while minimizing off-target effects, and single AAV9 vector delivery has successfully introduced disease-associated mutations in mouse tissues.¹⁵⁹ To overcome sequence motif constraints, directed evolution approaches (PACE and PANCE) produced improved DddA variants, particularly DddA6 and DddA11, which increased TC editing by ~4.3-fold and expanded targeting to AC and CC motifs, thereby broadening all-protein DdCBE applications for diverse C-G to T-A conversions.¹⁶⁰ Functionally, DdCBEs have been applied to introduce premature stop codons in mtDNA genes such as Atp8 and ND1, enabling efficient disruption of OXPHOS and controlled in vivo studies via Cre/loxP regulation.¹⁶¹ In addition, mitoBEs, which operate independently of DddA and combine TALEs with nickases and TadA8e or APOBEC1-UGI, achieved up to 77% editing efficiency and partially corrected the LHON m.11778G>A mutation (~20%), restoring ATP production and cellular respiration.¹⁶²

Mitochondrial Organelle Transfer Therapies

Researchers have extensively investigated mitochondrial replacement strategies to restore normal mitochondrial function by transferring nuclear genetic material into enucleated donor oocytes containing healthy mitochondria. These approaches aim to prevent the maternal transmission of pathogenic mtDNA mutations responsible for severe mitochondrial disorders while also addressing ovarian aging and female infertility associated with impaired oocyte quality.¹⁶³ Mitochondrial replacement strategies encompass two major modalities: mitochondrial replacement therapy (MRT) and mitochondrial transplantation therapy (MTT), with MRT recognized as the principal approach for preventing inherited mtDNA mutations. MRT, including pronuclear transfer (PNT), maternal spindle transfer, and polar body transfer (PBT), replaces defective oocyte mitochondria during in vitro fertilization to generate embryos containing nDNA from both parents and healthy donor-derived mtDNA, a process termed three-parent technology. MRT, also known as mitochondrial donation, is regulated differently across countries. In the United Kingdom, MRT is permitted under the Human Fertilisation and Embryology (Mitochondrial Donation) Regulations 2015 and overseen by the Human Fertilisation and Embryology Authority, with treatment authorized on a case-by-case basis for eligible women at high risk of transmitting serious mitochondrial disease.¹⁶⁴ Australia introduced a similar regulatory framework through the Mitochondrial Donation Law Reform (Maeve’s Law) Act 2022, which established a staged approach through a licensed pilot program.¹⁶⁴ In contrast, MRT research remains restricted in the United States because federal appropriations legislation prevents the FDA from considering applications for such clinical research.¹⁶⁴ These differences reflect ongoing ethical, safety, and regulatory concerns regarding heritable mitochondrial genetic modification. In contrast, autologous approaches, including autologous germline mitochondrial energy transfer (AUGMENT), have been explored. However, randomized clinical trials have demonstrated no significant advantage over conventional in vitro fertilization, raising concerns regarding their clinical effectiveness.¹⁶³

Approved in the United Kingdom in 2015 for the prevention of mitochondrial disorders, PNT replaces the mitochondrial genome by transferring the parental pronucleus from zygotes with mutated mtDNA into enucleated zygotes containing healthy mitochondria.¹⁶³ Early studies demonstrated that PNT results in minimal mtDNA carryover (<2%) while preserving normal blastocyst development.¹⁶⁵ However, even low heteroplasmy levels may facilitate mtDNA drift and reversion independent of mitochondrial function.¹⁶⁶ Notably, induced mitophagy during MRT reduces donor mtDNA persistence and enhances embryonic and offspring outcomes.¹⁶⁷ In 2024, healthy cynomolgus monkeys generated through PNT exhibited stable heteroplasmy, with 3.8% to 6.7% mtDNA carryover, providing evidence of the potential of PNT to prevent transmission of inherited mtDNA diseases.¹⁶⁸ Importantly, clinical evidence for PNT in humans has emerged from the Newcastle group. In two New England Journal of Medicine reports published in July 2025, 22 women commenced or completed PNT, resulting in eight live births. All eight infants were healthy at birth, had undetectable to low levels of maternal pathogenic mtDNA heteroplasmy in blood (undetectable to 16%), and showed normal developmental progress at the time of reporting.^169,170 These findings provide clinical evidence that PNT can substantially reduce the transmission of pathogenic mtDNA variants and support its use as a risk-reduction strategy for selected women at high risk of transmitting mitochondrial disease, while continued long-term monitoring remains necessary.

Maternal spindle transfer replaces mutated mtDNA by transferring the metaphase II spindle-chromosome complex from a patient’s oocyte into an enucleated donor oocyte before fertilization. First demonstrated in rhesus monkeys, it produced healthy offspring with spindle-derived nDNA, donor mitochondria, and less than 3% maternal mtDNA carryover, indicating prevention potential.¹⁷¹ In human oocytes, spindle transfer replaced mtDNA and generated normal stem cells despite occasional fertilization abnormalities.¹⁷² Spindle-chromosome transfer between unfertilized oocytes prevented disease transmission with less than 1% mtDNA carryover, preserving genomic integrity.¹⁷³ Okamoto et al. (2023) reported aggregated chromosome transfer using 3-isobutyl-1-methylxanthine–induced clustering, reducing carryover and improving fertilization and embryonic development.¹⁷⁴

