Mitochondrial encephalomyopathies that manifest in early infancy with severe, progressive neurological impairment include subacute necrotizing encephalomyelopathy, or Leigh syndrome, described by Professor Denis Archibald Leigh in 1951 following a postmortem examination of a 7-month-old boy with progressive neurological symptoms [1].

Leigh syndrome is a rare hereditary disorder, with an incidence of approximately 1 in 34,000-36,000 newborns, it affects both boys and girls equally. It is an extremely heterogeneous disease, caused by defects in multiple genes located on autosomes, the X chromosome, and mitochondrial DNA. It is known that the most common biochemical defect in this syndrome is a disruption of the amino acid sequence of proteins involved in the assembly of the subunits of complex IV of the electron transport chain (ETC) of the inner mitochondrial membrane — cytochrome c oxidase (COX), which catalyzes the transfer of electrons in the ETC from cytochrome c to molecular oxygen and formation of water, thus, plays a key role in mitochondrial oxidative phosphorylation.

The most common mutations are in the gene encoding the SURF1 (Sea Urchin Retroposon Family 1) protein, which is embedded in the inner mitochondrial membrane and plays a crucial role in the assembly of the COX complex [2]. Normally, the 35 kDa precursor of SURF1 is imported into the mitochondria, where its mature form (30 kDa) is formed after processing. However virtually all mutations (more than 60) in the SURF1 gene result in protein biosynthesis that is already shortened at the precursor stage and damage to the COX complex, destabilized by defective SURF1. About 50% of mutations in the SURF1 gene are missense, nonsense, and small deletions, with the most common being a 10-nucleotide deletion with a 2-nucleotide insertion (311-321del10insAT) in exon 4 and a 2-nucleotide deletion (845delCT) in exon 9 of the gene [3, 4].

Although it has been established that SURF1-associated Leigh syndrome has a typical course, leading to early mortality before the age of 10 in most cases, approximately 10% of cases exhibit milder symptoms with a longer lifespan [5]. In all cases, the onset of the disease occurs in infancy and is characterized by impaired feeding associated with the onset of damage to the central nervous system, hypotonic and poor coordination of sucking and swallowing, followed by further regression of already established psychomotor development [6]. According to a number of researchers, gastroparesis and delayed gastric emptying, as well as gastroesophageal reflux, in case of Leigh syndrome caused by mutations in the SURF1 gene and most likely due to neuromuscular and mitochondrial dysfunction caused by disturbances in the ETC with cytochrome oxidase deficiency [7]. Typical manifestations of the developing disease also include such disorders as neuropathy, ataxia, ophthalmoplegia and hypertrichosis [8]. To investigate the molecular mechanisms and potential therapies for Leigh syndrome, mouse models deficient in the SURF1(–/–) gene were created. These mice exhibit significant reduction in birth weight, which recovers by the age of one month, with manifestation of moderate motor developmental delay against the background of defective assembly and deficiency of a number of critical COX subunits, including mitochondria-encoded MT-CO1 [9]. Despite the described impairments, these mice have a lifespan similar to the wild type, and it has been suggested that the loss of a COX assembly factor such as Surf1 in mice leads to compensatory responses, including activation of mitochondrial biogenesis and the Nrf2-dependent signaling cascade, which contribute to the alleviation of the manifestations of the pathological phenotype [10]. Furthermore, in this model, AAV serotype 9 (AAV9) adenoviral vector therapy for SURF1associated Leigh syndrome, in which mice were injected with a codon-optimized human SURF1 gene (hSURF1opt) (AAV9/ hSURF1v1), was shown to restore not only SURF1 function but also complex IV activity and correct exercise-induced lactic acidosis without any toxicity to wild-type mice [11, 12].

Here, we used a similar approach with aim to explore the potential of gene therapy for Leigh syndrome in a model of subcutaneous fibroblasts from patients with a SURF1 gene mutation transfected with an adenoviral construct containing a codon-optimized human SURF1 gene.

