An intrinsic mechanism of metabolic tuning promotes cardiac resilience to stress

Co-immunoprecipitation experiments coupled with mass spectrometry led us to identify the interaction of the muscle-specific chaperone Melusin with the α and β subunits of the mitochondrial trifunctional protein (α-MTP and β-MTP) (Fig. EV1A,B). Melusin, as MTP subunits, was found as a component of the mitochondrial matrix in fractionation assays (Fig. EV1C,D), present both in subsarcolemmal and intermyofibrillar mitochondria (Fig. EV1E).

This interaction suggested a role for Melusin in the assembly of MTP multimeric complexes. Gel filtration experiments on mouse cardiac extracts showed Melusin in the same fractions containing both MTP α and β subunits, but did not show any changes in the multimeric organization of the MTP holoenzyme among Melusin-null (Brancaccio et al, 2003), Melusin overexpressing (De Acetis et al, 2005) and wild-type hearts (Fig. EV1F). Thus, Melusin is not required for MTP assembly.

We considered that Melusin might interact with the MTP to promote its catalytic activity and, ultimately, cardiac FAO. In contrast to this hypothesis, α-MTP activity was significantly higher in Melusin-null hearts (Fig. 1A) and lower in Melusin over-expressing hearts (Fig. EV2A) than in wild-type controls, indicating that Melusin unexpectedly limits FAO. To consolidate this finding, we evaluated the degree of palmitate oxidation in cardiac isolated mitochondria from mice either lacking or over-expressing Melusin. Palmitate oxidation capacity in Melusin-null mitochondria was ~35% higher than in wild-type controls (Fig. 1B), supporting our evidence that Melusin inhibits FAO and that the loss of Melusin leads to increased LCFA β-oxidation. Moreover, in mitochondria from hearts over-expressing ~20-fold higher Melusin (De Acetis et al, 2005), palmitate oxidation capacity was found ~55% lower than wild-type controls (Fig. 1B). The increased palmitate oxidation capacity detected in Melusin-null mitochondria was reversed by treatment with the MTP inhibitor trimetazidine (TMZ) (Fig. 1C,D), thus indicating that TMZ acts as an analog of Melusin on MTP.

Melusin levels did not affect the expression of MTP α and β subunits, as assessed by both qRT-PCR (Fig. EV2B) and western blot analyses (Fig. EV2C). The amount of the main lipid families (Fig. 1E; Table EV1) and the level of malonyl-CoA (Fig. EV2D), an important regulator of mitochondrial uptake of fatty acids, were comparable in the hearts of different genotypes, ruling out the possibility of a modification in lipid composition due to uncoupling between LCFA uptake and utilization (Yamamoto and Sano, 2022). We did not detect any difference in key molecules involved in glucose metabolism, such as Glucose transporter 1 and 4 (Glut1 and Glut4) mRNAs (Fig. EV2E) or pyruvate kinase and phosphorylated pyruvate dehydrogenase protein levels (Fig. EV2F). The preserved NADH/NAD⁺ ratio also excluded an imbalance in the fueling of the mitochondrial respiratory chain (Fig. EV2G). Overall, these results indicate that Melusin affects FAO directly through its binding to MTP.

To assess the relevance of this finding in human cells, we generated Melusin-null and over-expressing human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) (Figs. 1F and 2H–L). A proximity ligation assay (PLA) confirmed the interaction of Melusin and MTP also in these cells. (Fig. 1G). We determined a ~4.9-fold increase in palmitate β-oxidation capacity and a ~1.5-fold decrease in glucose flux through tricarboxylic acid (TCA) cycle in hiPSC-CMs compared to hiPSCs (Fig. 1H,I), in line with the expected change in metabolism between pluripotent and differentiated cells.

Melusin-null hiPSC-CMs boosted palmitate β-oxidation capacity by ~50% compared to wild-type controls, and this was matched by a ~3-fold reduction in glucose flux through the TCA cycle (Fig. 1H,I), possibly determined by the Randle cycle regulatory mechanism. These findings were observed in two separate Melusin-null hiPSC clones (Mel Null cl1 and cl2), ruling out the possibility of artefactual differences due to clonal variability. We further demonstrated that the higher palmitate oxidative capacity of Melusin-null cells reflects a real increase in palmitate utilization, by detecting a higher oxygen consumption rate in Melusin-null hiPSC-CMs than wild-type hiPSC-CMs when cultured in a media added with palmitate (Fig. 1J).

