OPA1 mutation affects autophagy and triggers senescence in autosomal dominant optic atrophy plus fibroblasts

Abstract

In several cases of mitochondrial diseases, the underlying genetic and bioenergetic causes of reduced oxidative phosphorylation (OxPhos) in mitochondrial dysfunction are well understood. However, there is still limited knowledge about the specific cellular outcomes and factors involved for each gene and mutation, which contributes to the lack of effective treatments for these disorders. This study focused on fibroblasts from a patient with Autosomal Dominant Optic Atrophy (ADOA) plus syndrome harboring a mutation in the Optic Atrophy 1 (OPA1) gene. By combining functional and transcriptomic approaches, we investigated the mitochondrial function and identified cellular phenotypes associated with the disease. Our findings revealed that fibroblasts with the OPA1 mutation exhibited a disrupted mitochondrial network and function, leading to altered mitochondrial dynamics and reduced autophagic response. Additionally, we observed a premature senescence phenotype in these cells, suggesting a previously unexplored role of the OPA1 gene in inducing senescence in ADOA plus patients. This study provides novel insights into the mechanisms underlying mitochondrial dysfunction in ADOA plus and highlights the potential importance of senescence in disease progression.

OPA1, ADOA plus, Autophagy, mitophagy, senescence, mitochondria

Introduction

Mitochondria play a crucial role in maintaining cellular health as they are involved in various cellular functions and signaling pathways, serving as a central hub for numerous cellular activities [1]. A key aspect of their role is the continuous fusion and division, enabling the formation of a functional network within the cells. The fusion and fission of mitochondria are driven by guanosine triphosphatases (GTPases), including Optic Atrophy 1 (OPA1). OPA1 is instrumental in regulating mitochondrial fusion, the formation of cristae junctions, and their maintenance [2]. The membrane fusion process and the preservation of cristae architecture necessitate the participation of two OPA1 forms—a long uncut isoform (L-OPA1) and a short, soluble form (S-OPA1) [3]. L-OPA1, a 960 amino acid protein, is akin to proteins in the dynamin family, but with an added N-terminal mitochondrial localization sequence that aids its import into the inner mitochondrial membrane (IMM). The OPA1 gene (OMIM^*605290) spans over 60 kb on chromosome 3 and consists of 31 exons capable of producing eight protein isoforms through alternative splicing [4]. Previously, all identified OPA1 mutations were associated with Autosomal Dominant Optic Atrophy syndrome (ADOA), a prevalent form of mitochondrial optic neuropathy that starts in early childhood and results in progressive visual impairment due to Retinal Ganglion Cells (RGCs) degeneration [5]. In ADOA plus (OMIM #125250), individuals suffer from vision loss and additional symptoms [6, 7].

OPA1 mutations have been associated with reduced fusion activity, increased mitochondrial network fragmentation, and mitochondrial DNA (mtDNA) depletion [8, 9] at the cellular level due to their dominant negative effect [10–17]. Mutations in the GTPase domain heighten the risk of developing a severe multisystem disorder, especially with missense mutations compared to truncating mutations [18]. There has been a single reported case of a homozygous OPA1 mutation, which led to early onset encephalomyopathy, cardiomyopathy, and infant mortality [19].

Typically, when mitochondrial function is impaired, damaged mitochondria are removed through a process called mitophagy, which selectively degrades these damaged organelles [20]. In the case of OPA1 mutations, which result in increased mitochondrial fragmentation, there is an observed enhancement of the mitophagy response. This has been observed in fibroblasts [9, 16, 21] and RGCs isolated from OPA1 mutant mouse models [9, 16, 22–25]. Studies have shown that point mutations or deletion of OPA1 in mice can lead to uncontrolled mitophagy. Interestingly, inhibiting autophagy has been found to protect two types of neurons [24, 26]. On the other hand, some cases of OPA1 haploinsufficiency have been associated with a significant reduction in mitochondrial turnover [16, 27]. Different OPA1 mutations can have varying effects on the general autophagy process, and this variability in the autophagic response in DOA is attributed to the type or location of the mutations and their impact on mitochondrial network connectivity [16]. The specific signals that connect alterations in mitochondrial function and morphology to mitophagy and autophagy are still not fully understood, particularly in neurons, where defective mitophagy is a common feature of various neurodegenerative diseases [26].

It is important to note that these mechanistic pathways are interconnected, and disturbances in cell viability can result from the disruption of multiple key outcomes [25, 28]. During mitochondrial stress, multiple cross-pathways are activated in cells, including the induction of programmed cell death [29] and increased susceptibility to apoptosis [30, 31]. Several studies have demonstrated that cells that cannot produce S-OPA1 are more vulnerable to oxidative damage under obligatory respiratory conditions, which results in necrotic death [3]. A recent study revealed that fibroblasts with mutations in MFN2, another vital component of the mitochondrial fusion system, have increased cell proliferation by activating the mTORC2/AKT signaling pathway [32–34]. Additionally, other in vitro studies have shown that mitochondria-related processes, metabolites, and ROS can trigger senescence pathways and phenotypes [35].

In our study, we focused on the mitochondrial and cellular characteristics of fibroblasts carrying a c.124C > T/p.His42Tyr mutation in the mitochondrial targeting pre-sequence of OPA1. Cultured primary fibroblasts with an OPA1 mutation in the N-terminal domain have never been reported by previous studies, as most studies focus their attention on the GTPase domain. Our investigation aimed to verify whether the presence of the mutation in the patient’s fibroblasts significantly affected mitochondrial morphology and function. Mutated fibroblast, not only displayed a compromised mitochondrial network and function but also showed decreased autophagy and premature senescence, impairing both mitochondrial and extra-mitochondrial activities.

Results

Genetic investigation of ADOA plus patient

The OPA1 missense variant c.124C > T/p.His42Tyr has been reported by Baldacci et al. (2022) and has been recently submitted by a single reporter in the ClinVar database (ID = 1000915). The variant is in exon 2 and affects a conserved amino acid in the OPA1 protein as predicted by various evolutionary conservation scores (GERP score: 5, phyloP100way_vertebrate 4.1, phastCons100way_vertebrate 1, Fig. 1A). The variant has evidence of pathogenicity, such as PM2 (Pathogenic Moderate criteria 2; absent from controls in ExAC), PP2 (Pathogenic Supporting criteria 2; missense in a gene where most variants are loss-of-function), and pathogenic tool queries (Sift: 0—Affect protein function; MutationTaster: 0.9969—Disease causing; Align-GVGD: High deleterious risk; MetaLR: 0.71—Damaging; Mutation Assessor: 0.581—Medium) that suggest the variant affects protein function. However, it also has some evidence of benignity, such as PP3 (Pathogenic Supporting Criteria 3; multiple lines of computational evidence suggest no impact on gene or gene product) and BP4 (Benign Supporting Criteria 4; a silent variant with no predicted impact on splicing). Therefore, the variant is classified as ‘Uncertain Significance’ because it does not meet the rules for combining criteria to reach a higher or lower classification.

Figure 1

Missense OPA1 variant (c.124C > T/p.His42Tyr) in a conserved region containing matrix targeting sequences (MTS). (A) Aminoacidic conservation plot from UCSC Genome browser (https://genome.ucsc.edu/), showing the result of PhyloP conservation score calculated multiple alignments of 100 vertebrate genomes, the result of multiple alignments of 30 mammals. The 42^nd codon of OPA1 encodes a highly conserved histidine residue. (B) The highlighted variant of interest p.His42Tyr is found in a region of high MTS propensity (confidence score > 4). This variant is in proximity to the cleavage site predicted by MitoFates and TargetP sites. (C) Representative Western Blot image of OPA1 and MFN2 in total lysates from control and patient-derived cells in basal conditions. Each band was normalized to GAPDH band density and a densitometrical analysis was performed. Densitometrical analysis showed total OPA1 (l-OPA1 + s-OPA1) levels and ratios of the l-OPA1 to the s-OPA1 isoforms and MFN2 in untreated (Untr) conditions. Data are presented as mean ± SD (n = 3). Student’s t-test; ^*P < 0.05; ^**P < 0.01 and ^***P < 0.001, ^****P < 0.0001. (D) RNA-seq reads from OPA1^His42Tyr fibroblasts were used to assess the ratio of mutant mRNA to wt mRNA. Out of 164 total reads, 74 (45%) supported the mutated variant T, and 89 (54%) supported the wild-type nucleotide C. (E) 100% RNA-seq reads from control fibroblasts supported the wild-type. Alt text: Panel A displays a plot of the genome alignments of 100 vertebrate species, which reveals that the 42nd codon of OPA1 encodes a highly conserved histidine residue. Panel B is a line graph that represents the missense OPA1 variant, situated in a region with high matrix targeting sequence propensity. Panel C contains bar graphs indicating that there are no differences in the protein abundances of OPA1 and MFN2 in OPA1-mutated fibroblasts compared to the control. Panels D and E demonstrate that RNA-sequencing reads support the mutated variant in OPA1-mutated fibroblasts.

