The peptide-mimetic compounds were designed to have the following features: i) therapeutic activity and safety at least as high as the β32_33 peptide in cell models; ii) higher stability than the β32_33 peptide in human plasma; iii) ability to reach the target organs and tissues, and in particular heart, skeletal muscle, and central nervous system (CNS) upon in vivo administration. To reach these goals, we designed three peptide-mimetic compounds comprising only amino acids with a D configuration, which is rarely used by living organisms, as opposed to the naturally occurring L configuration. Therefore, neither peptide-mimetic is expected to undergo either proteolytic cleavage or other stereospecific enzyme-catalysed transformation. Specifically, the three peptides are: the PMT, which comprises the D-enantiomers of the L-amino acids present in the β32_33 peptide, in the same order; and the PMT-8a and PMT-8b, which are fragments of the PMT, comprising the D-amino acids 1-8 and 5–12 of the PMT, respectively.

All constructs were synthesized with purity >85% by Pepscan, now part of the Biosynth Group (Biosynth B.V., Zuidersluisweg 2, 8243 RC Lelystad, The Netherlands). The compounds used for the study are listed in Table 1.

The experimentally determined 3D structures of bacterial LeuRS in complex with tRNA^Leu were downloaded from the Protein Data Bank (PDB) (Berman et al., 2000). Their PDB identifier (ID), Resolution and source are the following: PDB ID: 2BTE, 2.90 Å, T. thermophilus (Tukalo et al., 2005) and PDB ID: 4AQ7, 2.5 Å, E. coli (Palencia et al., 2012).

The sequences of human mt-LeuRS and mt-tRNA^Leu(UUR) were downloaded from the UniProt database (uniprot.org; ID: Q15031) (UniProt Consortium, 2005) and the National Center for Biotechnology Information (NCBI) database of gene sequences (https://www.ncbi.nlm.nih.gov/; ID: NC_012920.1:3230-3304), respectively.

Molecular docking of the 3D models of human LeuRS and mt-tRNA^Leu(UUR) was performed using by the freely available LZerD (https://lzerd.kiharalab.org/; Christoffer et al., 2021) and AlphaFold3 (https://alphafoldserver.com/welcome; Abramson et al., 2024) web servers.

Swiss-PDBViewer (http://www.expasy.org/spdbv/; Guex and Peitsch, 1997) was used for 3D structures visualization, analysis and modelling and FACE2FACE (https://face2face.ibpm.cnr.it/; Di Micco et al., unpublished) to calculate intermolecular contacts and generate 2D contact maps.

Previously established osteosarcoma-derived (143B.TK-) cybrid cell lines from patients, which bear either the m.3243A>G mutation in mt-tRNA^Leu(UUR) or the m.8344A>G mutation in mt-tRNA^Lys and controls (Perli et al., 2014; Perli et al., 2016; Perli et al., 2020) were used. The pathological mt-tRNA^Leu(UUR) mutant had a mutation load >98%. The mt-tRNA^Lys mutant had a mutation load of either ∼80% (H-8344) or ∼30% (L-8344). The H-8344 high mutation load mutant was pathological, whereas the L-8344 low mutation load mutant did not show any detectable phenotype and was used as control (Perli et al., 2020).

Cybrid cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 4.5 g/L D-glucose, 10% foetal bovine serum (FBS), 2 mM L-glutamine, 50 μg/ML uridine, 100 U/mL penicillin, and 100 mg/mL streptomycin (referred to as glucose medium) in a humidified atmosphere of 95% air and 5% CO₂ at 37°C. For cell viability experiments, cells were grown either in glucose medium or in glucose-free DMEM, supplemented with 5 mM galactose, 110 mg/mL sodium pyruvate, and 10% FBS (referred to as galactose medium). The reason for using the latter medium is that a pathological phenotype can be appreciated in cells growing on galactose, which forces cells to rely on mitochondrial respiration, but not in cells growing on glucose.

