{"paper_id":"581a4ec6-d816-47a9-afe2-b6ee10517a42","body_text":"1 \n \nPathogenic variants in autophagy -tethering factor EPG5 drive neurodegeneration \nthrough mitochondrial dysfunction and innate immune activation  \n \nKritarth Singh 1*, Hormos Salimi Dafsari 2,3,4*, Olivia Gillham 1, Haoyu Chi 1, Ivet \nMandzhukova1, Ioanna Kourouzidou1, Preethi Sheshadri 1, Chih -Yao Chung 1, Valeria  \nPingitore5,7, Fleur Vansenne 6, David L. Selwood 7, Diana Pendin 8,9, Gyorgy Szabadkai 1,9, \nManolis Fanto10†, Heinz Jungbluth4,11†, Michael R. Duchen1† \n \n1Department of Cell and Developmental Biology and Consortium for Mitochondrial Research, UCL, \nGower Street, London WC1E 6BT, UK  \n2Department of Pediatrics, Faculty of Medicine and University Hospital Cologne, University of \nCologne \n3Max-Planck-Institute for Biology of Aging and Cologne Excellence Cluster for Ageing -associated \nDiseases, Cologne, Germany \n4Department of Paediatric Neurology, Evelina London Children’s Hospital, Guy’s & St Thomas’ NHS \nFoundation Trust, London, UK  \n5 Department of Health and Biomedical Sciences, Universidad Loyola Andalucía, Seville, Spain \n6Department of Genetics, University Medical Center Groningen, University of Groningen, Groningen,  \nNetherlands \n7Drug Discovery, UCL Wolfson Institute for Biomedical R esearch, University College London, \nLondon WC1E 6BT, UK  \n8Neuroscience Institute, National Research Council, Padua 35131 Italy \n9Department of Biomedical Sciences, University of Padua, Padua 35131 Italy \n10Division of Basic and Clinical Neuroscience, IoPPN, King’s College London \n11 Randall Centre for Cell and Molecular Biophysics, Muscle Signalling Section, Faculty of Life \nSciences and Medicine (FoLSM), King’s College London, London, UK  \n \n*equal contribution   †joint supervision \n \nCorrespondence to: Michael R. Duchen, email: m.duchen@ucl.ac.uk \n \nKey words: EPG5-Related Disorder,  neurodegeneration, mitochondrial calcium, mPTP, \nexcitotoxicity, mtDNA, STING- IFN signalling \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n2 \n \nAbstract \nThe autophagy-tethering factor, ectopic P-granule 5 autophagy protein  (EPG5), plays a key \nrole in autophagosome -lysosome fusion. Impaired autophagy associated with pathogenic \nvariants in EPG5 cause a rare devastating multisystem disorder  known as Vici syndrome, \nwhich includes neurodevelopmental defects, severe progressive neurodegeneration and \nimmunodeficiency. The pathophysiological mechanisms driving disease presentation and \nprogression are not understood.  In patient-derived fibroblasts and iPS cells differentiated to \ncortical neurons, we found that impaired mitophagy leads to mitochondrial bioenergetic \ndysfunction. Physiological Ca 2+ signals resulted in paradoxical mitochondrial Ca 2+ overload \nattributed to downregulation of MICU1/3. Ca 2+ signals caused mitochondrial depolarisation, \nmtDNA release and activation of the cGAS-STING pathway, reversed by pharmacological \ninhibition of the mitochondrial permeability transition pore (mPTP) or of the STING pathway. \nThus, we have identified multip le potential therapeutic targets driving disease progression \nassociated with pathogenic EPG5 mutations, including impaired mitochondrial bioenergetics, \nmitochondrial Ca2+ overload, vulnerability to mPTP opening and activation of innate immune \nsignalling.  \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n3 \n \nIntroduction \nAutophagy is an evolutionarily conserved, fundamental intracellular homeostatic process with \nessential roles in metabolic  adaptation, defence against infections and the quality control of \ndefective proteins and organelles including mi tochondria (Klionsky et al., 2021) . Ectopic P-\ngranules 5 autophagy tethering factor (EPG5), initially reported in Caenorhabditis elegans, \nfunctions as a tethering protein in concert with specific SNARE complexes to facilitate specific \nfusion events between autophagosomes and lysosomes  and the formation of degradative \nautolysosomes (Tian et al., 2010 ; Guo et al., 2014; Wang et al., 2016 ; Hori et al., 2017 ). \nRecessive mutation s in human EPG5 cause a spectrum of neurodevelopmental disorder s \ntermed as EPG5-related disorders ( EPG5-RD) ranging from the severe early-onset \nmultisystem disorder Vici syndrome to relatively milder  neurodevelopmental manifestations \nwith progressive neurodegeneration  (Cullup et al., 2013; Byrne et al., 2016b; Dafsari et al., \n2022). Vici syndrome  patients are characterized by the key diagnostic features of \nmicrocephaly, callosal agenesis, cataracts, hypopigmentation, cardiomyopathy, (combined) \nimmunodeficiency and failure to thrive,  followed by severe progressive neurodegeneration \nwith a median life expectancy of approximately 2 4 months (Byrne et al., 2016a; Deneubourg \net al., 2022). While most bi-allelic loss-of-function EPG5 mutations are associated with Vici \nsyndrome, pathogenic missense variants in EPG5 can lead to less severe clinical presentations. \nAn attenuated EPG5 function is predicted to underlie the phenotypic variability seen in EPG5-\nRD (Vansenne et al., 2022 ; Dafsari et al., 2024 ). However, it is unclear whether the residual \nfunction of pathogenic EPG5 variants correlates directly to a corresponding defect in \nautophagy alone or if other downstream cellular processes are also disrupted. Interestingly, \npreliminary observations indicate an increased frequency of adult -onset neurodegenera tive \ndisorders in (putative) heterozygous EPG5 variant carriers, suggesting a potential  dosage \neffect.    \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n4 \n \nA p revious study has reported that  the histopathological appearance of  EPG5-related Vici \nsyndrome, often including marked ultrastructural mitochondrial abnormalities and decrease d \nrespiratory chain enzyme activity, may mimic primary mitochondrial disorders (McClelland et \nal., 2010; Byrne et al., 2016b) and some patients have been suspected to have a mitochondrial \ncytopathy before the causative pathogenic EPG5 variants were genetically confirmed \n(Balasubramaniam et al. , 2018; Waldrop et al., 2018) . Our recent findings have shown  that \ndefective mitophagy and associated mitochondrial dysfunction play a significant role in EPG5-\nrelated Vici syndrome and that some features of the disorder may be a direct consequence of \nmitochondrial dysfunction (Dafsari et al., 2024) . Impaired mitochondrial energy metabolism  \nhas been recognized as a key facto r in the pathogenesis of many common adult -onset \nneurodegenerative disorders including amyotrophic lateral sclerosis  (ALS), Parkinson’s and \nAlzheimer's disease (Duchen and Szabadkai, 2010; Schon and Przedborski, 2011; Area-Gomez \net al., 2019; Chu, 2022; Suomalainen and Nunnari, 2024). Mitochondrial dysfunction resulting \nfrom impaired mitochondrial quality control can lead to impaired ATP homeostasis, oxidative \nstress, impaired Ca2+ buffering and altered mitochondrial Ca 2+ signalling, all of which \ncontribute to neuronal dysfunction and cell death characteristic of many neurodegenerative \ndisorders (Duchen, 2012; Devine and Kittler, 2018;  Boyman et al., 2020; Plotegher et al., \n2020). \n \nMitochondria actively maintain neuronal  Ca2+ homeostasis through Ca2+ uptake and efflux \npathways during physiological cytosolic  [Ca2+] ([Ca2+]c) signals (Rizzuto et al., 2012; \nPlotegher et al., 2021; Verma et al., 2022) . Mitochondrial Ca2+ uptake is mediated by a Ca2+-\nselective ion channel, the mitochondrial calcium uniporter (MCU) located in the inner \nmitochondrial membrane (IMM) (Kamer and Mootha, 2015; Giorgi et al., 2018). This hetero-\noligomeric channel is composed of the pore -forming protein, MCU (Baughman et al., 2011; \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n5 \n \nDe Stefani et al., 2011), a scaffold, essential MCU regulator (EMRE) (Sancak et al., 2013), and \nCa2+-sensitive gatekeepers, MICU1, MICU2, and MICU3 (Perocchi et al., 2010; Plovanich et \nal., 2013; Patron et al., 2019) . The dimer of MICU proteins in concert with EMRE exert s a \ntight control on mitochondrial Ca2+ entry during changes in local [Ca2+]c (Csordas et al., 2013; \nPatron et al., 2014; Fan et al., 2020). A rise in mitochondrial matrix [Ca2+] ([Ca2+]m) increases \nthe rate of TCA cycle -driven NADH generation  and the rate of oxidative ATP production  \n(Duchen, 1992; Jouaville et al., 1999; Denton, 2009). Accumulation of Ca2+ in mitochondria is \nbalanced by Ca2+ efflux through the NLCX exchanger which normally maintains a low [Ca2+]m \n( Palty et al., 2010; Garbincius and Elrod, 2022) . Supraphysiological accumulation of Ca2+, \nhowever, can trigger the opening of a large conductance channel in the IMM, the mitochondrial \npermeability transition pore (mPTP), resulting in the collapse of ΔΨm, ATP hydrolysis by the \nFOF1 ATPase and mitochondrial osmotic swelling (Brenner and Moulin, 2012; Bhosale et al., \n2015; Briston et al., 2019) . This mitochondrial catastrophe seems to be a common final path \ndriving mitochondrial Ca2+ overload-induced excitotoxic injury and neuronal cell death as \nevident in MICU1-KO mice (Singh et al., 2022) and implicated in neurodegenerative diseases \nsuch as ALS (Lautenschlager et al., 2013) , Alzheimer's (Jadiya et al., 2021)  and Parkinson's \ndisease (Choi et al., 2022) as well as in several muscular dystrophies and myopathies (Palma \net al., 2009; Rao et al., 2014).  \n \nMitochondrial Ca2+ overload may also trigger the release of mtDNA via selective or limited \nmPTP opening in a subset of damaged mitochondria , activating a cytoprotective innate \ninflammatory response and e vade cell death  similar to  the phenomenon described recently \nduring cellular senescence (Victorelli et al., 2023). Cytosolic mtDNA is sensed as genotoxic \nstress by the DNA-sensing cyclic GMP –AMP synthase (cGAS) –stimulator of interferon \nresponse cGAMP interactor 1 (STING) pathway that activates type I/III IFN and nuclear factor \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n6 \n \n(NF)-κB signalling (West et al., 2015; Newman and Shadel, 2023) . Recent findings indicate \nthat a mtDNA-driven inflammatory response may trigger a more insidious, low-grade chronic \nneuronal loss , a feature of many late-onset neurodegenerative diseases , especially when \ncoincident with impaired mitophagy  (Sliter et al., 2018; Jauhari et al., 2020; Yu et al., 2020; \nJimenez-Loygorri et al., 2024). \n \nMitochondrial Ca2+ overload culminating in the loss of mitochondrial function, Ca2+-dependent \ncell death, inflammation , and progressive cell loss  has been implicated in major \nneurodegenerative diseases (Filadi and Pizzo, 2020). Impaired mitochondrial Ca2+ homeostasis \nmay underlie progressive infantile epileptic encephalopathy  found in nearly two -thirds of \npatients with EPG5-RD (Deneubourg et al., 2025)  and an associated chronic inflammatory \nresponse may explain  high levels of cytokines measured in Vici syndrome patients  (Piano \nMortari et al., 2018) . However, evidence for a specific mechanism driving aberrant \nmitochondrial Ca2+ signalling leading to mitochondrial dysfunction and inflammatory response \nis scarce and poorly understood.  \n \nOur r ecent findings point towards mitochondrial dysfunction and the failure of autophagic \nremoval of damaged mitochondria allowing progressive accumulation of mitochondrial defects \nas a major factor underlying the neurodegenerative disease and inflammatory dysregulation \nseen in patients with Vici syndrome and other EPG5-RDs (Dafsari et al., 2024) . We now \nidentify impaired mitochondrial Ca 2+ homeostasis as the causative mechanism for mtDNA-\ndriven STING- type I IFN inflammatory response in patient-derived fibroblasts with truncating \nand missense EPG5 variants. iPSC-derived cortical neurons carrying a founder EPG5 \npathogenic variant showed increased sensitivity to excitotoxicity in response to low \nconcentrations of glutamate that were innocuous in control cells, with responses characterised \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n7 \n \nby mitochondrial Ca 2+ overload, delayed calcium deregulation  (DCD), loss of ΔΨm and \nneuronal cell death. We demonstrate that impaired mitochondrial bioenergetic function and \nCa2+ overload found in patient-derived cells and EPG5 mutant neurons were attributable to the \ndownregulation of MICU1 and increased susceptibility to mPTP opening.  Remarkably, \npharmacological inhibition of mPTP  opening attenuated inflammatory signals and rescued \nmitochondrial bioenergetic function in patient-derived cells and EPG5 mutant neurons. These \nfindings highlight a key pathophysiological mechanism in Vici syndrome and other EPG5-RD \nand signpost potential therapeutic strateg ies for the rapidly expanding clinical and genetic \nspectrum of patients with EPG5- and other autophagy-related disorders. \n \nResults \nMitochondrial bioenergetic function is impaired in pat ient-derived fibroblasts bearing \npathogenic EPG5 mutations  \nTo characterise the impact of pathogenic EPG5 mutations on mitochondrial metabolism and \nbioenergetic function, we examined human dermal fibroblasts derived from patients  \n(Supplementary Table S1) either carrying homozygous missense mutations (p.Gln336Arg, \np.Arg1621Gln, hereafter referred to as patient 1 and patient 2 respectively ) or  compound \nheterozygous and homozygous truncating mutations (p.Arg299*/Pro1827Ala, \np.Phe1604Glyfs*20, hereafter referred to as patient 3 and patient 4 respectively )  as well as  \nhealthy control cell lines from donors matched for age and sex. Protein expression analysis by \nimmunoblotting in control and patient fibroblasts reve aled an almost complete absence of \nnormal EPG5 protein in all patient cells  (Fig. 1A-B), even in missense mutations, while in \ncomparison the EPG5 mRNA expression showed a significant reduction in patient cells (Fig. \nS1A).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n8 \n \nIn order to  assess mitochondrial bioenergetic function in patient fibroblasts, we first used \nequilibration of tetramethyl rhodamine methyl ester (TMRM) to measure ΔΨm by confocal \nimaging (Fig. 1C and Fig. S1B). Single cell analysis of mitochondrial TMRM fluorescence \nintensity showed reduced ΔΨm in all the patient cells compared to controls  (Fig. 1D) . \nQuantitative analysis of  mitochondrial volume  occupancy by co -labelling the cells with \nCalcein AM, to measure cytosolic area, showed no change between the cell lines (Fig. S1C). \nHowever, quantification of mtDNA copy number revealed a significant increase in mtDNA \ncopy number in all patient fibroblasts (Fig. S1D). Morphometric analysis of TMRM-labelled \nmitochondria showed that a large proportion of the mitochondrial network was fragmented \nwhile a small pool showed swollen morphology in patient cells  (Fig. 1E). To investigate the \ncause of the reduced ΔΨm in patient fibroblasts , we measured the  redox state of the NADH \npool, the main substrate for the mitochondrial electron transport chain (ETC). The resting level \nof NADH autofluorescence was quantified through the experimental determination of the \n“redox index,” a ratio of the signal representing the maximally oxidized pool (response to 1 \nμM FCCP) and the maximally reduced pool (the response to 1 mM NaCN) (Fig. 1F and Fig. \nS1E). Patient fibroblasts exhibited a significant increase in the NADH redox state (i.e. less \noxidised NADH/more reduced NAD+) in comparison to control cells, suggesting an impaired \nfunction of the mitochondrial ETC (Fig. 1G) consistent with the previous reports (Byrne et al., \n2016b). To confirm this, the oxygen consumption rate was measured using the Seahorse XFe96 \nextracellular flux analyser (Fig. 1H). Both the ATP-linked respiratory rate and spare respiratory \ncapacity were substantially reduced in patient cells compared to controls (Fig. 1I-J).  \n \nIn order to further identify the cause of the mitochondrial bioenergetic defect observed in the \npatient cells, steady-state levels of ETC components were analysed by SDS-PAGE (Fig. 1K) \nas well as the native assembly of these components into respiratory supercomplexes by BNGE \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n9 \n \nand immunoblotting (Fig. S1F). The expression levels of CI, CII and CIV were all significantly \nreduced as was the assembly of CI, CIII and CIV  compared to control cells except for CV \nwhich remained unaffected in all patient cells (Fig. 1L and Fig. S1G). Patient cells also showed \nan increased rate of production of intracellular reactive oxygen species (ROS), measured using \ndihydroethidium (DHE) fluorescence  (Fig. 1M) . Together, t hese data  show an impaired \nmitochondrial bioenergetic function and a  respiratory chain defect associated with  the \ndecreased expression and assembly of the OXPHOS complexes in patient fibroblasts carrying \npathogenic EPG5 mutations.  \n \nPathogenic EPG5 variants causes dysregulation of mitochondrial calcium signalling  \nThe Ca2+-dependent regulation of  mitochondrial metabolism , mediated by Ca2+ entry into \nmitochondria through the MCU complex, plays a key role in driving respiratory chain activity \nthat maintains the rate of ATP synthesis in response to increased energy demand (Szabadkai \nand Duchen, 2008) . As mitocho ndrial Ca 2+ uptake is potential -dependent, we investigated \nwhether impaired mitochondrial Ca 2+ uptake due to the reduced ΔΨm might amplify the \nbioenergetic defect in the patient cells. We therefore measured the mitochondrial matrix \n[Ca2+]m response in control and patient cells expressing mitochondrial-targeted aequorin  \nfollowing a challenge with 10 µM histamine (Fig. 2A). Surprisingly, mitochondrial Ca2+ uptake \nwas increase d in patient cells with significantly larger peak of [Ca2+]m upon physiological \nhistamine stimulation (Fig. 2B). This striking finding suggested that the increase in histamine-\ninduced [Ca2+]m in patient fibroblasts could be due to altered endoplasmic reticulum (ER) Ca2+ \ncontent and ER-mitochondria proximity or dysregulation of the MCU/NCLX complex. To test \nthis, cells were labelled with Fluo-4 AM and mito-Fura-2 AM (De Nadai et al., 2021; Pendin \net al., 2019)  to simultaneously monitor the changes in [Ca2+]c and [Ca2+]m (Fig. 2C -D). \nApplication of 10 µM histamine to both cell types produced a comparable rise in [Ca2+]c with \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n10 \n \nno significant changes in total Ca2+ released from the ER (Fig. S2A) or its clearance as \nmeasured by the time taken from  peak [Ca2+]c to half the final baseline value  (Fig. S2B). \nHowever, the increase in both resting and histamine-stimulated [Ca2+]m was confirmed by \nratiometric measurement of mito-Fura-2 intensity in patient cells (Fig. 2E-G). Notably, the \ninitial rate of histamine-induced mitochondrial Ca2+ uptake was also significantly increased in \npatient fibroblasts (Fig. 2F). We further examined the ER-mitochondria contact sites  by \nanalysing the fraction of mitochondrial  surface involved in contact with ER by transmission \nelectron microscopy (TEM) (Fig. S2C). We found no significant change in the frequency and \ntotal number of contacts in the 0-20 nm range, the gap width relevant for Ca2+ transfer, between \nboth the cell types (Fig. S2D).  \n \nThe increased velocity of mitochondrial Ca2+ uptake observed in patient fibroblasts prompted \nus to measure the expression levels of regulatory components of the MCU complex (Fig. 2H). \nMCU and NCLX protein levels were  not altered but surprisingly, the expression of MICU1, \nMICU2 and EMRE was significantly reduced in patient cells indicating that the ‘gatekeeper’ \nfunction of MICU1/MICU2 in mitochondrial Ca 2+ uptake is compromised in patient cells \nallowing rapid entry of Ca 2+ into the mitochondria  (Fig. 2I). Mitochondrial Ca2+ signalling \ndirectly impacts TCA cycle intermediates by the allosteric activation of pyruvate \ndehydrogenase (PDH)  and α -ketoglutarate dehydrogenase (αKGDH). Accumulation of \nmitochondrial Ca2+ activates PDH phosphatase (PDP1), which dephosphorylates the PDH E1α \nsubunit and thereby increases PDH activity to convert pyruvate to acetyl -CoA (Denton et al., \n1972). Immunoblotting analysis of phosphorylated PDH (p -PDH E1α, inactive) revealed a \nsignificantly reduced p-PDH E1α/PDH ratio in patient cells compared to controls (Fig. 2J and \nFig. S2E), consistent with the chronic elevation of resting [Ca2+]m and enhanced PDP activity \nin EPG5-deficient patient cells.  