Given the fundamental role of the N-pre-sequence terminal for mitochondrial protein import [36], in silico analysis was performed to estimate the probable consequence of the p.His42Tyr mutation. The MitoFates tool [37] predicted that the location of the mutation was approximately 38 amino acids from the Translocase of the Outer Membrane (TOM) recognition motif, potentially allowing the import of mutated proteins into the mitochondrion. Furthermore, we generated an iMLP curve with MTSviewer [38], these are curves that show the propensity of internal matrix targeting-like sequences (iMTS-Ls) in mitochondrial proteins. iMTS-Ls are regions within the protein that share structural features with matrix targeting sequences (MTSs), which direct proteins from the cytosol into mitochondria [39]. The curve is generated by a deep learning approach and is based on the long short-term memory (LSTM) recurrent neural network architecture. The higher the iMLP score, the more likely the region is an iMTS-L. Considering the OPA1 gene, the variant of interest p.His42Tyr is found within a region of high MTS propensity (confidence score > 4) near the predicted MitoFates and the TargetP cleavage site (Fig. 1B).

OPA1  ^His42Tyr fibroblasts have normal OPA1 and increased MFN2 protein level

We examined whether fibroblasts with the OPA1^His42Tyr mutation exhibited changes in OPA1 protein expression levels. Western blot analysis of whole-cell lysates from primary fibroblasts carrying the heterozygous missense p.His42Tyr mutation showed an equivalent amount of total OPA1 protein level, as well as a normalized l-OPA1:s-OPA1 ratio for GAPDH. We also measured the steady-state levels of MFN2 protein, another crucial component of mitochondrial fusion, in total homogenates from OPA1-mutated and normal control cells using western blot. The total MFN2 protein level was significantly higher (about a 1.4-fold change) in fibroblasts with the OPA1^His42Tyr mutation compared to the control (Fig. 1C).

Alteration of mitochondrial morphology and membrane potential, and increased ROS production in OPA1^His42Tyr fibroblasts

We investigated the appearance of mitochondria in OPA1^His42Tyr compared to control fibroblasts. The cells were grown in either glucose or galactose medium, with galactose specifically highlighting any potential issues with the OxPhos system. We categorized the fibroblasts into three groups based on their mitochondrial shape: tubular, intermediate, and fully fragmented. The OPA1^His42Tyr fibroblasts showed striking differences compared to the control cells in terms of both shape and the ability to restore the normal mitochondrial network. In the glucose medium, the OPA1^His42Tyr fibroblasts had a higher percentage of intermediate mitochondria, while in the galactose medium, they had a higher percentage of both intermediate and fragmented mitochondria. In contrast, the control fibroblasts maintained a filamentous and interconnected mitochondrial network. To assess the ability of cells to restore a normal mitochondrial network, we treated both mutant and control cells with carbonyl cyanide m-chlorophenyl hydrazine (CCCP). This treatment caused changes in the shape of the mitochondria, resulting in a decrease in the population of intermediate mitochondria and a significant increase in fragmented mitochondria in OPA1^His42Tyr fibroblasts compared to control fibroblasts. After removing the CCCP, we observed a reduction in the population of fragmented mitochondria in OPA1^His42Tyr fibroblasts, although the recovery did not fully return to the pre-CCCP level in the glucose medium. The fragmentation of the mitochondrial network and the inability to restore its normal shape were even more pronounced in the galactose medium (Fig. 2A). This fragmented network morphology in OPA1 fibroblasts may be a direct result of excessive fission events without proper fusion, which can lead to a decrease in the membrane potential (ΔψM) of the mitochondria [40]. To evaluate the mitochondrial membrane potential, we stained the cultured fibroblasts with a fluorescent dye called JC-1 (Fig. 2B). In OPA1^His42Tyr cells, the red/green fluorescence intensity ratio was approximately 40% lower compared to control cells, indicating mitochondrial depolarization in OPA1^His42Tyr fibroblasts.

Figure 2

Mitochondrial morphology, membrane potential and bioenergetics are altered in OPA1^His42Tyr fibroblasts. (A) Primary fibroblast cultures were incubated under normal conditions (DMEM-glucose or DMEM-galactose medium) and treated for 4 h with 10 μM of CCCP (CCCP4h) and with CCCP followed by 4 h in the initial medium (CCCP4h + Re). Bar graphs show fibroblast distribution incubated in DMEM-glucose or DMEM-galactose medium into three different categories based on mitochondrial morphology: tubular, intermediate and fragmented mitochondria. Data are presented as mean ± SD of three independent experiments (n = 3). Student’s t-test; ^*P < 0.05; ^**P < 0.01; ^***P < 0.001 and ^****P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed to compare the mean of fragmented mitochondria following CCCP induction or CCCP with Re to the mean of fragmented mitochondria in the base condition in OPA1^His42Tyr fibroblasts. C: Control fibroblasts; O: OPA1^His42Tyr fibroblasts. (B) Mitochondrial membrane potential (ΔΨm) was measured in control and OPA1^His42Tyr via JC-1 staining. Mitochondrial membrane depolarizationis represented by an increase of the fluorescence intensity ratio.Scale bars equal to 25 μm. (C) Measurement of endogenous respiration rates in digitonin-permeabilized cells was performed to compare OCRs in control and OPA1^His42Tyr fibroblasts. OCRs were measured under basal conditions (routine respiration) and substrate-driven conditions, as described in Materials and Methods. Bars represent the mean respiration rates expressed as nmoles O₂/min/mg protein. Data are presented as mean ± SD of three independent experiments (n = 3). CI: Complex I; CII: Complex II; CIII: Complex III; CIV: Complex IV. (D) Bar graph quantifying by fluorescence microplate reader the ROS formation in fibroblasts loaded with ROS detection reagents in untreated conditions (DMF, vehicle) and treated with pyocyanin (PYO). (E) Representative Western Blot image of mtTFA, VDAC1 and subunits of complex I (ND4 and NDUFS1), complex II (SDHA) and complex IV (COX IV) in total lysates from control and patient-derived cells. Densitometrical analysis was performed normalizing each band to GAPDH band density and it’s shown in the clustered bar. (F) Evaluation of mtDNA integrity performed by long-range PCR resulted in a 16.5 kb PCR fragment representing a full-sized mitochondrial genome in both control and OPA1^His42Tyr fibroblasts, as visualized on 0.8% agarose gel. (G) qRT-PCR analysis of mtDNA measures mtDNA copy number expressed as mtDNA: nDNA. This ratio is indicative of the relative number of mtDNA per cell. For more details, see Materials and Methods. Data are presented as mean ± SD of three independent experiments (n = 3). Student’s t-test; ^*P < 0.05; ^**P < 0.01 and ^***P < 0.001, ^****P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Panel A displays a bar graph that demonstrates the classification of mitochondrial morphology in the OPA1 mutated and control fibroblast grown in glucose or galactose medium. Panel B shows a bar graph of the mitochondrial membrane depolarization found in OPA1-mutated cells compared to the control. Panel C is a bar graph that exhibits reduced oxygen consumption by respiratory chain complexes in OPA1-mutated cells compared to control cells. Panel D illustrates the increased ROS production in OPA1-mutated compared to control cells. Panel E shows an increased level of mitochondrial transcription factor A (mtTFA) and some oxidative phosphorylation complexes subunits in OPA1-mutated compared to control cells. Finally, Panel F and G demonstrate normal mitochondrial DNA integrity and increased mitochondrial DNA copy number in OPA1-mutated cells compared to control.

As mutations in the OPA1 gene are known to interfere with mitochondrial respiration [16], we evaluated the mitochondrial respiratory function by using an oxygen electrode to measure the oxygen consumption rate (OCR) in both control and patient cells that had been permeabilized with digitonin. We calculated the basal respiration rate and the maximal respiration rate following the addition of the uncoupling agent 2,4-dinitrophenol. Our findings showed a diminished bioenergetic capacity in OPA1^His42Tyr fibroblasts, with a 25% decrease in baseline respiration and reductions of approximately 55% for Complex I, 50% for Complex II + III, and 80% for Complex IV activities compared to control cells (Fig. 2C).

We considered that defects in Oxphos function may result in the pathological production of ROS, an important by-product of mitochondrial oxidative metabolism. We examined ROS production in both OPA1^His42Tyr and control fibroblasts. Our findings revealed an increase in ROS levels in patient fibroblasts under basal conditions (about a 2.39-fold change). When ROS production was stimulated by the introduction of pyocyanin, ROS production levels significantly increased in mutated compared to the control fibroblasts (an approximately 1.8-fold change) (Fig. 2D).

