Introduction

Telomeres form protective caps at chromosome ends thanks to the binding of a six-member protein complex termed shelterin (1). With cell divisions, telomeres of normal somatic cells progressively shorten, a process that likely contributes to cellular aging and the development of various age-related pathologies (2,3). Accordingly, cohort studies established that shorter leucocyte telomere length (TL) was associated with a higher risk of diseases affecting specific organs, such as the lung or the liver, and an increased incidence of esophageal cancer, as well as lymphoid and myeloid leukemia (4).

While aging is a key determinant of TL, genetic variation also plays a significant role in the variability of leukocyte TL among healthy individuals, as evidenced by the identification of nuclear genomic SNPs associated with TL (5). Heritability estimates for human TL reach 82% (6). Consistently, TL in the zygote appears to play a key role in determining TL later in life (5), indicating that it is largely established during early embryonic development. Telomere length resetting during embryogenesis is a telomerase-dependent process that occurs at the morula-to-blastocyst transition (7). Recent findings in mice show that experimentally disrupting mitochondrial function in zygotes significantly impairs telomere elongation between the 8-cell and blastocyst stages, ultimately leading to shorter telomeres in the offspring (8). Similar outcomes have been observed in offspring from reproductively aged or obese female mice -two conditions known to compromise oocyte mitochondrial integrity (8). Collectively, these findings underscore the critical role of proper mitochondrial function in telomere regulation during embryogenesis. Further evidence comes from individuals with mitochondrial disorders such as MELAS syndrome and Leber’s hereditary optic neuropathy (LHON). These individuals exhibit significantly shorter leukocyte telomeres at birth, likely reflecting impaired TL establishment during early development (9).

While damaged mitochondria are associated with shorter TL in mouse embryos and may impair TL regulation in patients with mitochondrial diseases (9), the impact of naturally occurring mitochondrial genome SNPs on TL remains largely unexplored. With a total of 29 haplogroups and a myriad of subhaplogroups characterized by distinct variants compiled in the Mitomap database (10), human mtDNA genomes display large heterogeneity. Some variants within the coding regions of mtDNA, such as the NADH dehydrogenase subunits (ND1, 2, 3, 4, 4L and 5) of complex I (CI), Cytochrome b, the COXI, II and III subunits of CIV or the ATP6 and 8 subunits of the ATP synthase (CV) lead to distinct ETC properties that impact mitochondrial metabolism and ROS production. In line with the hypothesis that the mitochondrial genome may contribute to the observed stronger maternal influence on TL at birth (6,11) -given that mtDNA is maternally inherited-, we hypothesized that non-pathogenic variation in the mtDNA sequence could affect TL.

Building on reference curves for TL in the Belgian population, obtained through fluorescence in situ hybridization (FISH) combined with flow cytometry (Flow-FISH), we identified mtDNA variants associated with either very long or short leukocyte TL. We leveraged a transmitochondrial cybrid approach to directly test the effects of these mtDNA variants on mitochondrial function in mtDNA-depleted recipient cells and uncovered an inverse relationship between leukocyte TL in the donors and mitochondrial ROS levels measured in the cybrids. Our results further highlight the critical role of mtDNA-encoded CI activity in sustaining NAD⁺ metabolism and telomere integrity, particularly under oxidative stress conditions arising during the metabolic transition from glycolysis to oxidative phosphorylation in early cybrid development. Collectively, these findings reinforce the link between mtDNA-encoded mitochondrial function and telomere regulation in human cells.

Results

Establishment of reference curves for TL using Flow-FISH

To establish reference curves for TL in Belgium, we applied the gold standard Flow-FISH approach (12) on leukocytes isolated from a cohort of 491 healthy donors aged 7 days to 99 years old. Percentile curves for TL in lymphocytes (Fig 1A) and granulocytes (Fig 1B) were comparable to those obtained previously (13,14) (Fig S1A) and showed large heterogeneity among individuals of similar age. Although still controversial, obesity may negatively impact TL (5,15). In our cohort, average BMI was not statistically different between donors with telomeres below the 5^th or above the 90^th percentile (Fig S1B). Given previous evidence supporting a stronger maternal inheritance of TL in humans (6,11), with similar trend observed in birds (16), we sought to investigate the potential contribution of the human mitochondrial genome to TL heterogeneity. To this end, we selected individuals with long, average or short telomeres from our Flow-FISH curves (Fig 1C). All were of European origin, thus excluding a possible bias linked to African ethnic groups (5), and the seven donors belonged to distinct mitochondrial subhaplogroups (Fig 1D). Strikingly, donor #6, with lymphocyte telomeres at the 95^th percentile, turned out to belong to the K1a subhaplogroup that encompasses the centenarian-enriched A177T variant (m.G9055>A) in the ATP6 subunit of ATP synthase (CV) (17,18) with an allele frequency of 5.4% in Mitomap database (Fig 1D).