Remarkably, a high resting [Ca2+]m and the associated \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n11 \n \nmitochondrial bioenergetic deficiency are the characteristic cellular features of fibroblasts \nderived from patients with  MICU1 loss -of-function mutation s, a rare childhood disorder \ncharacterised by proximal myopathy, cognitive impairment and a progressive \nneurodegeneration (Logan et al., 2014; Lewis-Smith et al., 2016). \n \nSupraphysiological mitochondrial Ca 2+ accumulation can trigger the opening of mPTP and \nCa2+ dependent cell death  (Briston et al., 2019) . We therefore asked whether diminished \nexpression of the MCU regulatory proteins renders patient cells vulnerable to [Ca2+]m overload \nand mPTP opening. To examine this, isolated mitochondria from control and patient fibroblasts \nwere challenged with a series of 5 μM Ca2+ boluses to evoke [Ca2+]m overload. We found that \nmitochondria from patient cells tolerated fewer Ca 2+ pulses before the precipitous  increase in \nfluorescence signal indicating mPTP opening (Fig. 2K). Consistent with the above results, the \nrate of Ca2+ uptake was also increase d in patient cells in this assay, as measured by the time \ntaken from peak Ca2+ to half the final baseline value  (Fig. S2F). Interestingly, preincubation \nwith JP1-138, a highly specific novel mitochondrial cyclophilin D (CypD) targeting molecule \nand a potent inhibitor of PTP (Pingitore et al., 2024), significantly increased the mitochondrial \nretention c apacity in patient cells  (Fig. 2 K-L). Together, these data demonstrate that the \ndownregulation of Ca2+-sensing regulators, MICU1and MICU2 lowers the threshold for Ca2+ \nuptake and renders patient cells more vulnerable to [Ca2+]m overload and mPTP opening.  \n \nMitochondrial calcium overload triggers  mtDNA release in patient-derived fibroblasts \nbearing pathogenic EPG5 mutations  \nPathological excessive mitochondrial Ca2+ uptake can irreversibly lead to bioenergetic \ncollapse, mitochondrial swelling, and mPTP opening. This catastrophic loss of mitochondrial \nfunction is a major trigger for acute cell death (Bauer and Murphy, 2020). Alternatively, limited \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n12 \n \nmPTP opening in a subset of damaged mitochondria may lead t o a chronic low -grade innate \ninflammatory response. Previous studies have suggested that mPTP opening may allow \nmtDNA release into the cytosol thus driving activation of innate immune signalling and \ninflammation, a disease hallmark of ALS and other chronic  metabolic pathologies (Yu et al., \n2020; Xian et al., 2022). We therefore asked whether [Ca2+]m overload in patient cells drives \nthis pathway. Control and patient fibroblasts immunolabelled with anti-TOM20, anti-Citrate \nsynthetase (CS) and DNA antibodies to label OMM, IMM and mtDNA nucleoids, respectively, \nwere imaged using Airyscan microscopy which provides near super -resolution imaging, \nsufficient to visualise the integrity of the mitochondrial outer and inner membrane and mtDNA \nnucleoids (Fig. 3A). Both cell types displayed mtDNA nucleoids contained within a continuous \nIMM and OMM colocalizing together (Fig. S3A). However, we found cytosolic localisation of \nDNA puncta in over 30 % of patient cells as well as a significant increase in the average \ncytosolic DNA puncta per cell but none at all in control cells (Fig. 3B-C). Notably, patient cells \nalso displayed a fragmented mitochondrial network with swollen morphology (Fig. 3A). The \npresence of cytosolic DNA puncta in patient cells under normal condition s supported our \nhypothesis that increased resting [Ca2+]m in patient cells could cause chronic mtDNA release.  \n \nTo recapitulate the progressive chain of events which could lead to [Ca2+]m overload-induced \nmtDNA release, we first monitored ΔΨm in both control and patient fibroblasts following \nchallenge with 10 µM histamine  (Fig. 3D). Control cells maintained ΔΨm over time until the \nuncoupler-induced loss of ΔΨm. In contrast, patient cells exhibited a steep decrease in \nmembrane potential after the histamine challenge leading to more than a 30% reduction in ΔΨm \nover time (Fig. 3E). Loss of ΔΨm is one of the hallmark features of [Ca2+]m overload-induced \nmPTP opening.  Therefore, we next  co-labelled the cells with TMRM, mito -Fura-2 and \nPicoGreen (a potential -dependent DNA binding probe)  to monitor the intra-mitochondrial \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n13 \n \ndynamics of ΔΨm, [Ca2+]m and mtDNA respectively, over a prolonged time interval of 30 min \nto 1 h. Live-cell imaging of control fibroblasts following histamine challenge showed a slow \nincrease in [Ca2+]m which coincided with a small increase in ΔΨm which was maintained over \ntime (Fig. S3B-C). An increased fragmentation of the mitochondrial network was also observed \nafter histamine stimulation (Video S1 and Fig. S3D) suggesting a Ca2+-induced remodelling of \nmitochondrial morphology as previously reported (Chakrabarti et al., 2018). In comparison to \ncontrol cells , histamine challenge in patient cells caused a rapid  increase in  peak [Ca2+]m \nfollowed by a simultaneous fall in ΔΨm and mitochondrial swelling (Fig. 3F-G and Video S2). \nAnalyses of individual mitochondria in patient cells revealed the release of PicoGreen-labelled \nmtDNA associated with depolarization and swelling of mitochondria (Fig. 3H) that were never \nseen in control cells . Together, these data demonstrate the mechanism of [Ca2+]m overload-\ninduced mtDNA release in patient cells by identifying a progressive cascade of events starting \nfrom uncontrolled [Ca2+]m uptake and overload to loss of ΔΨm and mitochondrial swelling that \nculminates in mPTP opening and mtDNA release.  \n \nRelease of mtDNA drives the  activation of the cGAS-STING pathway and interferon \nresponse in patient-derived fibroblasts  \nTo characterize the cell -intrinsic changes in gene expression which may impact cellular \nfunction after mtDNA release, we performed bulk RNA sequencing (RNA -seq) on a control \n(control 1) and a pair of patient fibroblasts (patien t 1 and patient 3).  Gene set enrichment \nanalysis (GSEA)  of the gene signature enriched in patient fibroblasts revealed the top \nupregulated genes involved in the immune response (Fig. 4A-B and Fig. S4A-B). Among the \nmost significantly upregulated pathways, nine from the Gene Ontology biological process and \nseven from the KEGG pathway are related to immune activation and response. A gene \nexpression profile of the differentially expressed genes associated with the enriched pathways, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n14 \n \ncollected using a significance level of false discovery rate (FDR) <0.05 , uncovered a striking \nupregulation of the type I/III IFN signature gene set in the patient fibroblasts (Fig. 4C and Fig. \nS4C). Among the list of 45 IFN-stimulated genes (ISGs), we observed an upregulation of the \ngenes involved in direct antiviral activity ( Ifit1, Ifit3, Ifi44, Oasl), members of the ISG \ntranscription factors (Irf9 and Stat1) which are activated downstream of type I/III IFN receptor \nand chemokines (Ccl2 and Cxcl10) and pro-inflammatory cytokines (Tnf and Il-1b) involved \nin adaptive immune response.  Upregulation of these ISGs and pro-inflammatory genes was \nalso confirmed in the remaining patient fibroblasts with pathogenic EPG5 variants (Fig. 4D-E \nand Fig. S4D) suggesting that the induction of the IFN response pathway is likely a common \nfeature of EPG5 deficiency.   \n \nTo investigate the signalling events upstream of IFN induction in patient fibroblasts, we asked \nif the presence of cytosolic mtDNA could induce type I IFN by activating the canonical cGAS-\nSTING pathway.  cGAS is predominantly a nuclear protein, however, immunofluorescence \nanalysis revealed a significant increase in the intensity of the cytosolic cGAS in patient cells  \n(Fig. S4E-F) indicating its cytosolic relocalization  in response to immunogenic intracellular \nDNA (de Oliveira Mann and Hopfner, 2021). Live-cell imaging of fibroblasts over-expressing \nBFP-cGAS and co -labelled with TMRM and PicoGreen showed cytosolic cGAS puncta \ncolocalizing with free mtDNA devoid of mitochondria in patient cells (Fig. 4F-G) supporting \nthe notion that cytosolic cGAS observed in patient cells binds to mtDNA after its release from \nmitochondria. Cytosolic detection of mtDNA by cGAS activates classic STING signalosomes \nat the Golgi apparatus which in turn activates TBK1, a central kinase involved in the integration \nof innate immune signals from cytosolic DNA/RNA sensors to IFN induction by activating \ninterferon-regulatory factors (IRFs) (Liu et al., 2015; Zhang et al., 2019). In control and patient \nfibroblasts treated with histamine or thapsigargin, to increase [Ca 2+]c and amplify [Ca2+]m \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n15 \n \noverload, immunoblotting for active pTBK1 showed a substantial increase both under resting \nand histamine/thapsigargin-stimulated conditions in patient compared to control cells (Fig. 4H \nand Fig. S4G-H). Consistent with TBK1 activation, a significant increase in STING expression \nwas also detected in patient fibroblasts which further increased with histamine challenge (Fig. \n4I and Fig. S4I). Notably, we observed a robust activation of downstream transcription factors \nIRF3 and STAT1 in patient cells under resting condition which increased by more than two -\nfold after histamine challenge  (Fig. 4J and Fig. S4J-K).  Together, these results support the \nRNA-Seq and mRNA expression data and suggests that the cGAS-STING signalling is \nconstitutively active in patient -derived fibroblasts and further amplified by increased [Ca2+]m \nduring physiological histamine stimulation.  \n \nInhibition of mPTP opening by JP1-138 attenuates STING-dependent IFN response in  \npatient-derived fibroblasts  \nGiven that the constitutive activation of cGAS -STING signalling drives the downstream IFN \npathway in patient fibroblasts , we reasoned that  STING inhibition might curtail this \ninflammatory response. Treatment with H-151, a covalent inhibitor that blocks activation -\ninduced palmitoylation of STING (Haag et al., 2018), dampened the IFN response signalling \nin patient cells by reducing STING levels and normalizing active IRF3 and STAT1 comparable \nto the levels seen in control cells  (Fig. S5A-D). However, t reatment with G140, a cGAS-\nspecific inhibitor (Lama et al., 2019)  showed only a marginal effect on STING, pIRF3 and \npSTAT1 levels in patient cells (Fig. S5A-D). Furthermore, H-151 treatment also substantially \nreduced the expression levels of ISG s in patient cells (Fig. S5E-F) indicating that STING \ninhibition might represent a potential therapeutic  target in patients with pathogenic EPG5 \nvariants. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n16 \n \nWe next asked whether the STING-IFN response pathway might act in a feedback manner to \nfurther impair mitochondrial bioenergetic function in patient fibroblasts . Patient fibroblasts \nexhibited no significant change in resting ΔΨm after 24 h treatment with H -151 (Fig. 5A).  \nSimilarly, the NADH redox index remained unaltered in all the patient cells  (Fig. 5B). Long-\nterm treatment with H-151 for three days failed to show any significant increase in resting and \nmaximal respiration compared to untreated patient cells (Fig. 5C and Fig. S5G-I). These results \nsuggest that the inflammatory signature observed in patient fibroblasts is a consequence of \nimpaired mitochondrial Ca2+ homeostasis but does not contribute to the underlying \nbioenergetic defect of EPG5-related disorders. \n \nBased on our finding that mPTP inhibition by JP1-138 led to more than 50%  increase in \nmitochondrial calcium retention capacity of patient cells, we hypothesized that treatment with \nJP1-138 could also suppress downstream activation of  the cGAS-STING signalling and IFN \nresponse by protecting patient cells from [Ca2+]m overload and mtDNA release.  To test this, \nwe again monitored ΔΨm in both control and patient fibroblasts following a challenge with 10 \nµM histamine (Fig. 5D and Fig. S6A). As expected, patient cells showed a rapid los s of ΔΨm \nin comparison to control cells. However, p reincubation with JP1-138 completely rescued the \nhistamine-induced mitochondrial depolarization in patient cells  (Fig. 5E) suggesting  that \ntreatment with JP1-138 increases [Ca2+]m buffering capacity in intact patient fibroblasts. Next, \ncontrol and patient fibroblasts were treated with JP1 -138 for different time points and \nimmunoblotted to detect changes in the activation of the cGAS-STING signalling (Fig. 5F). \nInterestingly, JP1 -138 treatment was associated with a time -dependent decrease in the \nexpression of STING, pIRF3 and pSTAT1  reaching levels seen in control cells after 60 h. \nUsing this time point, we further tested the efficacy of JP1 -138 in supressing cGAS -STING \nsignalling by comparing it to another classic mPTP inhibitor, cyclosporin A (CsA)  (Fig. 5G \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n17 \n \nand Fig. S6B). Treatment with JP1-138 (100 nM) restored the activation of TBK1, IRF3 and \nSTAT1 and reduced the expression of STING to normal levels in patient cells  (Fig. 5H-I and \nFig. S6C-F). In comparison, CsA treatment showed  only a partial response and failed to \neffectively suppress the cGAS-STING signalling when used at a higher concentration (1 µM) \ndemonstrating the  specificity and potent activity of JP1 -138 in comparison to CsA.   \nImportantly, JP1 -138 treatment also  substantially reduced the expression levels of ISGs in \npatient cells (Fig. 5J-K). Together, these data  show that JP1-138 treatment increases [Ca2+]m \nbuffering capacity and protects patient cells from [Ca2+]m overload-induced mPTP opening and \nactivation of the cGAS-STING signalling.   \n \n JP1-138 suppresses mtDNA release and improves  mitochondrial bioenergetic function \nin patient fibroblasts bearing pathogenic EPG5 mutations  \nTo determine whether inhibition of mPTP by JP1-138 treatment also supressed chronic mtDNA \nrelease, control and patient fibroblasts immunolabelled with TOM20, CS and DNA were \nexamined using Airyscan imaging (Fig. 6A and Fig. S7A). As expected, patient cells showed \na significant  increase in  the cytosolic localization of DNA puncta as well as an increased \nproportion of the mitochondrial pool with fragmented and swollen morphology. Swollen \nmitochondria often displayed mtDNA extrusion in patient cells. Long term treatment with JP1-\n138 lead to a marked reduction in both the average number and the total percentage of cells \nwith cytosolic DNA puncta in patient fibroblasts  (Fig. 6B -C). Notably, we also observed \nremodelling of mitochondrial morphology in patient cells characterized by a s ignificant \nreduction in the swollen mitochondria pool (Fig. 6D) suggesting that JP1-138 treatment rescues \nthe mitochondrial bioenergetic defect in patient fibroblasts. Resting ΔΨm in patient cells \nincreased to the levels seen in control cells after long -term treatment with JP1-138 (Fig. 6E-F \nand Fig. S7B) indicating a significant repolarization of mitochondria in patient fibroblasts. The \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n18 \n \nincrease in ΔΨm was accompanied by a decrease in the NADH redox index (i.e. more oxidised \nNADH/less reduced NAD +) in co mparison to untreated patient fibroblasts  (Fig. 6G) . \nImprovement in these bioenergetic parameters was associated with a concomitant increase in \nmitochondrial respiration in patient fibroblasts  (Fig. 6H and Fig. S7C). Long-term treatment \nwith JP1-138 led to a small but significant increase in basal or ATP -linked respiration rate \nalong with a large increase in the spare respiratory capacity when compared to the untreated \npatient group  (Fig. S7D-E). However, compared to the ch anges in ΔΨm and NADH redox \nindex, the increase in mitochondrial respiration after JP1 -138 treatment was partial when \ncompared with untreated control fibroblasts. Together these data show that JP1-138 treatment \neffectively suppresses mtDNA release and part ially rescues mitochondrial bioenergetic \nfunction and morphology in patient fibroblasts carrying pathogenic EPG5 mutations.  \n \nImpaired mitochondrial Ca 2+ signalling sensitises Q336R iPSC -derived neurons to \nexcitotoxicity \nTo investigate the impact of impaired mitochondrial bioenergetic function and mitochondrial \nCa2+ signalling more specifically on the pathophysiology of neurodegeneration that is a severe \nfeature of patients with EPG5-related disorders, we used an  iPSC-derived cortical neuronal \nmodel, reprogrammed from patient 1 fibroblasts bearing the pathogenic founder Q336R \nmissense mutation and a CRISPR/Cas9-edited isogenic control.  In vitro neurodifferentiation \nof isogenic and Q336R iPSCs yielded mixed cultures of neurons and glial cells with no \nsignificant change in the proportion of cell types (Fig. S8A-B). The increased bioenergetic \ndemand imposed during  metabolic workload such as  physiological glutamate -induced \nexcitation may sensitize neurons with impaired mitochondria l function to glutamate \nexcitotoxicity (Plotegher et al., 2021) . To test th is, mixed cultures of neurons and glial cells, \nloaded with low-affinity [Ca2+]c indicator FuraFF-AM were challenged with near physiological \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n19 \n \n(10 µM) and toxic (100 µM) glutamate concentrations (Fig. 7A -B). Early peak [Ca2+]c \n(ΔFuraFFearly) remained unchanged between isogenic and Q336R neurons in response to 10 \nµM glutamate (Fig. 7C). However, in the majority of Q336R neurons but never in the isogenic \ncontrol cells, the initial transient response was followed by  a progressive secondary increase \nof [Ca2+]c (Fig. 7D-E). This characteristic response, referred to as delayed calcium deregulation \n(DCD) was evident only at 100 µM in control cells (Fig. 7A). Glutamate-induced DCD drives \nexcitotoxic neuronal cell death through a cascade of pathways  (Tymianski, 2011; Llorente-\nFolch et al., 2015). Disruption of mitochondrial Ca2+ homeostasis is the key contributing factor \nduring glutamate excitotoxicity  (Plotegher et al., 2021) . Therefore, we first measured the \nexpression of the regulatory components of the MCU complex using immunoblotting (Fig. 7F). \nAs seen in the patient fibroblasts, protein expression of MICU1 and MICU3, a brain- specific \nCa2+ sensing regulator, were significantly reduced in Q336R neurons along with EMRE, while \nthe relative levels of MCU and NCLX remained unchanged (Fig. 7G). BN-PAGE analysis of \nmitochondria isolated from isogenic and Q336R neuronal cultures revealed  the ~1.1 MDa \nassembly representing MCU complex containing gatekeeper subunits in isogenic mitochondria \nand a prominent constitutively active ~400 kDa MCU-EMRE complex devoid of gatekeeper \nproteins in Q336R mitochondria (Fig. S8C-E) as previously described in cortical mitochondria \nof Afg3l2 neuron-specific knockout mice with spastic ataxia-neuropathy (Konig et al., 2016). \nIn addition to defective neuronal MCU assembly, native complex I assembly was also reduced \nin Q336R mitochondria  (Fig. S8C). Analysis of mRNA expression  showed no significant \nchange among the components of the MCU complex (Fig. S8F) suggesting a post-translational \nregulation of MICU1/MICU3 expression/assembly in Q336R neurons.  \n \nConsistent with these results, the measurement of [Ca2+]m in isogenic and Q336R neurons \nloaded with mito-Fura-2 AM, showed a higher resting [Ca2+]m in Q336R neurons (Fig. S8G). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n20 \n \nMoreover, confocal images analysis of ratiometric images (F 355/F405) revealed a swollen \nmitochondrial pool with a high F355/F405 ratio in Q336R neurons compared to isogenic control \nneurons which showed elongated mitochondrial morphology with a low F355/F405 ratio (Fig. \nS8H). Exposure to a physiological concentration of glutamate evoked an early rise in [Ca2+]m \nand a larger [Ca2+]m peak in Q336R neurons compared to isogenic control neurons (Fig. 7H-\nL). Ratiometric image analysis also showed a rapid [Ca2+]m rise in Q336R neurons in response \nto glutamate challenge (Fig. S8I). Glutamate-induced [Ca2+]m overload in Q336R neurons was \nconfirmed using the low -Ca2+ affinity mito -aequorin probe  (Fig. S8J). These results are \nconsistent with [Ca2+]m overload as a cause of excitotoxic glutamate -induced DCD and \nneuronal cell death. Therefore, we analysed cytochrome c (Cyt c) release in neurons exposed \nto 10 µM glutamate overnight (12 h) by immunolabelling neurons for the TOM20 and Cyt c \n(Fig. 7M). The release of cytochrome c from m itochondria (TOM20+ and Cyt c-) is the major \ntrigger for caspase activation during the initial events of apoptotic cell death  cascade. After \nglutamate challenge, Q336R neurons showed a marked increase in  TOM20 intensity without \nCyt c intensity suggesting Cyt c release in comparison to isogenic control neurons (Fig. 7N). \nNotably, Q336R neurons also showed an aberrant remodelling of mitochondrial morphology \naccompanied by a significant increase in swollen mitochondria po ol in the somas and a more \nfragmented form in  the axons (Fig. S8K). Altogether, these results show that impaired \nmitochondrial Ca2+ signalling and the ensuing [Ca2+]m overload increases the vulnerability to \nglutamate-induced DCD and cell death in Q336R neurons. \n \nJP1-138 improves mitochondrial Ca 2+ buffering capacity and bioenergetic function in \nQ336R neurons  \nAn increase in [ Ca2+]m is the most potent inducer of mPTP opening and therefore provides a \ndirect link between sustained in creases in [ Ca2+]m and mitochondrial dysfunction during \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n21 \n \nexcitotoxic injury in neurons (Keelan et al., 1999; Abramov and Duchen, 2008). Given that the \npharmacological inhibition of mPTP in patient fibroblasts sup presses mtDNA-induced IFN \nresponse and partially restores mitochondrial bioenergetic function, we first determined \nmitochondrial Ca2+ retention capacity of neurons which were digitonin permeabilised after co-\nlabelling with Fluo-4 AM (which is lost from the cytosol but retained in mitochondria after \npermeabilization (McKenzie et al., 2017) and TMRM to monitor dynamic changes in [Ca2+]m \nand ΔΨm, respectively, using sequential additions of exogenous CaCl2 as described in Fig. S9A. \nThe highest fold-change in Fluo-4 intensity was evident only after the fourth successive Ca2+ \naddition (6.91 µM final Ca2+) in permeabilised isogenic neurons , indicating the Ca2+ uptake \nthreshold and cooperative activation of MCU by Ca2+ (Fig. 8A). Subsequent Ca2+ addition \n(13.3 µM final Ca2+) induced permeability transition culminating in the collapse of ΔΨm as \nindicated by the rapid loss of the TMRM signal. In contrast, mitochondria from Q336R neurons \nshowed an almost linear increase in Fluo-4 intensity with successive Ca2+ additions indicating \na lower threshold for activation and a constitutively active MCU complex. As a consequence \nof [ Ca2+]m overload, mPTP opening and rapid dissipation of ΔΨm was observed at lower \nconcentrations of Ca2+ (between 1.32 and 3.28 µM) in the Q336R cells (Fig. 8B). Notably, the \npreincubation of Q336R mitochondria with JP1-138 significantly increased the threshold of \nmPTP opening and the loss of ΔΨ m without affecting the rate of Ca2+ uptake indicating the \nprotective effect of JP1-138 from [Ca2+]m overload-induced mPTP opening and mitochondrial \ndepolarisation (Fig. 8A-B). To demonstrate this under (patho)physiological state, isogenic and \nQ336R neurons were loaded with Rhodamine-123 in ‘dequench mode’  to monitor time-\ndependent changes in ΔΨm following exposure to glutamate . After an initial transient \ndepolarization indicated by an early rise in fluorescence in response to glutamate , ΔΨ m \nrecovered to the baseline in almost all the isogenic neurons  (Fig. 8C and 8F ). In contrast , \nQ336R neurons showed a delayed secondary depolarisation in the majority of the neurons (Fig. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n22 \n \n8D), coincident with the glutamate -induced DCD response  (Abramov and Duchen, 2008) . \nMoreover, an uncoupler-induced mitochondrial depolarization caused a smaller additional \ndepolarization in comparison to isogenic neurons confirming near total dissipation of ΔΨ m \ninduced by  glutamate in the Q336R neurons .  Importantly, preincubation with JP1 -138 \ncompletely protected the Q336R neurons from glutamate -induced collapse of ΔΨm (Fig. 8E-\nG). To determine the bioenergetic differences under unstimulated resting state, we measured \nbasal and maximal respiration in isogenic and Q336R neurons. Both ATP -linked respiration \nrate and spare respiratory capacity were reduced in Q336R neurons (Fig. 8H). However, long-\nterm treatment with JP1 -138 resulted in a small but significant increase in both basal and \nmaximal respiration rate in Q336R neurons (Fig. 8I-J). Together, these results demonstrate that \nJP1-138 effectively increases mitochondrial Ca2+ buffering capacity and protects Q336R \nneurons from glutamate -induced loss of ΔΨ m, and improves mitochondrial respiration, thus \nhighlighting a potential therapeutic strategy to target neurodegeneration  in EPG5-related \ndisorders. \n \nDiscussion \nWhilst defective autophagy has been identified as a key mechanism in EPG5-related disorders, \ndownstream pathogenic effects remain poorly understood . Our recent findings suggested \nimpaired mitochondrial homeostasis due to defective PINK1/PARKIN -dependent mitophagy \nin patient-derived cells (Dafsari et al., 2024). Here, we show that impaired mitochondrial Ca2+ \nsignalling and the associated bioenergetic insufficiency could be major contributing factors to \nthe neurodegeneration observed in patients with EPG5-RD. We found that mitochondrial Ca2+ \noverload and increased vulnerability to mPTP opening drives mtDNA release and  the cGAS-\nSTING dependent IFN response in patient fibroblasts bearing pathogenic EPG5 variants. \nPatients with either truncating or missense mutations showed almost complete deficiency of \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n23 \n \nnormal EPG5 protein (~280 kDa) suggesting the reduced stability of dysfunctional prot ein. \nSimilar results were also reported in hESC (Vidyawan et al., 2025) and LCL (Piano Mortari et \nal., 2018)  model with EPG5 haploinsufficiency. An impaired mitochondrial bioenergetic \nfunction and morpholog y was found in all patient fibroblasts. The mitochondrial respiratory \ndefect was associated with downregulation of OXPHOS subunit expression despite an increase \nin mtDNA copy number.  Gene enrichment analysis showed upregulation of PPARGC1A \ntranscription factor encoding PGC1α, a master regulator of mitochondrial biogenesis  \n(Supplementary Table S2), indicating an adaptive reaction that could compensate in part for \nimpaired mitophagy in patient fibroblasts.  Mitophagy defects may also impact downstream \nmitochondrial quality control regulated by mitochondrial proteases involved in remodelling of \nIMM proteome and initiating mitochondrial unfolded protein response  (Uoselis et al., 2023; \nChakrabarty et al., 2024; Yamada et al., 2025). \n \nUnexpectedly, we found downregulation of MICU1/2 in patient fibroblasts and MICU1/3 in \nEPG5-mutant neurons as the causative mechanism for [Ca2+]m overload and ensuing mPTP \nopening. Pathogenic mutations in MICU1 itself are associated with a rare childhood disorder \nthat include neurodevelopmental, neurodegenerative and neuromuscular phenotypes with some \nclose parallels to EPG5-RD (Logan et al., 2014; Bhosale et al., 2017; Kohlschmidt et al., 2021). \nPatients with loss -of-function MICU1 mutations present with a progressive extrapyram idal \nneurodegenerative disorder (Logan et al., 2014), features that have a significant overlap with \nthe clinical phenotype associated with EPG5-RD, including ataxia, dystonia, and tremor as \nwell as basal ganglia abnormalities in magnet ic resonance imaging. Similarly, microcephaly, \nmyopathy, seizures and immunodeficiency also present in MICU1 deficiency as well as in \nEPG5 deficiency.  Furthermore, [Ca2+]m overload and mitochondrial dysfunction due to \ndownregulation of NLCX and MCU remodelling has been implicated in the neuropathology \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n24 \n \nand memory decline in Alzheimer’s disease (Jadiya et al., 2019). Although the mechanism of \nMCU remodelling in patient fibroblasts and EPG5 mutant neurons defined by downregulation \nof MICUs and EMRE protein levels has not been investigated here, it seems likely that stress-\ninduced activity of quality control proteases in IMM may control the protein turnover of MCU \nregulatory subunits.  Proteostasis stress  driven by impaired mitophagy may lead to the \nprocessing of DELE1 by OMA1 protease , culminating in activation of the integrated stress \nresponse (ISR) (Uoselis et al., 2023; Yamada et al., 2025) as evident by the ISR signature from \nRNA-seq data in patient fibroblasts  (Supplementary Table S2). It is possible that  the activity \nof alternative IMM or matrix proteases such as YME1L1 (Tsai et al., 2022) or AFG3L2 (Konig \net al., 2016)  respectively may control the processing and degradation of MICUs and EMRE  \nrespectively during the ISR.  \n \nWidespread mPTP opening causes pathological collapse of mitochondrial bioenergetic \nfunction, ATP hydrolysis  and depletion, mitochondrial swelling and rupture culminating in \nneuronal cell death as evident during ischemic stroke (Schinzel et al., 2005). In contrast, limited \nor selective opening of mPTP, a feature also evident d uring histamine -induced [ Ca2+]m \noverload in patient fibroblasts, may cause insufficient damage to lead to cell death, but enable \nthe release of mtDNA in to the cytosol, where it triggers the activation of  the cGAS-STING \npathway, driving an inflammatory IFN response. This protective inflammatory program could \nbe constitutively driven by patient fibroblasts in a chronic low -grade manner t hat maintains \nproliferation and evade s acute cell death. A recent study described a similar process ter med \nminority mitochondrial outer membrane permeabilization  (miMOMP) where MOMP \noccurring in a subset of mitochondria enables mtDNA efflux through BAK/BAX macropores \nand activates the cGAS-STING pathway to drive age-associated inflammation in mice (Ichim \net al., 2015; Victorelli et al., 2023). Mice deficient in Epg5 and other primary autophagy genes \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n25 \n \nsuch as Atg5 and Atg7 exhibit elevated cytokine expression and  cellular lung inflammation \nsufficient to inhibit influenza pathogenesis and evade cell death  (Lu et al., 2016) . Notably, \nthese cytokine levels are also increased in patients with EPG5-RD (Piano Mortari et al., 2018) \nand show a hyperimmunity against influenza.  Constitutive activation of STING-mediated \nantiviral signalling arising as a result of defective mitophagy and mtDNA release may provide \nthe explanation for the lack of influenza cases in patients with EPG5-RD. Consistent with these \nreports, we also found upregulated expression of pro-survival NF-κB target genes  such as  \ncFLIP and cIAP (Supplementary Table S2) in patient RNA-seq datasets in addition to the type \nI/III IFN signature genes.  \n \nThe pathogenic Q336R EPG5 mutation in iPSC -derived neurons impaired mitochondrial \nrespiration and reduced CI assembly similar to patient fibroblasts . The exposure to glutamate \nat concentrations that are innocuous for isogenic neurons  caused a rapid collapse of  \nbioenergetic capacity, disruption of [ Ca2+]c homeostasis and excitotoxic death in Q336R \nneurons. Glutamate-induced excitotoxicity in Q336R neurons is attributed to a lower \nmitochondrial Ca2+ uptake threshold and [Ca2+]m overload which combined with increased \nROS production leads to increased susceptibility to mPTP opening caused by downregulation \nof MICU1/3 holocomplex and a constitutive ly active MCU complex . Thus, an increased \nbioenergetic demand during neuronal metabolic workload may amplify the compromised \nmitochondrial function due to [ Ca2+]m overload and trigger a rapid pathological cascade \nculminating in cell death. Our data demonstrate that an impaired mitochondrial metabolism \nand Ca2+ signalling sensitises neurons to excitotoxic injury which may play an important role \nin neurodegeneration in EPG5-RD and may also contribute to epileptogenesis observed in \naround two -thirds of patients  (Deneubourg et al., 2025) . Loss of AFG3L2 activity, a \nmitochondrial m-AAA protease associated with spinocerebellar ataxia  and loss of  Purkinje \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n26 \n \nneuron, results in  impaired processing of EMRE and assembly of MCU complex enabling \n[Ca2+]m overload, mPTP opening and neuronal death  in mice (Konig et al., 2016). Moreover, \nneuronal loss of MICU1 increased [ Ca2+]m overload-induced excitotoxicity and caused \nprogressive degeneration of motor neurons in MICU1 -KO mice (Singh et al., 2022) . These \nfindings further highlight the impaired mitochondrial Ca2+ homeostasis as a key driver of \nexcitotoxic neuronal cell death in early-onset neurodegenerative diseases.  \n \nPrevention of [ Ca2+]m overload-induced mPTP opening by JP1 -138 effectively increased \nmitochondrial Ca2+ buffering capacity and partially rescued mitochondrial bioenergetic \nfunction in both patient fibroblasts and Q336R neurons. Prolonged treatment with JP1-138 also \nsuppressed mtDNA release and downstream activation of the cGAS-STING pathway and IFN \nresponse in patient fibroblasts in comparison to CsA, the prototypical inhibitor of mPTP with \noff-target effects (Briston et al., 2019). A significant but partial improvement of mitochondrial \nbioenergetic function by JP1 -138 underscores the pathophysiological impact of impaired \nmitophagy in regulating the neuronal energy homeostasis. Further investigations based on the \npharmacological modulatio n of impaired mitophagy are need ed to examine the role of \nmitophagy in mitigating inflammation and restoring mitochondrial function in EPG5-deficient \ncells. Inhibition of STING  using H-151 equally supressed STING -dependent IFN response, \nalthough it failed to rescue mitochondrial dysfunction. Several studies have demonstrated that \ngenetic ablation or pharmacological inhibition of STING protects against inflammation -\nmediated neuronal loss in mouse models of PD, ALS and lysosomal storage diseases (Sliter et \nal., 2018; Yu et al., 2020; Wang et al., 2024 ). Our findings suggest that the pharmacological \neffects of cGAS/STING inhibitors could be dependent on the mechanism mediating the \npathogenesis which, in patient fibroblasts with EPG5 mutations, is a consequence of impaired \nmitochondrial homeostasis. Submicromolar concentrations of JP1-138 were highly effective at \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n27 \n \nmaintaining ΔΨm even after [Ca2+]m overload in both patient fibroblasts and Q336R neurons. \nOur group recently reported improved mitochondrial activity and preclinical safety of JP1-138 \ncompound with 20-fold greater brain concentration following a single dose in mice (Pingitore \net al., 2024), highlighting its therapeutic potential in targeting Ca2+ mishandling and \nexcitotoxicity in EPG5-RD and other neurodegenerative diseases.  \n \n In summary, this  study expands our understanding of how impaired mitochondrial Ca2+ \nsignalling and dysfunction can contribute to the presentation and progression of EPG5-RD by \nidentifying a cascade of events starting from bioenergetic deficiency, unregulated \nmitochondrial Ca2+ uptake and Ca2+ overload to mPTP opening that culminates in low-grade \nmtDNA-driven chronic inflammation in patient fibroblasts and excitotoxicity and cell death in \nneurons. This progressive chain of events  may account for the clinically variable  and \nmultisystem pathophysiology of EPG5-RD but also help to identify multiple sites for \npharmacological intervention as potential novel therapeutic targets . These findings will \ntogether serve to improve our understanding of the progressive neurodegeneration in patients \nwith EPG5-related disorders and further aid patients and families in genetic counselling. \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n28 \n \nReferences                                                                                                                                                                                                                                          \n1. Abramov, A.Y., and Duchen, M.R. (2008). Mechanisms underlying the loss of mitochondrial \nmembrane potential in glutamate excitotoxicity. Biochim Biophys Acta 1777, 953-964. \n \n2. Area-Gomez, E., Guardia -Laguarta, C., Schon, E.A., and Przedborski, S. (2019). Mitochondria, \nOxPhos, and neurodegeneration: cells are not just running out of gas. J Clin Invest 129, 34-45. \n \n3. Balasubramaniam, S., Riley, L.G., Vasudevan, A., Cowley, M.J., Gayevskiy, V., Sue, C.M., \nEdwards, C., Edkins, E., Junckerstorff, R., Kiraly -Borri, C. , et al.  (2018). EPG5 -Related Vici \nSyndrome: A Primary Defect of Autophagic Regulation with an Emerging Phenotype Overlapping with \nMitochondrial Disorders. JIMD Rep 42, 19-29. \n \n4. Bauer, T.M., and Murphy, E. (2020). Role of Mitochondrial Calcium and the Permeability Transition \nPore in Regulating Cell Death. Circ Res 126, 280-293. \n \n5. Baughman, J.M., Perocchi, F., Girgis, H.S., Plovanich, M., Belcher -Timme, C.A., Sancak, Y., Bao, \nX.R., Strittmatter, L., Goldberger, O., Bogorad, R.L. , et al.  (2011). Integrative genomics identifies \nMCU as an essential component of the mitochondrial calcium uniporter. Nature 476, 341-345. \n \n6. Bhosale, G., Sharpe, J.A., Koh, A., Kouli, A., Szabadkai, G., and Duchen, M.R. (2017). Pathological \nconsequences of MICU 1 mutations on mitochondrial calcium signalling and bioenergetics. Biochim \nBiophys Acta Mol Cell Res 1864, 1009-1017. \n \n7. Bhosale, G., Sharpe, J.A., Sundier, S.Y., and Duchen, M.R. (2015). Calcium signaling as a mediator \nof cell energy demand and a trigger to cell death. Ann N Y Acad Sci 1350, 107-116. \n \n8. Boyman, L., Karbowski, M., and Lederer, W.J. (2020). Regulation of Mitochondrial ATP \nProduction: Ca(2+) Signaling and Quality Control. Trends Mol Med 26, 21-39. \n \n9. Brenner, C., and Moulin, M. (2012). Physiological roles of the permeability transition pore. Circ Res \n111, 1237-1247. \n \n10. Briston, T., Selwood, D.L., Szabadkai, G., and Duchen, M.R. (2019). Mitochondrial Permeability \nTransition: A Molecular Lesion with Multiple Drug Targets. Trends Pharmacol Sci 40, 50-70. \n \n11. Byrne, S., Dionisi -Vici, C., Smith, L., Gautel, M., and Jungbluth, H. (2016a). Vici syndrome: a \nreview. Orphanet J Rare Dis 11, 21. \n \n12. Byrne, S., Jansen, L., JM, U.K.-I., Siddiqui, A., Lidov, H.G., Bodi, I., Smith, L., Mein, R., Cullup, \nT., Dionisi-Vici, C., et al. (2016b). EPG5-related Vici syndrome: a paradigm of neurodevelopmental \ndisorders with defective autophagy. Brain 139, 765-781. \n \n13. Chakrabarti, R., Ji, W.K., Stan, R.V., de Juan Sanz, J., Ryan, T.A., and Higgs, H.N. (20 18). INF2-\nmediated actin polymerization at the ER stimulates mitochondrial calcium uptake, inner membrane \nconstriction, and division. J Cell Biol 217, 251-268. \n \n14. Chakrabarty, Y., Yang, Z., Chen, H., and Chan, D.C. (2024). The HRI branch of the integrate d \nstress response selectively triggers mitophagy. Mol Cell 84, 1090-1100 e1096. \n \n15. Choi, M.L., Chappard, A., Singh, B.P., Maclachlan, C., Rodrigues, M., Fedotova, E.I., Berezhnov, \nA.V., De, S., Peddie, C.J., Athauda, D. , et al.  (2022). Pathological structural conversion of alpha -\nsynuclein at the mitochondria induces neuronal toxicity. Nat Neurosci 25, 1134-1148. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n29 \n \n16. Chu, C.T. (2022). Mitochondria in neurodegeneration. Curr Opin Physiol 26. \n \n17. Csordas, G., Golenar, T., Seifert, E.L., Kamer, K.J., Sancak, Y., Perocchi, F., Moffat, C., Weaver, \nD., Perez, S.F., Bogorad, R., et al. (2013). MICU1 controls both the threshold and cooperative activation \nof the mitochondrial Ca(2)(+) uniporter. Cell Metab 17, 976-987. \n \n18. Cullup, T., Kho, A.L., Dionisi-Vici, C., Brandmeier, B., Smith, F., Urry, Z., Simpson, M.A., Yau, \nS., Bertini, E., McClelland, V. , et al.  (2013). Recessive mutations in EPG5 cause Vici syndrome, a \nmultisystem disorder with defective autophagy. Nat Genet 45, 83-87. \n \n19. Dafsari, H.S., Deneubourg, C., Singh, K., Maroofian, R., Suprenant, Z., Kho, A.L., Ingham, N.J., \nSteel, K.P., Sheshadri, P., Baur, F. , et al.  (2024). Mutations in EPG5 are associated with a wide \nspectrum of neurodevelopmental and neurodegenerative disorders. 2024.2006.2012.24308722. \n \n20. Dafsari, H.S., Ebrahimi -Fakhari, D., Saffari, A., Deneubourg, C., Fanto, M., and Jungbluth, H. \n(2022). EPG5-Related Disorder. In GeneReviews((R)), M.P. Adam, J. Feldman, G.M. Mirzaa, R.A. \nPagon, S.E. Wallace, and A. Amemiya, eds. (Seattle (WA)). \n \n21. De Nadai, A., Vajente, N., Pendin, D., and Mattarei, A. (2021). Mt -fura-2, a Ratiometric \nMitochondria-Targeted Ca(2+) Sensor. Determination of Spectroscopic Properties and Ca(2+) Imaging \nAssays. Methods Mol Biol 2275, 187-215. \n \n22. de Oliveira Mann, C.C., and Hopfner, K.P. (2021). Nuclear cGAS: guard or prisoner? EMBO J 40, \ne108293. \n \n23. De Stefani, D., Raffaello, A., Teardo, E., Szabo, I., and Rizzuto, R. (2011). A forty -kilodalton \nprotein of the inner membrane is the mitochondrial calcium uniporter. Nature 476, 336-340. \n \n24. Deneubourg, C., Ramm, M., Smith, L.J., Baron, O., Singh, K., Byrne, S.C., Duchen, M.R., Gautel, \nM., Eskelinen, E.L., Fanto, M., et al. (2022). The spectrum of neurodevelopmental, neuromuscular and \nneurodegenerative disorders due to defective autophagy. Autophagy 18, 496-517. \n \n25. Deneubourg, C., Salimi Dafsari, H., Lowe, S., Martinez -Cotrina, A., Mazaud, D., Park, S.H., \nVergani, V., Almacellas Barbanoj, A., Maroofian, R., Averdunk, L. , et al.  (2025). Epg5 links \nproteotoxic stress due to defective autophagic clearance and epileptogenesis in Drosophila and Vici \nsyndrome patients. Autophagy 21, 447-459. \n \n26. Denton, R.M. (2009). Regulation of mitochondrial dehydrogenases by calcium ions. Biochim \nBiophys Acta 1787, 1309-1316. \n \n27. Denton, R.M., Randle, P.J., and Martin, B.R. (1972). Stimulation by calcium ions of pyruvate \ndehydrogenase phosphate phosphatase. Biochem J 128, 161-163. \n \n28. Devine, M.J., and Kittler, J.T. (2018). Mitochondria at the ne uronal presynapse in health and \ndisease. Nat Rev Neurosci 19, 63-80. \n \n29. Duchen, M.R. (1992). Ca(2+) -dependent changes in the mitochondrial energetics in single \ndissociated mouse sensory neurons. Biochem J 283 ( Pt 1), 41-50. \n \n30. Duchen, M.R. (2012). Mitochondria, calcium -dependent neuronal death and neurodegenerative \ndisease. Pflugers Arch 464, 111-121. \n \n31. Duchen, M.R., and Szabadkai, G. (2010). Roles of mitochondria in human disease. Essays Biochem \n47, 115-137. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n30 \n \n32. Fan, M., Zhang, J., Tsai, C.W., Orlando, B.J., Rodriguez, M., Xu, Y., Liao, M., Tsai, M.F., and \nFeng, L. (2020). Structure and mechanism of the mitochondrial Ca(2+) uniporter holocomplex. Nature \n582, 129-133. \n \n33. Filadi, R., and Pizzo, P. (2020). Mitochondrial calcium handling and neuro degeneration: when a \ngood signal goes wrong. Current Opinion in Physiology 17, 224-233. \n \n34. Garbincius, J.F., and Elrod, J.W. (2022). Mitochondrial calcium exchange in physiology and \ndisease. Physiol Rev 102, 893-992. \n \n35. Giorgi, C., Marchi, S., and Pinton, P. (2018). The machineries, regulation and cellular functions of \nmitochondrial calcium. Nat Rev Mol Cell Biol 19, 713-730. \n \n36. Guo, B., Liang, Q., Li, L., Hu, Z., Wu, F., Zhang, P., Ma, Y., Zhao, B., Kovacs, A.L., Zhang, Z., et \nal. (2014). O-GlcNAc-modification of SNAP -29 regulates autophagosome maturation. Nat Cell Biol  \n16, 1215-1226. \n \n37. Haag, S.M., Gulen, M.F., Reymond, L., Gibelin, A., Abrami, L., Decout, A., Heymann, M., van der \nGoot, F.G., Turcatti, G., Behrendt, R. , et al.  (2018). Targeting STING with covalent small -molecule \ninhibitors. Nature 559, 269-273. \n \n38. Hori, I., Otomo, T., Nakashima, M., Miya, F., Negishi, Y., Shiraishi, H., Nonoda, Y., Magara, S., \nTohyama, J., Okamoto, N. , et al.  (2017). Defects in autophagosome -lysosome fusion underlie Vici \nsyndrome, a neurodevelopmental disorder with multisystem involvement. Sci Rep 7, 3552. \n \n39. Ichim, G., Lopez, J., Ahmed, S.U., Muthalagu, N., Giampazolias, E., Delgado, M.E., Haller, M., \nRiley, J.S., Mason, S.M., Athineos, D. , et al. (2015). Limited mitochondrial permeabilization causes \nDNA damage and genomic instability in the absence of cell death. Mol Cell 57, 860-872. \n \n40. Jadiya, P., Garbincius, J.F., and Elrod, J.W. (2021). Reappraisal of metabolic dysfunction in \nneurodegeneration: Focus on mitochondrial function and calcium signaling. Acta Neuropathol \nCommun 9, 124. \n \n41. Jadiya, P., Kolmetzky, D.W., Tomar, D., Di Meco, A., Lombardi, A.A., Lambert, J.P., Luongo, \nT.S., Ludtmann, M.H., Pratico, D., and Elrod, J.W. (2019). Im paired mitochondrial calcium efflux \ncontributes to disease progression in models of Alzheimer's disease. Nat Commun 10, 3885. \n \n42. Jauhari, A., Baranov, S.V., Suofu, Y., Kim, J., Singh, T., Yablonska, S., Li, F., Wang, X., Oberly, \nP., Minnigh, M.B. , et al.  (2020). Melatonin inhibits cytosolic mitochondrial DNA -induced \nneuroinflammatory signaling in accelerated aging and neurodegeneration. J Clin Invest 130, 3124-3136. \n \n43. Jimenez-Loygorri, J.I., Villarejo -Zori, B., Viedma -Poyatos, A., Zapata -Munoz, J., Ben itez-\nFernandez, R., Frutos -Lison, M.D., Tomas -Barberan, F.A., Espin, J.C., Area -Gomez, E., Gomez -\nDuran, A., et al. (2024). Mitophagy curtails cytosolic mtDNA -dependent activation of cGAS/STING \ninflammation during aging. Nat Commun 15, 830. \n \n44. Jouaville, L.S., Pinton, P., Bastianutto, C., Rutter, G.A., and Rizzuto, R. (1999). Regulation of \nmitochondrial ATP synthesis by calcium: evidence for a long-term metabolic priming. Proc Natl Acad \nSci U S A 96, 13807-13812. \n \n45. Kamer, K.J., and Mootha, V.K. (2015). The molecular era of the mitochondrial calcium uniporter. \nNat Rev Mol Cell Biol 16, 545-553. \n \n46. Keelan, J., Vergun, O., and Duchen, M.R. (1999). Excitotoxic mitochondrial depolarisation requires \nboth calcium and nitric oxide in rat hippocampal neurons. J Physiol 520 Pt 3, 797-813. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n31 \n \n \n47. Klionsky, D.J., Abdel-Aziz, A.K., Abdelfatah, S., Abdellatif, M., Abdoli, A., Abel, S., Abeliovich, \nH., Abildgaard, M.H., Abudu, Y.P., Acevedo -Arozena, A., et al.  (2021). Guidelines for the use and \ninterpretation of assays for monitoring autophagy (4th edition)(1). Autophagy 17, 1-382. \n \n48. Kohlschmidt, N., Elbracht, M., Czech, A., Hausler, M., Phan, V., Topf, A., Huang, K.T., Bartok, \nA., Eggermann, K., Zippel, S., et al. (2021). Molecular pathophysiology of human MICU1 deficiency. \nNeuropathol Appl Neurobiol 47, 840-855. \n \n49. Konig, T., Troder, S.E., Bakka, K., Korwitz, A., Richter -Dennerlein, R., Lampe, P.A., Patron, M., \nMuhlmeister, M., Guerrero-Castillo, S., Brandt, U., et al. (2016). The m-AAA Protease Associated with \nNeurodegeneration Limits MCU Activity in Mitochondria. Mol Cell 64, 148-162. \n \n50. Lama, L., Adura, C., Xie, W., Tomita, D., Kamei, T., Kuryavyi, V., Gogakos, T., Steinberg, J.I., \nMiller, M., Ramos-Espiritu, L., et al.  (2019). Development of human cGAS -specific small-molecule \ninhibitors for repression of dsDNA-triggered interferon expression. Nat Commun 10, 2261. \n \n51. Lautenschlager, J., Prell, T., Ruhmer, J., Weidemann, L., Witte, O.W., and Gr osskreutz, J. (2013). \nOverexpression of human mutated G93A SOD1 changes dynamics of the ER mitochondria calcium \ncycle specifically in mouse embryonic motor neurons. Exp Neurol 247, 91-100. \n \n52. Lewis-Smith, D., Kamer, K.J., Griffin, H., Childs, A.M., Pysde n, K., Titov, D., Duff, J., Pyle, A., \nTaylor, R.W., Yu-Wai-Man, P., et al. (2016). Homozygous deletion in MICU1 presenting with fatigue \nand lethargy in childhood. Neurol Genet 2, e59. \n \n53. Liu, S., Cai, X., Wu, J., Cong, Q., Chen, X., Li, T., Du, F., Ren, J., Wu, Y.T., Grishin, N.V., et al. \n(2015). Phosphorylation of innate immune adaptor proteins MAVS, STING, and TRIF induces IRF3 \nactivation. Science 347, aaa2630. \n \n54. Llorente-Folch, I., Rueda, C.B., Pardo, B., Szabadkai, G., Duchen, M.R., and Satrustegui, J. (2015). \nThe regulation of neuronal mitochondrial metabolism by calcium. J Physiol 593, 3447-3462. \n \n55. Logan, C.V., Szabadkai, G., Sharpe, J.A., Parry, D.A., Torelli, S., Childs, A.M., Kriek, M., Phadke, \nR., Johnson, C.A., Roberts, N.Y., et al. (2014). Loss-of-function mutations in MICU1 cause a brain and \nmuscle disorder linked to primary alterations in mitochondrial calcium signaling. Nat Genet  46, 188-\n193. \n \n56. Lu, Q., Yokoyama, C.C., Williams, J.W., Baldridge, M.T., Jin, X., DesRochers, B., Bricker,  T., \nWilen, C.B., Bagaitkar, J., Loginicheva, E. , et al.  (2016). Homeostatic Control of Innate Lung \nInflammation by Vici Syndrome Gene Epg5 and Additional Autophagy Genes Promotes Influenza \nPathogenesis. Cell Host Microbe 19, 102-113. \n \n57. McClelland, V., Cullup, T., Bodi, I., Ruddy, D., Buj-Bello, A., Biancalana, V., Boehm, J., Bitoun, \nM., Miller, O., Jan, W. , et al.  (2010). Vici syndrome associated with sensorineural hearing loss and \nevidence of neuromuscular involvement on muscle biopsy. Am J Med Genet A 152A, 741-747. \n \n58. Newman, L.E., and Shadel, G.S. (2023). Mitochondrial DNA Release in Innate Immune Signaling. \nAnnu Rev Biochem 92, 299-332. \n \n59. Palma, E., Tiepolo, T., Angelin, A., Sabatelli, P., Maraldi, N.M., Basso, E., Forte, M.A., Bernardi, \nP., an d Bonaldo, P. (2009). Genetic ablation of cyclophilin D rescues mitochondrial defects and \nprevents muscle apoptosis in collagen VI myopathic mice. Hum Mol Genet 18, 2024-2031. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n32 \n \n60. Palty, R., Silverman, W.F., Hershfinkel, M., Caporale, T., Sensi, S.L., Parnis, J., Nolte, C., Fishman, \nD., Shoshan -Barmatz, V., Herrmann, S. , et al.  (2010). NCLX is an essential component of \nmitochondrial Na+/Ca2+ exchange. Proc Natl Acad Sci U S A 107, 436-441. \n \n61. Patron, M., Checchetto, V., Raffaello, A., Teardo, E., Vecellio Reane, D., Mantoan, M., Granatiero, \nV., Szabo, I., De Stefani, D., and Rizzuto, R. (2014). MICU1 and MICU2 finely tune the mitochondrial \nCa2+ uniporter by exerting opposite effects on MCU activity. Mol Cell 53, 726-737. \n \n62. Patron, M., Granatiero, V., E spino, J., Rizzuto, R., and De Stefani, D. (2019). MICU3 is a tissue -\nspecific enhancer of mitochondrial calcium uptake. Cell Death Differ 26, 179-195. \n \n63. Pendin, D., Norante, R., De Nadai, A., Gherardi, G., Vajente, N., Basso, E., Kaludercic, N., \nMammucari, C., Paradisi, C., Pozzan, T., et al. (2019). A Synthetic Fluorescent Mitochondria-Targeted \nSensor for Ratiometric Imaging of Calcium in Live Cells. Angew Chem Int Ed Engl 58, 9917-9922. \n \n64. Perocchi, F., Gohil, V.M., Girgis, H.S., Bao, X.R., McCombs, J.E., Palmer, A.E., and Mootha, V.K. \n(2010). MICU1 encodes a mitochondrial EF hand protein required for Ca(2+) uptake. Nature 467, 291-\n296. \n \n65. Piano Mortari, E., Folgiero, V., Marcellini, V., Romania, P., Bellacchio, E., D'Alicandro, V., Bocci, \nC., Carrozzo, R., Martinelli, D., Petrini, S. , et al. (2018). The Vici syndrome protein EPG5 regulates \nintracellular nucleic acid trafficking linking autophagy to innate and adaptive immunity. Autophagy 14, \n22-37. \n \n66. Pingitore, V., Pancholi, J., Hornsby, T.W., Warne, J., Pryce, G., McCormick, L.J., Hill, J., Bhosale, \nG., Peng, J., Newton, L.S. , et al.  (2024). Delocalized quinolinium -macrocyclic peptides, an atypical \nchemotype for CNS penetration. Sci Adv 10, eado3501. \n \n67. Plotegher, N., Filadi, R., Pizzo , P., and Duchen, M.R. (2021). Excitotoxicity Revisited: \nMitochondria on the Verge of a Nervous Breakdown. Trends Neurosci 44, 342-351. \n \n68. Plotegher, N., Perocheau, D., Ferrazza, R., Massaro, G., Bhosale, G., Zambon, F., Rahim, A.A., \nGuella, G., Waddington, S.N., Szabadkai, G. , et al. (2020). Impaired cellular bioenergetics caused by \nGBA1 depletion sensitizes neurons to calcium overload. Cell Death Differ 27, 1588-1603. \n \n69. Plovanich, M., Bogorad, R.L., Sancak, Y., Kamer, K.J., Strittmatter, L., Li, A.A. , Girgis, H.S., \nKuchimanchi, S., De Groot, J., Speciner, L., et al. (2013). MICU2, a paralog of MICU1, resides within \nthe mitochondrial uniporter complex to regulate calcium handling. PLoS One 8, e55785. \n \n70. Rao, V.K., Carlson, E.A., and Yan, S.S. (2014).  Mitochondrial permeability transition pore is a \npotential drug target for neurodegeneration. Biochim Biophys Acta 1842, 1267-1272. \n \n71. Rizzuto, R., De Stefani, D., Raffaello, A., and Mammucari, C. (2012). Mitochondria as sensors and \nregulators of calcium signalling. Nat Rev Mol Cell Biol 13, 566-578. \n \n72. Sancak, Y., Markhard, A.L., Kitami, T., Kovacs-Bogdan, E., Kamer, K.J., Udeshi, N.D., Carr, S.A., \nChaudhuri, D., Clapham, D.E., Li, A.A. , et al.  (2013). EMRE is an essential component of the \nmitochondrial calcium uniporter complex. Science 342, 1379-1382. \n \n73. Schinzel, A.C., Takeuchi, O., Huang, Z., Fisher, J.K., Zhou, Z., Rubens, J., Hetz, C., Danial, N.N., \nMoskowitz, M.A., and Korsmeyer, S.J. (2005). Cyclophilin D is a component of mitochondrial \npermeability transition and mediates neuronal cell death after focal cerebral ischemia. Proc Natl Acad \nSci U S A 102, 12005-12010. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n33 \n \n74. Schon, E.A., and Przedborski, S. (2011). Mitochondria: the next (neurode)generation. Neuron  70, \n1033-1053. \n \n75. Singh, R., Bar tok, A., Paillard, M., Tyburski, A., Elliott, M., and Hajnoczky, G. (2022). \nUncontrolled mitochondrial calcium uptake underlies the pathogenesis of neurodegeneration in \nMICU1-deficient mice and patients. Sci Adv 8, eabj4716. \n \n76. Sliter, D.A., Martinez, J. , Hao, L., Chen, X., Sun, N., Fischer, T.D., Burman, J.L., Li, Y., Zhang, \nZ., Narendra, D.P. , et al.  (2018). Parkin and PINK1 mitigate STING -induced inflammation. Nature  \n561, 258-262. \n \n77. Suomalainen, A., and Nunnari, J. (2024). Mitochondria at the crossroads of health and disease. Cell \n187, 2601-2627. \n \n78. Szabadkai, G., and Duchen, M.R. (2008). Mitochondria: the hub of cellular Ca2+ signaling. \nPhysiology (Bethesda) 23, 84-94. \n \n79. Tian, Y., Li, Z., Hu, W., Ren, H., Tian, E., Zhao, Y., Lu, Q., Huang, X., Yang, P., Li, X. , et al.  \n(2010). C. elegans screen identifies autophagy genes specific to multicellular organisms. Cell  141, \n1042-1055. \n \n80. Tsai, C.W., Rodriguez, M.X., Van Keuren, A.M., Phillips, C.B., Shushunov, H.M., Lee, J.E., \nGarcia, A.M.,  Ambardekar, A.V., Cleveland, J.C., Jr., Reisz, J.A. , et al.  (2022). Mechanisms and \nsignificance of tissue-specific MICU regulation of the mitochondrial calcium uniporter complex. Mol \nCell 82, 3661-3676 e3668. \n \n81. Tymianski, M. (2011). Emerging mechanisms of disrupted cellular signaling in brain ischemia. Nat \nNeurosci 14, 1369-1373. \n \n82. Uoselis, L., Lindblom, R., Lam, W.K., Kung, C.J., Skulsuppaisarn, M., Khuu, G., Nguyen, T.N., \nRudler, D.L., Filipovska, A., Schittenhelm, R.B. , et al.  (2023). Temporal landscape of mitochondrial \nproteostasis governed by the UPR(mt). Sci Adv 9, eadh8228. \n \n83. Vansenne, F., Fock, J.M., Stolte -Dijkstra, I., Meiners, L.C., van den Boogaard, M.H., Jaeger, B., \nBoven, L., Vos, Y.J., Sinke, R.J., and Verbeek, D .S. (2022). Phenotypic expansion of EGP5 -related \nVici syndrome: 15 Dutch patients carrying a founder variant. Eur J Paediatr Neurol 41, 91-98. \n \n84. Verma, M., Lizama, B.N., and Chu, C.T. (2022). Excitotoxicity, calcium and mitochondria: a triad \nin synaptic neurodegeneration. Transl Neurodegener 11, 3. \n \n85.Victorelli, S., Salmonowicz, H., Chapman, J., Martini, H., Vizioli, M.G., Riley, J.S., Cloix, C., Hall-\nYounger, E., Machado Espindola -Netto, J., Jurk, D. , et al.  (2023). Apoptotic stress causes mtDNA \nrelease during senescence and drives the SASP. Nature 622, 627-636. \n \n86. Vidyawan, V., Puspita, L., Juwono, V.B., Deline, M., Pieknell, K., Chang, M.Y., Lee, S.H., and \nShim, J.W. (2025). Autophagy controls neuronal differentiation by regulating the WNT-DVL signaling \npathway. Autophagy 21, 719-736. \n \n87. Waldrop, M.A., Gumienny, F., Boue, D., de Los Reyes, E., Shell, R., Weiss, R.B., and Flanigan, \nK.M. (2018). Low -level expression of EPG5 leads to an attenuated Vici syndrome phenotype. Am J \nMed Genet A 176, 1207-1211. \n \n88. Wang, A., Chen, C., Mei, C., Liu, S., Xiang, C., Fang, W., Zhang, F., Xu, Y., Chen, S., Zhang, Q. , \net al.  (2024). Innate immune sensing of lysosomal dysfunction drives multiple lysosomal storage \ndisorders. Nat Cell Biol 26, 219-234. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n34 \n \n \n89. Wang, Z., Miao, G., Xue, X., Guo, X., Yuan, C., Wang, Z., Zhang, G., Chen, Y., Feng, D., Hu, J. , \net al.  (2016). The Vici Syndrome Protein EPG5 Is a Rab7 Effector that Determines the Fusion \nSpecificity of Autophagosomes with Late Endosomes/Lysosomes. Mol Cell 63, 781-795. \n \n90. West, A.P., Khoury-Hanold, W., Staron, M., Tal, M.C., Pineda, C.M., Lang, S.M., Bestwick, M., \nDuguay, B.A., Raimundo, N., MacDuff, D.A. , et al.  (2015). Mitochondrial DNA stress primes the \nantiviral innate immune response. Nature 520, 553-557. \n \n91. Xian, H., Watari, K., Sanchez -Lopez, E., Offenberger, J., Onyuru, J., Sampath, H., Ying, W., \nHoffman, H.M., Shadel, G.S., and Karin, M. (2022). Oxidized DNA fragments exit mitochondria via \nmPTP- and VDAC -dependent channels to activate NLRP3 inflammasome and interferon signaling. \nImmunity 55, 1370-1385 e1378. \n \n92. Yamada, T., Ikeda, A., Murata, D., Wang, H., Zhang, C., Khare, P., Adachi, Y., Ito, F., Quiros, \nP.M., Blackshaw, S. , et al.  (2025). Dual regulation of mitochondrial fusion by Parkin -PINK1 and \nOMA1. Nature 639, 776-783. \n \n93. Yu, C.H., Davidson, S., Harapas, C.R., Hilton, J.B., Mlodzianoski, M.J., Laohamonthonkul, P., \nLouis, C., Low, R.R.J., Moecking, J., De Nardo, D., et al. (2020). TDP-43 Triggers Mitochondrial DNA \nRelease via mPTP to Activate cGAS/STING in ALS. Cell 183, 636-649 e618. \n \n94. Zhang, C., Shang, G., Gui, X., Zhang, X., Bai, X.C., and Chen, Z.J. (2019). Structural basis of \nSTING binding with and phosphorylation by TBK1. Nature 567, 394-398. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n35 \n \nFigure 1 \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n36 \n \nFigure S1 \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n37 \n \nFigure 2 \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n38 \n \nFigure S2 \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n39 \n \nFigure 3 \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n40 \n \nFigure S3 \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n41 \n \nFigure 4 \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n42 \n \nFigure S4 \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n43 \n \nFigure 5 \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n44 \n \nFigure S5 \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n45 \n \nFigure S6 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n46 \n \nFigure 6 \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n47 \n \nFigure S7 \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n48 \n \nFigure 7 \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n49 \n \nFigure S8 \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n50 \n \nFigure 8 \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n51 \n \nFigure S9 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n52 \n \nFigure legends \nFigure 1: Patient-derived fibroblasts bearing pathogenic EPG5 mutations show impaired                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              \nmitochondrial bioenergetic function and respiratory defect.  \n(A) Immunoblot image of EPG5 protein expression in controls and patient-derived fibroblasts. \nActin was used as a loading control.  \n(B) Protein levels normalized to those in control 1, (nexp = 4, ****p <0.0001).  \n(C) Representative confocal image of TMRM labelled cells showing ΔΨm in control 1, control \n2, patient 1 and patient 2 fibroblasts. Scale bars: 20 μm.   \n(D) Mean TMRM values showing steady state ΔΨm, (nexp = 4, ncells analysed for each control \nand patient = 95-110, ****p <0.0001).  \n(E) Mitochondrial morphometric analysis of all TMRM confocal images classified into \nnetworked, fragmented and swollen mitochondria represented as percentage of total \nmitochondrial population, (nexp = 4, ncells analysed for each control and patient = 80-91, **p= \n0.0034, ***p= 0.0002).  \n(F) Representative NADH confocal image of control 1 and patient 1 fibroblasts at baseline and \nfollowing sequential application of the protonophore FCCP and cyanide (NaCN; complex IV \ninhibitor), Scale bars: 20 μm.  \n(G) NADH redox index plotted as percentage of the minimum (with FCCP) and maximum \n(with NaCN) values , ( nexp = 3, n cells analysed for each control and patient = 33 -40, ****p \n<0.0001).  \n(H) Normalised OCR traces from controls and patient fibroblasts , (nexp = 3, nrep = 4, **p= \n0.0043 and 0.0038, ****p <0.0001).  \n(I-J) Normalised ATP-linked respiration and spare reserve capacity of  controls and patient \nfibroblasts, (nexp = 3, nrep = 4, ****p <0.0001).  \n(K) Immunoblot image of OXPHOS protein subunits expression from whole cell lysates  of \ncontrols and patient-derived fibroblasts. TOM20 and Actin were used as loading controls.  \n(L) Protein expression levels normalized and plotted as fold difference relative to control, (nexp \n= 3-4, ***p= 0.0003 and 0.0001, ****p <0.0001).  \n(M) DHE oxidation rates plotted as ROS production rate and normalized to control 1,  (nexp = \n3, **p= 0.0015).  \nData (B, D, E, G, I, J, L and M ) are expressed as mean ± S D and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using \none/two-way ANOVA followed by posthoc Tukey’s test.     \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n53 \n \nFigure S1: Mitochondrial dysfunction and increased mtDNA copy number in EPG5-\ndeficient fibroblasts.  \n(A) EPG5 mRNA expression levels in patient fibroblasts determined with qRT-PCR on cDNA \ntranscript of isolated mRNA and normalized to the levels measured in control, (nexp = 4, ***p= \n0.0007).  \n(B) Representative confocal image of TMRM labelled cells showing ΔΨm in control 3, patient \n3 and patient 4 fibroblasts. Scale bars: 20 μm.  \n(C) Quantification of mitochondrial volume occupancy calculated from total cytosolic volume \n(Calcein AM intensity value) and represented as percentage mitochondrial mass, (nexp = 4, ncells \nanalysed for each control and patient = 95-110).   \n(D) Quantification of mtDNA copy number measured by qPCR using a mtDNA -specific \nprimer pair, from total genomic DNA isolated from control and patient fibroblasts, (nexp = 4-5, \n***p= 0.0008).  \n(E) Representative trace of quantified (mean ± SD) NADH autofluorescence in control and \npatient fibroblasts. Following the acquisition of baseline autofluorescence, application of 2.5 \nμM FCCP maximises the rate of respiration and oxidises all the mitochondrial NADH resulting \nin the lowest fluorescence signal (this level is considered as the minimum = 0% for NADH). \nApplication of 1 mM NaCN blocks respiration and prevents NADH oxidation. This allows the \nNADH pool to be regenerated and the highest fluorescence signal is obtained (this level is \nconsidered as the maximum = 100% for NADH). The NADH redox index can be calculated \nusing the obtained traces, (nexp = 3).  \n(F) Immunoblot image of native respiratory chain protein expression and supercomplex \nassembly from isolated mitochondria  of control and patient fibroblasts analysed using blue \nnative gel electrophoresis (BNGE). Supercomplex Vn was detected using ATP5A and used as \na loading control.  \n(G) Protein expression levels normalized and plotted as fold difference relative to control 1, \n(nexp = 3, ***p= 0.0005, 0.006 and 0.0001, ****p <0.0001).  \nData ( A, C, D, E, F, and H ) are expressed as mean ± S D and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test  (non-significant p values are denoted with \nnumeric values). \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n54 \n \nFigure 2: Patient-derived fibroblasts bearing pathogenic EPG5 mutations show impaired \nmitochondrial Ca2+ signalling.  \n(A) Mean traces for [Ca2+]m uptake measured using a mitochondria-target aequorin plate reader \nassay in response to 10 μM histamine, (nexp = 3, nrep = 5).  \n(B) Maximum [Ca2+]m induced by 10 μM histamine in control and patient fibroblasts, (n exp = \n3, nrep = 5, ****p <0.0001).  \n(C) Mean [Ca2+]c traces measured in control 1 and patient 1 fibroblasts upon 10 μM histamine \nstimulation by Fluo-4 AM, (nexp = 5-7).  \n(D) Mean [Ca2+]m traces measured in control 1 and patient 1 fibroblasts upon 10 μM histamine \nstimulation by mito-Fura-2 AM, inset shows rate of [Ca2+]m rise, (nexp = 5-7).  \n(E) Measurement of resting [Ca 2+]m levels in the control  and patient fibroblasts loaded with \nmito-Fura-2 AM before histamine stimulation, calculated from the traces in D, (**p= 0.0065).  \n(F) Quantitative analysis of time when [Ca 2+]m rise was at the half maxima upon histamine \nstimulation, calculated from the traces in D, (***p= 0.0003).  \n(G) Quantification of normalised areas under the curve (AUC) of the mito-Fura-2 AM traces, \nrepresenting total mitochondrial Ca2+ uptake over time, calculated from the traces in D, (**p= \n0.0094). \n(H) Immunoblot images of proteins involved in mitochondrial Ca2+ signalling from whole cell \nlysates of all the control and patient fibroblasts. ATP5A and Actin w ere used as loading \ncontrols.  \n(I) Protein levels relative to loading control ATP5A were normalized to those in control 1, (nexp \n= 4, *p= 0.0243 and 0.0105, ***p= 0.0002 and 0.0005).  \n(J) Immunoblot images for pPDH (PDH -E1α pS293) and total PDH (PDH -E1α) from whole \ncell lysates of control and patient fibroblasts. Actin was used as a loading control.  \n(K) Ca2+ retention capacity measured on isolated mitochondria from control and patient \nfibroblasts, mean traces showing the extra -mitochondrial calcium measured using Calcium \nGreen-5N after repetitive addition of 5 μM CaCl 2 boluses in the presence or absence of JP 1-\n138 (100 nM, added at 0 s), inset shows rate of Ca2+ in control and patient mitochondria.  \n(L) Mitochondrial Ca2+ retention capacity calculated as the percentage inhibition as compared \nto untreated mitochondria, (nexp = 4, ****p <0.0001).  \nData (A, B, C, D, E, F, G, I, and L) are expressed as mean ± SD and individual data points \nfrom independent experiments are shown in each plot. Statistical analysis was carried out using \ntwo-way ANOVA followed by posthoc Tukey’s test (non-significant p values are denoted with \nnumeric values).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n55 \n \nFigure S2: Both the ER Ca 2+ stores and ER-mitochondria contact sites distribution are \nunaltered in control and patient-derived fibroblasts.  \n(A) Quantification of normalised areas under the curve (AUC) of the Fluo -4 AM traces in \nresponse to 10 μM histamine, representing total ER Ca 2+ released over time, calculated from \nthe mean traces in Fig. 2B.  \n(B) Measurement of time when 50% of released ER Ca 2+ was cleared from the cytosol , \ncalculated from the mean traces in Fig. 2B.  \n(C) Representative TEM images of control and patient fibroblasts, with red segmented regions \nbetween ER and mitochondria marking ER-mito contact sites. Scale bars: 0.5 μm. Histograms \nof ER-mito contact widths distribution across control and patient fibroblasts , calculated from \nthe segmented regions, (nexp = 3, nrep analysed for eachcontrol and patients =181-220).  \n(D) Quantification of ER-mito contact widths from dataset in C.  \n(E) The ratio of the pPDH (PDH -E1α pS293) and total PDH (PDH -E1α) band intensities \nnormalised to the average control 1 ratio in Fig. 2J, (nexp = 4, ***p= 0.0004).  \n(F) Time when [Ca 2+]m uptake was at the half maxima after the first 5 μM CaCl 2 bolus, \ncalculated from the inset in Figure 2K (***p= 0.0001).  \nData ( A, B, D, E and F) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test  (non-significant p values are denoted with \nnumeric values).  \n \n \nFigure 3: Mitochondrial Ca2+ overload induces the release of cytosolic mtDNA in patient-\nderived fibroblasts.  \n(A) Representative super-resolution A iryscan images of control and patient fibroblasts \nimmunolabelled with DNA (green), TOM20 (red) and citrate synthetase (blue). The magnified \nimages on the right show areas in which DNA (green) does not co-localize with TOM20 (red) \nin patient cells. 3D representations of the inset show OMM (red), IMM ( blue) and mtDNA \n(green), where all the mt DNA puncta  (green spots)  are located within the mitochondria of \ncontrol fibroblasts  with some mtDNA puncta (red spots indicated with arrowheads)  in the \ncytoplasm of patient fibroblasts . mtDNA nucleoids partially outside the OMM and IMM \nsurface are shown in magenta and indicated by arrows . Overview scale bars, 10 μm and inset \nscale bars, 0.5 μm.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n56 \n \n(B) Percentage of control and patient fibroblasts showing cytosolic DNA puncta , (nexp = 4, \n****p <0.0001).  \n(C) Quantification of number of cytosolic DNA puncta release d per cell , (nexp = 4, ncells \nanalysed for each control and patient = 61-71, ****p <0.0001).  \n(D) Traces showing mean change ± SD in TMRM (representing ΔΨm) fluorescence intensity \nin response to 10 μM histamine  challenge and FCCP-induced depolarization in control and \npatient fibroblasts, (nexp = 4).  \n(E) Measurement of the mitochondrial depolarization at 500 s after 10 μM histamine \nstimulation, (nexp = 4, nrep analysed for each control and patients = 12-14, ****p <0.0001).  \n(F) Mean traces of TMRM fluorescence intensity and [Ca2+]m change after 10 μM histamine \nchallenge in patient 1 fibroblasts co-labelled with TMRM, mito-Fura-2 and PicoGreen, (nexp = \n3).  \n(G) Snapshots from time -lapse confocal imaging of patient fibrobla sts co -labelled TMRM \n(red) and PicoGreen (green), upper panels and m ito-Fura-2 AM, ratiometric lower panels, \nquantified in F. Elapsed time after 10 μM histamine challenge is indicated (Video S2). White \narrowheads denote nucleoid externalization events. Scale bars, 5 μm.  \n(H) Quantitative analysis of the average number of PicoGreen puncta released from TMRM -\nlabelled mitochondria into the cytosol in response to 10 μM histamine, (ncells analysed for each \ncontrol and patient = 15, ***p= 0.0006).  \nData ( B, C, D, E, F and H ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test. \n \n \nFigure S3: Physiological stimulation with histamine does not induces mitochondrial \ndepolarisation or swelling in control fibroblasts.  \n(A) Pearson’s R (colocalization) value of TOM20 (red) and citrate synthase (blue) labelled \nmitochondria from confocal images of cont rol and patient fibroblasts represented in Fig. 3A, \n(nexp = 4, ncells analysed for each control and patient = 55-66).  \n(B) Mean traces of TMRM fluorescence intensity and [Ca2+]m change after 10 μM histamine \nchallenge in control 1 fibroblasts co-labelled with TMRM, m ito-Fura-2 AM and PicoGreen, \n(nexp = 3).  \n(C) Snapshots from time-lapse confocal imaging of control 1 fibroblasts co-labelled TMRM \n(red) and PicoGreen (green), upper panels and m ito-Fura-2 AM, ratiometric lower panels, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n57 \n \nquantified in B. Elapsed time after 10 μM histamine challenge is indicated (Video S1). White \narrowheads denote  mitochondrial fragmentation  events after 10 μM and 20 μM histamine  \nchallenge. Scale bars, 5 μm.  \n(D) Quantitative morphometric analysis of TMRM labelled mitochondria in control and patient \ncells after stimulation with 10 μM histamine . Mitochondrial pool classified into networked, \nfragmented and swollen mitochondria are represented as a percentage of the total mitochondrial \npopulation, (nexp = 3, ncells analysed for each control and patients = 18-28, *p= 0.0313, **p= \n0.0050, ****p <0.0001).  \nData (A, B and D ) are expressed as mean ± SD and individual data points from independent \nexperiments are shown in each plot. Statistical analysis was carried out using one /two-way \nANOVA followed by posthoc Tukey’s test (non-significant p values are denoted with numeric \nvalues). \n \n \nFigure 4: Activation of cGAS-STING pathway and interferon response  triggered by \ncytosolic release of mtDNA in patient-derived fibroblasts.  \n(A-B) Pathway analysis of RNA-seq data from patient 1 compared to control 1. Dot plots and \nGSEA performed using two types of gene set: the Gene Ontology (GO) biological process and \nKEGG pathway. NES, normalized enrichment score.  \n(C) Heatmap of RNA-seq data displaying the top 50 upregulated differentially expressed type \nI/III IFN genes (DEGs) in control 1 and patient 1 fibroblasts, (nexp = 3).  \n(D-E) qRT-PCR analysis of ISGs expression in all control and patient fibroblasts , (nexp = 3, \n****p <0.0001).  \n(F)   Representative confocal images of control  1 and patient 1 fibroblasts transfected with \nBFP-cGAS and co-labelled with TMRM (red) and PicoGreen (green). The magnified images \non the right show areas in  which a mtDNA nucleoids (green) devoid of TMRM fluorescence \nco-localize with cytosolic cGAS puncta (grey) in patient cells indicated by a white arrowhead. \nOverview scale bars, 10 μm and inset scale bars, 5 μm.  \n(G) Quantification of the number of cGAS(+) and TMRM (-) PicoGreen puncta per cell, (nexp \n= 3, ncells analysed for each control and patient = 26-30, ***p= 0.0004).  \n(H) Immunoblot images of proteins involved in  the cGAS-STING signalling and ISGs \ninduction from whole cell lysates of control and patient fibroblasts  treated with 10 μM \nhistamine and 1 μM thapsigargin for 24 h. Actin was used as a loading control.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n58 \n \n(I) Protein expression levels of STING normalized and plotted as fold difference relative to \ncontrol, (nexp = 4, **p= 0.0056 and 0.0096, ***p= 0.0006).  \n(J) The ratio of pSTAT1 (p Tyr701) and total STAT1 band intensities normalised to the control \n1 ratio, (nexp = 4, *p= 0.124 and 0.0161, ***p= 0.0008).  \nData (D, E, G, I and J) are expressed as mean ± SD and individual data points from independent \nexperiments are shown in each plot. Statistical analysis was carried out using two-way \nANOVA followed by posthoc Tukey’s test. \n \n \nFigure S4: Additional characterisation of cGAS-STING activation and ISGs induction in \npatient-derived fibroblasts bearing EPG5 mutations.  \n(A-B) Pathway analysis of RNA-seq data from patient 3 compared to control 1. Dot plots and \nGSEA performed using two types of gene set: the Gene Ontology (GO) biological process and \nKEGG pathway. NES, normalized enrichment score.  \n(C) Heatmap of RNA-seq data displaying the top 50 upregulated differentially expressed type \nI/III IFN genes (DEGs) in control 1 and patient 3 fibroblasts, (nexp = 3).  \n(D) qRT-PCR analysis of pro-inflammatory ISGs expression in all control and patient \nfibroblasts, (nexp = 3, **p= 0.0029, ****p <0.0001).  \n(E)   Representative confocal images of control  and patient fibroblasts immunolabelled with \ncGAS (red) antibody and stained with DAP1 to quantify cytoplasmic and nuclear localization \nof cGAS.  \n(F) Quantitative analysis of relative cGAS fluorescence intensity from cytoplasmic and nuclear \narea to determine cytosolic cGAS translocation in control and patient fibroblasts, (nexp = 3, ncells \nanalysed for each control and patient = 15-20, ****p <0.0001).  \n(G) Immunoblot images of proteins involved in  cGAS-STING signalling and ISGs induction \nfrom whole cell lysates of control and patient fibroblasts  treated with 10 μM histamine and 1 \nμM thapsigargin for 24 h. Actin was used as a loading control.  \n(H) The ratio of pTBK1 (pSer172) and total TBK1 band intensities normalised to the control \n1 ratio, (nexp = 4, ****p <0.0001).  \n(I) Protein expression levels of STING normalized and plotted as fold difference relative to \ncontrol, (nexp = 4, **p= 0.0096, ***p= 0.0003, 0.0006).  \n(J-K) The ratio of pIRF3 (pSer396) and total IRF3 and pSTAT1 (p Tyr701) and total STAT1 \nband intensities normalised to  the control 1 ratio, (n exp = 4, ****p <0.0001, ***p= 0.0004, \n0.0002).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n59 \n \nData ( D, F, H, I, J and K ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test.   \n \n \nFigure 5: Attenuation of STING-dependent interferon response by JP1-138 treatment in \npatient-derived fibroblasts.  \n(A) Quantitative analysis of steady state  ΔΨm by measuring TMRM fluorescence intensity in \ncontrol and patient fibroblasts either untreated or treated with STING inhibitor, H-151 (1 μM, \n24 h) (nexp = 3, ncells analysed for each control and patient = 12-15, ****p <0.0001).  \n(B) NADH redox index of control and patient fibroblasts either untreated or treated with H -\n151, (nexp = 3, ncells analysed for each control and patient = 7-10, ***p= 0.0001).  \n(C) Normalised OCR traces from controls and patient fibroblasts  either untreated or treated \nwith H-151 for longer durations (0.5 μM, 3 days) (nexp = 3, nrep = 4).  \n(D) Traces showing mean change ± SD in ΔΨm in response to 10 μM histamine challenge and \nFCCP-induced depolarization in the absence (upper panel) and presence of JP1-138 (100 nM) \n(lower panel) in control and patient fibroblasts, (nexp = 4).  \n(E) Quantification of the mitochondrial depolarization at 1000 s after 10 μM histamine \nstimulation, (nexp = 4, nrep analysed for each control and patient = 36-44, ****p <0.0001).  \n(F) Immunoblot images of cGAS-STING signalling proteins from whole cell lysates of control \n1 and patient 1 fibroblasts treated with JP1 -138 (100 nM) for the indicated time . Actin was \nused as a loading control. Relative expression levels of pSTAT1 (p Tyr701), STING and pIRF3 \n(pSer396) normalized and plotted as fold difference relative to control, (nexp = 3, *p= 0.0235, \n**p= 0.0049, ***p= 0.0007, ****p <0.0001).  \n(G) Immunoblot images of the cGAS-STING signalling proteins from whole cell lysates of \ncontrol and patient fibroblasts treated with either CsA (1 μM) or JP1-138 (100 nM) for 3 days. \nActin was used as a loading control.  \n(H) Protein expression levels of STING normalized and plotted as fold difference relative to \ncontrol, (nexp = 4, *p= 0.0265, ***p= 0.0004).  \n(I) The ratio of pSTAT1 (p Tyr701) band intensities normalised to the control 1 ratio, (n exp = \n4, *p= 0.0197, ***p= 0.0007).  \n(J-K) qRT-PCR analysis of ISGs expression in all the control and patient fibroblasts  either \nuntreated or treated with JP1-138 (100 nM) for 3 days (nexp = 3, ***p= 0.0001, ****p <0.0001). \nData (A, B, C, D, E, F, H, I, J and K) are expressed as mean ± SD and individual data points \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n60 \n \nfrom independent experiments are shown in each plot. Statistical analysis was carried out using \none/two-way ANOVA followed by posthoc Tukey’s test (non-significant p values are denoted \nwith numeric values). \n \n \nFigure S5: Inhibition of STING activation by H -151 attenuates the STING-dependent \ninterferon response  but does not improve mitochondrial function in p atient-derived \nfibroblast.  \n(A) Immunoblot images of the cGAS-STING signalling proteins from whole cell lysates of \ncontrol and patient fibroblasts  treated with either STING inhibitor, H -151 (1 μM) or cGAS \ninhibitor, G-140 (100 nM) for 24 h. Actin was used as a loading control.  \n(B) Protein expression levels of STING normalized and plotted as fold difference relative to \ncontrol, (nexp = 3, ***p= 0.0008).  \n(C-D) The ratio of pIRF3 (pSer396) and total IRF3 and pSTAT1 (p Tyr701) and total STAT1 \nband intensities normalised to the control 1 ratio, (nexp = 3, **p= 0.0034, ***p= 0.0002, ****p \n<0.0001).  \n(E-F) qRT-PCR analysis of ISGs expression in the control and patient fibroblasts  either \nuntreated or treated with H-151 (1 μM) for 24 h (nexp = 3, ***p= 0.0006, ****p <0.0001).  \n(G) Normalised OCR traces from controls and patient fibroblasts  either untreated or treated \nwith H-151 for longer durations (0.5 μM, 3 days) (nexp = 3, nrep = 4).  \n(H-I) Normalised ATP-linked respiration and spare respiratory capacity of all the controls and \npatient fibroblasts measured from traces in G and Fig. 5C (nexp = 3, nrep = 4, ****p <0.0001). \nData (B, C, D, E, F, G, H and I ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test  (non-significant p values are denoted with \nnumeric values). \n \n \nFigure S6: Additional characterisation of the effect of JP1-138 treatment on the STING-\ndependent interferon response in patient-derived fibroblasts. (A) Traces showing mean \nchange ± SD in ΔΨm in response to 10 μM histamine challenge  and FCCP -induced \ndepolarization in the absence (upper panel) and presence of JP1 -138 (100 nM) (lower panel) \nin control and patient fibroblasts, (nexp = 4).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n61 \n \n(B) Immunoblot images of  proteins involved in the cGAS-STING cascade from whole cell \nlysates of control and patient fibroblasts treated with either CsA (1 μM) or JP1-138 (100 nM) \nfor 3 days. Actin was used as loading a control.  \n(C) The ratio of pTBK1 (pSer172) and total TBK1 band intensities normalised to the control \n1 ratio in Fig. 5G, (nexp = 4, ***p= 0.0001, ****p <0.0001).  \n(D-F) Relative expression levels of STING, ratio of pIRF3 (pSer396) and total IRF3 and \npSTAT1 (p Tyr701) and total STAT1 band intensities normalized and plotted as fold difference \nrelative to control, (nexp = 3, *p= 0.0395, **p= 0.0020, ***p= 0.0002, ****p <0.0001).  \nData ( A, C, D, E and F ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using two-\nway ANOVA followed by posthoc Tukey’s test (non-significant p values are denoted with \nnumeric values, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). \n \n \nFigure 6: JP1-138 treatment reduces cytosolic mtDNA release and improves  \nmitochondrial bioenergetic function in patient fibroblasts bearing pathogenic EPG5 \nvariants.  \n(A) Representative Airyscan images of control  1 and patient 1 fibroblasts treated with either \nJP1-138 (100 nM) or vehicle (DMSO) for 3 days  and immunolabelled with  DNA (green), \nTOM20 (red) and citrate synth etase (blue). The magnified images on the right show areas in  \nwhich DNA (green) does not co-localize with TOM20 (red) in patient cells marked by white \narrow heads which is significantly reduced in JP1 -138-treated patient fibroblasts. Overview \nscale bars, 10 μm and inset scale bars, 5 μm and 2.5 μm.  \n(B) Percentage of control and patient fibroblasts  treated with either JP1 -138 (100 nM) or \nvehicle (DMSO) for 3 days and analysed for cytosolic DNA puncta, (nexp = 5, ***p= 0.0001, \n****p <0.0001).  \n(C) Quantification of the number of cytosolic DNA puncta release d per cell, calculated from \nB (nexp = 4, ncells analysed for each control and patient = 16-21, ***p= 0.0001, ****p <0.0001). \n(D) Mitochondrial morphometric analysis of all Airyscan confocal images in A classified into \nnetworked, fragmented and swollen mitochondria represented as percentage of total \nmitochondrial population, ( nexp = 4, ncells analysed for each control and patient = 16-21, *p= \n0.0140, **p= 0.0028, ***p= 0.0006).  \n(E) Representative confocal image of TMRM labelled control and patient fibroblasts treated \nwith either JP1-138 (100 nM) or vehicle (DMSO) for 3 days. Scale bars: 20 μm.   \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n62 \n \n(F) Mean TMRM values showing steady state ΔΨm, (nexp = 4, ncells = analysed for each control \nand patient = 42-59, ****p <0.0001).  \n(G) NADH redox index of control and patient fibroblasts treated with either JP1-138 (100 nM) \nor vehicle (DMSO) for 3 days  (nexp = 3, ncells analysed for each control and patient =  27-31, \n***p= 0.0001, ****p <0.0001).  \n(H) Normalised OCR traces from controls and patient fibroblasts  treated with either JP1-138 \n(100 nM) or vehicle (DMSO) for 3 days (nexp = 3, nrep = 4, **p= 0.0077, ***p= 0.0009).  \nData ( B, C, D, F, G, and H ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using \none/two-way ANOVA followed by posthoc Tukey’s test. \n \n \nFigure S7: Additional characterisation of the effe ct of JP1 -138 treatment on \nmitochondrial respiration and cytosolic mtDNA release in patient-derived fibroblast.  \n(A) Representative Airyscan images of control and patient fibroblasts treated with either JP1-\n138 (100 nM) or vehicle (DMSO) for 3 days and immunolabelled with DNA (green), TOM20 \n(red) and citrate synth etase (blue). The magnified images on the right show areas in  which \nDNA (green) does not co-localize with TOM20 (red) in patient cells marked by white arrow \nheads which is significantly reduced in JP1 -138-treated patient fibroblasts. Overview s cale \nbars, 10 μm and inset scale bars, 5 μm and 2.5 μm.  \n(B) Representative confocal image of TMRM labelled control and patient fibroblasts treated \nwith either JP1-138 (100 nM) or vehicle (DMSO) for 3 days . Scale bars: 20 μm. Normalised \nOCR traces from controls and patient fibroblasts  treated with either JP1 -138 (100 nM) or \nvehicle (DMSO) for 3 days (nexp = 3, nrep = 4).  \n(D-E) Normalised ATP-linked respiration and spare reserve capacity of  controls and patient \nfibroblasts calculated from traces in C and Fig. 6H (nexp = 3, nrep = 4, ****p <0.0001).  \nData (C, D and E ) are expressed as mean ± SD and individual data points from independent \nexperiments are shown in each plot. Statistical analysis was carried out using two-way \nANOVA followed by posthoc Tukey’s test.  \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n63 \n \nFigure 7: Enhanced susceptibility to glutamate-induced delayed Ca2+ dysregulation and \ncell death in Q336R neurons.  \n(A-B) Cytosolic calcium concentration in neurons labelled with the low -affinity calcium \nsensor, FuraFF AM and measured by fluorescence imaging for the indicated time intervals. \nChanges in [Ca 2+]c measured from individual (grey) and mean (black) traces, following \nsequential exposure of isogenic neurons to 10 μM and 100 μM  glutamate are plotted as a \nfunction of time. The traces reveal a significant difference between the responses of isogenic \nand Q336R neurons upon 10 μM glutamate stimulation which was followed by delayed \ncalcium deregulation (DCD) in a large proportion of Q336R neurons, (nexp = 3, ncells analysed \nfor isogenic and Q336R = 91-120).  \n(C) Baseline subtracted peak values of the early response at about 150  s to 10 μM glutamate \nfor isogenic and Q336R neurons.  \n(D)  Baseline subtracted peak values of DCD at 2000 s in response to 10 μM glutamate, (****p \n<0.0001).  \n(E)  FuraFF ratiometric images for isogenic and Q336R neurons shown at the start of the \nexperiment (t = 0  s) and at 2000  s after exposure to glutamate, showing the full recovery of \n[Ca2+]c in isogenic control neurons, while the sustained very high [Ca 2+]c levels in the Q336R \nneurons reflecting DCD, similar to response after 100 μM glutamate in isogenic control in A . \nScale bars, 50 μm.   \n(F) Immunoblot images of proteins involved in mitochondrial Ca2+ signalling from whole cell \nlysates of isogenic and Q336R neurons. ATP5A was used as a loading control.  \n(G) Protein levels relative to loading control ATP5A were normalized to those in control 1, \n(nexp = 4, ****p <0.0001).  \n(H-I) Mitochondrial Ca2+ concentration in neurons labelled with the mito-Fura-2 AM and \nmeasured by fluorescence imaging for the indicated time intervals. Changes in [Ca 2+]m \nmeasured from individual grey traces in isogenic control and red traces in Q336R neurons  \nfollowing exposure to 10 μM glutamate are plotted as a function of time, (nexp = 3, ncells analysed \nfor isogenic and Q336R = 55-75).  \n(J) Mean [Ca2+]m traces measured with mito -Fura-2 upon exposure to 10 μM  glutamate and \ncalculated from individual traces in H and I.  \n(K) Time when [Ca2+]m rise was at the half maxima upon glutamate stimulation, calculated \nfrom the traces in H and I, (****p <0.0001).  \n(L) Maximum [Ca2+]m rise induced exposure to 10 μM  glutamate, calculated from the traces \nin H and I, (**p= 0.0041).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n64 \n \n(M) Representative confocal image of isogenic control and Q336R neurons treated with either \n10 μM glutamate or vehicle (PBS) for 12 h and immunolabelled with TOM20, Cytochrome c \n(Cyt c) and neuronal β-Tubulin III, Clone TUJ1. The magnified images on the right show areas \nin which TOM20 (green) does not co-localize with Cyt c (red) in Q336R neurons. Overview \nscale bars, 20 μm, inset scale bars, 5 μm.  \n(N) Quantitative analysis of percentage of total mitochondrial area lacking Cyt c, (nexp = 3, ncells \nanalysed for isogenic and Q336R = 32-40, **p= 0.0059).  \nData ( C, D, G, K, L and N ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out either \nusing two-tailed unpaired Student’s t-test or one-way ANOVA followed by posthoc Tukey’s \ntest (non-significant p values are denoted with numeric values).   \n \n \nFigure S8: Altered mitochondrial Ca2+ homeostasis in Q336R neurons.  \n(A) The neural cells were differentiated from NPCs and immunostained with the neuronal \nmarker, MAP2 and the glial cell marker, GFAP. Scale bars, 20 μm.  \n(B) The MAP2 (green) expression and GFAP (magenta) fluorescence levels were used to \ndetermine the percentage of targeted cells, while DAPI (blue) staining was used to determine \nthe total number of cells in the field of view, (nexp = 3).  \n(C) Immunoblot image of native MCU complex and respiratory chain protein expression and \nsupercomplex assembly identified using indicated antibodies from isolated mitochondria  of \nisogenic and Q336R neurons and analysed using BNGE.  \n(D-E) Quantitative analysis of high and low molecular weight MCU and MICU1 containing \ncomplexes analysed using BNGE in C, (nexp = 3, *p= 0.0258, ****p <0.0001).  \n(F) mRNA levels of genes involved mitochondrial Ca2+ signalling using qPCR on cDNA \ntranscript of isolated mRNA and normalized to the levels measured in isogenic control, (nexp = \n3).  \n(G) Quantification of resting [Ca2+]m in isogenic and Q336 neurons, calculated from calculated \nfrom the traces in Fig. 8H and I before glutamate stimulation, (**p= 0.0084).  \n(H) Representative mito-Fura-2 ratiometric images for isogenic and Q336R neurons (the image \nexcited at 355 nm divided by that excited at 405 nm) showing a higher steady-state [Ca2+]m as \nwell swollen morphology of mitochondrial in Q336R neurons. Scale bars, 10 μm.  \n(I)  Representative mito-Fura-2 ratiometric images for isogenic and Q336R neurons obtained \nat the start of the experiment (t = -5 s) and at 5 s and 10 s after exposure to glutamate, showing \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n65 \n \nthe respective rate of increase in [Ca2+]m in isogenic control and Q336R neurons. Scale bars, \n50 μm.  \n(J) Mean traces for [Ca 2+]m uptake measured  in isogenic and Q336R neurons  using a \nmitochondria-target aequorin plate reader assay in response to 10 μM glutamate. Inset plot \nshows maximum [Ca2+]m induced by 10 μM glutamate (nexp = 3, nrep = 5, ***p= 0.0006).  \n(K) Quantitative morphometric analysis of mitochondrial population in soma s and axons of \nisogenic and Q336R neurons treated with either 10 μM glutamate or vehicle (PBS) for 12 h. \nAll immunofluorescence images including representative images shown in Fig. 7M were \nclassified into networked, fragm ented and swollen mitochondria represented as a percentage \nof the total mitochondrial population, (nexp = 3, ncells analysed for isogenic and Q336R = 32-45, \n****p <0.0001).  \nData (B, D, E, F, G, J and K ) are expressed as mean ± SD and individual data point s from \nindependent experiments are shown in each plot. Statistical analysis was carried out either \nusing two-tailed unpaired Student’s t -test or one-way ANOVA followed by posthoc Tukey’s \ntest.   \n \n \nFigure 8: JP1-138 increases mitochondrial Ca2+ buffering capacity and partially rescues \nbioenergetic function in Q336R neurons.   \n(A) Traces showing mean change ± SD in Fluo -4 AM (representing [Ca2+]m) and TMRM \n(representing ΔΨm) fluorescence intensity in response to increasing concentrations of \nexogenous Ca2+ (upward tick on the x -axis) to digitonin -permeabilised isogenic and Q336R \nneurons bathed in a pseudo-intracellular recording solution with or without JP1-138 (0.5 μM).  \n(B) Quantitative analysis of Fluo -4 intensity (upper panel) and TMRM flu orescence (lower \npanel) showing peak change in permeabilised isogenic and Q336R neurons  treated with or \nwithout JP1-138 in response to ascending concentrations of Ca2+ in the recording buffer, (nexp \n= 3, n cells analysed for  isogenic and Q336R  = 32-45, **p= 0.0050, ***p= 0.0005, ****p \n<0.0001).  \n(C-E) Mitochondrial membrane potential measured using Rhodamine-123 (with the ‘dequench \nprotocol’) in responses to exposure to 10 μM glutamate stimulation and plotted as a function \nof time for isogenic and Q336R treated with either JP1-138 (0.5 μM) or vehicle (DMSO). An \nincrease in Rhod amine-123 fluorescence intensity reports mitochondrial depolarization. The \nresponses to 10 μM glutamate were significantly different between isogenic and Q336R \nrevealing a large de polarization in the Q336R neurons . Pre -incubation with JP1 -138 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n66 \n \ncompletely rescues glutamate -induced depolarisation in Q336R neurons. FCCP -induced \nmitochondrial depolarization represents total ΔΨm after glutamate stimulation, (nexp = 3, ncells \nanalysed for isogenic and Q336R = 42-55).  \n(F-G) Quantitative analysis of the mitochondrial depolarization at early time point and 1500 s \nafter 10 μM glutamate stimulation, expressed as Rhod123 F/F0 for isogenic and Q336R neurons \nwith or without JP1-138 pretreatment, (****p <0.0001).  \n(H) Normalised OCR traces from isogenic and Q336R neurons treated with either JP1-138 (0.5 \nμM) or vehicle (DMSO) for 3 days (nexp = 3, nrep = 4).  \n(I-J) Normalised ATP-linked respiration and spare reserve capacity of  isogenic and Q336R \nneurons calculated from traces in G (nexp = 3, nrep = 4, *p= 0.0177, ****p <0.0001).  \nData ( B, F, G, H, I and J ) are expressed as mean ± SD and individual data points from \nindependent experiments are shown in each plot. Statistical analysis was carried out using one-\nway ANOVA followed by posthoc Tukey’s test. \n \n \nFigure S9: Mitochondrial Ca 2+ buffering capacity deter mined by simultaneous \nmeasurement of [Ca2+]m and ΔΨm in Q336R neurons.  \n(A) Representative confocal images of permeabilised isogenic control neurons show changes \nin fluorescence of TMRM (red) and Fluo-4 AM (grey) in response to increasing concentrations \nof calcium (3.28  and 13.3 μM). Before digitonin permeabilization, Fluo -4 intensity shows \ncytosolic localization of the dye along with TMRM signal localizing to tubular mitochondria \nboth in somas and axons of isogenic control neurons. The punctate staining pa ttern of Fluo 4 \nand TMRM following permeabilization confirms mitochondrial localisation of the dyes.  (B) \nTraces showing mean change ± SEM in Fluo -4 (representing [Ca2+]m) and TMRM \n(representing ΔΨm) fluorescence intensity in response to increasing concentr ations of Ca2+ in \nthe media. Quantified data from one representative control experiment depict the characteristic \nresponses. The final free Ca2+ in the media is indicated for each respective addition.  \n \n \nVideo description: \n \nVideo S1: Time-lapse imaging of ΔΨm, [Ca2+]m and mtDNA dynamics in response to \nhistamine in control 1 fibroblasts. Control 1 fibroblasts co-labelled with TMRM, mito-Fura-\n2 and PicoGreen were imaged every 12.5 s to simultaneously monitor the change in ΔΨm, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n67 \n \n[Ca2+]m and mtDNA dynamics , respectively, in response to the successive application of 10 \nμM and 20 μM histamine. Movie playback 30fps. Scale bar, 10 μm. \n \nVideo S2: Time-lapse imaging of ΔΨ m, [Ca2+]m and mtDNA extrusion in response to 10 \nμM histamine in patient 1 fibroblasts.  Patient 1 fibroblasts co -labelled with m ito-Fura-2, \nTMRM and PicoGreen were imaged every 12.5 s to simultaneously monitor the increase in \n[Ca2+]m, collapse of ΔΨm, and mtDNA extrusion, respectively, in response to 10 μM histamine \nchallenge. Movie playback 30fps. Scale bar, 10 μm. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n68 \n \nMaterials and Methods \nCell culture and treatments \nFibroblasts carrying specific EPG5 mutations (Table S1) were isolated from skin biopsies \nobtained from patients as part of the diagnostic process. Age and sex matched healthy control \nfibroblasts were obtained from the MRC Centre for Neuromuscular Disorders Biobank, \nLondon. All cell lines were cultured and ma intained in Dulbecco’s modified Eagle’s medium \n(10566016, Gibco) supplemented with 10% fetal bovine serum (16140071, Gibco), and 1% \nAntibiotic-Antimycotic (15240096, Gibco) and incubated at 37°C with 5% CO 2. Fibroblasts \nlines were maintained at maintained at sub-confluence (80%) and cultured between 3 to 10 \npassages. All cell cultures were routinely checked for mycoplasma using MycoAlert™ \nMycoplasma Detection Kit (LT07 -118, Lonza). For transfection, Human Dermal Fibroblasts \nNucleofector Kit (VPI -1002, Lonza) was used according to the manufacturer's instructions. \nFibroblasts were treated with either freshly prepared histamine (CAY33828, Cambridge \nBioscience) or a thapsigargin (586005, Merck), H -151 (S6652, Selleck) and G140 (S9945, \nSelleck) stock solutions ma de in DMSO, at indicated working concentrations for 24 h. \nFibroblasts and neuronal cultures were preincubated with JP1 -138 in imaging media for 10 \nminutes before live cell imaging. For long -term treatments with either JP1 -138, CsA (1101, \nTocris) or H-151, the culture medium was changed every two days.  \n \nhiPSCs derivation and culture \nThe iPSC lines, GM27291 and GM28930 used in this study were generated by Coriell Institute. \niPSC cell line GM27291 was derived from GM26636 fibroblasts (Patient 1) (Mitchell et al., \n2022) and iPSC cell line GM28930 is the isogenic control for the patient -derived line \nGM27291. iPSCs were thawed, expanded, and maintained using hESC -qualified matrigel -\ncoated plates (356277, Corning) with mTeSR1 iPSC medium (85850, STEMCELL \nTechnologies) as  per the commercially available protocol from STEMCELL Technologies. \nThe iPSCs were passaged when they reached approximately 70% –80% confluency using 0.5 \nmM EDTA (S311 -500, Fisher Scientific) prepared in PBS (10010023, Thermo Fisher \nScientific). \n \nDifferentiation of human iPSCs to neuronal networks \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n69 \n \nCortical neuronal cultures were generated from human iPSCs according to (Shi et al., 2012) . \nNeural maintenance media (NMM) comprised a 1:1 mixture of DMEM/F‐12 Gl utaMAX \n(10565018, Thermo Scientific) with Neurobasal media (12348 -017, Thermo Scientific), \nsupplemented with 2.5 µg/ml insulin (I9278, Sigma -Aldrich), 1 mM L‐glutamine (25030 -\n024Thermo Scientific), 50 µm nonessential amino acids (NEAA) (11140 -050, Thermo \nScientific) , 50 µM β‐mercaptoethanol (31350010, Gibco),  0.5x N‐2 supplement (17502001, \nThermo Scientific), 0.5x B‐27 (17504001, Thermo Scientific), and 0.25% \npenicillin/streptomycin (15140122, Gibco). Neural induction media (NIM) consisted of NMM \nwith 1 µM Dorsomorphin (3093, Tocris) and 10 µM SB431542 (1614, Tocris). In brief, iPSCs \nwere cultured as above transferred at a seeding ratio of 2:1 to hESC‐qualified Matrigel coated \nwells, in mTeSR1 supplemented with 10 µM ROCK inhibitor (1254/10, Bio -techne). Cells \nwere incubated at 37°C for 24 h‐48 h to reach 100% confluence. Cell layers were gently washed \nwith PBS and medium replaced with Neural Induction Media (NIM), which was then replaced \ndaily, until a dense neuro‐epithelial sheet formed at around 8‐12 day s after neural induction. \nAt this point, cells were passaged using 1 mg/ml dispase (17105041, Thermo Scientific) in \nNIM, and lifted as aggregates. These aggregates were washed and seeded at a ratio of 1:2 in \nfresh NIM, onto laminin (L2020-1MG, Sigma-Aldrich) wells. Cells were incubated overnight \nto allow attachment, and clumps that had not been attached were transferred to fresh laminin \ncoated wells. Upon the formation of neural rosettes, culture medium was replaced with NMM \nsupplemented with 20 ng/ml recom binant human FGF2 (100‐18B, PeproTech), which was \nchanged every 48 h. FGF2 was withdrawn from the culture medium after 4 days, and cells were \nexpanded using dispase to lift neural rosettes. On day 25 (±1 day) after induction, cultures were \ndissociated into  single cells using Accutase (A1110501, Thermo Scientific), and seeded in \nNMM at a ratio of 1:1 in fresh laminin coated wells. From this point onwards, cells were \nexpanded as single cell suspensions with Accutase, and media was replaced at least every 48 \nh. At ~ day 36 of neural induction, iPSC‐NPCs (neural progenitor cells) were seeded in poly‐\nL‐lysine (P4707, Sigma) and laminin coated plates at a density of 5x104 cells/cm2 for terminal \nneuronal differentiation, and NMM was replaced at least every 48 h. Ne uronal cultures were \nmatured until at least day 62. \nFor neuronal characterization, Neuronal cultures in glass -bottom 96 -well plate (655892, \nSensoPlateTM) were fixed for 15 min at room temperature using a 4% paraformaldehyde \nsolution, washed and permeabilized for 5 min with 0.1% Triton X-100 (AC215682500, Fisher \nScientific) in PBS. Fixed cells were blocked with 3% BSA (A6003, Sigma) in PBS to block \nthe nonspecific bindings for 60 min. Cells were then incubated overnight at 4°C with the GFAP \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n70 \n \nand MAP2 primary antibodies prepared in blocking solution. The following day, the cells were \nwashed with PBS and incubated with secondary antibodies, Alexa Fluor 488 or Alexa Fluor \n568 based on the primary antibody host species, for 1 h. Then, the cells were washed with PBS \nand incubated with 4,6-diamidino-2-phenylindole (DAPI) solution (1:10,000, D1306, Thermo \nScientific) for 5 min to detect all nuclei and imaged using confocal microscope as described \nbelow. Details of the antibodies used in this study can be found in Supplementary Table S3.     \n \nRNA-sequencing  \nFibroblasts for RNA-sequencing were seeded at a density of 1 × 10 6 cells/dish and cultured in \n10 cm dishes with regular media before harvesting on day two. RNA sequencing was \nperformed at the UCL Genomics Core Facil ity using the Mag -Bind® Total RNA 96 Kit for \nRNA-sequencing polyA capture on the Illumina NextSeq 2000 P2 sequencer. Reads were \nmapped and analyzed by the SARTools R package. Differential expression analysis was carried \nout with the Bioconductor package DE Seq2 (v.1.48.0) (Love et al., 2014) . Genes were \nannotated using the org.Hs.eg.db package (Genome wide annotation for Human, v.3.21.0) and \ndifferentially expressed genes (DEG) with an adjusted p -value cut-off of 0.05 were identified \nas statistically significant. Gene ontology (GO) and Kyoto Encyclopedia of Genes and \nGenomes (KEGG) enrichment analysis were performed using the ClusterProfiler package (Yu \net al., 2012) . Results of DEG analysis were visualized with heatmaps plotted in ClustVis \n(Metsalu and Vilo, 2015).  \n \nTransmission Electron Microscopy \nFibroblasts cultured on coverslips were fixed in EM fixative (2% glutaraldehyde + 2% \nparaformaldehyde in 0.1M sodium cacodylate) for 1 h followed by washings with 0.1M \nCacodylate Buffer. The Coverslips were then fixed in 1% osmium tetroxide and 1% potassium \nferricyanide in 0.1M sodium cacodylate, followed by sequential dehydration in ethanol. The \ncoverslips were embedded in Epoxy Resin (Araldite CY212) Kit (Agar Scientific Ltd.) \naccording to the standa rd protocol. Embedded coverslips were sectioned to 50 nm using an \nultra-microtome fitted with a diamond knife, mounted onto TEM compatible copper grids. The \ngrids were then stained with Lead Citrate for 3 min before proceeding for Imaging. Images \nwere acquired using Jeol 1400 Transmission Electron Microscope and Gatan software. Images \nwere taken at a magnification of 1200X (digital magnification). Interfaces between ER and \nmitochondria were segmented using DeepMIB software (Belevich and Jokitalo, 2021)  \nanalyzed using a custom ImageJ script: https://sites.imagej.net/MitoCare/ . \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n71 \n \nROS measurements \nThe rate of intracellular ROS production in fibroblasts was measured using a superoxide \nindicator,  dihydroethidium (DHE; D11347, Thermo Scientific) as described previously \n(Chung and Duchen, 2022). Following the measurements of fluorescence intensity, cells were \nstained with Hoechst 33 342 (62249, Thermo Scientific) for 10 min to label and count the \nnumbers of cell nuclei representing cell numbers in each well using an automated fluorescent \nimage acquisition system (ImageXpress MicroXL). Subsequently, the fluorescence intensity \nwas normalized based on the relative cell numbers obtained. \n \nSDS-PAGE and immunoblotting \nFibroblast cultures from 60 mm dishes were washed with PBS and collected by trypsinization \n(0.5% trypsin-EDTA; 15400054, Gibco). Similarly, mixed neuronal cultures were gently lifted \nand collected using Accutase. To prepare the samples for Western blot analysis, cell pellets \nwere homogenized in 100 –150 μl RIPA buffer (R0278, Sigma -Aldrich) containing 1x Halt \nProtease and Phosphatase Inhibitor Cocktail (78440, Thermo Scientific).  