The primary symptoms are neurologic and occur at largely similar frequencies between both nuclear-LSS and mtDNA-LSS forms. Developmental delay and/or regression are core symptoms of LSS. Motor system involvements include hypotonia and movement disorders such as dystonia and chorea, spasticity, and cerebellar ataxia. Additionally, seizures/epilepsy is a common manifestation, and peripheral neuropathy can occur in patients.

Importantly, brain stem dysfunction and lesions may lead to a variety of critical issues. These include respiratory abnormalities such as apnea, central hypoventilation or hyperventilation, and irregular respiration, as well as bulbar problems which often lead to abnormal swallowing (dysphagia)—in up to 80% mtDNA-LSS and 25% in nuclear-LSS, and speech difficulties (dysarthria). Other brain stem-related signs include ophthalmoparesis and abnormalities of thermoregulation.

Neuroimaging is crucial for LSS diagnosis and typically reveals characteristic patterns. The hallmark finding is the presence of bilateral and typically symmetric lesions in the central nervous system. On T2-weighted magnetic resonance imaging (MRI), these lesions appear as hyperintensities (bright signals). On computed tomography (CT), they may be seen as hypodensities. These lesions are most found in the basal ganglia, thalamus, and brain stem. The periaqueductal gray matter, medulla, and spinal cord can also be involved.

While LSS is primarily a neurologic disorder, various other systems can be affected, particularly in mtDNA-LSS. (1) Gastrointestinal manifestations that include poor weight gain and other gastrointestinal issues, (2) Cardiac involvement that may include cardiomyopathy and conduction defects, and (3) Other systems that include hepatic (liver), renal (kidney), and endocrine (hormonal) manifestations.

Mitochondrial gene-encoded Leigh Syndrome Spectrum (mtDNA-LSS) is maternally inherited, for a minority of the patients (about 30%). About 40% of mtDNA-LSS patients have mtDNA pathogenic variants affecting complex V subunit MT-ATP6, with the most common variants being m.8993T$>$G, m.8993T$>$C, and m.9176T$>$C. Other mtDNA-LSS patients are mostly caused by variants in the complex I subunits, encoded by genes MT-ND3, MT-ND5, MT-ND6, MT-ND4, and MT-ND1. There is a total of 15 mtDNA genes reported, but rarely in other mtDNA genes than listed above [9].

Nuclear gene-encoded Leigh Syndrome Spectrum (nuclear-LSS) accounts for about 70% of the cases, caused by over 120 LSS-related genes reported to date. Most of the genes are associated with autosomal recessive inheritance. There are only a few genes associated with LSS in autosomal dominant modes: DNM1L, OPA1, SLC25A4, and SSBP1, which are all de novo pathogenic variants. Interestingly, they are all involved in mitochondrial fission/fusion (DNM1L, OPA1, SLC25A4) and maintenance (SSBP1) [6, 10].

X-linked inheritance is reported for LSS caused by a heterozygous or hemizygous pathogenic variants in genes PDHA1, AIFM1, NDUFA1, and HSD17B10, where PDHA1 is the first discovered nuclear LSS gene [5].

Clinical criteria below for mtDNA-LSS and nuclear-LSS are adapted from the Mito-GCEP expert panel consensus [6].

Typical neuroradiologic findings: Must include the bilateral and typically symmetric hyperintensities on T2-weighted magnetic resonance imaging (MRI) or hypodensities on CT. Hyperintensities or hypodensities must be found in the brain stem and/or basal ganglia with or without bilateral. Alternatively, Hyperintensities or hypodensities must be found in the thalamus, cerebellum, subcortical white matter, and/or spinal cord.

AND at least one of the following characteristic neurologic clinical features: Developmental regression, developmental delay, or neurobehavioral/psychiatric features.

AND at least one of the following biochemical and/or mitochondrial abnormalities: Elevated concentration of lactate and/or glycine in relevant tissues, respiratory chain enzyme activity deficiency, pyruvate dehydrogenase complex in respective tissues, or mitochondrial fission/fusion defect.

For nuclear-LSS, molecular genetic testing approaches include comprehensive whole genome (WGS) or whole exome (WES) tests, PMD-targeted multi-gene panel tests, or single-gene tests. For mtDNA-LSS, molecular genetic testing may be an integral part—either as by-product or specially augmented with mitochondrial genome content—of the comprehensive WGS or WES tests or a pure whole mtDNA genome sequencing.