METHODS

The study included three patients with a genetically confirmed diagnosis of Leigh syndrome and typical clinical manifestations. The inclusion criteria for the study were: 1) a genetically confirmed diagnosis of Leigh syndrome with frequently occurring mutations in the SURF1 gene (tab. 1); 2) age of disease manifestation up to 3 years; 3) informed consent from the patient’s legal representative; exclusion criteria were the presence of another genetically determined disease, including a syndromic form of the pathology. Two patients were born fullterm, in satisfactory condition at birth (Apgar score 8/9 points), with normal weight and length for gestational age; one patient was born at 30 weeks. All patients exhibited psychomotor developmental delays from the first months of life. Over time, their condition progressively worsened, with loss of motor and speech skills. The first two patients experienced metabolic crises — acetonemic states characterized by vomiting, food refusal, weakness, and lethargy. Severe bulbar dysfunction necessitated gastrostomy placement in the first and second patients, a tracheostomy in the third patient, and the first patient required respiratory support (NIVL). All patients were diagnosed after 1.5 years of age. MRI data revealed changes in brain tissue characteristic of the disease: the first patient, aged 2 years, showed MRI signs of bilateral symmetrical structural changes in the substantia nigra, red nuclei, along the conduction pathways in the medulla oblongata and in the middle cerebellar peduncles. The second patient (aged 14 months) showed bilateral symmetrical increases in signal intensity in the basal ganglia, corticospinal tracts and brainstem structures. The third patient (aged 12 months) showed increased signal intensity in the lower parts of the brainstem (pons, medulla oblongata) without involvement of the basal ganglia. The diagnosis was confirmed by molecular genetic research methods (tab. 1).

At the time of the study, all patients' neurological status included impaired eye coordination and fixation, rotatory nystagmus with a vertical component, floating eye movements, bulbar-pseudobulbar syndrome, diffuse generalized trunk and limb hypotonic/atonic associated with mixed tetraparesis, decreased tendon-periosteal reflexes, polyneuropathy syndrome with distal muscle wasting of the limbs, and myogenic contractures in large joints.

The control group consisted of healthy children (3–13 years old) with undetected mutations in the SURF1 gene and no symptoms who presented to the children's outpatient department of the Kulakov National Research Center for Obstetrics, Gynecology and Perinatology as part of a routine pediatric examination.

Collection of blood samples

Whole blood obtained in vacuette with EDTA was treated with ammonium chloride and potassium lysing buffer (AbiLyse ACK Lysing Buffer, USA) for 15–60 minutes after intravenous collection from patients, following the manufacturer's protocol for subsequent sample collection for Western blot analysis.

Fibroblast Cultivation

Primary subcutaneous fibroblast cell cultures were kindly provided by the Pirogov Russian National Research Medical University. Starting from passage 3, the cells were cultured in a complete nutrient medium consisting of DMEM (PanEco, Russia): F12 (PanEco, Russia) at a 1 : 1 ratio, supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin (PenStrep, GIBCO, USA), and 10% fetal bovine serum (FBS, GIBCO, USA). The culture medium was maintained at 37 °C in a 5% CO₂ atmosphere. Upon reaching 60% confluence, the cells were subcultured at a 1:3 ratio using 0.05% trypsin/EDTA (PanEco, Russia) in culture flasks. Trypsin activity was suppressed by adding an appropriate amount of fresh nutrient medium. Cell counts were determined at each passage using a Goryaev chamber (MINIMED, Russia). Cell morphology was assessed using a NIKON ECLIPSE TS100 inverted light microscope (JAPAN).

Construction of the AAV vector carrying the hSURF1 gene

Total RNA was isolated using the commercial RNA-solo kit (Eurogen) according to the manufacturer's protocol. Tissue samples were homogenized and lysed with ExtractRNA reagent (Eurogen, Russia), followed by RNA extraction and purification. The concentration and purity of the obtained RNA were assessed spectrophotometrically at 260/280nm on “Nanodrop spectrophotometer” (USA). RNA integrity and the absence of degradation were verified by electrophoresis in a 2% agarose gel on horizontal system (Helicon, Russia) according standard protocol (http://molbiol.ru/protocol/07_01.html). The results were visualized using the GenoSens 2250 Touch gel documentation system (Clinx). First-strand cDNA synthesis was performed on the isolated RNA template using the Magnus reverse transcriptase reagent kit (Eurogen). To amplify the complete coding sequence of the SURF1 gene, specific primers containing additional nucleotide sequences at the 5'-ends complementary to the terminal regions of the linearized pAAV-CAG plasmid were designed for subsequent assembly by the TLTC method. The primers had the following structure: forward primer SURF1-dir:5'-tgtccaggcggccgccATGGCGGCG GTGGCT-3' reverse primer SURF1-rev: 5'-aggcacagtcgagg cagatctTCACACACCAGGTGTCCCAC-3' (lowercase letters denote sequences homologous to the vector, and uppercase letters denote regions complementary to the beginning and end of the SURF1 gene open reading frame). Primer annealing temperatures and specificities were calculated using Primer3. The SURF1 gene was amplified using polymerase chain reaction (PCR). The amplification protocol included an initial denaturation at 95 °C for 5 minutes, 30 cycles (denaturation at 95 °C for 30 seconds, annealing at 62 °C for 30 seconds, elongation at 72 °C for 1 minute), and a final elongation at 72 °C for 5 minutes. PCR products were analyzed by electrophoresis in a 1% agarose gel.