Following \nhomogenization, cell lysates were centrifuged at 16,000g at 4 °C for 30 min and the protein \nconcentration in the supernatant was quantified using the Pierce BCA Assay Kit (23227, \nThermo Scientific). Equivalent amounts of total protein (30 µg) samples in NuPAGE 4x LDS \nSample Buffer (NP0007, Invitrogen) and 2% β-mercaptoethanol (63689, Sigma-Aldrich) were \nboiled at 95 °C for 10 min. Proteins were separated on either 12% Bolt Bis -Tris Plus \n(NW04127, Invitrogen) or 4 –12% NuPAGE Bis -Tris polyacrylamid e gels (NP0335, \nInvitrogen) immersed in MOPS running buffer (NP0001, Invitrogen) and transferred onto \nPVDF membranes (1620175, Bio -Rad). Membranes were then blocked in SuperBlock \nBlocking Buffer (37545, Thermo Scientific) for 1 h at room temperature, and probed overnight \nat 4°C using indicated primary antibodies. After incubation with appropriate secondary \nantibodies, protein bands were detected using a chemiluminescent reagent (Luminata Forte \nWestern HRP substrate; WBLUF0100, Merck) and imaged using a Chem iDoc system (Bio -\nRad). Membranes were further stripped using Restore Western Blot Stripping Buffer (21059, \nThermo Scientific) and re -probed with additional primary antibodies. Quantification of the \nprotein bands was performed using with ImageJ (NIH) and Im age Lab software, v6.0.1 (Bio-\nRad). Details of all the antibodies used in this study can be found in Table S3.     \n \nBlue native gel electrophoresis (BNGE) and immunoblotting  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n72 \n \nMitochondria from the fibroblasts were isolated using mitochondria isolation buffer 1 (MIB1; \n225 mM Mannitol, 75 mM Sucrose, 5 mM HEPES, 1 mM EGTA and 1 mg/ml fatty acid free \nBSA) and MIB2 (same as MIB1 but without BSA) according to the method described \npreviously (Singh and Duchen, 2022). Mitochondria isolation from the neuronal cultures were \nperformed similarly, with the exception of homogenization step, performed using a Dounce \ntissue grinder tube. Protein concentration of the isolated mitochondria was quantified using the \nPierce BCA Assay Kit (23227, Thermo Scientific) and an equivalent amounts of total protein \n(50 µg) was solubilized with digitonin followed by centrifugation at 20,000g for 20 min at 4°C. \nDigitonin-solubilized mitochondria were separated on 3 –12% NativePAGE Bis -Tris gels \n(BN1001, Invitrogen) and electroblotted onto PVDF membrane (1620175, Bio-Rad) according \nto the manufacturer’s instructions. Membranes were then blocked and probed with indicated \nprimary antibodies as described above. Quantification of the protei n bands was performed \nusing ImageJ. Details of all the antibodies used in this study can be found in Table S3.     \n \nQuantitative reverse transcription PCR (RT-qPCR) \nTotal RNA was extracted from fibroblasts and neuronal culture using the RNeasy Plus Mini \nKit (74104, Qiagen) according to the manufacturer’s instructions. Quality check and the \nquantification of isolated mRNA was done on the Nanodrop 2000c (ND -2000, Thermo \nScientific). cDNA was synthesized from 1 μg of total RNA using SuperScript IV First -Strand \nSynthesis System Kit (18091050, Invitrogen) and quantitative PCRs was performed using \nSYBR Green JumpStart Taq ReadyMix (Sigma -Aldrich) on a CFX96 Real -Time PCR \nDetection System (Bio-Rad). Data were analysed using the comparative 2 −ΔΔCt method. Ct of \nthe gene of interest was normalized to that of β-actin. \nFor the quantitative analysis of the relative mtDNA copy number, total genomic DNA was \nextracted from fibroblasts using the DNeasy  Blood & Tissue Kit (69506, Qiagen) and the \nquantitative PCR was performed similarly with primers for the mtDNA tRNALeu (UUR) and \nfor the nuclear B2M (β -2-microglobulin) to determine the relative mtDNA copy number of \ncells (Rooney et al., 2015) . The following equation was used to determine the relative \nmitochondrial DNA content, 2 x 2 ΔCt, where ΔCt is nuclear DNA Ct value subtracted by \nmtDNA Ct value. All primer pairs used can be found in Table S4. \n \nMitochondrial oxygen consumption rate (OCR) \nMeasurements of mitochondrial respiration in fibroblasts and neurons were conducted with the \nSeahorse Bioscience XFe96 bioanalyzer using the Seahorse XF Cell Mito Stress Test Kit \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n73 \n \n(103015-100, Agilent). Fibroblasts were seeded at a density of 1 × 104 cells/well on XF96 cell \nculture microplates (102416 -100, Agilent) and cultured for 1 -2 days and/or treated for the \nindicated time. For neuronal cultures, iPSC-derived NPCs were seeded at a density of 5 × 104 \ncells/well on XF96 cell culture microplates coated with poly‐L‐lysine and laminin for terminal \nneuronal differentiation and maturation using the method describe above. On the day of the \nexperiment, the culture medium was replaced with Seahorse  XF Base medium (103334 -100, \nAgilent) supplemented with 1mM pyruvate (11360070, Gibco), 2 mM glutamine (25030081, \nGibco) and 10mM glucose (A2494001, Gibco) and incubated in a CO 2-free incubator for 30 \nmin at 37 °C before loading into the Seahorse Analyser. After the measurement of basal \nrespiration, the drugs oligomycin (5 μM), FCCP (1 μM, 2 μM), and rotenone/antimycin A (0.5 \nμM/ 0.5 μM) were added to each well in sequential order. Data was analysed using the XF Cell \nMito Stress Test Report Generator. The OCR data from fibroblasts were normalized to the cell \nnumber obtained after counting the numbers of cell nuclei with ImageXpress MicroXL as \ndescribed above. The OCR data from neurons were also normalized to the cell counts obtained \nafter constructing a calibration curve and calculating the normalized cell number per well with \nCyQuant Direct Cell Proliferation Assay Kit (C35011, Invitrogen) and expressed as a \npercentage of the baseline measurement of untreated line. \n \nMitochondrial membrane potential (ΔΨm) in fibroblasts and neurons \nThe steady -state and time -lapse measurement of ΔΨ m in fibroblast was carried out using \ntetramethylrhodamine methyl ester (TMRM) in the redistribution mode  where a lower \nfluorescence intensity indicates reduced ΔΨm and vice versa. For the steady-state measurement \nof ΔΨm, cells were seeded at a density of 1 × 10 5 cells/dish on fluorodishes (FD35-100, WPI) \nand cultured for 1 day and/or treated for the indicated time. On the day of the experiment, cells \nwere washed once with recording buffer 1 (RB1; 150mM NaCl, 4.25mM KCl, 4 mM NaHCO3, \n1.25mM NaH2PO4, 2 mM CaCl2, 1.2 mM MgCl2, 10mM D-glucose, and 10mM HEPES at pH \n7.4) and incubated with 25 nM TMRM (T668, Invitrogen) and 1 μM Calcien-AM (C3100MP, \nInvitrogen) in RB1 for 30 minutes at 37°C. Following incubation, cells were washed twice with \nRB1 and TMRM was added in the RB1 to avoid its depletion while imaging. z-stacks with 0.45 \nµm thickness with a pixel dwell time  of 1.54 μs were acquired using a Zen Black software -\ncontrolled LSM 880 confocal microscope (Carl Zeiss) equipped with a plan -Apochromat \n63x/1.4 oil DIC objective lens and an Ar (λex = 488 nm, λem = 500–550 nm for Calcien-AM)and \nDPSS laser  source (λ ex = 561 nm, λ em = 575–625 nm for TMRM) at 37 °C. Mean TMRM \nfluorescence intensity was quantified using the same threshold across all the samples and the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n74 \n \npercentage mitochondrial mass was calculated from the area of the binarized images of \nCalcienAM and TMRM using ImageJ.   \nFor the time-lapse measurement of ΔΨm, images were acquired from a single z -plane every 2 \ns interval with a pixel dwell time of 1.54 μs. After acquiring baseline TMRM images, 10 uM \nhistamine was applied directly into the fluorodishes using a m icropipette and subsequent \nimages were obtained to monitor the effect of histamine on ΔΨ m. At the end of each \nexperiment, 2.5 μM Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP; C2920, \nSigma-Aldrich) was added as a positive control which depola rized the mitochondria. Time \nseries were analysed using ImageJ by selecting and measuring the mean fluorescence intensity \nin the regions of interests (ROIs) in each field. Individual ΔΨm traces were normalized to their \nbaseline intensity obtained before stimulation.  \nMeasurement of ΔΨm in neurons was carried out using Rhodamine-123 in the dequench mode \nwhere dequenching or increase in Rhodamine -123 fluorescence indicates mitochondrial \ndepolarization and loss of ΔΨ m. Neuronal cultures seeded in a glass -bottom 96 -well plate \n(655892, SensoPlate) were labelled with 10 μg/ml Rhodamine-123 (R8004, Sigma-Aldrich) in \nBrainPhys Imaging Optimized Medium (05796, STEMCELL Technologies) at 37 °C and 5% \nCO2 for 20 min. After incubation, cells were washed thrice and imaged as above using a Plan-\nNeofluar 40x/1.30 oil objective and an Ar laser source (λex = 488 nm, λem = 510–600 nm). For \nall experiments, the laser illumination intensity was kept to a minimum (max 0.5%) to avoid \nphototoxicity and photobleaching. After acquiring baseline Rhodamine -123 images, neurons \nwere stimulated using 10 μM glutamate (G1626, Sigma-Aldrich) and subsequent images were \nacquired. At the end of each experiment, 1 μM FCCP was added to evaluate the Rhodamine -\n123 fluorescence intensity corresponding to the loss of ΔΨm. Time series were analysed using \nImageJ as described above. \n \nMitochondria morphology analysis \nMorphometric analysis of TMRM -labelled or T OM20-immunolabelled raw confocal images \nwas performed using a 2D cell segmentation model, MitoSegNet (Fischer et al., 2 020). \nNeuronal somas and axons were manually segmented in ImageJ using β-tubulin III staining for \nmorphological distinction. Briefly, images were pre -processed to 8 -bit format and \nmitochondria were first segmented using the MitoS basic toolbox \n(https://github.com/MitoSegNet). Segmentation masks were then run on MitoA analyzer tool \nfor the quantification of the morphological features broadly categorized into shape descriptors \nsuch as area, eccentricity and perimeter and the network descriptors such as branch length, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n75 \n \nnumber of branches and curvature index. The shape descriptors values were then used to \ncalculate the percentage of elongated, fragmented or swollen mitochondrial pool.  \n \nMeasurement of mitochondrial NAD(P)H \nNAD(P)H autofluorescence imaging was carried out to investigate the mitochondrial redox \nstate in fibroblasts. Cells were seeded at a density of 1 × 10 5 cells/dish on fluorodishes and \ncultured for 1 day and/or treated for the indicated time. On the day of  the experiment, cells \nwere washed once with RB1 and imaged using the Carl Zeiss LSM 880 confocal microscope \nequipped with an ultraviolet (UV) laser (λ ex = 355 nm, λ em = 410–480 nm) and quartz Plan -\nApochromat 63x /1.4 oil objective at 37 °C. Images were ac quired from a single z-plane with \nthe pinhole wide open to maximize signal and laser illumination intensity kept to a minimum \n(0.1-0.2%) to avoid phototoxicity. After acquiring baseline images, cells were first exposed to \n2.5 μM FCCP to depolarize the mitochondria completely and achieve maximal respiration. The \noxidation of the mitochondrial pool of NADH into non -fluorescent NAD + led to the lowest \nfluorescence signal, which was considered as 0%. Thereafter, 1 mM cyanide (NaCN), an \ninhibitor of mitochondrial respiration, was added to allow the regeneration of the mitochondrial \npool of NADH (the highest fluorescence signal was considered 100%). The baseline \nautofluorescence acquired from each cell was normalized between the minimal (0%) and \nmaximal (100%) fluorescent signals to calculate the redox index. \n \nMitochondrial Ca2+ concentration measurements with aequorin \n[Ca2+]m measurements in fibroblasts and neuronal cultures were carried out using the \nmitochondria-targeted luminescent aequorin probe, mtAEQ  as previously described (Bonora \net al., 2013).  Fibroblasts were seeded at a density of 2 × 104 cells/well on white 96-well plate \n(6005680, PerkinElmer) and cultured for 1-2 days. Similarly, iPSC-derived NPCs were seeded \nat a density of 5 × 104 cells/well on white 96-well plate coated with poly‐L‐lysine and laminin \nfor terminal n euronal differentiation and maturation using the method describe above. Two \ndays before the experiment, cells were transduced with mtAEQ adenovirus. Following \nincubation, media was replaced with 5 μM coelenterazine (C2944, Invitrogen) in Krebs Ringer \nBuffer (125mM NaCl, 5.5mM D -Glucose, 5 mM KCl, 20mM HEPES, 1 mM Na 3PO4, 1mM \nGlutamine, 100mM Pyruvate, and 1.2 mM CaCl 2 at pH 7.4). The plate was then incubated in \nthe dark for 2 h at 37 °C. Baseline luminescence signals were acquired using a plate reader \n(CLARIOstar, BMG Labtech) every 1 s followed by fluidic additions of either 10 μM histamine \nor 10 μM glutamate using integrated syringe injectors. At the end of each experiment, maximal \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n76 \n \naequorin signal was obtained by permeabilizing the cells with 1 mM digitoni n and exposing \nthe cells to a saturating Ca2+ concentration of 10 mM CaCl2. For analysis, luminescence values \nwere converted into Ca2+ concentration as previously described (Bonora et al., 2013). \n \nSimultaneous measurement of [Ca2+]c and [Ca2+]m in fibroblasts \nDynamic measurements of cytosolic and mitochondrial Ca 2+ concentrations were carried out \nin fibroblasts loaded with 2.5 μM Fluo-4 AM and 1 μM mt-Fura-2.3 AM (a modified version \nof mt-Fura-2 (Pendin et al., 2019), referred here as mito-Fura-2 AM) in RB1 containing 0.005% \nPluronic F-127 for 30 minute s at 37°C. Following incubation, cells were washed thrice with \nRB1 and imaged using the Carl Zeiss LSM 880 confocal microscope equipped with an UV (λex \nCa2+ bound mito-Fura-2 = 355 nm, λ ex Ca2+ unbound mito-Fura-2 = 405 nm,  λ em = 470–600 \nnm) and Ar laser  (λex = 488 nm, λem = 505–560 nm for Fluo-4) and quartz Plan-Apochromat \n63x /1.4 oil objective at 37 °C. Images were acquired sequentially from a single z-plane every \n12.5 s interval with a pixel dwell time of 1.54 μs. For all experiments, the laser illumi nation \nintensity was kept to a minimum (0.25- 0.5%) to avoid phototoxicity. After acquiring baseline \nimages for about 100 s, 10 μM histamine was applied directly into the fluorodishes using a \nmicropipette and subsequent images were obtained. A total of 1 μ M ionomycin was added at \nthe end of each course as a positive control. Time series were analyzed in ImageJ as described \nabove. After background subtraction, the change in Fluo-4 intensity was calculated relative to \nthe baseline and shown as ΔF/F 0. For mito -Fura-2, ratios between the fluorescence intensity \nexcited at 405 nm and at 355 nm were calculated at each time point and plotted as 355/405 \nratio representing [Ca2+]m. The values for area under the curve (AUC), basal mito-Fura-2 ratio, \nt0.5 to [Ca 2+]m peak and t 0.5 of stabilization for [Ca 2+]c peak were calculated for each \nindependent experiment in GraphPad Prism. \n \nCalcium retention capacity assay \nThe capacity of isolated mitochondria to accumulate Ca 2+ until the mPTP opens was \ndetermined using the method described earlier (Bhosale and Duchen, 2019) . Isolated \nmitochondria from the fibroblasts (0.5mg/ml), using the method described above, were \nresuspended in the RB2 (75 mM D -mannitol, 25 mM sucrose, 5 mM KH 2PO3, 20 mM Tris -\nHCl, 100 mM KCl, and 0.1% BSA fatty acids free at pH 7.4) supplemented with 10 mM \nsuccinate and 1 μM rotenone to energize the mitochondria. 100 μl of the mitochondrial \nsuspension was plated in a glass -bottom 96 -well plate in triplicate for each condition. \nExtramitochondrial Ca 2+ levels were quantified by measuring fluorescence intensity of the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n77 \n \nCa2+-sensitive dye, Calcium Green -5N (C3737, Invitrogen) at 1μM concentration The \nfluorescence intensity was recorded using a plate reader (CLARIOstar, BMG Labtech) at 30 ̊C \nwith the following  filters: ex/em: 480 nm/520 nm. Ca 2+ additions were achieved using \nintegrated syringe injectors, where subsequent 10 μl additions of 50 μM CaCl2 were added for \na total of 12 injections. The area under the curve was used as a measure of extramitochondrial \nCa2+, which was expressed as a proportion of total Ca 2+ added (Ca2+ free condition was used \nfor background subtraction). This value was used to calculate the proportion of buffered Ca2+, \nand subsequent percentage inhibitions were calculated compared to untreated. \n \nCytosolic and mitochondrial calcium imaging in neurons \nMeasurements of cytosolic Ca2+ concentrations were performed in neuronal cultures seeded on \nfluorodishes and loaded with 2.5 μM FuraFF AM (CAY20416, Cambridge Bioscience) in \nBrainPhys Imaging Optimized Medium containing 0.001% Pluronic F -127 for 30 minutes at \n37°C. Following incubation, cells were washed thrice and imaged on a custom-made Olympus \nIX71 inverted epifluorescence microscope equipped with a UAPO/340 20x/0.70 water \nobjective and a Xen on arc lamp. FuraFF was excited alternately at 340 nm ± 20 nm and 380 \nnm ± 20 nm and emitted light was collected through a dichroic T510lpxru (Chroma). Images \nwere acquired with a Zyla CMOS camera (Andor) every 2 s using MetaFluor 7.8.12.0 \n(Molecular Devices). After acquiring baseline images for about 100 s, neurons were stimulated \nwith 10 μM and/or 100 μM glutamate applied directly into the fluorodishes using a \nmicropipette. A total of 2 μM ionomycin was added at the end of each experiment as positive \ncontrol. For analysis, time series were imported and processed in ImageJ as described above. \nRatios between the fluorescence intensity excited at 380 nm and at 340 nm were calculated at \neach time point and plotted as 340/380 ratio representing [Ca 2+]c. Represe ntative FuraFF \nratiometric images at indicated time were obtained using Metamorph 7.8.12.0 (Molecular \nDevices). \nMitochondrial Ca2+ concentrations were measured in neurons loaded with 1 μM mito -Fura-2 \nAM in BrainPhys imaging medium containing 0.001% Pluronic F-127 for 30 minutes at 37°C. \nFollowing incubation, cells were washed thrice and imaged using the Carl Zeiss LSM 880 \nconfocal microscope equipped with an UV laser and quartz Plan -Apochromat 63x/1.4 oil \nobjective, as described above. Images were acquired from a single z -plane every 5 s interval \nwith a pixel dwell time of 2.05 μs. After acquiring baseline images, 10 μM glutamate was \napplied directly into the fluorodishes using a micropipette and subsequent images wer e \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n78 \n \nobtained.  