To further investigate mitochondrial bioenergetics in fibroblasts, we determined the levels of mitochondrial OxPhos complexes in OPA1^His42Tyr fibroblasts using western blotting (Fig. 2E). We used antibodies against selected subunits of complex I (ND4 and NDUFS1), complex II (SDHA), and complex IV (COXIV). Interestingly, OPA1^His42Tyr cells exhibited a significant increase in the expression of nuclear-encoded OxPhos subunits NDUFS1, SDHA, and COX IV (about 1.9, 1.84, and 1.99-fold change, respectively), as well as the mitochondria-encoded OxPhos subunit ND4 (an approximately 1.39-fold change), compared to control cells. Additionally, we measured the expression of the key regulator factor of mtDNA transcription and replication, the Mitochondrial transcription factor A (mtTFA or Tfam). Consistent with the elevated expression levels of certain nuclear and mitochondrial-encoded OxPhos complexes, we observed an increased level of mtTFA protein (approximately a 2.05-fold change) in OPA1^His42Tyr compared to control fibroblasts (Fig. 2E). Additionally, we assessed the level of VDAC1, which serves as an indicator of mitochondrial mass and found no significant changes in OPA1 fibroblasts compared to the control.

As OPA1 plays a role in reshaping cristae and coordinating mitochondrial bioenergetics, which in turn regulates the maintenance of mitochondrial DNA, we decided to examine the mitochondrial genome. We used a long-range PCR assay to evaluate the integrity of the mitochondrial genome in fibroblasts derived from the patient. The PCR amplification of total DNA from OPA1^His42Tyr fibroblasts produced major products of 16 080 bp, corresponding to the full-length intact genomic mtDNA (Fig. 2F). These findings suggest that there were no significant rearrangements in these cells. However, it’s important to note that skin fibroblast cultures may not be the most effective method for identifying major mtDNA deletions. Additionally, we performed quantitative PCR to measure the ratio of mitochondrial to nuclear DNA (mtDNA/nDNA) in OPA1^His42Tyr fibroblasts. The mtDNA/nDNA ratio was slightly higher in OPA1^His42Tyr compared to control fibroblasts (Fig. 2G).

OPA1^His42Tyr fibroblasts exhibit hindered autophagy

To investigate the autophagy process in OPA1^His42Tyr fibroblasts, we examined the levels of microtubule-associated protein 1 light chain 3 (MAP1LC3/LC3) and SQSTM1/p62, both of which are recognized markers of autophagy and related structures. When autophagy is induced, LC3B-I attaches to phosphatidylethanolamine on the autophagosomes’ surface, forming LC3B-II. The quantity of LC3B-II is indicative of the number of autophagic vesicles/vacuoles; hence, it is used as an autophagy marker [41]. SQSTM1/p62 can bind LC3 and the proteins targeted for degradation in lysosomes. Typically, SQSTM1/p62 is significantly reduced when autophagic activity is heightened, but it accumulates in the cytoplasm when autophagy is compromised.

Autophagy levels were evaluated through Western blot analysis under basal conditions. Due to the potential sensitivity and instability of LC3-I compared to LC3-II [42], the levels of LC3-II were compared to GAPDH rather than LC3-I. OPA1^His42Tyr fibroblasts exhibited reduced expression of both LC3B-II and SQSTM1 compared to control fibroblasts (lane 1 vs. lane 5). As autophagic flux is a dynamic and multi-step process, the decreased expression of LC3B-II could indicate either an induction of LC3B-II degradation by activated autolysosomes or an inhibition of LC3B-II synthesis. To investigate this further, we used chloroquine, which neutralizes lysosomal pH and prevents lysosomal degradation [42]. This inhibition halts the autophagic process and allows for the measurement of intrinsic autophagy in cells. Even in the presence of chloroquine, LC3B-II levels were lower in OPA1^His42Tyr compared to control cells (approximately 36% lower) (lane 2 vs. lane 6), suggesting a reduction in autophagosome formation at upstream steps rather than an increase in LC3 turnover (Fig. 3). To further assess the autophagic process, we induced autophagy using torin1, an ATP-competitive inhibitor of the mechanistic target of rapamycin kinase complex (mTORC). Following autophagy induction, OPA1^His42Tyr fibroblasts showed higher levels of LC3B-II compared to baseline conditions (lane 5 vs. lane 7), like control fibroblasts (lane 1 vs. lane 3), indicating functional autophagic machinery. However, torin1 induction resulted in a 32% decrease in LC3B-II levels in OPA1^His42Tyr fibroblasts compared to control fibroblasts (lane 3 vs. lane 7), suggesting reduced autophagosome formation in the mutant cells. These findings confirmed the data obtained under basal and chloroquine conditions, indicating an overall reduction in autophagosome formation in OPA1^His42Tyr fibroblasts. Furthermore, torin1-induced autophagy in the presence of chloroquine led to a significantly lower level of LC3B-II (approximately 33% decrease) in OPA1^His42Tyr fibroblasts compared to control fibroblasts (lane 4 vs. lane 8). This confirmed that the mutant cells have impaired autophagy in the early stages of the process. However, the decrease in LC3B-II levels was not accompanied by the accumulation of SQSTM1, which is an indicator of autophagic degradation. It is worth noting that changes in SQSTM1 levels can be cell-type and context-specific and may be influenced by autophagy-independent mechanisms [42]. In some cell types, there may be no change in the overall amount of SQSTM1 despite strong autophagy induction [42] as observed in our case. Additionally, a significant loss of SQSTM1 does not always correlate with increased autophagic flux [43], as observed in our study (Fig. 3).

Figure 3

Autophagosome synthesis is reduced in OPA1^His42Tyr fibroblasts. Representative Western Blot for the investigation of autophagy-related protein expression in OPA1^His42Tyr fibroblasts. Fibroblast cells were incubated under normal conditions or treated with 50 μM CQ for 3 h with or without 1 μM Torin1. Western blotting was performed for LC3B and SQSTM1 in total cell lysates from control and patient-derived cells for each condition. Each band was normalized to GAPDH band density and densitometrical analysis of LC3B-II and SQSTM1 levels for each condition was performed. Data are presented as mean ± SD, n = 3. Student’s t-test; ^*P < 0.05; ^**P < 0.01 and ^***P < 0.001, ^****P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Western blot experiment showing decreased expression of the autophagy-related markers LC3B and SQSTM1 in OPA1 mutated compared to control fibroblasts.

Evaluation of mitophagy in OPA1^His42Tyr fibroblasts

To investigate whether impaired autophagosome formation affects mitophagy flux, we examined mitophagosome and autolysosome formation in OPA1^His42Tyr fibroblasts. Mitophagy was induced using CCCP, and the number of mitophagosomes was assessed through immunofluorescent labeling with LC3B and TOMM20 (Translocase of Outer Mitochondrial Membrane 20) antibodies. In OPA1^His42Tyr fibroblasts, autophagosomes effectively engulfed mitochondria, as indicated by an increase in the LC3B/TOMM20 (LTc): TOMM20 (T) points/cell under baseline conditions compared to control cells (about a 2.8-fold change). This suggests that although there were fewer autophagosomes, a higher proportion of them were involved in engulfing mitochondria in OPA1^His42Tyr fibroblasts. Additionally, the LTc: T dots/cell significantly increased after CCCP treatment in both OPA1^His42Tyr and control fibroblasts compared to basal conditions, indicating that mitophagosome formation was increased but not maximally in OPA1^His42Tyr fibroblasts. Interestingly, the number of mitophagosomes decreased after CCCP treatment and subsequent recovery, suggesting a reduction in mitophagosome formation compared to CCCP treatment alone (Fig. 4A).

Figure 4

Mitophagosomes and autolysosomes can form normally in OPA1^His42Tyr cells. Fibroblasts were incubated under normal conditions (Untreated), treated for 4 h with 10 μM of CCCP (CCCP 4 h) and treated with CCCP followed by 4 h in the initial medium (CCCP4h + Re). (A) Cells were fixed and labeled with anti-LC3B and anti-TOMM20 and counterstained with DAPI for each condition. Quantification of the ratio between LC3B and TOMM20 colocalization (LTc, in yellow) and TOMM20 (T) expressed as LTc:T dots/cell. (B) Cells were fixed and labeled with anti-LC3B and anti-LAMP1 and counterstained with DAPI for each condition. Quantification of the ratio between LC3B and LAMP1 colocalization (LLpc) and LAMP1 (Lp) expressed as LLpc: Lp dots/cell. Scale bars equal to 25 μm. Values are expressed as mean ± SD (n = 3 independent experiments). Student’s t-test; ^*P < 0.05; ^**P < 0.01 and ^***P < 0.001, ^****P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Immunofluorescence experiment shows similar formation of mitophagosomes and autolysosomes in OPA1 mutated and control fibroblasts.