Melusin overexpression in hiPSC-CMs, by means of AAV vectors serotypes 6 and 9, led to dose-dependent increase in glucose processing by the TCA cycle and decrease in palmitate oxidative capacity (Fig. 1H,I). Overall, these experiments indicate that the MTP inhibitory function of Melusin in cardiomyocytes is conserved between mouse and humans.

To further probe the activity of Melusin in other contexts, we also forced its expression in the human fibro-cardiac cell line AC16, a cell line that normally do not express this protein (Fig. 1K). PLA experiments demonstrated Melusin interaction with MTP also in these cells (Fig. EV2M). We found that Melusin expression in the AC16 cell line reduces the palmitate oxidation capacity and increases the glucose flux through the TCA cycle (Fig. 1L). In addition, the expression of Melusin decreased the real-time oxygen consumption rate when AC16 were cultured in media containing palmitate (Fig. 1M). Overall, these findings indicate that Melusin acts as a general inhibitor of MTP, even when expressed in non-contractile cells.

Increased levels of FAO induce the production of more ROS by direct and indirect mechanisms (Battault et al, 2020; Isei et al, 2022; Mylonas et al, 2023; Speijer et al, 2014; St-Pierre et al, 2002), thus, in line with Melusin ability to inhibit FAO, we detected higher levels of ROS in Melusin-null hearts and lower levels in Melusin over-expressing hearts than in the respective controls (Fig. 2A). Specifically, ROS were 4-times higher in mitochondria from Melusin-null hearts than wild-type controls and almost three times lower in mitochondria from Melusin over-expressing cardiac muscles than their matched wild-type organs (Fig. 2A). Similar results were obtained by comparing the amount of ROS in subsarcolemmal and intermyofibrillar mitochondria from hearts of different genotypes (Fig. 2B). This regulation did not depend on differences in lipid peroxidation (Fig. EV3A), on the NADPH oxidative state (Fig. EV3B) and the level of NADPH oxidase 4 (Fig. EV3C), neither on relevant alterations in the expression of key antioxidant enzymes (Fig. EV3D).

Moreover, treatment of Melusin-null mice with the MTP inhibitor TMZ restored ROS amounts to the wild-type levels (Figs. 1C and 2C), thus confirming that Melusin limits ROS generation by tuning MTP activity and FAO.

In further agreement, we found that a high-fat diet, boosting FAO, leads to a ~1.6-fold increase in palmitate oxidative capacity (Fig. 2D) and ~4-fold elevation in ROS production (Fig. 2E) in murine heart mitochondria. However, this induction of FAO and ROS generation did not occur when Melusin is over-expressed in the heart (Fig. 2D,E).

Next, we tested whether Melusin-mediated FAO repression, and the consequent reduction of ROS, could be confirmed in human cardiomyocytes. In line with murine models, ROS were ~2-fold higher in Melusin-null hiPSC-CMs (Fig. 2F), and, conversely, ~3-fold lower in Melusin over-expressing hiPSC-CMs (Fig. 2G) than controls. Results in hiPSC-CMs were also confirmed in AC16 cells expressing Melusin where ROS levels were reduced by ~3.7-fold (Fig. 2H). We detected an analogous modulation by analyzing specifically the level of superoxide anion through MitoSOX fluorescence in living wild-type and Melusin-null hiPSC-CMs (Fig. 2I) and in AC16 expressing or not Melusin (Fig. 2J). All these data support the idea that Melusin regulates the generation of mitochondrial ROS.