Constructs made of compounds listed in Table 1, comprising D-peptides linked to the Cy5 fluorophore via maleimide cross-linker, were administered to sub-confluent cybrid cell cultures at 0.25 µM in glucose medium. About 24 h after treatment with each construct, cells were incubated with 200 nM Mitotracker Red FM (LifeTechnologies Italia, Monza, Italy) for 30 min at 37°C.

Subsequently, cells were visualized by confocal microscopy. Images of 800 × 800 px (at 88 nm/px) were acquired at the Olympus iX83 FluoView1200 laser scanning confocal microscope using a 60 × NA1,35 oil objective (Olympus Italia SRL Milano, Italy), zoom 3×, 559 nm, and 635 nm lasers and filter setting for Mitotracker Red and Cy5. The fluorescence images were analysed with the ImageJ software (Rasband, 2018) to determine the Pearson’s correlation coefficient.

For a subset of experiments, we evaluated the mt localization of constructs up to 72 h after treatment. PMT and PMT fragments were administered to sub-confluent cybrids at a concentration of 0.25 µM in the glucose medium. At 24-, 48- and 72-h after treatment, cells were incubated with Mitotracker Red and visualized by confocal microscopy. After each confocal imaging section, a fresh glucose medium was added to the cells, which were maintained under standard conditions (a humidified atmosphere of 95% air and 5% CO₂ at 37°C).

To assess growth capability, cells were harvested and seeded at 30 × 10⁴ in 60 mm dishes in glucose medium for 24 h with the addition of one of the compounds. Cells were switched in glucose or galactose and, after 24 h, cell viability was measured by the Trypan blue dye exclusion assay. Cells were harvested with 0.25% trypsin and 0.2% EDTA, washed, suspended in PBS in the presence of Trypan blue solution (Sigma–Aldrich) at 1:1 ratio and counted using a hemocytometer. The number of viable cells in the galactose medium was expressed as a percentage of the number of cells in glucose medium.

Oxygen consumption rate (OCR) of cybrids was evaluated with Clark type oxygen electrode (Hansatech Instruments, Norfolk, United Kingdom). After cybrid incubation with compounds at 5 µM concentration, both control and mutant cybrids were maintained in glucose medium for 36 h, then OCR was measured in intact cells (3 × 10⁶) in 1 mL DMEM lacking glucose supplemented with 10% sodium pyruvate.

Mitochondrial toxicity exerted by PMT was measured using the Mitochondrial ToxGlo™ Assay (Promega Italia S. r.l., Milano, Italy) according to the manufacturer’s protocol. Cybrids were plated on a 96-well plate and treated with different PMT concentrations (0.5, 2, 5, 10, 20, 30, 40 and 50 µM). Twenty-four hours after treatment, control cells (both wild-type and mutated) were incubated with either 400 μg/mL digitonin (a cytotoxic agent) or 50 µM sodium azide (a mitotoxic agent) for 3 h, as positive controls for cyto- or mito-toxicity, respectively. Subsequently, cells were incubated with specific reagents and fluorescence or luminescence were measured with a GloMax Multi + Luminometer (Promega Italia Srl., Milano, Italy).

To assess whether the β32_33 peptide, PMT, PMT-8a and PMT-8b peptide-mimetics are stable in blood, we set up a chromatographic assay able to evaluate compound concentration after incubation in human plasma.

Four different samples from healthy volunteers were used. Blood was drawn by venepuncture in vacutainer containing EDTA as an anticoagulant. Plasma was separated by centrifugation and immediately used for the experiments. Each compound was dissolved in 500 μL plasma at a final concentration of 0.2 mM and split into two aliquots, one of which was immediately analysed to assess the basal compound level (T₀); the other was incubated at 37°C under gentle shaking, and the amount of each compound was measured at different time points (i.e., 1.5, 3, 6 and 72 h).

To perform chromatographic analyses, the samples were treated with 3 volumes of acetonitrile containing 1% formic acid and then extracted by using the Ostro™ pass-through sample preparation system to remove proteins and phospholipids. Samples were dried under vacuum, resuspended in 100 μL of 0.1% formic acid containing 5% acetonitrile, and then injected directly onto the chromatographic column.