Figure 1.

Distinct mitochondrial genomes differently modulate TL in transmitochondrial cybrids

To evaluate the impact of distinct mitochondrial genomes on TL in human cells, we used a transmitochondrial cybrid strategy in which mtDNA-depleted Rho⁰ (ρ⁰) cells are fused with donor platelets, thereby restoring functional mitochondria carrying donor-specific mtDNA variants (19). The presence of variants in cybrids was checked by sequencing and mtDNA content was evaluated by qPCR (Fig S2). Southern blot analysis revealed that telomeres in 143B ρ⁰ cells were shorter than in parental 143B cells (Fig 2A). To rule out clonal artifacts, mtDNA depletion was independently repeated, and multiple ρ⁰ clones were isolated, all of which exhibited a consistent reduction in TL (Fig S3A). This telomere shortening was not associated with decreased expression of the telomerase subunits hTR or hTERT (Fig S3B), but correlated with a reduction in telomerase activity measured in vitro using cell extracts (Fig S3C).

Figure 2.

In line with a role for mitochondrial function in supporting telomerase activity, fusion of 143B ρ⁰ cells with platelets from donors #6 and #7 (Cyb6 and 7) restored telomerase activity and rescued TL in ρ⁰ cells to levels comparable to those of parental 143B cells (Fig 2A-B and Fig S3D). Telomere elongation in these cybrids was accompanied by a reduced frequency of short telomeres, as assessed by TeSLA (Fig S3E). Unexpectedly, TL was not restored in Cyb1-5 cybrids. Instead, fusion with platelets from donors #1 and #2 induced a marked and reproducible telomere shortening in ρ⁰ recipient cells (Fig 2A and Fig S3D). This effect was not associated with reduced telomerase activity or decreased hTR or hTERT transcript levels, but rather with pronounced telomeric damage and telomere fusions following fusion (Fig 2B-E).

Together, these findings demonstrate that mitochondrial DNA sequence variation exerts a strong influence on telomere length in 143B ρ⁰ recipient cells, ranging from telomere elongation to severe shortening.

Low complex I activity shortens the telomeres of ρ⁰ recipient cells

We next investigated why different mtDNA sequences exerted such divergent effects on TL in 143B ρ⁰ recipient cells. When mitochondria are replenished in ρ⁰ cells during the formation of transmitochondrial cybrids, a shift from glycolysis to OXPHOS occurs (Fig S4A-B) and cells need to adapt to this metabolic change. Complex I (CI) plays a crucial role in regulating the NAD⁺/NADH balance and this is crucial for the activity of NAD⁺-dependent poly (ADP-ribose) polymerase 1) (PARP1) repair enzyme, an enzyme required for the repair of telomeres undergoing oxidative stress in cultured cells (20–22). We thus hypothesized that optimal levels of CI activity may be required for telomere protection in cells experiencing acute oxidative stress during the in vitro metabolic shift before cellular adaptation. Consistent with this, we found significantly lower CI activity in Cyb1 and Cyb2 cybrids (Fig 3A) and a strong positive correlation between TL measured in the cybrids at the first passages and CI activity (Fig 3B). Supporting a lower CI activity in in Cyb1 and Cyb2 cybrids, we noticed a significant up-regulation of the NAD⁺ salvage pathway genes NAMPT and NAPRT1, likely reflecting a compensatory mechanism (Fig 3C).

Figure 3.

Building on the idea that, in the context of lower CI activity, telomeres may not be able to resist oxidative stress-induced shortening, we repeated the fusion of donor #2’s platelets to 143B ρ⁰ cells in the presence of N-acetyl-L-cysteine (NAC) antioxidant and/or nicotinamide riboside (NR), a precursor of NAD⁺. Strikingly, while telomeres were shortened in all clones (5/5) selected from cybrids obtained in the presence of N-acetyl-L-alanine (NAA) used as control, only one out of the eight selected cybrids obtained in the presence of both NAC and NR showed telomere shortening (Fig 3D-E). Treatment with either NAC or NR alone showed intermediate phenotype (Fig 3D-E). The detection of telomeric parylation signals further supported a role for NAD⁺-dependent PARP1 activity in repairing oxidative damage at shortened telomeres during the process of cybrid formation (Fig 3F).