A functional analysis of mitochondria isolated from mouse hearts showed that the electron flux through the respiratory chain correlates with the rate of FAO, being >60% higher in Melusin-null mice (Fig. 2K). ATP levels are similarly anti-correlated with Melusin expression (Fig. 2L). These differences were detected both in subsarcolemmal and intermyofibrillar mitochondria (Fig. 2M). This evidence was not explained by the different abundance of mitochondria, as indicated by comparable amounts of the mitochondrial proteins Vdac1 and Tom20 in cardiac protein lysates (Fig. EV2E). Similarly, the expression of the five mitochondrial respiratory chain complexes was equivalent in all genotypes (Fig. EV3E). Of note, no significant differences were detected in measuring the ATP content in the entire heart tissues, suggesting that cytoplasmic processes may mostly compensate for the differential ATP production in mitochondria (Fig. EV3F), in line with the healthy phenotype of Melusin overexpressing mice.

The physiologic restriction of FAO mediated by Melusin did not block the increase in lipid metabolism associated with the higher energetic demand induced by β-adrenergic stimulation (Banke et al, 2010). Indeed, simulating intense exercise by treating the isolated perfused heart for 10 min with a mild concentration of isoproterenol (Fig. EV3G), we detected an increase in FAO in wild-type hearts (Fig. EV3H), while no further increase was observed in Melusin-null hearts (Fig. EV3H). To note, the 20-fold overexpression of Melusin blocks the increase in FAO induced by mild β-adrenergic stimulation (Fig. EV3H). This suggests that, in the absence of Melusin, cardiomyocytes metabolize lipids to their highest potential, while physiological levels of Melusin tune this process, restraining ROS production. The amount of ROS observed in isoproterenol-treated hearts of different genotypes (Fig. EV3I) likely depends on the basal level of ROS in these hearts added to ROS generated by isoproterenol per se (Dey et al, 2018).

Overall, these data demonstrate that by inhibiting MTP-dependent FAO, Melusin limits the over-production of ROS and regulates mitochondrial function.

We previously highlighted how Melusin overexpression protects the heart from pressure overload induced by transverse aortic constriction (TAC) (De Acetis et al, 2005), while Melusin-null mice develop left ventricle dilation and heart failure in 4 weeks (Brancaccio et al, 2003). To evaluate whether the ability of Melusin to regulate metabolism plays a role in TAC protection, we subjected Melusin-null and over-expressing mice, together with wild-type controls, to 4-day TAC or sham surgery (Fig. 3A). This early time point was chosen since it precedes the TAC-induced cardiac remodeling (Brancaccio et al, 2003; De Acetis et al, 2005). At 4 days after TAC, FAO mildly (~37%) decreased in wild-type mice, in line with the metabolic rewiring occurring in the heart during stress. Unexpectedly, FAO appeared potently reduced in Melusin-null hearts (Fig. 3B), apparently in contrast with the role of Melusin in limiting MTP activity. We posited that this result might be due to a mitochondrial dysfunction triggered by the ROS increase induced by ROS generated by pressure overload per se (Suetomi et al, 2018; Wang et al, 2018), added to the basal ROS level derived from FAO. Indeed, we detected an abnormally high ROS production (Fig. 3C). In contrast, FAO rate and ROS levels in Melusin over-expressing mice did not vary in response to pressure overload (Fig. 3B,C). ROS levels did not depend on differences in the level of NADPH oxidase 4 among genotypes (Fig. EV4A).

In line with these observations, the mitochondrial electron flux, reduced by ~2.7 and ~2.4-fold, respectively in wild-type and Melusin over-expressing hearts (Fig. 3D), is ~8.6-fold reduced in Melusin-null hearts (Fig. 3D) in response to TAC. We further detected no changes in mitochondrial ATP in all genotypes, except for Melusin-null hearts where a ~7.5-fold drop was observed (Fig. 3E). These alterations occurred in the absence of significant variations in MTP expression (Fig. EV4A) and differences in mitochondrial content (Fig. EV4B,C) or structure (Fig. EV4D,E). In line, we measured the ATP in the cardiac total homogenate of all the mouse genotypes and found a 3-fold drop in total ATP content in Melusin-null hearts subjected to TAC (Fig. EV4F).

Overall, these results show that, at early stages during pressure overload, Melusin protects the heart from the ROS over-production and the consequent drop of the mitochondrial function.