Molecular testing with LSS coverage is widely available across the world, and the choice of approach may be dependent on clinical suspicion, age of presentation, and availability of genomic testing. A search of NCBI Genetic Test Registry (GTR) website returned 271 tests from 32 labs across the world, using the following keyword filters: Test purpose = “Diagnosis” and Test method = “Sequence analysis of the entire coding region” (https://www.ncbi.nlm.nih.gov/gtr/all/tests/?term=Leigh%20syndrome&filter=testpurpose:diagnosis, accessed on July 22, 2025), many of them are multi-gene panels and some of them are single LSS gene tests. Understandably, WGS and WES can also cover the test for nuclear-LSS and most mtDNA-LSS.

The initial decades of Leigh syndrome (LS) research was comprised mostly of clinical observations and foundational investigations in biochemistry and imaging. LS was first described as a distinct neuropathological entity as “subacute necrotizing encephalomyelopathy”, characterized by symmetrical lesions in the brainstem and basal ganglia. Early research focused on establishing LS as an inborn error of metabolism, with strong evidence pointing towards defects in cellular energy production.

Studies from the 1970s through the 1990s frequently linked LS to elevated lactate and pyruvate levels, suggesting a primary defect in pyruvate metabolism. The X-linked form of Leigh syndrome was inferred from an unexplained male/female ratio in Leigh’s syndrome of 1.83/1 [14]. These led to the identification of Pyruvate Dehydrogenase Complex (PDHC) deficiency as a significant cause, with mutations in the X-linked PDHA1 gene being the first nuclear gene mutation discovered as a cause of LSS [5, 15]. Deficiencies in mitochondrial respiratory chain complexes, particularly Cytochrome c Oxidase (Complex IV), were also reported [16], though the underlying genetic causes remained elusive. Functional studies were limited to biochemical assays on patient tissues, primarily muscle and fibroblasts.

The first mitochondrial DNA (mtDNA) mutation linked to LS, the m.8993T$>$G in the MT-ATP6 gene (Table 1, Supplementary Table 1), was a landmark discovery in 1992 [4]. This finding established the existence of maternally inherited LS and introduced the concept of heteroplasmy, where the percentage of mutant mtDNA correlated with disease severity. The subsequent identification of the milder severity m.8993T$>$C variant further solidified MT-ATP6 genotype-disease correlation [17].

Towards the end of this era, the first nuclear genes encoding respiratory chain components and assembly factors (complex II) were identified for gene SURF1. The study used complementation assays in patient cell lines of rodent/human rho0 hybrids (Table 2) to confirm that this nuclear gene caused LS. This led to the landmark discovery of mutations in SURF1 as a major cause of COX-deficient LS [18]. Of note, Rho-0 ($\rho$0) immortalized cell lines which are devoid of mitochondrial DNA was invented in 1989 [19] and had been primarily used to create cytoplasmic hybrids (cybrids) to study the functional effects of specific mtDNA variants. But Tiranti et al’s work [18] used it to introduce complementing chromosomes without exogeneous mtDNA to discover SURF1. The first reports of mutations in nuclear-encoded Complex I subunits were NDUFS8 [20] and NDUFS7 [21].

Therapeutic approaches during this period were entirely supportive, focusing on managing symptoms and metabolic crises with “mitochondrial cocktails” consisting of vitamins and cofactors such as thiamine, riboflavin, and coenzyme Q10, often with limited and anecdotal success (Table 3).

The potential of sodium dichloroacetate (DCA) to lower lactate levels was explored. The lactic and pyruvic acid concentrations in serum and cerebrospinal fluid (CSF) were decreased by the oral DCA treatment in some patients but not in other patients, thus its clinical efficacy was controversial [22, 23].

The advent of Sanger sequencing technologies dramatically accelerated the pace of gene discovery. This decade saw the identification of numerous novel mutations in both mtDNA and nuclear DNA genes. The list of causative nuclear genes for Complex I deficiency expanded significantly to include NDUFS4, NDUFS3, NDUFS2, NDUFA2, and the assembly factor NDUFAF2 (Table 1). Mutations in other respiratory chain complexes and related pathways were also identified, including SDHA (Complex II), BCS1L (Complex III), SCO2 and COX10 (Complex IV assembly), SUCLA2 and SUCLG1 (Krebs cycle), and PDHB (PDHC).

A key development was the creation of yeast [24] model and the first animal model of mouse with knockout nuclear gene Surf1 [25]—though newer study questioned that this mode did not develop LS, rather displayed prolonged lifespan, which reflects the complexity in creating LS animal model [26] (Table 2, Ref. [19]). Cell-based models, primarily patient-derived fibroblasts, remained crucial for biochemical characterization and for confirming the pathogenicity of newly discovered genetic variants through complementation assays.