Altogether, we propose that the in vitro process of cybrid formation exacerbates the impact of low CI activity on telomeres during the acute oxidative stress linked to a metabolic shift from glycolysis to OXPHOS.

Lymphocyte TL inversely correlates with ROS levels in transmitochondrial cybrids

Having identified CI activity as a strong regulator of TL during in vitro cybridisation, we asked whether mtDNA-encoded CI activity could also account for the differences in TL observed in donors’ blood cells. However, analysis of age-adjusted lymphocyte TL versus CI activity revealed no association between these parameters (Fig 4A).

Figure 4.

In mouse oocytes, impaired mitochondrial function, associated with elevated mitochondrial ROS production, adversely affects telomere elongation (8). In human cells, experimentally induced mitochondrial dysfunction was also reported to lead to telomere damage through the production of ROS (23). We therefore tested whether human mitochondrial genomes, through their impact on ROS levels, may modulate TL in blood cells. As transmitochondrial cybrids receive the donors’ mitochondria but no other genetic information, this cellular model allows to specifically assess the impact of the mitochondrial genome on ROS production. Mitochondrial superoxide production was measured by electron paramagnetic resonance (24) in clones #1 of all cybrids (Fig 4B). Differences in ROS levels across the cybrids were not linked to distinct changes in oxidative vs glycolytic metabolism in the cybrids that were characterized by similar lactate/glucose and ATP/ADP ratios (Fig S5A-B). Strikingly, ROS levels in the cybrids inversely correlated with the age-adjusted TL measured by Flow-FISH in the corresponding donors’ lymphocytes (Fig 4C). Elevated levels of ROS were notably detected in Cyb3 cybrid. The mtDNA from donor #3 harbors the extremely rare F63L variant (m.T6090>C; Mitomap frequency: 0.002%) in the COX1 subunit, located within a transmembrane helix in close proximity to the heme A group and associated with reduced CIV efficiency (Fig 4D and Fig S6). Together, these findings support a role for mitochondrial ROS in human TL regulation in vivo.

Rare mtDNA variants may account for the maternal inheritance of long telomeres

To further assess the relevance of our findings to a potential role of mitochondria in TL inheritance in humans, we searched for evidence of maternal inheritance of TL within families. Such analyses are most informative in families with extremely long (>P90) or short (<P10) telomeres. We therefore focused on three individuals with lymphocyte TL above the 90th percentile: donors #1, #6, and #7 (Fig 1C).

Despite donor #1 exhibiting lymphocyte TL at the 99^th percentile, TL in lymphocytes from both of this donor’s children was close to the 50^th percentile or only slightly above, arguing against a dominant maternal inheritance of the exceptionally long telomeres observed in donor #1 (Fig 4E). Donor #6 belongs to the K1a subhaplogroup Identified by the MitoAging project as one of the four mtDNA variants significantly associated with longevity, with the A177T variant also showing protective effect against Parkinson’s disease (25). Although we could not assess TL in donor #6’s progeny, as she does not have children, the fact that her mother showed telomeres at the 90^th percentile and that two European male donors, identified in independent cohorts (26) as having telomeres at, respectively, the 97.5^th and above the 99^th percentile (Fig 4E), were found to display K1a mtDNA, support the notion that this subhaplogroup is associated with long telomere inheritance. This led us to screen additional unrelated donors with varying TL for the presence of the m.G9055>A variant. The variant was identified in 1 of 11 donors with lymphocyte TL below the 50^th percentile, 2 of 26 donors with TL between the 50^th and 90^th percentiles, and 3 of 13 donors with TL at or above the 90^th percentile (Fig 4F and Fig S7), suggesting a trend toward higher frequency of the m.G9055>A (ATP6 A177T) variant among individuals with lymphocyte TL at or above the 90^th percentile (p=0.153).

Donor #7 does not belong to the K1a haplogroup; nevertheless, both her mother and her daughter exhibited telomere lengths at the 95^th and near the 90^th percentile, respectively (Fig 4E). In contrast, the telomere lengths of her husband and her father were slightly above the 10^th and around the 50^th percentile, respectively (Fig 4E). These findings pointed to a maternal inheritance pattern associated with the H1b1e mtDNA subhaplogroup, identified through blood sequencing in these women but not previously linked to longevity. While the causal link between the rare m.A3796>G variant in donor #7 (ND1 T164A, CI) and the long telomere phenotype remains to be determined, this variant was absent in all 49 unrelated donors with varying telomere lengths that we screened.