The rise of cardiac ROS production is thought to occur in response to doxorubicin, a commonly used and effective antineoplastic drug with cardiotoxic side effects (Berthiaume and Wallace, 2007; Brancaccio et al, 2020). We thus tested whether Melusin, by controlling FAO and ROS production, could affect the cardiac response to doxorubicin.

First, we analyzed the effect of doxorubicin on FAO and ROS production in acute, 6 h after doxorubicin treatment (Fig. 3F). At this time point, FAO decreased in wild-type and Melusin overexpressing mice (~40% and 50%, respectively) (Fig. 3G), while, despite its higher starting point, drastically dropped (~88%) in Melusin-null hearts (Fig. 3G), similarly to the TAC model.

On the other hand, total ROS content was significantly increased by the doxorubicin treatment itself both in wild-type and Melusin-null mice proportionally to their respective basal states, with Melusin-null mice showing consistently higher ROS than wild-type controls (Fig. 3H). Conversely, Melusin over-expressing hearts showed lower ROS levels in basal conditions that did not increase after doxorubicin treatment (Fig. 3H). The high level of ROS in doxorubicin-treated Melusin-null mice can explain the mild significant increase in the level of NADPH oxidase 4 in these mice, while no differences were detected in treated Melusin over-expressing mice compared to wild-type mice (Fig. EV5A).

We further analyzed the mitochondrial function after 6 h from doxorubicin administration, detecting alterations like those observed in pressure overload. Indeed, in Melusin-null mice the electron flux rate was reduced by ~71% after doxorubicin treatment, while no changes were detected in wild-type and Melusin over-expressing mice (Fig. 3I). Mitochondrial ATP levels showed a ~96% drop in Melusin-null mice, in accordance with the decrease in FAO, while no significant differences were detected in wild-type and over-expressing mice (Fig. 3J). Similarly to TAC, this finding did not correlate with detectable changes in the expression of MTP (Fig. EV5A), mitochondrial content (Fig. EV5B,C) or structure (Fig. EV5D,E), and mitochondrial integrity (Fig EV5F). By measuring the total amount of ATP in heart tissues, we detected a decrease of ~60% in Melusin-null mice in response to doxorubicin treatment (Fig. EV5G).

We then treated with doxorubicin AC16 cells wild-type and expressing Melusin and hiPSC-CMs wild-type and knockout for Melusin (Fig. 4A), showing that Melusin protects cells from the drop of FAO and the rise of ROS (Fig. 4B,C), in line with data obtained in mice. In particular, Melusin absence in AC16 and hiPSC-CMs correlates with a lower number of mitochondria with normal membrane potential after doxorubicin treatment (Fig. 4D,E), indicating that Melusin protects cardiomyocytes from mitochondrial dysfunction.

Finally, given the protective effect of Melusin overexpression after an acute doxorubicin challenge, we studied whether increased Melusin abundance could mitigate cardiotoxicity induced by a repeated doxorubicin treatment (Fig. 4F). We found that Melusin overexpression leads to preserved fractional shortening (Fig. 4G) and ejection fraction (Fig. 4H), as well as reduced cumulative mortality (Fig. 4I).

These results demonstrate that in response to doxorubicin treatment, similarly to what we observed after pressure overload, the modulatory role of Melusin on FAO preserves the heart from an excessive ROS generation and from mitochondrial dysfunction.

The sharp decline in FAO observed in Melusin-null mitochondria after challenges, like TAC and doxorubicin, suggested that the higher basal levels of ROS present in these hearts in steady state, boosted by the ROS induced by the stress itself, induce mitochondrial dysfunction and the consequent drop of FAO. Therefore, we tested if keeping FAO low before stress, reduces the ROS levels generated after the cardiac challenge. We treated Melusin-null mice with the MTP inhibitor trimetazidine (TMZ) before either TAC (Fig. 5A,B) or doxorubicin administration (Fig. 5D,E) and we revealed that, after both challenges, pre-treatment with TMZ potently reduced ROS in Melusin-null hearts (Fig. 5C, F). Next, we tested if reducing ROS with the MitoQ scavenger concomitantly to the challenge could mitigate the mitochondrial dysfunction and the observed drop of FAO (Fig. 5G,J). In line with ROS being critical in this event, MitoQ blocked the raise of ROS levels (Fig. 5H,K) and dampen the severe decline of FAO seen in Melusin-null hearts (Fig. 5I,L).