To further understand autolysosome formation under mitochondrial stress conditions, we performed immunolabeling with LC3B and LAMP1 (lysosomal-associated membrane protein 1) during CCCP treatment. The ratio of LC3B/LAMP1 (LLpc): LAMP1 (Lp) dots/cell, representing the fusion of lysosomes with autophagosomes, did not significantly change in OPA1^His42Tyr fibroblasts compared to control fibroblasts. This confirms that the decrease in LC3B-II levels observed in Western blot analysis was not due to increased autolysosome formation. Moreover, the LLpc: Lp dots/cell significantly increased after CCCP treatment in both OPA1^His42Tyr and control fibroblasts, indicating that autolysosome formation was not maximally activated in OPA1^His42Tyr fibroblasts (Fig. 4B). These findings suggest that when mitophagy is stimulated, the formation of mitophagosomes and autolysosomes in OPA1^His42Tyr fibroblasts occurs similarly to control fibroblasts.

Furthermore, the colocalization of LC3B and parkin RBR E3 ubiquitin-protein ligase (PRKN) was increased but not maximally in untreated OPA1^His42Tyr fibroblasts compared to control cells (about 1.73-fold change) (Fig. 5A and D). We also examined the expression and localization of proteins involved in the mitochondrial recycling process, PTEN-induced kinase 1 (PINK1) and PRKN. The number of PINK1-recruited mitochondria, indicated by PINK1/TOMM20 colocalization (PTc): PINK1 (P) dots/cell (Fig. 5B and E), was increased in untreated OPA1^His42Tyr fibroblasts compared to control cells (about 1.6-fold change) and did not change upon CCCP treatment. Additionally, the colocalization of PINK1 and PRKN (PPrc) dots/cell was increased in untreated OPA1^His42Tyr fibroblasts compared to control cells (about a 3-fold change) and remained unchanged after CCCP treatment (Fig. 5C and F). Overall, these results suggest the correct recruitment of mitophagy markers in OPA1^His42Tyr fibroblasts. However, the reduced capacity for autophagosome synthesis in OPA1^His42Tyr fibroblasts may limit mitophagy, as the mitophagy rate was not maximal in the mutated cells.

Figure 5

The recruitment of mitophagy markers on mitochondria occurs normally in OPA1^His42Tyr. Fibroblasts were incubated under normal conditions (Untreated) and treated for 4 h with 10 μM of CCCP (CCCP 4 h), fixed and labeled with anti-LC3B and anti-PRKN (A); anti-PINK1 and anti-TOMM20 (B); or anti-PINK1 and anti-PRKN (C). Scale bars equal to 25 μm. (D) Quantification of LC3B and PRKN colocalization (LPc) expressed as LPc dots/cell. (E) Quantification of the ratio between PINK1-TOMM20 colocalization (PTc) and PINK1 (P) expressed as PTc:P dots/cell. (F) Quantification of PINK1 and PRKN colocalization (PPrc) expressed as PPrc dots/cell. Values are expressed as mean ± S.D. (n = 3 independent experiments). Student’s t-test; ^*P < 0.05; ^**P < 0.01; ^***P < 0.001, ^****P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Immunofluorescence experiment shows proper mitophagy marker recruitment on mitochondria and correct association of PRKN-labeled mitochondria with autophagosomes in OPA1-mutated fibroblasts.

Transcriptome profile of OPA1^His42Tyr fibroblasts

To elucidate molecular pathways in mutated OPA1^His42Tyr fibroblasts, we performed whole RNA sequencing of OPA1^His42Tyr  versus control fibroblasts. We found 644 DEG with a generalized fold change (GFOLD) > 1.5. These genes were mainly enriched in the biological processes (BP), cellular components (CC), and molecular functions (MF) involved in cell adhesion, extracellular matrix organization and central nervous system development. The CC of most DE genes was the extracellular region, extracellular matrix, an integral component of the plasma membrane, and extracellular space and their MF was the extracellular matrix structural constituent (Table S1). No significant KEGG pathways were associated with the DEG. Protein-protein interaction (PPI) network analysis was constructed with STRING which integrates both known and predicted PPIs and can be applied to predict functional interactions of proteins [44, 45]. A network was generated containing 384 nodes and 1020 interactions with a minimum required interaction score > 0.4 and only query proteins were displayed (Fig. S1); the most significant three sub-modules of the network (MCODE score > 5) have been evidenced with pink circles. BiNGO analysis showed that the first sub-module was enriched with statistically significant GO Biological Processes mostly involved in the cell cycle, microtubule-based process, M phase, microtubule cytoskeleton organization, and mitotic cell cycle, all of them were found to be downregulated (Table S2). The downregulation of these key cell cycle processes is typically associated with reduced proliferation and is suggestive of cell cycle arrest. The second cluster involves processes related to the nucleoside monophosphate catabolic process and cyclic nucleotide catabolic processes. Up-regulated genes in this cluster include phosphodiesterases (PDEs) such as PDE3A, PDE10A, and PDE5A (Fig. S1 and Table S2). These enzymes hydrolyze cyclic adenosine monophosphate (cAMP), which has been shown to regulate various physiological processes, including autophagy. Additionally, PDE5 specifically hydrolyzes cyclic guanosine monophosphate (cGMP) [46, 47]. The third cluster contains HIST1H1D, HIST1H2AJ, HIST1H3J, HIST1H1A, and HIST1H1B genes, which belong to the HIST1 cluster and appear to consist of silenced genes associated with chromatin assembly, nucleosome organization, protein-DNA complex assembly, nucleosome assembly, and DNA packaging. The downregulation of the HIST1 cluster has been linked to the presence of a senescent cell phenotype [48] (Fig. S1 and Table S2). All clusters have been reported in Table S2.

We also used RNA-seq data to assess the ratio of mutant to wild-type mRNA and to infer the likelihood of mutant protein expression. We analyzed the reads obtained from OPA1^His42Tyr fibroblasts and identified the variant in the OPA1^His42Tyr mRNA. Specifically, out of a total of 164 reads, 74 (45%) exhibited the mutated variant T, and 89 (54%) displayed the wild-type nucleotide (Fig. 1D). In contrast, mRNA from control fibroblasts showed no mutation (Fig. 1E). The presence of the mutant allele was further confirmed through Sanger sequencing using the reverse primer of the PCR-amplified cDNA region encompassing the mutation (data not shown).

Phosphorylation of Ser473 in the protein kinase AKT is reduced in fibroblasts with the OPA1^His42Tyr mutation

Transcriptomic analysis revealed that genes involved in cell cycle and proliferation were downregulated, prompting us to investigate the activation status of AKT, a crucial regulator of cell survival and proliferation [49]. Based on the OPA1^His42Tyr gene expression signature, AKT was predicted to be inactivated (Blue shaded nodes; Fig. S2). The full activity of AKT relies on its total protein concentration and the phosphorylation levels of Thr308 and Ser473 residues, which reflect the activity of upstream kinases PDK1 and mTORC2, respectively [50–56]. Our findings showed an increase in total AKT levels in mutant fibroblasts (about 57%). The phosphorylation of AKT at Thr308 was not significantly changed. However, the phosphorylation of AKT at Ser473 was significantly reduced compared to control cells (about 49%) (Fig. 6A). The significant reduction in AKT phosphorylation at Ser473 is indicative of mTORC2 inactivation.

Figure 6

OPA1  ^His42Tyr fibroblasts show decreased phosphorylation of AKT(Ser473), low growth rate and senescence activation. (A) Representative Western Blot images of p-AKT and t-AKT in total cell lysates from control and patient-derived cells in basal conditions. Each band was normalized to GAPDH band density and densitometrical analysis of p-AKT: t-AKT was performed. (B) Growth curves and population doubling time and level for OPA1^His42Tyr, and control fibroblasts grown in 24-well plates over 3 days. 6000 cells were seeded in 10% FBS serum supplemented with complete DMEM culture media. Doubling time was calculated from day 1 to day 2 by cell counting method and using the following formula: PD = t × Log2/(LogC2 − LogC1). (PD = Population doubling time, t = 24 h (interval of cell count), Log = 10 based Log, C1 = 1st cell count, C2 = 2nd cell count). The doubling level (PDL) was calculated from day 0–3 according to the equation: PDL = 3.32 × logN/No (N = harvest cell number and No= initial cell number). (C) The quantification of SA-βGal activity-positive cells. (D) Representative Western Blot image of p53 and p21 in total lysates from control and patient-derived cells. Each band was normalized to GAPDH band density and a densitometrical analysis was performed. All values shown are the mean of three independent experiments ± SD. Student’’s t-test; significance is indicated within the figures using the following scale: ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Graphs show that OPA1 mutation reduces growth rate, increases senescence, and enhances the p53/p21 pathway compared to control while exhibiting decreased AKT(Ser473) phosphorylation.