Collectively, these findings support the idea that multiple mitochondrial genome SNPs may contribute to the inheritance of telomere length.

Discussion

While long telomeres are often viewed as markers of good cellular health, shorter leukocyte TL is linked to a higher risk of diseases affecting organs like the lungs and liver, as well as various cancers (4). Although aging is a major factor influencing TL, genetic variation also plays a key role and helps explain why TL established in the zygote strongly influences TL throughout life (5).

Unraveling the mechanisms that determine TL in the zygote is therefore highly relevant, particularly for optimizing in vitro fertilization strategies. A recent study in mice suggests that mitochondrial dysfunction impairs telomere elongation during embryonic development-a defect that was largely reversed by antioxidant treatment, highlighting the harmful effect of ROS on TL regulation in the embryo (8). In humans, prenatal exposure to psychological stress has also been linked to shorter leukocyte TL in offspring (27). Given the established connection between psychological and oxidative stress (28,29), these findings point to a similar influence of mitochondrial function on TL regulation during human embryogenesis. Building on these findings -and previous evidence suggesting stronger maternal inheritance of TL in humans (6,7)-we hypothesized that the mitochondrial genome may influence human TL. To investigate this, we employed a transmitochondrial cybrid model using platelets from seven donors with known leukocyte TL and distinct mitochondrial subhaplogroups. Notably, we observed an inverse correlation between donor leukocyte TL and mitochondrial ROS levels in the cybrids, indicating that mitochondrial subhaplogroup affects mitochondrial function and suggesting a potential role for mtDNA variation in shaping leukocyte TL in vivo. Our findings further implicate CI activity in this process, as, while donors #6 and #7, with signs of maternal inheritance of their long telomeres, both display mitochondria with high CI activity, donor #1, with long telomeres but low mtDNA-encoded CI activity, did not transmit her long telomeres to her progeny. In summary, our data indicate that low levels of mitochondrial ROS, combined with potent CI activity, may be key for a mtDNA-driven maternal inheritance of long telomeres. The Japanese m.C5178A longevity-associated variant in the ND2 gene (30) further supports an important role for CI in promoting mitochondrial fitness, and possibly TL regulation.

Whether and how the centenarian-linked K1a mtDNA contributes to reduced ROS levels awaits further investigation. As a defective aerobic production of ATP, linked to the ATP6 A177T variant, was proposed to account for the lack of mtDNA K haplogroup in Finnish elite endurance athletes (31), we hypothesize that, by reducing ATP aerobic production, ROS production decreases, thus providing a possible explanation to this longevity-associated mtDNA (32). Consistent with this, Cyb6 cybrids showed a distinct response to FCCP uncoupling agent as, unexpectedly, respiration could not be restored in cells with oligomycin-inhibited ATP synthase, suggesting a defective ability in increasing oxygen consumption rate to face an energy demand increase (Fig S4A-B). This could be related to the location of the A177T variant in a crucial region of ATP6, along the proton pore (33). Strikingly, 32.7% of the genetically homogeneous population of Ashkenazi Jews express the ATP6 A177T variant (10). As telomeres from Ashkenazy centenarians blood cells are longer than in control groups (34), it will be interesting to test the link to the m.G9055>A variant encoding ATP6 A177T.