Chromatographic analyses were performed on a Water Acquity H-Class UPLC system (Waters, Milford, MA, United States), including a quaternary solvent manager (QSM), a sample manager with a flow-through needle system (FTN), a photodiode array detector (PDA) and a single-quadruple mass detector with electrospray ionization source (ACQUITY QDa). Analyses were performed on a reverse phase C18 column (75 mm × 3.2 mm i. d., 2.5 μm particle size). The mobile phase was solvent A, 0.1% formic acid in water, and solvent B, 0.1% formic acid in acetonitrile. The flow rate was 0.5 mL/min, the column temperature was set at 25°C and the elution was performed by linearly increasing the concentration of solvent B up to 70% in 7 min. Mass spectrometric detection was performed in the positive electrospray ionization mode, using nitrogen as the nebulizer gas. Analyses were performed in the Total Ion Current (TIC) mode with a mass range of 100–1200 m/z. The capillary voltage was 0.8 kV, cone voltage 8 V, ion source temperature 120°C and probe temperature 600°C. Quantification was performed by Selected Ion Recording (SIR): at m/z = 917.88, corresponding to the [M+2H]²⁺ ion obtained from either β32_33 peptide or PMT; at m/z = 961.42, corresponding to the [M + H] ⁺ obtained from PMT-8a; and at m/z = 869.48, corresponding to the [M + H]⁺ ion obtained from PMT-8b.

The radiolabelling protocol was optimized by investigating different conditions for pH, peptide amounts, temperature, and time of reaction. During the setup, the absence of unreacted copper by SPE tC18 cartridges was ascertained, therefore a subsequent purification step was not required. The final protocol was as follows: 370 MBq of [⁶⁴Cu]CuCl₂ in HCl 0.1M were added to 240–250 μg of PMT-NOTA peptide in buffer sodium acetate 0.4 M, pH 7. The reaction mixture was incubated at 70°C for 30 min. The mixture was allowed to cool down to room temperature before quality control. The radiochemical purity of ⁶⁴Cu-PMT was assessed by reversed phase HPLC on C18 column by applying a gradient elution with water/acetonitrile added with 0.1% TFA.

Two-month-old male C57BL/6J mice were purchased from ENVIGO RMS S. r.l. (San Pietro al Natisone, Italy). All animals were housed in the San Raffaele Research Institute (Milan) animal facility under constant temperature and humidity, with a 12-h light/dark cycle and access to food and water ad libitum and environmental enrichment. Experimental procedures involving the use of animals complied with the EU Directive 2010/63/EU for animal experiments and have been approved by the Ethical Committees of the San Raffaele Research Institute (Milan) and the Italian Ministry of Health (license n. 717/2021-PR). The study was conducted in compliance with the ARRIVE 2.0 guidelines.

Sixty adult C57BL/6 male mice were initially included in the study. Twenty animals were dedicated to in vivo/ex vivo biodistribution of the radiolabelled PMT ⁶⁴Cu-PMT. The evaluation was carried out at four time points: 1, 3, 24 and 48 h. Forty animals were dedicated to tolerability studies with increasing PMT concentration: 1, 5, 10 and 50 mg/kg PMT or saline, i. v. (n = 10 per group). The highest dose was evaluated only in two animals, since they died immediately after the administration. Therefore, only three groups of ten animals were dedicated to the tolerability test, and fifty C57BL/6 male mice were employed for the current study in total. Necropsy of the two animals injected with 50 mg/kg PMT did not highlight any obvious macroscopic effect, and no further molecular investigations were done. Due to the rapidity of death, and the hydrophobic character of the C-terminal region of the PMT (see the Discussion section), it is possible that PMT aggregates obstructed one or more small vessels, causing micro-infarctions in vital organs (e.g., lung, heart, brain).