These results indicate that inhibiting MTP and lowering FAO reduce the burden of ROS accumulation during stress-limiting mitochondrial dysfunction.

The reduction of FAO timely during stress might improve overall cardiac resilience. To test this hypothesis, we treated wild-type mice with TMZ concomitantly to pressure overload as well as doxorubicin administration. TMZ dissolved in drinking water (Fig. 6A) preserved fractional shortening and ejection fraction for 28 weeks after TAC (Fig. 6B,C). Similarly, the oral administration of TMZ concomitantly with three doses of doxorubicin (Fig. 6D) protected hearts from the decrease in fractional shortening and ejection fraction seen in vehicle-treated mice (Fig. 6E,F). In conclusion, these experiments demonstrate that approaches aimed at mitigating lipid oxidation concomitantly to challenge improves the ability of the heart to cope with stress (Fig. 7).

Mice received humane care in compliance with Italian law (DL-116, 27 January 1992) and in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). The mice were housed under controlled temperature (20–24 °C) and humidity (50–60%) with a regular 12/12-h light/dark cycle (7 AM–7 PM). They were provided with ad libitum access to food and water and were daily monitored for health status by trained personnel. Male mice 129/Sv wild-type and null for Melusin (Brancaccio et al, 2003) and male mice FVB wild-type and transgenic for human Melusin (De Acetis et al, 2005) of 8–13 weeks of age were used in conformity of the authorizations n. 856/2017-PR 03/11/2017 of the Italian Health Ministry and the guidelines of the University of Turin.

Mice were randomly grouped into treated and control groups. Blinding was not possible because of institutional requirements for labeling the transgenic mouse models and treatments given. When possible, experiments were blinded if conducted by collaborators or other group members.

Male healthy donor-derived WTC11 human induced pluripotent stem cells (hiPSCs) (Coriell Institute for Medical Research) were previously generated by Bruce Conklin’s laboratory at the Gladstones Institutes in San Francisco, USA (Kreitzer et al, 2013). hiPSCs were cultured on Geltrex (Life technologies)-coated tissue culture plates in Essential 8 medium (E8; Gibco) with 0.4% penicillin/streptomycin (Gibco). Cells were grown at sub-confluence, passaged using PBS + EDTA 0.5 mM, and replated in E8 with Thiazovivin 2 μM (Cayman Chemical Company). Melusin-null hiPSCs were generated by CRISPR/Cas9-mediated hemizygous deletion of the ITGB1BP2 promoter, as better discussed below.

hiPSC-derived cardiomyocytes (hiPSC-CMs) were differentiated from hiPSCs, using an established biphasic WNT modulation protocol (Bertero et al, 2019). Briefly, on day -1, subconfluent hiPSCs were primed for differentiation with 1 μM Chiron (Cayman Chemical Company) in E8. On day 0, cells mesoderm induction was achieved by adding RBA medium [RPMI 1640 (Gibco) + BSA (Sigma-Aldrich) 0.5 mg/mL + l-ascorbic acid 2-phosphate (Fujifilm) 213 ug/mL] supplemented with 4 μM Chiron. On day 2, cardiac mesoderm specification was achieved with RBA supplemented with 2 μM WNT-C59 (Cayman Chemical Company). On day 4, cells were fed with RBA alone. From day 6, cells were cultured with RPMI-B27 medium [RPMI 1640 supplemented with B27 (Thermo Fisher Scientific) and 0.4% penicillin/streptomycin]. Beating cardiomyocytes were observed from day 8. On day 16, hiPSC-CMs were dissociated and plated on Geltrex-coated dishes at a density of 3 × 10⁵ cells/cm². Cells were lactate-selected for 4 days by feeding with DMEM no glucose/sodium pyruvate (Gibco) with 4 mM sodium l-lactate (Thermo Fisher Scientific). Day 25–27 hiPSC-CMs were harvested for experimental analysis and flow cytometry. Melusin overexpressing hiPSC-CMs were obtained by transduction with AAV6 and AAV9 encoding the ITGB1BP2 cDNA with a myc tag, while a GFP or empty AAV was used as control.

hiPSCs and hiPSC-CMs were tested every 2 weeks for mycoplasma contamination by PCR.