Regarding treatment options, this era remained focused on supportive care and the use of mitochondrial cocktails. However, the advancements in molecular testing and genetic mechanism study were leading to the emergence of the first targeted strategies based on specific genetic diagnoses (Table 3). For instance, thiamine and biotin supplementation were shown to be highly effective for biotin-thiamine responsive basal ganglia disease caused by mutations in the thiamine transporter gene SLC19A3 [27].

The widespread adoption of Next-Generation Sequencing (NGS), particularly Whole Exome Sequencing (WES) and large multi-gene panels designed with LSS/PMD associated genes, revolutionized the diagnostic process for LSS. This high-throughput technology enabled the rapid and cost-effective identification of causative mutations in many patients for whom previous single-gene testing had been slow and lack sensitivity. This led to an explosion in the number of novel LSS-associated genes with dozens of new genes identified (Table 1, Supplementary Table 1), including those involved in mitochondrial translation (MRPS34, IARS2, NARS2, TSFM), coenzyme Q10 biosynthesis (COQ2, pDSS2), and novel respiratory chain assembly factors (NDUFAF3, NDUFAF4, NDUFAF5, NDUFAF6, ACAD9). Biotin-responsive Basal Ganglia disease is highlighted as a treatable differential diagnosis of Leigh syndrome [28]. The genetic landscape of LSS was revealed to be far more complex than previously imagined.

This period also saw the maturation of experimental models (Table 2), most notably the Ndufs4^-⁣/- knockout mouse [29], which recapitulated some core features of LS, including progressive neurodegeneration and brain stem lesions. This model became a cornerstone for investigating disease pathophysiology and for preclinical testing of potential therapies. Patient-derived induced pluripotent stem cells (iPSCs) emerged as a powerful tool, allowing researchers to generate disease-relevant cell types, such as neurons and cardiomyocytes, to study disease mechanisms in a human, patient-specific context [30]. Study on Ndufs4^-⁣/- knockout mouse led to the discovery that mTOR inhibition with rapamycin could significantly slow disease progression and extend lifespan [31]. This demonstrated that a mTOR signaling pathway could potentially be targeted for treating LS. Another discovery in this mouse model was that chronic hypoxia could both prevent and reverse established neurodegenerative disease [32].

These breakthroughs spurred early targeted clinical trials for LS (Table 3). The antioxidant EPI-743 (vatiquinone) was tested in a Phase 2a open-label trial, showing signs of neurological improvement in a small cohort of children [33]. These early trials, while small, marked a significant shift from purely supportive care towards incorporating mechanism-based therapeutic development. Of note, no positive results with EPI-743 have been published for Leigh Syndrome since 2012.

Piloting a new path of preventing LS occurrence, spindle-chromosome complex transfer (ST) method was used in a woman carrying the m.8993T$>$G mutation associated with Leigh syndrome. The transfer successfully led to the birth of a healthy child and stayed healthy in 7 months follow up assessment [34].

The current era is characterized by a continued expansion of the genetic landscape of LSS, driven by WES, WGS, and RNA Sequencing. Adoption of the long read sequencing with PacBio and Nanopore technologies can reveal complicated variations that are previously not accessible with the short read NGS sequencing. Novel genes such as FASTKD5, NDUFC2, SQOR, and TMEM126B were shown to be linked to LS (Tables 1,4). The focus of research is now shifting from gene discovery towards understanding the complex cellular consequences of these mutations.

Experimental models have become increasingly sophisticated. Human iPSC technology is being used to create 3D cerebral organoids that model features of LS brain development and pathology, revealing defects in corticogenesis [35]. Live animal models, particularly the Ndufs4^-⁣/- KO mouse, are continuously used to dissect the cell-type-specific contributions to the disease [36] and to explore novel therapeutic avenues, such as acarbose which can extend mice lifespan and suggested involvement of gut microbiome in disease pathogenesis [37].

The most exciting developments are in the therapeutic arena. The advancement in molecular genetics and mechanism study along with the preclinical successes of the previous decade have paved the way for more advanced and targeted strategies (Table 3). Gene therapy is a major new focus, with AAV-based vectors being developed to deliver functional copies of defective nuclear genes. Preclinical studies using AAV9 vectors to deliver SURF1 or NDUFS4 to mouse models have shown significant biochemical and clinical improvements [38, 39]. A double administration of self-complementary AAV9NDUFS4 prevents Leigh syndrome in Ndufs4^-⁣/- mice, leading to 9 months’ heathy lifespan [40]. An optimized mitoBEs single-base-editing could precisely edit and introduce LS-causing mtDNA variant mt-Atp6 T8591C into mouse and observed LS symptoms [41]. Another study used engineered mtDNA adenine base editor (eTd-mtABE), first introducing the pathogenic mtDNA variant into rat zygotes and later correcting the introduced variant at high efficiency—to 53% wild type allele—in a rat Leigh syndrome model [42].