PET-CT imaging was performed using β-Cube^® and γ-Cube^® (Molecubes, Gent, NE), respectively, equipped with the mouse Hotel insert hosting up to four mice simultaneously. A group of 5 animals were injected into the caudal vein with the radiopharmaceutical ⁶⁴Cu-PMT (mean injected radioactivity of 3.62 ± 1.03 MBq, corresponding to mean injected mass of 4.15 ± 0.53 µg) and subjected to PET-CT acquisition at 1, 3, 24 and 48 h after injection. For each time point, animals were anesthetized by isoflurane (2% in medical air) and placed prone on the CT scanner bed for CT acquisition (2 min after scout). Thereafter, the bed with the immobilized animals was moved in the PET scanner for the acquisition (30 min). Temperature was maintained constant, and respiratory and cardiac rates were monitored. At the end of the last acquisition (48 h), animals were sacrificed by decapitation, followed by immediate blood recovery by cardiac puncture. An aliquot of blood was centrifuged to obtain the plasma. The following tissues were also dissected: liver, heart, lungs, spleen, intestine, stomach, kidney, muscle, and brain. All samples were placed in pre-weighed tubes for gamma-counter counting. After decay, the tubes were weighted again for tissue net weight determination to calculate the percentage of injected dose of radioactive compound per gram of tissue (%ID/g).

Fifteen additional animals were injected with the same dose of radioactive peptide as the other five and sacrificed at 1, 3, and 24 h (n = 5 per time point) for gamma-counting of tissues, to compare ex vivo biodistribution with in vivo PET/CT results. Animals were anesthetized with isoflurane and sacrificed by cardiac puncture. An aliquot of blood was centrifuged to obtain the plasma. Liver, heart, lungs, spleen, intestine, stomach, kidney, muscle, and brain were also dissected and counted to gamma-counter as described above.

PET and CT data collected for each animal and at each time point were reconstructed and automatically co-registered. Images were then corrected for isotope decay and analysed using the software PMOD 4.1; areas of interest were quantified using Volumes of Interest (VOIs) analysis. VOIs were manually drawn on transaxial sections within liver, heart wall, kidney, muscle, cerebrum, and cerebellum, and radiolabelled PMT signal was automatically extracted. Data were expressed as SUV_mean (Standardized Uptake Value), the index usually adopted to express PET data, which is corrected for animals’ body mass.

Adult C57BL/6 male mice were used to determine the general tolerability of i. v. administered PMT and identify the Maximum Tolerated Dose (MTD). To this purpose, animals were divided into four groups (n = 10 each) and injected in a tail vein with either saline (controls) or PMT freshly dissolved in saline at three scalar doses (1, 5 or 10 mg/kg). Each group was randomly divided in two subgroups (n = 5 each). Blood ocular withdrawal was performed at 1 h and 48 h post-injection for each subgroup, respectively. During this procedure, animals were anaesthetized with isoflurane (2% in medical air) and monitored until awakening.

The effect of PMT administration was evaluated every 2 days by observation of the animals and application of the following criteria and scores (see Supplementary Table S1): body weight loss (score: 1–5); coat conditions (score: 1–3); dehydration (score: 1–3); respiration (score: 1–3); spontaneous behaviour (score: 0–5); reaction to handling (score: 0–3) (IACUC of the San Raffaele Research Institute and FELASA Glossary 2015 (Fentener van Vlissingen et al., 2015).

The observation was carried out for 2 weeks post peptide injection. At the end, all the animals underwent a grid-hanging test to measure muscular strength performances after PMT or saline treatment (Gayi et al., 2018).

After a period of habituation to the metallic grid, mice were placed individually on top of a metallic grid (40 × 40 cm, 1.3 cm square mesh) suspended 50 cm above an empty cage. The grid was turned upside down and latency to fall was measured up to a fixed limit of 600 s. Each mouse performed three trials and the longest holding time (s) among them was multiplied by the animal’s body weight (g) to calculate Holding impulse (g*s).

After the test, mice were sacrificed under general anaesthesia (4% isoflurane in medical air) by cardiac puncture to withdraw blood, brain, heart, liver, kidney, spleen, and lymph nodes, which were collected into tubes with buffered formalin and stored at +4°C.