Experimentally induced mitochondrial dysfunction leads to telomere damage (23). Mechanistically, the TTAGGG repeats of telomeres are exquisitely sensitive to 8-oxo-guanine formation, a lesion recognized by the base excision repair pathway that, after removal of the oxidized base, generates single-strand breaks eventually recruiting NAD+-dependent PARP1 repair enzyme (21,35). If left unrepaired in replicating cancer cells, these breaks can convert into double-strand breaks, resulting in distal telomere loss and chromosomal fusions (22). In normal fibroblasts and epithelial cells, telomeric 8-oxoG lesions induced by a chemoptogenetic tool were sufficient to activate p53-dependent senescence (36), highlighting the harmful effects of oxidative stress at telomeres on cellular function. During the generation of transmitochondrial cybrids, cells undergo a metabolic shift from glycolysis to OXPHOS, accompanied by increased ROS production. Before cells adapt to this shift, oxidative stress can be particularly damaging. Supporting the essential role of CI in maintaining PARP1 activity via regulation of the NAD⁺/NADH balance, mitochondrial platelets with low CI activity triggered significant telomere shortening in recipient cells during early cybrid formation -a loss that was prevented by co-treatment with an antioxidant and an NAD⁺ precursor. Although the observed telomere loss may be exacerbated by the in vitro oxidative stress conditions -since donors with low CI activity did not show short leukocyte telomeres-our findings are consistent with reports that NAD⁺ supplementation can alleviate telomere dysfunction in cells from dyskeratosis congenita patients (37) and improve hematopoietic function and telomere integrity in Tert⁻/⁻ mice (38). Data are also consistent with the observation that mice lacking PARP display telomere shortening and chromosomal instability (39). Together, these data bring support to the hypothesis that, when extremely pathogenic and associated with high oxidative stress levels, such as in MELAS or LHON patients (40), mutations in CI may impair telomere length resetting in utero.

Our study also unexpectedly revealed a reduction in telomere length in 143B cells following mitochondrial depletion. This telomere shortening was associated with decreased telomerase activity, as measured by ddTRAP, but not with reduced expression of the telomerase subunits hTERT or hTR. 143B osteosarcoma cells are deficient in thymidine kinase 1 and therefore lack an effective pyrimidine salvage pathway. In the absence of oxidative phosphorylation, the de novo thymidine synthesis pathway is also compromised, rendering 143B ρ⁰ cells dependent on uridine supplementation for survival. Although this hypothesis requires further investigation, we propose that the shorter telomeres observed in 143B ρ⁰ cells may be linked to the emerging role of thymidine nucleotide metabolism in the regulation of human telomere length, potentially through an allosteric effect of dTTP on telomerase activity (41–43).

Altogether, our findings that certain mitochondrial genomes may support telomere elongation are relevant to three-parent in vitro fertilization, a technique that uses donor eggs to prevent the transmission of mitochondrial diseases from mother to child (44). Future work will be necessary to investigate how mitochondrial activity may be involved and whether the nucleotide metabolism that recently emerged as an important regulator of telomerase activity may play additional role.

Materials and Methods

Human study participants

Blood samples were obtained from a total of 491 healthy donors. Cohort characteristics are as follows: 261 females and 230 males, median age of 26.3 y (7 days-99.2 y). Protocols were approved by the CHEF institutional Ethics committee and written informed consents were obtained from all 18 year-old or over volunteers. For children below 18, left-over blood samples were obtained through the Cliniques universitaires Saint-Luc after screening for inclusion.

Leukocyte purification

Blood samples were collected in EDTA tubes and kept at room temperature for a maximum of 48 h before processing. Leukocytes were purified after erythrocyte lysis with ice-cold 0.155 M NH₄Cl (Sigma-Aldrich), 0.01 M KHCO₃ (Sigma-Aldrich) and 0.1 mM EDTA (Sigma-Aldrich) solution as described (12). Leukocytes were then resuspended in hybridization buffer (5 % dextrose/10 mM HEPES/0.1 % BSA), and frozen at −80 °C in 80 % FBS (Cytiva)/20 % DMSO (Sigma-Aldrich).

Flow-FISH

The average telomere length (TL) was measured in peripheral blood leucocytes by flow cytometry combined with FISH (Flow-FISH), using bovine thymocytes as reference as previously described (12). TL of bovine thymocytes was measured by telomere restriction fragment (TRF) analysis (see below). TL was evaluated in lymphocyte and granulocyte populations distinguished by LDS751 staining of DNA as described (12). The telomeric PNA probe was as follows: FITC-OO-CCCTAACCCTAACCCTAA (Panagene). Values obtained for TL in lymphocytes and granulocytes are provided in Table S1.

Genomic DNA extraction

Genomic DNA was extracted by overnight digestion at 45°C with 100 µg/mL proteinase K (Sigma-Aldrich) in lysis buffer (10 mM Tris–HCl, 10 mM EDTA, 1 % SDS; pH 8.0) followed by DNA purification with 25:24:1 phenol–chloroform–isoamyl alcohol (Sigma-Aldrich) and precipitation in isopropanol (Thermo Fisher Scientific) and 0.3 M sodium acetate; pH 5.5 (Sigma-Aldrich). gDNA was subsequently treated with 0.2 mg/mL RNAse A (Thermo Fisher Scientific) for 1 h at 37 °C, purified, and precipitated again as described above. DNA concentration, purity and integrity were evaluated with Nanodrop (Thermo Fisher Scientific).