During oogenesis, meiotic divisions produce polar bodies with minimal cytoplasm and few mitochondria, thereby limiting the transmission of mutated mtDNA. PBT exploits this feature by transferring either the first polar body from a metaphase II oocyte into an enucleated donor oocyte before fertilization or the second polar body from a patient-derived zygote into a donor zygote lacking the female pronucleus, thereby allowing mitochondrial replacement with minimal mtDNA carryover. PBT offers advantages over PNT and maternal spindle transfer. In murine models, Wang et al. (2014) showed that PBT effectively prevented mtDNA variant transmission and supported normal offspring development.¹⁷⁵ An optimized second PBT approach utilizing spindle-protrusion retention allowed earlier transfer, thereby reducing DNA damage and mitochondrial carryover, and produced normal-karyotype blastocysts, embryonic stem cells, and healthy live-born mice.¹⁷⁶

MTT is an emerging regenerative strategy that delivers isolated mitochondria into injured tissues to restore cellular bioenergetic function by enhancing ATP production, reducing oxidative stress, and improving Ca²⁺ homeostasis, thereby supporting tissue repair across diverse disease contexts and demonstrating promising preclinical outcomes.¹⁷⁷ Intercellular MTT, particularly mediated by mesenchymal stem cells, has been shown to restore aerobic respiration and promote functional recovery in pulmonary, cardiac, renal, corneal, and neurological disorders, emphasizing its broad clinical potential while necessitating mechanistic elucidation to support future translation.¹⁷⁸ However, its clinical application remains constrained by challenges in isolation protocols, donor selection, delivery routes, dosing strategies, postisolation mitochondrial viability loss, immune activation, and the absence of reliable long-term storage methods, while delivery specificity still requires optimization. Despite sustained research efforts, its precise mechanisms of action remain incompletely understood.

Preclinical and translational studies have demonstrated therapeutic benefits across multiple disease models. In neurodegenerative and neurovascular conditions, MTT has shown protective effects, including limited but notable improvements in spinal cord injury and enhanced sensorimotor recovery in traumatic brain injury via human umbilical cord mesenchymal stem cell–derived mitochondria, as well as beneficial outcomes in stroke models.¹⁷⁹ In a cardiac arrest resuscitation model, MTT improved survival and neurological recovery in rats by restoring metabolic parameters, enhancing cerebral perfusion, reducing pulmonary injury, and supporting systemic recovery, thereby reinforcing its therapeutic potential.¹⁸⁰ Similarly, in rat traumatic brain injury models, transplantation of umbilical cord mesenchymal stem cell–derived mitochondria enhanced sensorimotor function by restoring mitochondrial content, reducing neuronal apoptosis, attenuating astrogliosis and microglial activation, and preserving brain cytoarchitecture.¹⁸¹ In Purkinje-cell–specific Drp1 knockout mice, MTT alleviated cerebellar ataxia by restoring mitochondrial function, reducing mitophagy, delaying apoptosis, and mitigating neurodegeneration for up to 3 weeks.¹⁸² In another study, umbilical cord mesenchymal stem cell–derived mitochondrial product PN-101 improved mitochondrial function, enhanced myogenesis, and reduced inflammation in idiopathic inflammatory myopathy models, while also demonstrating safety and clinical improvement in patients with refractory disease.¹⁸³

CONCLUSION

Mitochondrial diseases represent a highly heterogeneous group of inherited disorders caused by dysfunction in mitochondrial biology, predominantly arising from mutations in mtDNA and nuclear genes. Their clinical presentation varies markedly across newborns and older patients and involves multiple organ systems, driven by heteroplasmy, nuclear-mitochondrial cross-talk, oxidative stress, and impaired mitochondrial quality control. These factors collectively complicate genotype-phenotype correlations, early diagnosis, and therapeutic stratification. Consequently, clinical management has long remained largely supportive or palliative.

Advances in molecular genetics and high-throughput sequencing have substantially improved the identification of pathogenic variants and the elucidation of disease mechanisms. Research in mitochondrial diseases is gaining considerable momentum, and the therapeutic landscape is rapidly evolving, with advances in targeted molecular therapies, metabolic interventions, mitochondrial replacement strategies, and precision-medicine approaches offering increasing opportunities for disease-specific treatment.

Progress in precision medicine has introduced promising therapeutic avenues, including mitochondrial-targeted gene-editing tools (such as nucleases and systems), mitochondrial replacement techniques aimed at preventing maternal transmission, and emerging metabolic and pharmacological strategies. Despite encouraging preclinical and early clinical findings, their translation into routine clinical practice remains constrained by challenges in safe and efficient mitochondrial or genetic delivery, off-target effects, tissue specificity, and the marked clinical heterogeneity of patient populations. Ethical concerns, particularly regarding germline modification and long-term safety, further necessitate rigorous regulatory oversight.

Future precision-medicine approaches to mitochondrial disorders should also consider the specific molecular mechanisms underlying altered mitochondrial dynamics rather than relying primarily on morphological phenotypes. Similar phenotypes, including mitochondrial fragmentation, elongation, or disrupted networks, may result from distinct abnormalities in organelle tethering, fusion, fission, intracellular trafficking, or mitochondrial quality-control processes, with different functional and therapeutic implications.^184,185 Integrating genetic information with functional assessment of mitochondrial dynamics, bioenergetics, trafficking, and quality-control pathways may improve mechanistic classification and support the selection of mutation-specific therapeutic strategies.^186,187 Such an integrated approach could enhance therapeutic decision-making and facilitate the development of targeted interventions based on the underlying mechanisms of mitochondrial dysfunction.

Overall, continued interdisciplinary advances in mitochondrial biology, diagnostics, and translational therapeutics are essential. A personalized and integrated approach, supported by robust clinical validation and international collaboration, is required to transform current experimental strategies into safe and effective therapies, ultimately shifting mitochondrial disease management from symptomatic care toward disease modification and prevention.