Time series were analyzed in ImageJ and the ratios between the fluorescence \nintensity excited at 405 nm and at 355 nm were calculated at each time point and plotted as \n355/405 ratio representing [Ca 2+]m. Basal mito -Fura-2 ratio, t 0.5 to [Ca 2+]m peak and peak \namplitude values were calculated for each cell in GraphPad Prism. Representative mito -Fura-\n2 ratiometric images were generated using Metamorph. \n \nMitochondrial calcium retention assay in permeabilized neurons \nTo measure the capacity of neuronal mitochondria to accumulate Ca 2+ until the mPTP opens, \ndigitonin-permeabilized neurons were loaded with Fluo -4 AM and TMRM to simultaneously \nmeasure [Ca2+]m uptake and the loss of ΔΨm as a readout for mPTP opening. Briefly, neuronal \ncultures seeded on fluorodishes, were washed once with RB1 and loaded with 5 μM Fluo -4 \nAM, 25 nM TMRM and 0.1% Pluronic F -127 dissolved in RB1 for 30 minutes at room \ntemperature (RT). Following incubation, cells were permeabilized with 10 μg/ml digitonin in \nRB3 (5 mM NaC l, 130 mM KCl, 1 mM KH 2PO4, 20 mM HEPES, 6.5 mM MgCl 2, 1.5 mM \nEGTA, 6 mM EDTA, 0.4 mM CaCl2, 2 mM malate, 2 mM succinate, 2 mM glutamate and 2.5 \nmM ADP; pH adjusted to 7.3 with 1 M KOH) containing 250 nM TMRM and 2 μM \nthapsigargin for 10 minutes. After per meabilization, excess digitonin was washed off and the \nneurons were imaged in RB3 with 250 nM TMRM and 2 μM thapsigargin to inhibit \nsarco/endoplasmic reticulum Ca2+ ATPase. Imaging was performed as described above, after \nbaseline TMRM and Fluo-4 images, small volumes (5 or 10 μl) of 50 mM CaCl2 were carefully \nadded to the chamber using a P10 micropipette until the loss of TMRM signal. Images were \nacquired every 120 seconds over a total time of 30 minutes. The final free Ca 2+ ion \nconcentration in RB3 was cal culated using WEBMAXC Extended software: \nhttps://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcE.ht\nm. TMRM and Fluo-4 fluorescence intensities were calculated relative to baseline and shown \nas ΔF/F0 where ΔF is the difference in fluorescence between baseline and post CaCl2 addition \nand F0 is the basal fluorescence. Ratios of Fluo-4 and TMRM fluorescence intensities (ΔF/F0) \nwere plotted against time and fitted with a nonlinear sigmoidal curve function. \n \nLive cell imaging of mtDNA release in fibroblasts \nQuantitative analysis of mtDNA dynamics in fibroblasts was carried by co -labelling cells \nseeded in fluorodishes with 25 nM TMRM, 1 μM PicoGreen (P11495, Invitrogen), 1 μM mito-\nFura-2 and 0.005% Pluronic F -127 in RB1. After incubation for 30 min at 37 °C, cells were \nwashed thrice and imaged on Carl Zeiss LSM 880 confocal microscope equipped with an UV \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n79 \n \n(λex Ca2+ bound mito-Fura-2 = 355 nm, λ ex Ca2+ unbound mito-Fura-2 = 405 nm, λ em = 470–\n600 nm), Ar (λ ex = 488 nm, λ em = 505–560 nm for PicoGreen) and DPSS laser source (λ ex = \n561 nm, λ em = 570–700 nm for TMRM) and quartz Plan -Apochromat 63×/1.4 oil objective. \nImages were acquired from a single z -plane every 12.5 s interval with a pixel dwell time of \n0.85 μs. After acquiring baseline images for about 250 s, 10 μM histamine was applied directly \ninto the fluorodishes using a micropipette and subsequent images were obtained. At the end of \neach experiment, 1 μM FCCP was added at the end of the time course to completely depolarize \nmitochondria. Time series were processed and analyzed in ImageJ a nd Metamorph. TMRM \nfluorescence intensities were normalized relative to baseline and shown as ΔF/F0, mito-Fura-2 \nfluorescence intensities were plotted as 355/405 ratio representing [Ca2+]m as described above. \nFor the quantification of cytosolic PicoGreen puncta, number of PicoGreen puncta outside the \nnucleus and the TMRM-labelled mitochondrial perimeter were counted manually between the \ntime points after the challenge with 10 μM histamine and before the application of 1 μM FCCP. \nFibroblasts expressing mTagBFP-cGAS (Addgene:102603, a gift from Nicolas Manel) were \nco-labelled with 1 μM PicoGreen and 25 nM TMRM and imaged as described above with \nmTagBFP-cGAS visualized using an UV laser source (λ ex = 405 nm, λem = 410–470 nm). For \nthe quantification of the cyto solic PicoGreen puncta with cGAS, cells displaying puncta \npositive for cGAS and PicoGreen were selected and a binary image of TMRM and PicoGreen \nchannels was generated in ImageJ. A segmented mask of non-mitochondrial Picogreen puncta \nwas generated using image calculator and number of events with cGAS and PicoGreen overlap \nwithin each ROI were scored. \n \nImmunocytochemistry \n \nImmunofluorescence analysis fibroblasts \nImmunofluorescence analysis of cytosolic mtDNA puncta/ mtDNA release in fibroblasts was \nperformed using Airyscan imaging. Briefly, cells were seeded at a density of 2 × 104 cells/well \nin a glass-bottom 96-well plate and cultured for 1-2 days. On the day of the experiment, cells \nwere fixed in 5% PFA in PBS for 30 min at RT, then washed three times with  PBS, followed \nby quenching with 50 mM ammonium chloride in PBS. After fixation, cells were washed thrice \nwith PBS and permeabilized in 0.1% Triton X-100 in PBS for 10 min, followed by three washes \nin PBS. Permeabilized cells were then blocked with 10% FBS in PBS, followed by incubation \nwith primary antibodies in 5% FBS in PBS, for 2 h at RT. Cells were then washed three times \nwith 5% FBS in PBS and labelled with the corresponding secondary antibodies prepared in 5% \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n80 \n \nFBS in PBS for 1 h at RT. After three washes, super-resolution Airyscan images were acquired \non a Zeiss LSM 880 with Airyscan detector in SR mode using all 32 pinholes. At least 8 z -\nstacks with optimal slice sizes 0.45 µm (overview image) and 0.185 µm (inset image) with a \npixel dwell time of 4.10 μs were acquired. Prior to image analysis, raw confocal micrographs \nwere automatically processed into deconvoluted Airyscan images using the Zen Black \nsoftware. The number of DNA puncta outside the nucleus and mitochondrial perimeter were \ncounted in Image J by generating a nuclear and mitochondrial mask and subtracting this \nsegmented mask from DNA channel in image calculator. The colocalization coefficient \n(Pearson's R-value) was quantified using Coloc 2 plugin. For 3D rendering and quantification \nof cytoso lic mtDNA puncta or mtDNA release events, cells with enlarged nucleoids were \nselected and a segmented mask of IMM was created using surface function in Imaris 9.8 \n(Bitplane). Using this mask mtDNA nucleoids outside the surface were differentiated using \nspot function. A surface rendering of OMM was generated similarly and the transparency of \nIMM and OMM was adjusted to allow the visualization of mtDNA nucleoid spots. Details of \nall the antibodies used in this study can be found in Supplementary Table S3.     \nFor the cGAS immunofluorescence, cells were cultured, fixed and permeabilized as above. \nAfter permeabilization, cells were blocked with 5% BSA in PBS for 1 h at the RT and incubated \novernight with primary antibody diluted in PBS with 1% BSA and 0.1% Trito n X-100. After \nthree washes in PBS, cells were incubated with secondary antibody diluted in PBS with 1% \nBSA for 2 h at RT. Cells were then washed three times and incubated with DAPI for 5 min to \nlabel nuclei. After three washes, cells were imaged using the confocal microscope as described \nabove. The nuclear cGAS intensity in each cell was quantified from the DAPI positive area \nand the cytosolic cGAS intensity per cell was calculated from subtracting the DAPI positive \narea from total cGAS mask area in ImageJ. \n \nImmunofluorescence analysis neurons \nNeuronal cultures were treated overnight with 10 μM glutamate and fixed with 4% PFA in 2x \nmicrotubule stabilization buffer (160 mM PIPES, 5 mM EGTA, 1 mM MgCl 2, pH 7.2) for 10 \nmin at RT. After fixation, cells were gently washed three times with PBS and 0.1% Triton X -\n100 for 10 min. Following permeabilization, cells were blocked with buffer containing 3% goat \nserum and 0.1% Triton X -100 in PBS for 30 minutes and incubated with primary antibodies \nprepared in blocking buf fer for 2 h at RT. After three washes, cells were incubated with the \ncorresponding secondary antibodies prepared in the blocking buffer for 1 h at RT. After three \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n81 \n \nwashes, nuclei were stained using DAPI and imaged as described above. z -stacks were \nimported into ImageJ and somas and axons were manually segmented using β -tubulin III \nstaining. A colocalization mask of T OM20 and Cyt c channel was subtracted from the Cyt c \nchannel allowing the separation of pixel intensities occurring outside the mitochondria. These \nbinary images were used to quantify the area occupied by mitochondria without Cyt c. \n \nStatistical analysis:  \nNo formal statistical methods were used to predetermine sample sizes. Quantitative data are \nobtained from at least three independent biological replicates denoted by nexp, nrep indicates the \nnumber of technical replicates within each independent experiment, and n cells denotes the \nnumber of cells used for data analysis . Data are expressed as mean ± SD unless otherwise \nspecified. All confocal images are representative of at least three biological replicates. Where \napplicable, curve fitting was performed using linear or nonlinear regression functions. Tests of \nnormality were performed (Shapiro–Wilk test) to identify normal or non -normal populations. \nFor normally distributed data, a paired or unpaired, two -tailed Student’s t -test was used to \ncompare two groups. To compare more than two groups, multiple (paired or unpaired), two -\ntailed Student’s t -tests or a one -way or two -way analysis of variance (ANOV A) with \nappropriate multiple comparisons tests were used. Statistical differences were considered \nsignificant when the α -value of p was <0.05. Estimated p values are either stated as actual \nvalues or denoted by *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. For all experiments, \nthe analysis was not blinded. Microsoft Excel was used to store all raw data. All statistical \nanalyses were made using GraphPad Prism 10.4.1 (USA). \n \n \n \n \n \n \n \n \n \n \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n82 \n \nTable S1. EPG5 mutations (NM_020964.2) in patient-derived fibroblasts  \n \n \nTable S2. Upstream analysis of transcription regulators in patient fibroblasts  \nPatients  \nVariant 1                                              Variant 2 Source (Clinical \npresentation, \noutcome and \nFamily/patient ID) Nucleotide   amino acid    Exon       Nucleotide    amino acid    Exon \nPatient 1 \n(Pat1) \nc.1007A>G       p.Q336R            2 c.1007A>G       p.Q336R            2 \nCullup et al., 2013, \nPatient ID- 7.1 \nPatient 2 \n(Pat2) \nc.4862G>A       p.R1621G          28 c.4862G>A       p.R1621G          28 \nVansenne et al., \n2022, Patient ID- \n1.1 \nPatient 3 \n(Pat3) \nc.895C>T          p. R299*             2 c.5479C>G        p.P1827A          31 \nDafsari et al., \n2024, Patient ID-  \nPatient 4 \n(Pat4) c.4952+1G>A  p.F1604Gfs*20   28            c.4952+1G>A  p.F1604Gfs*20   28 \n  Byrne et al., \n2016, Patient ID- \n4.1 \nTranscription \nregulators \nPathway responsive genes \n(Han et al., 2018) \nZ-score adj. p-value \nPPARGC1A \nFASTK, CYP7B1, APOLD1, \nNFKBIA, PTGS2, PDK4, BCL2L1 \n1.41141849 0.030074675 \nATF4 \nBEX4, MTHFD1, MTHFD2, ATF3, \nIDH1, PHGHD, RGS2, PIM1, PSHP, \nEIF4EBP1, EIF1B, DNAJB9, HSPA2, \n1.457997196 0.031461968 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n83 \n \n \nTable S3: Key resources table \nHSP70, HSP90B1, SEC61A2, RELB, \nNFKB1, TNFSF10B, EIF2A \nRELA \nABCB4, BCL2, ABCA1, BIRC5, \nBCL2L1, CCL2, CXCL10, \nSERPINE2, SOD2, TRAF2, CFLAR,  \n1.859075413  0.028959004 \nSTAT1 \nCCL2, CXCL10, FGF2, IFIT3, IL6, \nIRF7 \n1.437937718  0.039748444 \nIRF3 CCL2, CXCL1O, ATF4 1.729626042 0.01774614 \nKLF5 \nBIRC5, CCNB1, CCND1, MMP3, \nMYH10, ITGB2,  \n1.566385318 0.026121499 \nMAF CCND2, GCLC, IL4, MMP13 1.811069918 0.022278662 \nFOSL1 MMP1, PLAUR, ITGB3, CLU 1.740310439 0.03584421 \nReagent or resource Source Identifier \nAntibodies \nRabbit pAb anti-EPG5 (1:1000) Abcam Cat# ab122186 \nMouse mAb anti-β-actin (1:10000) Cell Signaling Technology Cat# 3700  \nMouse Ab anti-OxPhos cocktail (1:1000) Invitrogen Cat# 45-8199 \nRabbit mAb anti-TOM20 (1:5000 IB, 1:200 IF) Abcam Cat# ab186735 \nMouse mAb anti-MCU (1:1000) Sigma-Aldrich Cat# AMAB91189 \nRabbit pAb anti-MCUB (1:1000) Proteintech Cat# 20387-1-AP \nRabbit pAb anti-MICU1 (1:1000) Thermo Scientific Cat# HPA037480 \nRabbit pAb anti-MICU2 (1:1000) Sigma-Aldrich Cat# HPA045511 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n84 \n \nRabbit pAb anti-MICU3 (1:1000) Thermo Scientific Cat# PA5-55177 \nRabbit pAb anti-EMRE (1:1000) Abcam Cat# ab157387 \nRabbit pAb anti-NCLX (1:1000) Sigma-Aldrich Cat# SAB2102181 \nMouse mAb anti-ATP5A (1:1000) Abcam  Cat# ab14748 \nRabbit pAb anti-phospho PDH E1α (Ser293) \n(1:1000) \nSigma-Aldrich Cat# AP1062 \nMouse mAb anti- PDH E1α (1:5000) Abcam Cat# ab110330 \nRabbit pAb anti-STAT1 (1:1000) Cell Signaling Technology Cat# 9172S \nRabbit mAb anti-phospho STAT1 (Tyr701) (1:1000) Cell Signaling Technology Cat# 9167S \nRabbit mAb anti-TBK1 (1:1000) Cell Signaling Technology Cat# 3504S \nRabbit mAb anti-phospho TBK1 (Ser172) (1:1000) Cell Signaling Technology Cat# 5483S \nRabbit mAb anti-STING (1:1000) Cell Signaling Technology Cat# 13647S \nRabbit pAb anti-IRF3 (1:1000) Abcam Cat# ab25950 \nRabbit mAb anti-phospho IRF3 (Ser396) (1:1000) Cell Signaling Technology Cat# 29047S \nHRP-Goat Anti-Rabbit IgG (H+L) Jackson ImmunoResearch Cat# 111-035-045 \nHRP-Goat Anti-Mouse IgG (H+L) Jackson ImmunoResearch Cat# 315-035-045 \nRabbit mAb anti-cGAS (1:100 IF) Cell Signaling Technology Cat# 15102S \nRabbit pAb anti-Citrate synthetase (1:100 IF) Abcam Cat# ab96600 \nMouse mAb anti-DNA (1:200 IF) Sigma-Aldrich Cat# CBL186 \nMouse mAb anti-Cytochrome C (1:200 IF) BD Pharmingen Cat# 556432 \nMouse mAb anti-Beta-Tubulin III (TUJ1) (1:100 IF) STEMCELL Technologies Cat# 60052 \nRabbit pAb anti-GFAP (1:100 IF) Sigma-Aldrich Cat# AB5804 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n85 \n \nChicken pAb anti-MAP2 (1:100 IF) Abcam Cat# ab92434 \nAlexa Fluor 647 Donkey anti-Rabbit IgG (H+L) Thermo Scientific Cat# A-31573 \nAlexa Fluor 488 Donkey anti-Mouse IgG (H+L)  Thermo Scientific Cat# A-21202 \nAlexa Fluor 568 Donkey anti-Rabbit IgG (H+L)  Thermo Scientific Cat# A-10042 \nAlexa Fluor 488 Goat anti-Rabbit IgG (H+L)  Thermo Scientific Cat# A-11008 \nAlexa Fluor 568 Goat anti-Mouse IgG (H+L) Thermo Scientific Cat# A-11004 \nAlexa Fluor 647 Goat anti-Chicken IgG (H+L)  Thermo Scientific Cat# A-21449 \nCommercial Kit Assays \nMycoAlert Mycoplasma Detection Kit  Lonza Cat# LT07-118 \nSeahorse XF Cell Mito Stress Test Kit Agilent Cat# 103015-100 \nSeahorse XFe96/XF Pro FluxPak Agilent Cat# 103793-100 \nHuman Dermal Fibroblasts Nucleofector Kit Lonza Cat# VPI-1002 \nNativePAGE Novex Bis-Tris Gel System Thermo Scientific Cat# BN2007, BN2008 \nExperimental Models: Cell Lines \nControl Fibroblasts This paper  \nPMID: 24336167 and \nMRC CNMD Biobank \nLondon \nPatient Fibroblasts This paper  Table S1 \nIsogenic iPSCs Coriell Institute RRID: GM28930 \nQ336R iPSCs Coriell Institute \nRRID: GM27291 and \nPMID: 35700637  \nCritical chemicals and reagents \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n86 \n \nJP1-138 \nValeria Pingitore et al., \n2024 \nPMID: 38985859 \nMito-Fura-2 AM (v2.3) Pendin et al., 2019 PMID: 31132197 \nPlasmid and AAV vector \npTRIP-CMV-mTagBFP2-2A-FLAG-ntcGAS \nAddgene, (Gentili et al., \n2015) \nCat# 102603 \nAde-mtaequorin \nGift from Rosario Rizzuto \nlab, (Tosatto et al., 2016) \nPMID:  27138568 \nSoftware and algorithms \nFIJI (ImageJ); Version 1.54p \nhttp://fiji.sc, (Schindelin et \nal., 2012)  \nRRID: SCR_002285 \nPrism 10  GraphPad RRID: SCR_002798 \nMitoSegNet \nhttps://github.com/MitoSeg\nNet, (Fischer et al., 2020) \nPMID:  33083756 \nImage Lab software; Version 6.1 Bio-Rad  RRID: SCR_014210 \nMicroscopy Image Browser; Version 2.81 \nhttps://mib.helsinki.fi/, \n(Belevich and Jokitalo, \n2021)  \nRRID:SCR_016560 \nImaris Version 9.8 Bitplane RRID:SCR_007370 \nMetaMorph Microscopy Automation and Image \nAnalysis Software \nMolecular Devices RRID:SCR_002368 \nMetaFluor Fluorescence Ratio Imaging Software Molecular Devices RRID:SCR_014294 \nZEISS ZEN Microscopy Software Carl Zeiss RRID:SCR_013672 \nSeahorse Wave Agilent RRID:SCR_014526 \nR Project http://www.r-project.org/ RRID:SCR_001905 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n87 \n \nTable S4. Primer pairs used in this study. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nTarget Forward Reverse Source \nmitochondria CACCCAAGAACAGGGTTTGT TGGCCATGGGTATGTTGTTA (Rooney et \nal., 2015) nucleus TGCTGTCTCCATGTTTGATGTATCT TCTCTGCTCCCCACCTCTAAGT \nACTB CACCATTGGCAATGAGCGGTTC AGGTCTTTGCGGATGTCCACGT \nThermo \nScientific \nEPG5 CCTTCTGTATCTTCACCGTCCG GAAGTCAGCCACCTCGGTCAAA \nNCLX ATGGTGGCTGTGTTCCTGACCT GGTGCAGAGAATCACAGTGACC \nMCU CAGCACTGTTGTGCCCTCTGAT  GGCTTGAGTGTGAACTGACAGC \nMICU1 GACAGTGGCTAAAGTGGAGCTC  CCTCTCATCAGCCGTTGCTTCA \nMICU2 GGATGGCAGTTTTACAGTCTCCG  GAAGAGGAAGTCTCGTGGTGTC \nMICU3 CACTGATGGCAATGAGATGGTGG  GACGCAGCATTGCACGCTTTTC \nEMRE GCCTAGCTTGAGGAAAGATGGC  ATGGAGAACACACGCAGAAGGC \nIL1B CCACAGACCTTCCAGGAGAATG  GTGCAGTTCAGTGATCGTACAGG \nIL6 AGACAGCCACTCACCTCTTCAG  TTCTGCCAGTGCCTCTTTGCTG \nTNF CTCTTCTGCCTGCTGCACTTTG  ATGGGCTACAGGCTTGTCACTC \nCCL2 AGAATCACCAGCAGCAAGTGTCC  TCCTGAACCCACTTCTGCTTGG \nCXCL10 GGTGAGAAGAGATGTCTGAATCC  GTCCATCCTTGGAAGCACTGCA \nIFIT3 CCTGGAATGCTTACGGCAAGCT GAGCATCTGAGAGTCTGCCCAA \nISG15 CTCTGAGCATCCTGGTGAGGAA  AAGGTCAGCCAGAACAGGTCGT \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n88 \n \nResource availability \nLead Contact  \nFurther information and requests for resources and reagents should be directed to and will be \nfulfilled by the Lead Contact, Michael R Duchen (m.duchen@ucl.ac.uk). \n \nMaterials Availability \nThis study did not generate new unique reagents. \n \nData and Code Availability \nRaw FastQ files for RNA -sequencing analyses will be deposited to GEO upon publication. \nGene-set Analyses using biological processes and the KEGG pathway analysis of the RNA -\nsequencing experiment as well as the unprocessed blot images deposited at Mendeley Data \nupon publication. This paper does not report original code. \n \nAcknowledgements  \nWe would like to thank the Coriell Institute for Medical Research (Camden, NJ, USA) for their \nkind gift of patient fibroblasts with EPG5 mutations. We are thankful for the valuable \ndiscussions and support from all the members of M. Duchen, G. Szabadkai and M. Fanto labs. \nWe acknowledge The MRC Centre for Neuromuscular Diseases Biobank (supported by the \nNational Institute for Health Research Biomedical Research Centres at Great Ormond Street \nHospital for Children, NHS Foundation Trust) for providing all age and sex-matched healthy \ncontrol fibroblasts used in this study. We also acknowledge UCL Genomics Core Facility for \nthe RNA -sequencing studies. T his work was supported by grants from Action Medical \nResearch (GN2959  to K .S. and M .R.D.) and Great Ormond Street Hospital for Children  \n(V4218 to M.R.D.).  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n89 \n \nAuthor contributions: \nK.S. and M.R.D.- Conceptualization, funding acquisition, investigation, visualization, \nmethodology, writing–original draft, project administration, writing–review and editing. H.J., \nH.S.D. and M.F.- Conceptualization, resources, investigation, formal analysis, supervision, \nwriting–review and editing. O.G. and H.C. - Formal analysis, methodology, investigation and \ndata curation. I.M., I.K., P.S. and C.Y.C.- Formal analysis, methodology and inves tigation, \nwriting-review and editing. V.P., D.L.S., F.V . and D.P.- Resources, writing -review and \nediting. G.S.- Conceptualization, writing-review and editing.  \n \nSupplementary references \n \n1. Belevich, I., and Jokitalo, E. (2021). DeepMIB: User-friendly and open-source software for training \nof deep learning network for biological image segmentation. PLoS Comput Biol 17, e1008374. \n \n2. Bhosale, G., and Duchen, M.R. (2019). Investigating the Mitochondrial Permeability Transition Pore \nin Disease Phenotypes and Drug Screening. Curr Protoc Pharmacol 85, e59. \n \n3. Bonora, M., Giorgi, C., Bononi, A., Marchi, S., Patergnani, S., Rimessi, A., Rizzuto, R., and Pinton, \nP. (2013). Subcellular calcium measurements in mammalian cells using jellyfish photoprotein aequorin-\nbased probes. Nat Protoc 8, 2105-2118. \n \n4. Chung, C.Y., and Duchen, M.R. (2022). A Plate Reader -Based Measurement of the Cellular ROS  \nProduction Using Dihydroethidium and MitoSOX. Methods Mol Biol 2497, 333-337. \n \n5. Fischer, C.A., Besora-Casals, L., Rolland, S.G., Haeussler, S., Singh, K., Duchen, M., Conradt, B., \nand Marr, C. (2020). MitoSegNet: Easy -to-use Deep Learning Segmentation f or Analyzing \nMitochondrial Morphology. iScience 23, 101601. \n \n6. Gentili, M., Kowal, J., Tkach, M., Satoh, T., Lahaye, X., Conrad, C., Boyron, M., Lombard, B., \nDurand, S., Kroemer, G. , et al.  (2015). Transmission of innate immune signaling by packaging of \ncGAMP in viral particles. Science 349, 1232-1236. \n \n7. Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion \nfor RNA-seq data with DESeq2. Genome Biol 15, 550. \n \n8. Metsalu, T., and Vilo, J. (2015). ClustVis: a web t ool for visualizing clustering of multivariate data \nusing Principal Component Analysis and heatmap. Nucleic Acids Res 43, W566-570. \n \n9. Mitchell, M.W., Grandizio, C., Turan, N., and Requesens, D.V. (2022). An induced pluripotent stem \ncell line (CIMRi001 -A) from a Vici syndrome donor with a homozygous recessive c.1007A>G \n(p.Q336R) mutation in the EPG5 gene. Stem Cell Res 63, 102833. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint \n\n90 \n \n10. Pendin, D., Norante, R., De Nadai, A., Gherardi, G., Vajente, N., Basso, E., Kaludercic, N., \nMammucari, C., Paradisi, C., Pozzan, T., et al. (2019). A Synthetic Fluorescent Mitochondria-Targeted \nSensor for Ratiometric Imaging of Calcium in Live Cells. Angew Chem Int Ed Engl 58, 9917-9922. \n \n11. Rooney, J.P., Ryde, I.T., Sanders, L.H., Howlett, E.H., Colton, M.D., Germ, K.E., M ayer, G.D., \nGreenamyre, J.T., and Meyer, J.N. (2015). PCR based determination of mitochondrial DNA copy \nnumber in multiple species. Methods Mol Biol 1241, 23-38. \n \n12. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., P reibisch, S., \nRueden, C., Saalfeld, S., Schmid, B. , et al. (2012). Fiji: an open-source platform for biological-image \nanalysis. Nat Methods 9, 676-682. \n \n13. Shi, Y., Kirwan, P., and Livesey, F.J. (2012). Directed differentiation of human pluripotent stem \ncells to cerebral cortex neurons and neural networks. Nat Protoc 7, 1836-1846. \n \n14. Singh, K., and Duchen, M.R. (2022). Analysis of Organization and Activity of Mitochondrial \nRespiratory Chain Complexes in Primary Fibroblasts Using Blue Native PAGE. Methods  Mol Biol \n2497, 339-348. \n \n15. Tosatto, A., Sommaggio, R., Kummerow, C., Bentham, R.B., Blacker, T.S., Berecz, T., Duchen, \nM.R., Rosato, A., Bogeski, I., Szabadkai, G. , et al.  (2016). The mitochondrial calcium uniporter \nregulates breast cancer progression via HIF-1alpha. EMBO Mol Med 8, 569-585. \n \n16. Yu, G., Wang, L.G., Han, Y., and He, Q.Y. (2012). clusterProfiler: an R package for comparing \nbiological themes among gene clusters. OMICS 16, 284-287. \n \n                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted May 28, 2025. ; https://doi.org/10.1101/2025.05.23.655806doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}