MtDNA subhaplogroup identification

PCR reactions were performed on 20-50 ng total DNA using GoTaq® G2 DNA Polymerase (Promega) and the primers provided in Table S2 so as to cover the entire mtDNA sequence (45). PCR program was as follows: 3 min at 95 °C followed by 40 cycles of 95 °C for 45 s, 47/50/55 °C for 45 s, and 72 °C for 1 min, with a final step at 72 °C for 5 min. PCR products were loaded on 1 % agarose gel (Eurogentec), stained with 0.5 µg/mL ethidium bromide (Sigma-Aldrich) and extracted from the gel using SmartPure Gel Kit (Eurogentec), according to the manufacturer’s instructions, followed by Sanger sequencing (Azenta/Genewiz). MtDNA subhaplogroups were identified using MITOMASTER tool on mitomap.org website.

MtDNA content quantification

MtDNA content quantification was performed by qPCR on 60 ng of total DNA, using KAPA SYBR FAST (Sigma-Aldrich) and primer sequences (45) provided in Table S2.

Cell culture and treatments

143B cells and cybrids were grown in DMEM high glucose (4.5 g/L) supplemented with 10 % FBS (Cytiva) and 1 % Penicillin/Streptomycin (Capricorn Scientific). 143B ρ⁰ cells were grown in DMEM high glucose supplemented with 0.11 g/L pyruvate (Thermo Fisher Scientific), 10 % FBS (Cytiva), 50 ρg/mL uridine (Sigma-Aldrich) and 1 % Penicillin/Streptomycin (Capricorn Scientific). For the subcloning of 143B ρ⁰ population, cells were treated exactly as if they would undergo fusion with platelets and kept throughout the cloning process in ρ⁰culture medium. All cell lines were regularly checked for mycoplasma contamination. For this, 100 µL of culture medium were heated at 95°C for 5 min before spinning down. qPCR reactions were performed on 2 µL of the supernatant as described previously (46) using primers listed in Table S2.

Isolation of human platelets

Blood was collected in 8.5 mL-CPDA tubes (Sarstedt) and centrifuged at 330 x g for 20 min at 22 °C. The upper platelet-rich plasma fraction was treated with 4 µg/mL eptifibatide (GlaxoSmithKline) and 1 U/mL apyrase (Sigma-Aldrich) on ice before centrifugation at 800 x g for 10 min at room temperature. The pellet was resuspended into 5 mL physiological saline solution containing 4 µg/mL eptifibatide and 1 U/mL apyrase. Platelets were counted and diluted with the saline solution to obtain 10⁵ platelets/µL and stored for 30-45 min at 37 °C. 40 × 10⁶ platelets were then transferred into a 2 mL Eppendorf tube, centrifuged at 800 x g for 10 min at room temperature and directly processed for fusion with 143B ρ⁰ cells.

Cybrid formation

Cybrids were obtained as previously described (19). A detailed protocol is provided as Supplementary information.

Generation of new 143B ρ⁰ clones

New 143B ρ⁰ clones were generated as previously described (47). Briefly, 1 × 10⁵ cells were seeded in 10-cm dishes and allowed to adhere overnight at 37 °C. Cells were then cultured in DMEM, High Glucose, Pyruvate (Gibco, 41966029) supplemented with 10% FBS, 1% penicillin–streptomycin, 50 µg/ml freshly added uridine (Sigma, U3003), and 50 ng/ml ethidium bromide (Invitrogen, 15585011). The treatment was maintained for 4 weeks prior to clonal isolation. MtDNA depletion in the resulting clones was checked by qPCR as described above.

Mitochondrial superoxide assessment by EPR spectroscopy

The electron paramagnetic resonance (EPR) assay is based on the oxidation of Mito-TEMPO-H, a diamagnetic cyclic hydroxylamine that accumulates inside the mitochondria and is transformed into a paramagnetic nitroxide. To isolate the contribution of superoxide to the oxidation of the probe, PEG-SOD2 was used as a control as described previously (24,48–50). A detailed protocol is provided as Supplementary information.