Fibroblasts carrying the OPA1^His42Tyr mutation exhibit a reduced growth rate and an increase in the proportion of senescent cells

Based on transcriptomic data, we have formulated a hypothesis that cell proliferation is impaired in OPA1^His42Tyr fibroblasts. This conclusion is based on the observation that many genes associated with the cell cycle process were found to be downregulated (Table S2 and Fig. S2). Additionally, both the transcriptomic data and the experimental results indicate a decreased AKT activation status. This, in turn, suggests a lower activity of mTORC2, which is a crucial regulator of cell survival, growth, and proliferation. To further explore this, we conducted a growth curve analysis comparing OPA1^His42Tyr and control fibroblasts in a glucose medium. We observed a delayed and reduced entry into the log phase in OPA1^His42Tyr compared to control cells. From day 1 to 3 after seeding, OPA1^His42Tyr fibroblasts had a significantly lower cell number than control cells. The doubling time of fibroblasts was approximately 55% longer, while the doubling level was about 47% lower compared to the control (Fig. 6B).

Since an increase in population doubling time is observed in arrested fibroblasts, we investigated cellular senescence (i.e. the permanent arrest of cell proliferation) [57]. To investigate cell senescence, we utilized the SA-βGal assay, which specifically detects senescent cells by detecting the production of senescence-associated β-galactosidase (SA-βGal) [58]. OPA1^His42Tyr cells exhibited a higher rate of senescence, as indicated by an increase in the percentage of blue-stained cells by β-Gal (Fig. 6C and Fig. S3A).

Based on the OPA1^His42Tyr gene expression signature, p21WAF1/CIP1(CDKN1A) and p53 were predicted to be activated (orange shaded nodes; Fig. S2). Consequently, the activation of the p53/p21WAF1/CIP1 pathway is known to mediate cell-cycle arrest in senescence [59]. p21WAF1/CIP1 acts downstream of p53 ultimately inhibiting cell cycle progression [60]. Therefore, we examined the levels of p53 and p21WAF1/CIP1 in OPA1^His42Tyr fibroblasts. Our results showed increased levels of p53 and p21 proteins in OPA1^His42Tyr fibroblasts compared to control cells (of about 57% and 32% respectively) (Fig. 6D). Collectively, these findings indicate that OPA1^His42Tyr cells exhibit enhanced activation of the senescence pathway and display an early senescent phenotype.

Pharmacological suppression of mTORC1 restores the initiation of autophagy and prevents OPA1^His42Tyr fibroblasts from becoming senescent

The mTOR pathway, known for regulating a broad range of crucial cellular functions such as cell growth, proliferation, survival, and autophagy, has been associated with lifespan extension and the enhancement of aging-related processes like cellular senescence [61]. The beneficial effects on lifespan attributed to mTOR inhibitors have been linked to the inhibition of mTORC1, while the inhibition of mTORC2 could potentially be detrimental [62, 63]. Given that OPA1^His42Tyr fibroblasts exhibited diminished autophagy, decreased levels of AKT (Ser473) phosphorylation (an indicator of mTORC2 activity), and increased senescence, we assumed that the pharmacological inhibition of mTORC1 could modulate these defects. This could be achieved by leveraging the molecular negative feedback mechanism that regulates mTORC2 activity. It has been shown that there is a reciprocal regulation of the different mTOR complexes [64], where an increase in complex-1 triggers the downregulation of complex-2, impacting downstream mTORC2 signaling. At the same time, inhibition can stimulate mTORC2 activity [65, 66].

To specifically inhibit mTORC1, both control and ^OPA1His42Tyr fibroblasts were exposed to Everolimus at a concentration of 5 μM for 3 h (Fig. 7A and Fig. S3B). The inhibition of mTORC1 by everolimus led to a simultaneous increase in mTORC2 activity, as indicated by the phosphorylation of AKT on Ser473 and a decrease in t-AKT level in both control and OPA1-mutated fibroblasts; in particular, we found an increase in the p-AKT(Ser473):t-AKT ratio in both control and OPA1-mutated fibroblasts (about 25-fold and 4.5-fold change, respectively). This suggests that the drug treatment selectively targeted mTORC1 rather than mTORC2.

Figure 7

Pharmacological mTORC1 inhibition restores autophagy and protects OPA1^His42Tyr fibroblasts against senescence. (A) Representative Western Blot images of p-AKT (Ser473), t-AKT, LC3B, p53 and p21 in total cell lysates from control and patient-derived cells treated both with vehicle (DMSO) and 5 μM Everolimus for 3 h. Each band was normalized to GAPDH band density and a densitometrical analysis was performed. (B) The tables represent the quantification of total cell number, SA-β-gal⁺ cells number and SA-β-gal⁺ cells percentage in control and patient-derived cells treated both with vehicle (DMSO) and 5 μM everolimus. All values shown are the mean of three independent experiments ± SD. Student’s t-test; significance is indicated within the figures using the following scale: ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001. A one-way ANOVA test with Bonferroni’s correction was performed for multiple comparisons. Alt text: Panel A displays bar graphs which indicate that when OPA1-mutated fibroblasts are treated with mTORC1 inhibitor everolimus, there is an increase in AKT(Ser473) phosphorylation and p53 expression, a decrease in p21 levels, and a restoration in the expression of autophagosome protein LC3B-II. In Panel B, there is a table that shows the decline in the number of β-galactosidase-positive cells in OPA1-mutated fibroblasts because of everolimus treatment.

We examined the expression of the autophagosome marker LC3B-II in cells treated with either everolimus or the vehicle only (DMSO). The results demonstrated a significant upregulation of LC3B-II expression upon everolimus treatment, indicating an enhancement in autophagosome formation. Additionally, we analyzed the expression of markers associated with cell cycle arrest and senescence in both everolimus and DMSO-treated cells. Western blot analysis was performed to assess the levels of p53 and p21 in the fibroblast cell lines. The findings revealed that everolimus significantly increased p53 expression in both control and OPA1-mutated cells (of about 88% and 31%, respectively). Conversely, there was a downregulation of downstream target p21 levels in response to everolimus treatment in both cell types (about 73% in control and 23% in OPA1-mutated fibroblasts). The Western blot analyses are presented in Fig. 7A.

To evaluate senescence, cells treated with everolimus or DMSO from both control and OPA1-mutated fibroblasts were stained for β-galactosidase activity. In everolimus-treated cells, a decrease in the number of β-galactosidase-positive cells was observed (Fig. 7B and Fig. S3B). Furthermore, the total number of cells treated with everolimus was reduced after 2 days compared to DMSO-treated cells (Fig. 7B). These findings demonstrate that everolimus treatment protected OPA1^His42Tyr fibroblasts from senescence without stimulating proliferation.

Discussion

In this study, we examined primary fibroblasts from a patient with ADOA plus, who had a c.124C > T/p.His42Tyr mutation in the OPA1 gene. We investigated the mitochondrial functions, dynamics, cellular phenotypes, outcomes, and fate. We evaluated OPA1 expression and found no change in OPA1^His42Tyr compared to the control fibroblasts. The presence of mutant mRNA in ADOA plus fibroblasts likely leads to the production of a defective OPA1 protein, manifesting a dominant negative effect, as suggested by the severe alteration of the mitochondrial network in OPA1-mutated fibroblasts. We discovered that the mitochondrial network was significantly fragmented and disrupted. The number of mitochondria or their mass does not appear to change, as suggested by VDAC1, mtDNA copy number, mtTFA (a factor intimately connected with the mitochondrial genome), and the OxPhos complexes (Fig. 8).

Figure 8

Schematic representation of the altered processes in OPA1 ^His42Tyr fibroblasts. The mutant OPA1 induces mitochondria depolarization (mitochondrion), increases ROS production, and reduces electron transfer chain (ETC) capacity. AMP-activated protein kinase (AMPK), a serine-threonine kinase, is an energy sensor that can respond to mitochondrial insult to maintain cellular energy homeostasis. OPA1-mutated fibroblasts also present a reduction of autophagy initiation and cellular proliferation. PI3K-AKT-mTOR pathway regulates both autophagy and proliferation. The phosphatidylinositol 3-kinase (PtdIns3K) activation recruits several proteins to the membrane, including 3-phosphoinositide-dependent kinase (PDK1) that phosphorylates AKT at Thr(308). A second AKT phosphorylation, required for its complete activation, is operated by mTORC2 at Ser(473) leading to cellular proliferation. AKT can activate mTORC1, a crucial inhibitor of autophagy that is instead suppressed by AMPK activation. All these pathways can contribute to p53 activation. p53 is widely known for its role as a transcription factor that regulates the expression of different stress response genes, e.g. genes involved in cell cycle regulation, such as the cell cycle inhibitor p21. p53 can upregulate p21 and p21 can act as a negative regulator of the cellular levels of p53. Expression of both p53 and p21 is increased in OPA1-mutated cells, thus contributing to cellular senescence (blue cells). Senescence and autophagy can be rescued by inhibiting mTORC1 through drug treatment with everolimus. Alt-text: Illustration depicting the PI3K-AKT-mTOR signaling pathway regulating autophagy, proliferation, and senescence. Treatment with mTORC1 inhibitor everolimus reduces senescence and induces autophagy in OPA1 mutated fibroblasts.