The human cardiomyocyte cell line (AC16) (#SCC109, Millipore) was cultured in DMEM medium (Gibco) containing 12% FBS (Gibco) and 1% penicillin/streptomycin (Thermo Fisher Scientific). AC16 cells expressing Melusin were obtained by infection with a pLVX lentiviral vector encoding the human ITGB1BP2 cDNA with a myc-tag. An empty pLVX lentiviral vector was used as a control. Differentiation was favored using the mitogen-free medium, as suggested by the manufacturer. AC16 cells were tested every 2 weeks for mycoplasma contamination by PCR.

We generated the CRISPR/Cas9 plasmids needed to knockout Melusin in hiPSCs (pX459-V2.0_ITGB1BP2_-302 and pX459-V2.0_ITGB1BP2_ + 135) according to an established procedure. Briefly, single guide RNAs (sgRNAs) were constructed by cloning 20-mer crispr RNAs (crRNAs) as double-stranded oligonucleotides in pSpCas9 (BB)-2A-Puro (PX459) V2.0 (a gift from Feng Zhang; Addgene plasmid #62988). Two crRNAs were cloned, with PAM sites mapping 302 bp upstream and 135 bp downstream of the transcriptional start site (TSS) of ITGB1BP2: 5′- GCTTAATCCTCATCGAGCTG-3′ and 5′-GACCCTAATACCAACCTTCC-3′, respectively.

hiPSCs knockout for Melusin were obtained through CRISPR/Cas9-mediated deletion of the promoter and a portion of the first exon of the ITGB1BP2 gene, similar to what we previously described (Bertero et al, 2019). About 3 × 10⁵ WTC11 hiPSCs were seeded in a six-well Geltrex-coated dish in E8 supplemented with 2 μM Thiazovivin, and immediately transfected with 2 μg of sgRNA-containing plasmids using 6 μL of GeneJuice (Millipore). The procedure co-delivered 1 μg for each of two sgRNAs targeting the ITGB1BP2 promoter up- and down-stream: pX459-V2.0_ITGB1BP2_-302 and pX459-V2.0_ITGB1BP2_ + 135, respectively. After 16 h, cells were rinsed in PBS, and fed with E8 containing 2 μM Thiazovivin and 0.5 μg/mL puromycin dihydrochloride (Sigma-Aldrich) to enrich for transfected hiPSCs. After 24 h, the media was replaced with conventional E8 and changed daily. After 72 h, hiPSCs were dispersed to single cells using PBS 0.5 mM EDTA and seeded at a low density of 5 × 10³ cells per 100 mm culture dish in E8 with 2 μM Thiazovivin. After daily media changes with E8 for 7–10 days, pseudo-clonal lines were isolated by picking well-separated individual colonies and expanded for genotyping. Genomic DNA was extracted using the QIAGEN DNeasy Blood & Tissue Kit and used as a template for PCR with LongAmp Taq (NEB). ITGB1BP2 promoter deletion was assessed using primers located 591 bp and 220 up- and downstream to the ITGB1BP2 TSS (fw 5′-GCAGTAATGCAGCATCAACCT-3′ and rev 5′-GGGCAGGTAGTTTCTTTCCCC-3′). The resulting product was run on a 1% Agarose-TBE gel and detected by SYBR Safe staining, and two clones (clone 1 and clone 2) with hemizygous deletion of the ~0.4 kb ITGB1BP2 promoter fragment were isolated. Both Melusin knockout hiPSC lines maintained a normal karyotype throughout the procedure, as assessed by standard G-banding.

hiPSC-CMs were differentiated and lactate-selected. On day 29, cardiomyocytes were washed twice with PBS (w/Ca²⁺ and Mg²⁺) and incubated with the adeno-associated viral vectors (AAV vectors) diluted in RPMI-B27 at a multiplicity of infection (MOI) of 100,000. The AAV vectors used are the following: AAV6-Melusin and AAV9-Melusin and relative controls GFP or empty, depending on the assay. AAVs for Melusin contain the pAAV-MCS plasmid carrying the cDNA for Melusin, preceded by a myc tag. GFP AVVs contain the pAAV-MCS plasmids carrying the GFP cDNA. A high-density AAVs production was obtained in packaging cells AAV-293T (Agilent) (Vitadello et al, 2020). Forty-eight hours after the infection cells were washed twice with PBS (w/Ca²⁺ and Mg²⁺) and fed with fresh RPMI-B27. On days 34–35, cardiomyocytes were utilized for experimental assays.