Seahorse analyses

Oxygen consumption rates (OCR) were measured using a Seahorse XFe96 plate reader (Agilent). All assays were carried out using a seeding density of 10,000 cells/well in non-buffered DMEM, adjusted at pH 7.4. Mitochondrial function parameters (i.e. basal and maximal respirations and ATP production-linked OCR) were evaluated, with the XF Cell MitoStress Test (Agilent), after sequential treatment with 1 µM oligomycin, 1 µM FCCP and 0.5 µM rotenone/antimycin A, as previously described (51). ETC complex activity was assessed as previously reported (52), with slight modifications. Briefly, 10⁴ cells per well were first permeabilized with 1 nM Seahorse XF Plasma Membrane Permeabilizer (Agilent) and respiration was stimulated by adding 10 mM pyruvate, 5 mM malate, and 2 mM ADP. After the baseline scan, 100 nM rotenone was then added to inhibit complex I and 10 mM succinate was injected to establish complex II respiration. Last, 4 μM antimycin A was added to inhibit complex III and 100 μM N,N,N’,N’-Tetramethyl-p-phenylenediamine dihydrochloride (TMPD) + 10 mM ascorbate was injected as a complex IV substrate. OCR data were then normalized according to protein content in each experimental well.

Telomere restriction fragment (TRF) length analysis

TRF analyses were performed as described previously (53) on 2-5 µg of RNAse-treated genomic DNA digested with 20 U HinfI and 20 U RsaI and using [γ-³²P] ATP-labelled (TAACCC)₄ probe. Smart Ladder (Eurogentec) was run together with the samples, stained with ethidium bromide and imaged with a ruler. TRF were quantified using the online available WALTER (Web-based Analyser of the Length of Telomeres) software (https://www.ceitec.eu/chromatin-molecular-complexes/rg51/tab?tabId=125#WALTER).

TeSLA analysis

The procedure described previously (54) was followed for TeSLA analysis. Primer sequences are provided in Table S2.

Immunofluorescence and fluorescence in situ hybridization

Immunofluorescence and FISH experiments were performed as described previously (53) using the following antibodies and telomeric probe: rabbit anti-53BP1 (Novus Biologicals), Alexa 488 goat anti-rabbit (Invitrogen), mouse anti-Poly(ADP-ribose) (Millipore), Alexa 488 Goat anti-mouse (Abcam), 5’(TYE 563)GGGTtAGGGttAGgGTTAGGGttAGGGttAGGGtTA(TYE 563) (small letters indicate LNA^TM modified bases) (Qiagen). For the detection of parylated proteins, cells were grown in the presence of 1 μM PDD00017273 PARG inhibitor (Sigma-Aldrich). For telomeric FISH on metaphase spreads, cells were treated as described previously (55) and telomeres were probed with 5’(TYE 563)GGGTtAGGGttAGgGTTAGGGttAGGGttAGGGtTA(TYE 563). Images were acquired with the Cell Observer Spinning Disk confocal microscope (Zeiss) with 100× objective and analyzed using ImageJ software (National Institute of Health), while maintaining the same threshold for samples from the same experiment.

Liquid chromatography-mass spectrometry analysis of metabolites

6-well plates with cultured cells were rapidly washed twice with ice cold water and submerged in liquid nitrogen. 300 µL of a solution of 90 % methanol and 10 % chloroform were added, lysates were scraped and transferred to microcentrifuge tubes for centrifugation for 15 min at 4 °C and 22, 000 x g. The supernatants were dried in a Speed-Vac vacuum concentrating system and resuspended into 50 µL of 50:50 water:methanol solution. Analyses by LC-MS were performed as previously described (56) based on a method by Coulier et al. (57). A detailed protocol is provided as Supplementary information.

Dosage of extracellular glucose and lactate

Glucose and lactate concentrations were measured in extracellular media by using specific enzymatic assays and an ISCUSflex microdialysis analyzer (M Dialysis), as previously described (51).

RNA isolation and reverse transcription quantitative real-time PCR

RNA was extracted using TriPure Isolation Reagent (Sigma-Aldrich) according to the manufacturer’s instructions. For cDNA synthesis, 1 μg of RNA was retro-transcribed with 200 U MMLV-RT (Thermo Fisher Scientific), 0.2 µg random hexamers (Thermo Fisher Scientific), 20 U RiboLock RNase inhibitor (Thermo Fisher Scientific) and 2 mM dNTP mix. qPCR reactions were performed using KAPA SYBR FAST (Sigma-Aldrich) and the following program: 95°C for 10 min followed by 40 cycles of 95 °C for 15 sec and 60 °C for 30 sec. Primer sequences are provided in Table S2.