Despite the preservation of mtDNA integrity, the respiratory function and mitochondrial membrane potential were significantly affected in ADOA plus cells, with a significant decrease in the respiratory activity of the OxPhos complexes, particularly impacting complex IV activity [16, 67]. This impairment may be caused by the deregulation of the respiratory chain, disorganization of the inner structure of mitochondria, or both. It could also be due to a direct interaction between OPA1 and respiratory chain complexes I, II, and III, as demonstrated in a study by Zanna et al. in 2008 [68]. The organization of respiratory chain complexes into super-complexes (RCS) is crucial for the stability and assembly of each complex, which can affect the efficiency of electron transfer. According to a recent study [69], the formation of RCS is important for optimal OxPhos activity. The observed increase in OxPhos subunits could be a compensatory mechanism to sustain the diminished OxPhos activity. However, it does not necessarily imply that the resulting complexes are functional.

We assessed the efficiency of cellular and mitochondrial quality control systems, specifically autophagy and mitophagy, in clearing damaged mitochondria. Autophagy plays a crucial role in maintaining neuronal functions, and its impairments have been linked to various neurodegenerative disorders [70, 71]. The role of OPA1 in autophagy/mitophagy has been the subject of several studies (see ref. in the introduction). We investigated the ability of these processes to eliminate defective mitochondria. Our study in OPA1 fibroblasts showed a reduction in autophagosome formation and a possible impairment in the clearance of depolarized and isolated mitochondria by mitophagy due to autophagy. Mitophagy targets mitochondria for degradation in lysosomes via autophagosomes. In mammals, this process involves PINK1 and PRKN [72, 73]. Typically, PINK1 kinase is imported into the mitochondria and immediately degraded by PARL proteolysis in healthy mitochondria. However, in damaged or stressed mitochondria, PINK1 degradation is inhibited, allowing PINK1 to accumulate. PINK1 then phosphorylates various mitochondrial substrates, promoting PRKN binding to the outer mitochondrial membrane and subsequent PRKN-mediated ubiquitination of outer membrane proteins [74]. These ubiquitinated proteins attract autophagosomes and stimulate their incorporation into the damaged organelle, initiating mitophagy [73]. Our investigation of mitophagy markers in ADOA plus fibroblasts revealed accumulated PINK1 and PRKN on damaged mitochondria, suggesting that OPA1 mutation does not affect the recruitment of mitophagy proteins to defective mitochondria. Despite the significant impact of OPA1 mutation on the mitochondrial network, the autophagic response in OPA1^His42Tyr fibroblasts was not enhanced.

To identify deregulated molecular pathways in OPA1 mutant fibroblasts, we conducted whole RNA sequencing of OPA1^His42Tyr and control fibroblasts. The data revealed a significant downregulation of the expression of genes/pathways related to the mitotic cell cycle and proliferation, which was confirmed by the reduced growth rate of OPA1-mutated compared to control fibroblasts. We then examined the activation status of serine/threonine protein kinase AKT/PKB, a key regulator of cell survival and proliferation [49]. In mammalian cells, mTORC2 primarily regulates cell proliferation and survival by phosphorylating Ser473 of AKT [75–77]. Given the close link between mTOR and AKT regulation, we used AKT (Ser473) phosphorylation as an indicator of mTORC2 activity. Our findings showed a significant decrease in mTORC2 activity in mutant fibroblasts, as evidenced by reduced AKT (Ser473) phosphorylation but increased total AKT expression. This increase in total AKT expression may serve as a compensatory mechanism to counteract the decline in AKT activity. Additionally, studies have reported that increased expression of AKT protein is linked to reduced abundance of phospho-AMP-activated protein kinase [78]. Although we did not directly examine the phosphorylation or expression levels of this protein, it is worth noting that AMPK activation is crucial in promoting cell survival, especially in cells with dysfunctional mitochondria, by activating mTORC2 [79]. Furthermore, it should be noted that AMPK responsiveness tends to decrease with age, which could explain the altered metabolic regulation observed, leading to impaired autophagosome formation and clearance (via mTOR), increased oxidative stress, and reduced resistance to cellular stress [80, 81].

Considering these observations, we could associate the low activation of mTORC2 in OPA1-mutated fibroblasts with low cellular survival, which is evident considering the proliferative dysfunction, the increased doubling time, and the decreased doubling level of these cells. Moreover, we examined the possibility of cell senescence, given the observed increase in population doubling in arrested cell populations and the findings [57].

Senescence, a process associated with aging and cellular dysfunction, is influenced by the signaling pathway of the mechanistic target of rapamycin (mTOR) [82]. The mTOR pathway can regulate the tumor suppressor pathways p53/p21 and Rb/p16, which are known to be involved in senescence-associated growth arrest [83, 84]. During senescence, the expression of p53 and its downstream target p21 is increased, leading to cell cycle arrest [85–87]. The OPA1^His42Tyr gene expression signature indicated activation of p21WAF1/CIP1and p53 activation, which was consistent with the elevated protein levels of p53 and p21 that we observed in OPA1-mutated cells, suggesting the emergence of an early senescent phenotype compared to control cells, further confirmed by the increase in the β-Gal activity. Mitochondrial dysfunction can impair energy production and cellular metabolism, and severe oxidative stress can suppress autophagy and trigger senescence [88–90]. In our study, we found that OPA1-mutated fibroblasts exhibited increased levels of reactive oxygen species (ROS), suggesting increased oxidative stress, which could contribute to the induction of cell senescence (Fig. 8) [91].

Previous research in mice has demonstrated that the deletion of OPA1 in muscle cells results in premature senescence and premature death, indicating a correlation between OPA1 mutation, mitochondrial dysfunction, and senescence [92, 93]. In contrast, a study conducted on Chang cells found that the overexpression of OPA1 protein, correctly directed toward mitochondria, promoted mitochondrial fusion, and augmented senescence-associated galactosidase activity [94]. Similarly, changes in mitochondrial fusion events have been observed in senescent mesenchymal stromal/stem cells (MSCs), based on slightly higher levels of MFN1 and OPA1 proteins [95]. The role of OPA1 in senescence is further supported by the finding of a lack of mitochondrial fusion and reduced levels of OPA1 in giant mitochondria, which accumulate within aged or diseased postmitotic cells due to insufficient autophagy [96]. Therefore, the involvement of the OPA1 gene in cell senescence, as demonstrated in OPA1 KO mice models, and Chang and MSCs cells, provides support for the role of OPA1 in senescence. Notably, this role has never been reported in ADOA plus cells.

We hypothesized that impaired autophagy in OPA1-mutated fibroblasts could exacerbate mitochondrial dysfunction, preventing the removal of damaged mitochondria and increasing susceptibility to stressors that lead to cellular senescence. Therefore, we investigated the effects of mTOR inhibition on autophagy and senescence in OPA1-mutated fibroblasts. Treatment with the mTOR inhibitor, everolimus, restored autophagosome formation and decreased the expression of the senescence-associated marker β-Gal in OPA1-mutated fibroblasts.

Additionally, everolimus increased p53 expression and decreased p21 levels. It is known that p53 can upregulate p21 [97], but in the context of mTORC1 inhibition, p21 can also be degraded by hypophosphorylated 4E-BP1 [98]. Therefore, our data suggest that everolimus may directly decrease p21 levels, leading to a secondary effect of p53 upregulation. Furthermore, this pattern of p53 and p21 expression (with high p53 and low p21 levels) could potentially induce cell quiescence. In addition to its function as a cell cycle inhibitor, p21 also appears to stabilize complexes involved in the G1/S cell cycle transition and is shown to be involved in cell cycle arrest induced by different mTORC inhibitors [99–102]. The suppression of geroconversion, a hyperfunction associated with age-related diseases, could prolong the lifespan, and provide benefits in the context of neurodegenerative diseases. Quiescent cells, unlike senescent cells, can resume proliferation in response to appropriate signals [103–106]. Our findings suggest that targeting the mTORC pathway could be a promising therapeutic strategy to ameliorate mitochondrial dysfunction and senescence in patients with OPA1 mutations.