The pAAV-CAG plasmid vector was prepared for cloning using PCR amplification with the Vec-dir (AGATCTGCCTCGACTGTGCCT) and Vec-rev (GGCGGCCGCCTGGACA) primers and Kapa polymerase (Rosche). This ensured the production of a linear molecule with terminal sequences homologous to the amplified SURF1 gene. Recombinant plasmid assembly was performed using the T5 exonuclease-dependent assembly method. The purified SURF1 gene PCR product was mixed with the linearized vector supplemented with T5 exonuclease and the mixture was incubated on ice for 5 minutes. Two microliters of the reaction mixture were used to transform chemically competent E. coli strain XL-Blue (Eurogen) cells according to the standard method [13].

Plasmid DNA was isolated from the cultures using the miniprep method and Plasmid Miniprep 2.0 kit (Eurogen). Primary screening of positive clones was performed using PCR with SURF1-specific primers. The accuracy of the resulting recombinant construct was further confirmed by restriction analysis with PstI endonuclease in buffer Orange (SibEnzyme), followed by restriction profile analysis by agarose gel electrophoresis.

Assessing the cytotoxic effect of the AAV vector carrying the hSURF1 gene

The cytotoxic effect of the virus was assessed using the MTT assay on the viability of subcutaneous fibroblasts from patients without the Surf1 mutation. Cell cultures from passage 4 were seeded in 96-well transparent flat-bottomed plates (Greiner Bio-One, Austria) at a concentration of 1 × 10⁵ cells/well. After 24 hours of cultivation at 37°C and 5% CO₂, the cells were treated with various concentrations of the vector: 10¹³, 10¹², 10¹⁰, and 10⁸ viral particles per ml of culture medium. Control wells were supplemented with culture medium in appropriate quantities instead of viral particles. After adding viral particles, the cells were incubated for 24 hours under the conditions described above. At the end of the incubation, 20 μl of a solution containing 4,5-dimethylthiazol-2-yl-2,5-diphenyltetrazolium bromide (MTT, Santa Cruz) at a concentration of 5 mg/ml were added to each well of the microplate. After 4 hours, the medium with the MTT reagent was removed, and 100 μl of dimethyl sulfoxide (DMSO) were added to the wells to solubilize formazan crystals. Optical density was measured using an ELISA microplate reader at 570 nm. The assay was performed in triplicate to calculate the mean half-maximal inhibitory concentration (IC₅₀) of viral particles and standard deviation.

Electrophoresis and Western blot analysis of target proteins in whole blood and fibroblasts lysates

To analyze the relative content of the studied proteins in whole blood and subcutaneous fibroblasts lysates with and without viral infection, samples were separated by SDS-polyacrylamide gel electrophoresis followed by electro transfer to a nitrocellulose membrane (Bio-Rad) [15]. The membranes were incubated with specific primary antibodies (anti-Surf1 (Abcam; ab110256, 1 : 1000), anti-MTCO1 (Abcam; ab14705, 1 : 1000) followed by incubation with secondary anti-species HRP-linked antibodies. The signals were analyzed by enhanced chemiluminescence assay using the Novex™ ECL Chemiluminescent Substrate Reagent Kit (INVITROGEN, USA) according to the manufacturer's instructions. The signal from the proteins of interest was normalized to the total protein content determined by the staining intensity using Ponceau's dye (0.5% in 1% acetic acid).

Target Genes Expression Assessment in Fibroblasts Lysates after Transfection

Gene expression was assessed by real-time reverse transcription using polymerase chain reaction (PCR) with transcript-specific primers (tab. 1). RNA isolation and purity analysis were performed as described above. For reverse transcription (RT), 2 μl of 20 μM random hexaprimer were added to 0.5 μg of total RNA and incubated for 2 min at 70 °C. The tube was then transferred to ice, and 4 μl of 5x RT buffer, 2 μl of 10 mM dNTP mixture, 2 μl of 20 mM DTT, and 1 μl of MMLV reverse transcriptase (50 U/μl) (Eurogen, Russia) were added. The mixture was incubated for 40 min at 40 °C and the reaction was stopped by heating for 10 min at 70 °C. Subsequent real-time PCR was carried out according to the following scheme: 50 ng of template cDNA, 0.4 μl of each forward and reverse primer (100 pmol), 2 μl of 5xqPCRmixHS SYBR. Amplification was monitored by electrophoresis in 3% agarose gel. Statistical data processing was performed using the QGENE software (2-ΔCt method). The expression of each target gene was normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene. In the course of the study, primers specific to the transcripts of the corresponding genes were selected (tab. 2).

RESULTS

The relative content of SURF1 and MTCO1 proteins in venous blood cells of 20 pediatric patients (3–11 years old) without mitochondrial pathologies (control group) was determined using Western blot analysis. Data for reference values are presented in fig. 1. It was shown that the relative content of the studied proteins is reduced in patients with mutations in the SURF1 gene compared to the control group. 