Telomerase repeat amplification protocol (TRAP) and droplet digital TRAP (ddTRAP)

For telomerase activity assessment in vitro using the TRAP assay, 500,000 cells were lysed in 100 µL CHAPS lysis buffer (TRAPeze S7700, Millipore) on ice for 30 min and proteins were quantified by Bradford assay (Bio-Rad). 500 ng of the lysis mixture were subjected to the TRAP assay following the Millipore kit instructions and TRAP products were separated on TBE/acrylamide:bisacrylamide (19:1) gel and visualized by staining with SYBR Gold Nucleic Acid GelStain (Invitrogen) with a PhosphoImager (Fujitsu). For ddTRAP (58), 500 ng of the lysis mixture were added to 50 µL of 1× TRAP buffer (200 mM Tris–HCl, pH 8.3, 15 mM MgCl₂, 2.5 mM each dNTPs, 200 mM TS primer (Table S2), and 0.4 mg/ml BSA), and extension reaction was performed by incubating the mix at 25°C for 40 min, followed by deactivation at 95°C for 5 min. For the ddPCR reaction, 3.75 μL of lysis-extension mix were added to 1x EvaGreen QX200 Super-mix (Bio-Rad), 50 nM TS primer, 50 nM ACX primer (Table S2) and H₂0 up to 25 µL. PCR reaction was performed as follows: initial denaturation for 5 min at 95°C followed by 40 cycles of 95°C/30 s, 54°C/30 s, and 72°C/30 s. Data were analyzed on Quantasoft software.

Statistics

Statistical analyses were performed using PRISM 8 software. Normality of the data was always checked before statistical tests. Details are provided in the Figure legends. Nomograms for Flow-FISH-based measurements of TL in the healthy population were drawn using the Age-related reference interval function of MedCalc Software (Belgium) based on methods described previously (59–61).

Data availability

Table S1 contains the numerical data used to generate the Flow-FISH curves of Figure 1A and Figure 1B.

Acknowledgements

We are grateful to all volunteers for providing blood samples. We thank Sandrine Horman, Marie Octave and Valentine Robaux for help with platelet purification and Géraldine Aubert and Mary Armanios for guidance in Flow-FISH. We are grateful to Victoria Van Regemorter and Caroline Huart for sharing unpublished data on TL in their cohort. We thank Chloé Despontin and Antoine Fattaccioli for technical support, Céline Coquette for help with blood collection and Mélina Vaurs, and Nausica Arnoult for critical reading of the manuscript. This work was supported by the Fonds National pour la Recherche Scientifique (FNRS) (Belgium), The Belgian Foundation against Cancer and WELBIO (Walloon Excellence in Life Sciences and Biotechnology) (grant X250624 to CC). Patrick Revy’s team is Labellisé Ligue (France). We are grateful to the de Duve Institute for constant support.

Additional information

Author contributions

M.M., J.-P.D. and A.D. generated the Flow-FISH curves. G.L., I.S., B.B. and A.D. recruited volunteers and helped collecting blood. M.M., B.M. and B.G. contributed to ROS measurement. M.M., E.R. and C.C. contributed to OCR and ETC complex activity measurement and to glycolytic flux evaluation. I.H. and G.B. contributed to metabolite analyses. E.F. and P.R. contributed to ddTRAP analyses. M.M. and T.A. contributed to transmitochondrial cybrid formation. G.L.B. contributed to PAR-FISH analysis. M.M. performed TRF, TeSLA, RT-qPCR, IF/FISH and qPCR analyses. A.D. and C.C. designed and supervised the study. A.D. wrote the paper.

Funding

Fonds National de la Recherche Scientifique, FNRS

• Anabelle Decottignies

Fonds National de la Recherche Scientifique, FNRS, Belgium (PDR-WEAVE: T.W017.23)

• Guido Bommer

Fonds National de la Recherche Scientifique, FNRS, Belgium (PDR: T.0034.24)

• Guido Bommer

Fonds National de la Recherche Scientifique, FNRS, Belgium (CDR J.0084.22)

• Bernard Gallez

Fonds National de la Recherche Scientifique, FNRS, Belgium (CDR J.0135.18)

• Bernard Gallez

Belgian Foundation against Cancer

• Cyril Corbet

WELBIO (X250624)

• Cyril Corbet

Labellisé Ligue

• Patrick Revy

Additional files

Supporting text, Figures S1 to S7, SI Reference, Tables S1 to S2