Echocardiographic analyses were performed on mice treated with doxorubicin, and relative controls, to follow the cardiac parameters and on mice subjected to TAC, and relative controls, to measure the degree of pressure load and follow the cardiac parameters. Mice anesthetized with isoflurane were analyzed using a Vevo 2100 Imaging System device (Visualsonics, Inc.). Two-dimensional-guided M-Mode echocardiography from the parasternal long-axis view was used to measure cardiac parameters. Fractional shortening (FS %) and ejection fraction (EF %) were calculated according to standard formulas. The degree of narrowing of the constricted aorta was evaluated via pulse wave Doppler echography downstream the constriction. Only mice with a pressure gradient between 30 mmHg and 60 mmHg were used for analyses.

Wild-type and Mel null mice were sacrificed to collect their hearts. Hearts were homogenized in IP buffer (25 mM Tris-HCl pH 8, 1 mM MgCl₂, 10% glycerol) added with proteases and phosphatases inhibitors and centrifuged at 16,000 × g for 10 min. Supernatants were quantified, and 40 μg of total extract per sample were diluted in Laemmli buffer and stored as input. About 1 mg of total cardiac extract was incubated overnight at 4 °C with 30 μl of anti-Melusin antibody conjugated to CNBr–Sepharose beads (GE Healthcare). The following day, beads were washed in the IP buffer and resuspended in Laemmli buffer.

Mice wild-type, null or over-expressing Melusin were sacrificed and the hearts were excised and homogenized in isolation buffer (100 mM KCl, 50 mM Tris pH 7.4, 5 mM MgSO₄7H₂O, 1 mM EDTA, 1.8 mM ATP, 0.5% BSA pH 7.4) added with proteases and phosphatases inhibitors in a glass potter tissue grinder. Samples were clarified by centrifugation at 400 × g for 5 min at 4 °C and the first supernatant and the first pellet were collected separately.

The first supernatant was used to purify subsarcolemmal mitochondria (SS). This was centrifuged at 9000 × g for 10 min at 4 °C obtaining the second supernatant, collected as cytosolic fraction for fractionation experiments. The SS mitochondrial pellet was washed in washing buffer (100 mM KCl, 50 mM Tris pH 7.4, 5 mM MgSO₄7H₂O, 1 mM EDTA pH 7.4) and resuspended in mitochondrial buffer (0,25 M sucrose, 20 mM Hepes, 1 mM EDTA pH 7.4).

The first pellet was used to purify intermyofibrillar mitochondria (IMF). The pellet was exposed to 5 μg/g trypsin for 10 min in agitation. The pellet was then diluted in lysis buffer and centrifuged at 5000 × g for 5 min at 4 °C. The second supernatant was discarded, while the second pellet was resuspended in lysis buffer and centrifuged at 800 × g for 10 min at 4 °C twice. The supernatants were combined in a new tube and centrifuged at 9000 × g for 10 min at 4 °C to obtain the final IMF mitochondrial pellet resuspended in mitochondrial buffer (0.25 M sucrose, 20 mM Hepes, 1 mM EDTA pH 7.4).

For fractionation experiments, the SS and IMF final mitochondrial pellets were quickly incubated in 2 mg/ml Proteinase K, to digest residues of cytosol, washed in washing buffer and lysed for western blotting analyses.