This is our second report focussing on the positive impact of targeting mTORC/AKT in fibroblasts from individuals with mitochondrial neuropathies. In our previous study, we discovered that fibroblasts carrying the MFN2 mutation distribute impaired or altered mitochondria among dividing cells, which govern cell growth via mTORC2/AKT activation. On the other hand, fibroblasts with the OPA1 mutation hinder mTORC2 activation and experience cellular senescence. Vam Lent et al. in 2021 demonstrated that in human iPSC-derived motor neurons harboring the most common MFN2 mutation (MFN2^R94Q), AMPK/AKT/mTORC signaling was altered. The authors suggested that cytosolic stress integrators like AKT and AMPK might sense mitochondrial dysfunction [107].

Our study presents several limitations, including low sample size and that the data derived could be due to the specific genetic variant or its background. Nevertheless, the involvement of the OPA1 gene in cell senescence is intriguing for its impact and has never been reported in human cells.

There are still unanswered questions that would be worthy of investigation: whether the inhibition of autophagy and the activation of cell senescence take place in the specific neuronal cells of individuals with ADOA plus; and whether there is a metabolic shift toward more anaerobic metabolism. Additionally, further research on the targets identified in our analysis will require suitable models for validation, such as in vitro co-cultures or the OPA1 knock-in mouse model. These investigations are outside the scope of this study.

Overall, our study highlights the importance of understanding the effects of mitochondrial proteins, such as OPA1, on cellular phenotypes and the potential for developing therapeutic strategies that target autophagy, senescence, and proliferation to treat primary mitochondrial neuropathies. Further research is needed to investigate these processes in neuronal cells and develop effective treatments for mitochondrial disorders.

Materials and methods

Case study

The case under study has already been published by Baldacci et al. in 2022. In brief, it involves a severely affected 46-year-old woman. The patient showed decreased visual acuity, bilateral palpebral ptosis, and ophthalmoparesis in both lateral and vertical gaze, and exhibited extra-ocular manifestations such as muscle stiffness and bilateral weakness. Muscle stiffness was mainly due to spasticity (e.g. upper motor neuron dysfunction) and weak legs were possibly linked to myopathy, neurogenic muscle changes, spasticity, or their combination. The patient was diagnosed with Autosomal Dominant Optic Atrophy (ADOA) plus carrying a point c.124C > T/p.His42Tyr mutation in OPA1. A muscle biopsy was analyzed using standard histochemical and histoenzymatic assays, including succinate dehydrogenase (SDH), cytochrome c oxidase (COX), and COX/SDH. The histochemical analysis of the muscle biopsy revealed COX-negative, SDH-positive, and ragged red fibers (RRF). The histochemical and histoenzymatic findings indicated significant neurogenic muscle pain with signs typical of primary mitochondrial myopathy. No mtDNA depletion nor multiple deletions were found in the patient’s muscle biopsy.

The patient provided informed consent for blood sampling, genetic testing, skin biopsy, fibroblast culture, and the publication of the case report.

Cell culture and analysis

Primary fibroblasts were obtained by explants from a skin punch biopsy of a patient with an OPA1 mutation, and from an individual who showed no histological and biochemical signs of mitochondrial disease and was used as a healthy control. The ages of the patient and control when the biopsy was performed were similar, 47ys and 46ys, respectively. Informed consent was obtained from both individuals before the procedure. The patient’s mutation is a heterozygous missense c.124C > T/p.His42Tyr at the exon 2, which was previously reported [108].

The cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM; EuroClone, ECB7501LX10) supplemented with 10% (v/v) fetal bovine serum (FBS; EuroClone, ECS5000L), 1% (v/v) L-glutamine (EuroClone, ECB3000D), 1% (v/v) penicillin/streptomycin (EuroClone, ECB3001D), and 50 μg/ml of uridine (Sigma-Aldrich, U3003) at 37°C in a humidified atmosphere of 5% CO₂. All experiments were performed on control and patient fibroblasts after two passages following the thawing of the cells. The thawed cells had passage numbers P9, P10, and P11, and three independent replicates were performed for both control and patient cells. For the experiments, growing cells were plated on sterile plastic dishes, on sterile plastic flasks, or sterile glass coverslips and allowed to adhere for at least 24 h before use. For mitochondrial network studies, where indicated, cells were grown in DMEM lacking glucose supplemented with 5.5 mM of galactose (DMEM galactose) for 72 h. For inhibition of autophagy, chloroquine (CQ; C6628) was added to the cell culture medium at a concentration of 50 μM, whereas to induce autophagy, torin1 (475991) was added at concentrations of 1 μM. CQ and torin1 were added for 3 h before the end of the treatments. For mitochondrial network and mitophagy studies, the uncoupler agent carbonyl cyanide p-(trifluoromethoxy) phenylhydrazone (CCCP; C2920) was used at 10 μM for 4 h. For autophagosome number, AKT/phospho-AKT, and senescence markers measurements by Western blot, OPA1, and control fibroblasts were treated with 5 μM everolimus or vehicle control (DMSO) 3 h before collection. For SA-βGal senescence staining cells were treated for two days with 5 μM of everolimus along with DMSO control. After two days of incubation, the cells were placed into normal growth media for three days. These chemicals were purchased from Sigma Aldrich unless otherwise indicated.

Mitochondrial DNA analysis

Total DNA (nuclear and mitochondrial) was extracted from patient-derived fibroblasts and control fibroblasts using the ‘Wizard® Genomic DNA Purification Kit’ (Promega, A1125) according to the manufacturer’s instructions. Large-scale mitochondrial deletion detection and mitochondrial DNA copy number were performed using long-range PCR and quantitative real-time PCR, respectively [34, 109].

Oxygen electrode polarography

Respiration measurements were performed in digitonin-permeabilized cells using a Clark-type oxygen electrode (Oxygraph; Hansatech Instruments, England) as previously described [34].

Western blotting analysis

Cell extracts were prepared as described [34], separated by 8% or 10% homemade SDS-polyacrylamide gel or 12% TGX Stain-Free FastCast Gels (Bio-Rad, 1610185) (20 to 50 μg of protein per line) and then electro-transferred to PVDF membranes using a Trans-Blot transfer apparatus (Bio-Rad, California, USA). The membranes were blocked as previously reported [34] and probed with the following antibodies: anti-OPA1 (BD Transduction Laboratories, 612606), anti-MFN2 (Abnova, H00009927) for mitochondrial morphology study; anti-VDAC1 antibody (Abcam, ab15895) for mitochondrial mass evaluation; anti-mtTFA (Novus Biologicals, 18G102B2E11) for mitochondrial biogenesis evaluation; anti-ND4 (A16-R, Santa Cruz Biotechnology, sc-20499-R), anti-NDUFS1 (H-300, Santa Cruz Biotechnology, sc-99232), anti-SDHA (Cell Signaling Technologies, 5839S) and anti-COX IV (Cell Signaling Technologies, 4844S) for respiratory chain subunits complexes studies; anti-p62 (SQSTM1, BD Transduction Laboratories, 610833); anti-LC3B (Sigma-Aldrich, L7543) for autophagy flux study; anti-p53 (DO-1, Santa Cruz Biotechnology, sc-126) and anti-p21 (C-19, Santa Cruz Biotechnology, sc-397) for senescence measurements (Table S3). AKT phosphorylation level on Thr308 and Ser473 was measured with phospho-specific antibodies (Cell Signaling Technology, D25E6, D95, respectively) and adjusted to total AKT using AKT total antibody (Cell Signaling Technology, 40D4) (Table S3). Membranes were re-probed with a specific antibody for GAPDH (ProteinTech, 60004-1-Ig) as an index of homogenate protein loading in the lanes. Then, suitable peroxidase-conjugated secondary antibodies were used for immunostaining and the ECL procedure for development by the LI-COR C-Digit blot scanner (Licor, Lincoln, NE) [34].

JC-1 staining of mitochondria fibroblasts

Detection of altered mitochondrial membrane potential (Δψm) was performed by 5,5′,6,6′-tetrachloro-1,1′,3,3′ tetraethyl-benzimidazolylcarbocyanide (JC-1) staining without fixation [34]. Cells were observed using a fluorescence Leica TCS SP8 microscope coupled to the Laser Scanning Confocal Microscopy (LSCM) system and equipped with narrow interference filters, using 40× oil immersion objectives with a pinhole airy unit of 1. For each independent experiment, the integrated density of red and green channels was analyzed in more than 30 cells using ImageJ software. Red fluorescence is an index of ΔΨm preservation, whereas green fluorescence is an index of mitochondrial membrane depolarization. The red/green fluorescence ratio was calculated considering that high mitochondrial membrane potential is associated with red JC-1 emission.