To determine the safety of viral construction, a dose of 104 per cell was found to be optimal based on the results of a study examining the relationship between cell viability and the number of viral particles used for transfection (fig. 2). Next, an adenoviral construct with a normal SURF1 gene analogue was added to fibroblasts from patients with a mutation in the SURF1 gene and to control fibroblasts at a dose optimized to ensure no reduction in cell viability in the MTT assay.

Expression of the SURF1 and MTCO1 genes in fibroblast lysates after infection with AAV9-SURF1

We used quantitative PCR to evaluate the expression levels of target genes in fibroblasts treated with the gene therapy construct (fig. 3). First, we found that fibroblasts from patients without SURF1 mutations had significantly higher expression levels than fibroblasts from patients with Leigh syndrome and SURF1 mutations (fig. 3, left panel). Furthermore, MTCO1 expression levels were not significantly different between fibroblasts from patients and controls (fig. 3, right panel). Second, SURF1 expression levels significantly increased in fibroblasts from patients with mutations and did not significantly differ from the expression levels of this gene in control fibroblasts treated with the gene therapy construct. 

We observed a similar pattern for the MTCO1 gene — a significant increase following transfection in diseased fibroblasts compared to pre-transfection levels. Interestingly, the MTCO1 gene expression level after transfection was several times higher than baseline for both diseased and healthy cells (fig. 3).

Western blot analysis of SURF1 and MTCO1 content in fibroblasts

When analyzing the levels of target proteins after exposure to the gene therapy construct, we identified significant differences between control cells and subcutaneous fibroblasts from a patient with the SURF1 mutation. Using Western blot analysis, we assessed changes in the relative amounts of the studied proteins and showed that 24 hours after virus introduction into the culture medium, both wild-type cells and SURF1 mutants showed a significant increase in protein levels (p < 0.05), but this increase was more pronounced for fibroblasts from patients with the mutation (fig. 4, left panel). We also demonstrated that the baseline SURF1 protein level in wild-type cells was significantly higher than in mutant cells. As for the MTCO1 protein, which is initially significantly more abundant in wildtype cells, no increase in its expression under the influence of the virus is observed in control fibroblasts, while for cells from patients with the SURF1 mutation, the MTCO1 level is significantly increased 24 hours after adding the virus to the cells (fig. 4, right panel) and even exceeds the protein level in wild-type cells (p < 0.05).

DISCUSSION

In frame of the generally accepted concept that SURF1 protein plays a critical role in the assembly of the cytochrome oxidase holoenzyme, from a mechanistic point of view it can be proposed that the loss of SURF1 protein affects COX activity by stabilizing the structure of the multiprotein complex in the mitochondrial membrane [15]. When Leigh syndrome manifests in patients with mutations in the SURF1 gene, usually no more than 20% of COX activity is retained, while in model animals with a knockout of this gene, at least 50% of the enzyme activity remains, which leads to such a significant difference between the phenotypes of patients with Leigh syndrome and mice deficient in the SURF1 protein [12]. On the other hand, although the use of such models does not seem promising due to the observed phenotypic differences, gene therapy using an AAV9-SURF1-based vector resulted in a significant (up to 30%) [12] improvement in COX activity in the brain, muscle, and liver of SURF1-deficient mice.

In our model, in which we compared the effect of AAV9SURF1 transfection of fibroblasts from patients with clinically diagnosed Leigh syndrome and fibroblasts from healthy children, we aimed to determine not only the expression level of SURF1 protein but also MTCO1, which is commonly used as a marker for the level and native structure of the entire cytochrome oxidase. The observed synchronous increase in the expression levels of both the SURF1 and MTCO1 genes and proteins in fibroblasts from patients, along with the previously observed improvements in the condition of model animals with similar changes in response to gene therapy, suggest the possibility of increased cytochrome oxidase activity. Measuring this activity is our goal for next study on cells from patients with Leigh syndrome to further substantiate the need to continue preclinical trials of the proposed adenoviral-based drug.

The proposed approach to studying the effectiveness of gene therapy in patient cells avoids the shortcomings of the mouse model, where, according to the researchers, a more severe phenotype must be created to detect an effect due to the absence of behavioral defects and neuromotor impairments. It has previously been shown that in existing model animals, the manifestation of the disease in a mild form is due to the activation of compensatory mechanisms in surviving individuals during natural selection, in contrast to humans, where medical care leads to survival even when a severe form of the disease manifests. [10].

CONCLUSION

The resulting AAV9-hSURF1 construct allows for safe transfection at a dose that, after just 24 hours, leads to the possibility of recording a reliable multiple increase in the expression of both genes and proteins SURF1 and MTCO1 in cells from patients with Leigh syndrome with an initially low content of these proteins compared to cells from healthy individuals.