Samples of blood collected at 1 h, 48 h and 15 days were centrifuged to obtain plasma that was transferred to cryovials and stored at −80°C.

At the end of the tolerability test, tissues collected from animals injected with saline (n = 5) or treated with 1, 5 and 10 mg/kg of PMT (n = 10), were formalin-fixed and paraffin-embedded. Five µm-thick sections were obtained from the paraffin blocks and stained with hematoxylin and eosin for histologic examination.

Confocal microscopy experiments showed that, upon exogenous administration to mutant cells, the fluorescent signal of PMT and PMT fragments (PMT-8a and PMT-8b) was detectable within cybrids and overlapped with the mt reticulum, indicating mt localization, as demonstrated by Pearson’s correlation coefficients (PCC values: 0.78 ± 0.06, 0.83 ± 0.02, 0.95 ± 0.01 and 0.93 ± 0.01 for βp, PMT, PMT-8a and PMT-8b respectively; Figure 1). The fluorescence signal remained constant over time; in fact, mt localization did not change up to 72 h, as indicated by Pearson’s correlation coefficients (Supplementary Figure S1).

To evaluate the effect of PMT and PMT fragments on viability, cybrids were grown in glucose-free medium supplemented with galactose (galactose medium), a condition that both forces cells to rely on the mt respiratory chain for ATP synthesis and causes a significant growth reduction in the presence of mutations. Administration of 5 µM PMT induced a significant improvement of cell viability in m.3243A>G and m.8344A>G mutant cybrids, as compared with non-treated mutant cells (Figures 2A,B). PMT rescuing activity was comparable to that of the β32_33 peptide and higher than that of either PMT-8a or PMT-8b at the same concentrations as PMT. PMT-8b significantly improved cell viability only in m.3243A>G mutant (Figure 2A).

To investigate whether increased cell viability was related to improved mt bioenergetics, we analysed the respiratory capability of mutant and control cells by using the Clark type electrode. PMT determined a significant increase in oxygen consumption rate in both the m.3243A>G and m.8344A>G pathological mutants (Figures 2C,D). This increase was higher than that determined by either PMT-8a or PMT-8b. The same treatment had no effect on the viability and oxygen consumption of control cells (Supplementary Figure S2A–D).

Given that PMT ameliorated cell viability and increased oxygen consumption to a higher extent with respect to either PMT-8a or PMT-8b, we performed subsequent experiments on PMT alone.

To evaluate the toxicity of increasing concentrations of exogenously administered PMT, we performed the Mitochondrial ToxGlo™ Assay. PMT resulted to be neither cytotoxic nor mitotoxic up to: 20 µM in m.3243A>G mutant cybrids; at least 50 µM in m.8344A>G mutant cybrids; and at least 40 µM in healthy control cells (Figures 3A,B).

We investigated the molecular reasons why the PMT has essentially the same therapeutic activity as the β32_33 peptide in cell models, in spite of their opposite stereochemistry. To this end, we analysed the interactions between LeuRS and tRNA^Leu in the experimentally determined structures of the complexes from T. thermophilus and E. coli available from the PDB (see Table 2). This analysis indicated that, in both experimental structures, LeuRS residues present in the β32_33 peptide regions interact with the tRNA^Leu only through their side-chain atoms.

We also analysed the molecular models of human mt-LeuRS and mt-tRNA^Leu(UUR) built by LZerD and AlphaFold3 (AF3) servers, two publicly available resources that provided accurate predictions of macromolecular assemblies in the 16th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP16: https://predictioncenter.org/casp16/). In the AF3 and LZerD models the two molecules are and are not placed in structurally homologous positions with respect to that observed in bacterial structures, respectively; therefore, only the AF3 best model was analysed (Supplementary Figure S3). The estimate of overall model accuracy provided by AF3, and expressed as predicted local distance difference test (pLDDT) values (Jumper et al., 2021), is very high (pLDDT >90) or high (70 > pLDDT ≥90) for most of the protein structure (49% and 39% of the atoms, respectively) and high for 81% of the tRNA structure, the remaining regions of which (18%) have low (50 > pLDDT ≥70) accuracy. In this model, the number of main chain contacts between the β32_33 region and the tRNA is very low (i.e., 2/94 = 2% of total contacts) (Table 2; Supplementary Figure S3).