Mitochondrial replacement/donation therapy (MRT) with meiotic spindle transfer or pronuclear transfer has been developed. In such so-called three-person IVF techniques, the nuclear DNA from an oocyte or a zygote (the first two persons) carrying a pathogenic mtDNA variant is transferred into an enucleated donor cell (the third person) that provides wild-type mtDNA background.

In one early human trial, spindle-chromosome complex transfer (ST) method was used on a female carrier of Leigh syndrome (mtDNA mutation 8993T$>$G), with a long history of pregnancy losses and deaths of offspring. The treatment successfully led to the birth of a healthy boy with neonatal mtDNA mutation load of 2.36–9.23%, who stayed healthy at 7 months’ follow up assessment [34].

In two recent back-to-back publications, the studies used mitochondrial donation through pronuclear transfer on 22 patients carrying pathogenic mtDNA variants with homoplasmy or elevated heteroplasmy. There have since been 8 live births from 22 patients (36%), with all infants reported as being healthy at birth, with no or low levels of mtDNA heteroplasmy in blood, being 77% to 100% lower than levels of the maternal pathogenic mtDNA variants [43, 44] (ClinicalTrials.gov number: NCT04113447).

Put together, mitochondrial donation through meiotic spindle transfer or pronuclear transfer was effective in reducing the transmission of homoplasmic and heteroplasmic pathogenic mtDNA variants. MRT is currently only approved in the United Kingdom. Despite regulatory concerns about the probability of mitochondrion—nuclear incompatibility after introduction of mitochondria from a different origin into an embryo and concerns about a low probability of pathogenic variant heteroplasmy reversion, these precision therapies represent a new frontier in the treatment of Leigh syndrome, offering real and proven hope for disease-preventing interventions against mtDNA-LSS.

The mitochondrial transplantation approach delivers healthy mitochondria to affected cells. While still in early stages, preclinical studies have shown its promise in LS mouse models. The recent study by Nakai et al. [45] demonstrated that both bone marrow transplantation (which releases mitochondria into circulation) and direct administration of isolated healthy mitochondria could significantly improve morbidity and mortality in the Ndufs4^-⁣/- LS mouse model. Further, even cross-species administration of human mitochondria to Ndufs4^-⁣/- mice improves LS symptoms [45]. This pre-clinical success provides a strong rationale for exploring this strategy in humans to treat more mtDNA-caused mitochondrial diseases including mtDNA-LSS. Clinical trials for mitochondrial transplantation are underway for more diseases, and the positive preclinical data in LS mouse models make this a compelling research direction.

To date, no FDA-approved small molecule therapies exist for Leigh Syndrome Spectrum (LSS) or other primary mitochondrial diseases [46]. Nevertheless, several promising candidates are emerging, with early evidence drawn from preclinical in vivo studies in animal or organoid models, compassionate use in affected patients, and ongoing clinical trials.

The diagnostic criteria for LSS, the approaches to gene discovery, and individual cases diagnosis practices by different clinicians have evolved over decades, which occasionally leads to inconsistencies across time and between healthcare centers. To standardize the evidence assessment for the LSS gene–disease relationship (GDR), the international ClinGen Mito-GCEP Expert Panel was established in 2017, with over 30 geneticists, neurologists, metabolic physicians, neuropathologists, bioinformaticians, researchers, and laboratory directors from 9 countries. Using a modified ClinGen framework for Gene–Disease Clinical Validity Curation, this multidisciplinary panel systematically evaluated 113 LSS genes for GDR. In the first phase, the panel classified 31 GDRs as “Definitive”, 38 as “Moderate” and 45 as “Limited” or “Disputed” [6]. The renewed funding supported the phase 2 work of expanding to all PMD genes, which continuously updated clinical cases and functional evidence thus leading to GDR upgrading 16 of the 43 “Limited” GDRs to “Definitive” or “Moderate” (Shen L, personal communication, July 2025).

These 114 curated LSS GDRs represent one of the largest resources systematically evaluated by expert consensus and serves as a reference dataset through the Leigh Syndrome Spectrum portal on the Mitochondrial Disease Sequence Data Resource Consortium (MSeqDR) website ([51]; Portal URL: https://mseqdr.org/leigh.php). This GDR resource supports accurate variant interpretation, facilitates genetic diagnosis of LSS, and advances precision medicine efforts.

Because LSS is a rare disease, case-level, genetic, and mutational data are often sparse and dispersed across various database resources. This creates a challenge for the disease community due to a lack of comprehensive and centralized resources. This section describes several web-based, comprehensive resources that aggregate heterogeneous data for LSS genetics and molecular diagnosis.