Immunofluorescence assay

The following primary antibodies were used: anti-VDAC1 (Abcam, ab15895); anti-LC3B (Sigma-Aldrich, L7543); anti-Human LAMP1 (BDPharmingen, 555798) anti-TOMM20 (Abcam, ab56783); anti-PINK1 (Abcam, EPR20730) and anti-Parkin (PRKN = PRK8) (Santa Cruz Biotechnology, sc-32282) for autophagy/mitophagy study; anti-LAMP1 (BD Pharmingen™; 555798) (Table S3). After three washes in PBS, cells were reacted with appropriate secondary antibodies conjugated with Alexa Fluor 488 or 568 (Thermo Fisher Scientific, A11001 and A11036, respectively) for 1 h at room temperature in a humid chamber. Nuclear chromatin was stained with the fluorescent dye 4,6-diamidino-2-phenylindole-dihydrochloride 5 μg/ml (DAPI; Sigma-Aldrich, D9542) for 10 min at room temperature and, after three washes in PBS, coverslips were mounted in Vectashield (Vector, H-100-10). Stained cells were examined with a Leica TCS SP8 microscope (Leica Microsystems, Buffalo Grove, IL, USA), coupled to the laser scanning confocal microscopy (LSCM) system. For mitochondrial network evaluation, images were acquired at 60× magnifications (Nikon×60 Plan Apocr λ NA = 1.40 oil immersion objective) with a Nikon Ti2-inverted microscope equipped with a structured illumination imaging system ViCo 2.0. Three different mitochondrial morphology groups were outlined: tubular mitochondria (circularity 0–0.3), intermediate mitochondria (circularity 0.3–0.6), and fragmented mitochondria (circularity 0.6–1) as described elsewhere [110]. Mitochondrial circularity was quantified by using ImageJ software (https://imagej.nih.gov/ij/). For each experiment, at least 20 cells were used to calculate the relative percentages of the mitochondrial areas for the different mitochondrial groups normalized on the total area of the mitochondrial compartment.

For mitophagy analysis, stained cells were examined using 40× and 63× oil immersion objectives with a pinhole airy unit of 1. Dot counting was performed using the default ‘Analyze Particles’ plugin in ImageJ software. Optical Z of the cells was taken by a threshold intensity; binary images were obtained. Puncta with area of 0.1–1.767 um² were quantitated for LC3B; 0.03–0.5 um² for LAMP1; 0.75–3 um² for TOMM20; >0.08 um² for PINK1 and PRKN. Colocalization measurement was obtained by counting the dots occupied by the overlapping element for the cell (i.e. LTc; LLpc, PPrc) or by the ratio between the dots occupied by the overlapping element and single-channel dots (i.e. LTc: LLpc: Lp; PTc:P). At least 30 cells were counted.

RNA-sequencing pre-processing and data analysis

RNA extraction and whole transcriptome analysis were performed on OPA1, and control fibroblasts as previously described [34]. We collected cells in three independent experiments and isolated total RNA with the RNeasy Plus Kit (Qiagen, 74034). We checked RNA purity and quantity with the Agilent 2100 Bioanalyzer RNA Nano Chip (Agilent, Santa Clara, CA, USA). We prepared directional RNA-Seq libraries from 100 ng of total RNA with the TruSeq Stranded Total RNA Sample Prep Kit (Illumina, 20020597) following the manufacturer’s protocol. We sequenced the libraries with the Illumina NextSeq 500 platform (Illumina, San Diego, CA) and sequencing raw data in FASTQ format, were quality-checked using the FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc) reads with less than 50 bases as well as low-quality regions (phred cutoff of 25) were discarded. Alignment was performed with STAR [111] and differential gene expression analysis between OPA1^His42Tyr, and control fibroblasts was carried out using a generalized fold change algorithm (GFOLD) as reported [34]. We considered only genes with a GFOLD (adjusted log2 fold change) value equal to or higher than 1.5 in absolute value. Differentially Expressed Genes (DEGs) were functionally annotated by online DAVID software (https://david.ncifcrf.gov/tools.jsp,2021Update) and significantly enriched pathways were identified (Benjamini-Hochberg False Discovery Rate (FDR) P < 0.001). We used STRING (https://string-db.org/) to assess protein-protein interactions (PPIs) among DEGs and built networks with STRINGApp in Cytoscape software (Version 3.9.1) as previously reported [34]. Ingenuity pathway analysis (IPA v.94302991) was used to evaluate the activation status of serine/threonine protein kinase AKT/PKB.

Growth curve and doubling time and numbers of OPA1 and control fibroblasts

To calculate the doubling time, we used the cell counting method and the following formula: PD = t × Log2/(LogC2-LogC1) (PD = Population doubling time, t = 24 h (interval of cell count), Log = 10basedLog, C1 = 1st cell count, C2 = 2nd cell count). To calculate the doubling level (PDL), we used the equation: PDL = 3.32 × logN/No (N = harvest cell number and No = initial cell number) [112, 113]. We seeded OPA1 fibroblasts and control cells in 24-well plates and grew them for 3 days under normal culture conditions. To obtain an accurate measurement of the PDL, we counted the number of cells attached to the dish after seeding, rather than the number of cells seeded [112]. After cell adhesion, we harvested the cells by trypsinization and counted them using a hemocytometer, which allowed us to calculate the PDL. To create a growth curve, we performed cell counts every 24 h from day 0 to day 3. We calculated the PD time using the above formula from day 1 to day 2, while the PDL time was calculated using the number of cells from day 0 (after adhesion) until day 3. We performed a total of 3 individual experiments.

Senescence-associated β-galactosidase (SA-βGal) assay

The cells were washed using PBS and then fixed for 3–5 min at room temperature using 3% formaldehyde. After washing, they were incubated at 37°C (without CO₂) with a fresh solution of senescence-associated β-Gal (SA-β-Gal) stain. This solution consisted of 1 mg of 5-bromo-4-chloro-3-indolyl P3-D-galactoside (X-Gal) per ml (stock = 50 mg of dimethylformamide per ml), 40 mM citric acid/sodium phosphate, pH 6.0, 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM NaCl, and 2 mM MgCl₂. After 24 h, staining became evident, and only senescent cells showed blue signals that could be detected using bright-field microscopy. To determine the percentage of SA-β-Gal-positive cells, we examined the cells in 3 random fields, with at least 100 cells each, using a Carl Zeiss Primovert microscope (10×).

Detection of fluorescent ROS production

To measure ROS generation levels in fibroblasts, we used the ROS-ID Total ROS/Superoxide Detection Kit (ENZ-51010) (Enzo Life Science, New York). We made sure to transfer the same number of cells into each well of a 96-well black wall/clear bottom plate to ensure 70%–80% confluency on the day of the experiment. After changing the media, we pre-treated the cells with N-Acetyl-L-cysteine (NAC), which is a ROS inhibitor, for 30 min. We used Pyocyanin (200 μM) to induce ROS production and had a control group that included vehicle alone (no inhibitor and no ROS inducer, only anhydrous DMF) and vehicle with inhibitor. We observed ROS production using an oxidative stress detection reagent (green) for 30 min to an hour and quantified fluorescence with a Victor 2030 multilabel reader (PerkinElmer, Waltham, MA, USA).

Statistical analysis

All statistical analyses were performed using PRISM® 7.04 in analytical software (GraphPad Software Inc., San Diego, CA) and Excel (Microsoft, Inc.). Results were expressed as average values±SD of at least three independent determinations, each performed in triplicate, if not otherwise specified using sex and age-matched control and OPA1 fibroblasts. Statistical significance was calculated using Student’s t parametric test set at ^*P < 0.05; ^**P < 0.01; ^***P < 0.001; and ^****P < 0.0001. One-way analysis of variance (ANOVA) test was performed to examine the differences between more than two dependent groups. Bonferroni correction was used for multiple comparisons.

Acknowledgements

We thank the family for participating in this study, the patients’ associations MITOCON, UILDM (Unione Italiana Lotta alla Distrofa Muscolare), and CollaGe-Associazione-Genitori-Manzoni-Poli-Molfetta.

Conflict of interest statement: None declared.

Funding

This work was supported by grants from REGIONE PUGLIA-MALATTIE RARE-Petruzzella, Uff. Pres. n. 246-10 2019 ‘NeurApulia’ and by donations of Parents’ Associations (to V.P.) and Fondazione Opera Pia Monte di Pietà e Confidenze, Molfetta (to P.Z. and A.A.). Financial support from the Italian Ministry of Health-Ricerca Corrente 2021 (to S.D., and F.M.S.).

Disclosure statement

All authors have read and agreed to the published version of the manuscript. The authors alone are responsible for the content and writing. No potential competing interest was reported by the authors.