Following incubation with human plasma, the amount of β32_33 peptide decreased by 90% after 1.5 h and was almost undetectable after 72 h (Figure 4). PMT had significantly higher plasma stability than β32_33 peptide at all time points, although the amount of detectable compound after 72 h was reduced to about 25%. Conversely, after 72 h the amount of detectable PMT-8a and PMT-8b were >90% and >85%, respectively.

Radiochemical purity of the ⁶⁴Cu-PMT construct was above 99%. Specific activity, considering labelled and unlabelled peptide, was about 1.4 GBq/mg at the End of Synthesis (EOS). The radiochromatogram of ⁶⁴Cu-PMT quality control has been inserted in the supplementary information (Supplementary Figure S4). The final radiolabelled peptide was stable in saline solution at room temperature, maintaining a radiochemical purity of 93.5% after 24 h (Supplementary Figure S4B). Data were not acquired for longer time points.

The ex vivo biodistribution of the radiolabelled ⁶⁴Cu-PMT construct, expressed as %ID/g, showed an increase of radioactivity concentration in blood, plasma, brain areas and all the peripheral organs over time. The only exception was represented by the kidneys, where values peaked at 1 h, dramatically dropped at 24 h and slightly increased at 48 h post-injection in healthy mice (Figure 5; Table 3). Due to the increase in radioactivity uptake in blood over time, we also expressed data as tissue-to-blood ratios of radioactivity concentration. These data showed a decrease over time in liver, kidney and spleen, the organs devoted to the elimination of labelled PMT, while it remained stable in the other districts (Figure 6; Table 4).

The in vivo PET/CT imaging data analysis performed on peripheral and central areas are consistent with the results obtained by counting tissues by gamma counter (Figure 7; Table 5). In fact, the concentration of radiolabelled PMT reported as SUV_mean or %ID/g (see Figure 7; Supplementary Figure S5, respectively) increased over time in the heart wall and brain areas, whereas it remarkably decreased in the excretory organs, i.e., liver and kidney, at 24 h and 48 h post injections. At variance with what is observed in ex vivo studies, muscle uptake remains stable. Consistent with high spleen uptake, images showed a clear PMT accumulation in lymph nodes as well (Figure 8). Interestingly, the lack of choroid plexus signal, which is the main exchange system for free copper in the brain, in the PET images indicates that free copper is present in small amounts, if any (Figure 8).

The i.v. injection of increasing concentrations of PMT solution did not produce any significant effect on the behaviour of animals evaluated at 14 days with the Grid-hanging test (Figure 9). Mice treated with 1, 5 and 10 mg/kg of PMT exhibited similar muscular strength to mice injected with saline.

In the days following the injection of 10 mg/kg of PMT, a subgroup of animals manifested slight distress increase, consisting in mild tachypnoea (score 1) in 5/10 mice, and body weight decrease (score 2–3) in 8/10 animals, as compared with initial values (Supplementary Figure S6). A slight decrease of body weight was observed also in animals treated with PMT 1 mg/kg (score 1 in 5/10 mice) and 5 mg/kg (score 1–2 in 5/10 mice and 4 in 1/10 mice). The spontaneous behaviour, as well as body weight, improved in the following days. In fact, no significant body mass differences were observed between animal groups at the end of the experiment (Figure 10). Moreover, the mean value of body weight between groups was not significantly different at any time point (p < 0.05, multiple comparison with one-way ANOVA).

Microscopic analysis of sections stained with hematoxylin and eosin did not show any morphological change in any organs or tissues of treated animals with respect to controls (Supplementary Figure S7).