Mitochondria were subjected to treatments with proteinase K, osmotic shock, and Triton X-100 (TX100) to progressively digest mitochondrial compartments. The SS final mitochondrial pellet was washed with washing buffer, resuspended in mitochondrial buffer (0.25 M sucrose, 20 mM Hepes, 1 mM EDTA pH 7.4), and divided in aliquots of 20–30 μg:

For all the PK treated mitochondria, proteins were precipitated in 10% final volume of TCA for 15 min, centrifuged at 18,000 × g for 20 min at 4 °C and resuspended in Laemmli buffer for western blot analysis.

Gel filtration chromatography was performed on a Superose-6 Increase 10/300 GL column (GE Healthcare) using the AKTA purifier system (GE Healthcare). Wild-type and Mel null mice were sacrificed, and the hearts were excised and homogenized in gel filtration buffer (25 mM Tris-HCl pH 8.0, 1 mM MgCl₂, 10% glycerol) added with proteases and phosphatases inhibitors. Extracts were centrifuged at 16,000×g for 20 min and diluted to 2 mg/ml. About 600 μl of extract was loaded into the machine and separated at a flow rate of 0.5 ml/min. Fractions of 0.5 ml were collected. Molecular mass standards (GE Healthcare) were used to calibrate the column: thyroglobulin (669 kDa), ferritin (440 kDa), catalase (240 kDa), and aldolase (158 kDa). The fractions were concentrated using Nanosep 3K and resuspended in Laemmli buffer for western blot analyses.

Fresh heart samples and mitochondrial samples were prepared as described above.

Frozen hearts were powdered and homogenized in TBS + 1% TX100 with proteases and phosphatase inhibitors.

Cells were washed from the medium and lysed in a plate with TBS + 1% TX100 with protease and phosphatase inhibitors.

Lysates were finally centrifuged twice for 10 min at 16,000 × g at 4 °C and ̴40 μg of total protein extract was resuspended in Laemmli buffer. Samples were boiled at 95 °C, separated by electrophoresis in reducing and denaturing polyacrylamide gels, and transferred onto nitrocellulose membranes. The loading was tested by Red Ponceau staining or Coomassie Blue staining. After saturation with TBS, 0.3% Tween 20 added with 5% BSA, blot strips were incubated overnight with primary antibodies and after with the proper secondary antibody. Signals were detected by the chemiluminescent reagent LiteAblot (Euroclone) at the Chemidoc Touch Imaging System (Bio-Rad). Band intensities were quantified using the Image Lab software (Bio-Rad).

Western blotting were performed using the following primary and secondary antibodies: anti-Melusin mouse monoclonal antibody (clone C3, 1:1000 from stock 1 µg/µl) generated in our Institute (Sbroggio et al, 2011), anti-vinculin (Sigma-Aldrich, SAB4200080, 1:1000), anti-HADHA (α-MTP) (Abcam, ab203114, 1:1000), anti-HADHB (β-MTP) (Santa Cruz Biotechnology, sc-365907, 1:1000), anti-VDAC1 (Abcam, ab15895, 1:1000), anti-actin (Santa Cruz Biotechnology, sc-1616-R, 1:1000), anti-tom20 (Santa Cruz Biotechnology, sc-11415, 1:1000), anti-total OXPHOS WB antibody cocktail (Abcam, ab110413, 1:1000), anti- pyruvate kinase PKM2 (Thermo Fisher Scientific, MA5-14890, 1:1000), anti-phospho-pyruvate dehydrogenase α1 (ser293) (Cell signaling, 31866S, 1:1000), anti- pyruvate dehydrogenase (Cell signaling, 3205, 1:1000), anti-catalase (Abcam, ab16731, 1:1000), anti-glutathione peroxidases 1 (Thermo Fisher Scientific, #PA5-26323, 1:1000), anti-glutathione peroxidases 4 (Santa Cruz Biotechnology, sc-166570, 1:1000), anti-peroxiredoxin 1 (Abcam, ab109498, 1:1000), anti-peroxiredoxin 3 (Abcam, ab222807, 1:1000), anti-superoxide dismutase 1 (Santa Cruz Biotechnology, sc-11407, 1:1000), anti-superoxide dismutase 2 (Santa Cruz Biotechnology, sc-133134, 1:1000), membrane integrity wb antibody cocktail (Abcam, ab110414, 1:1000), anti-Nox4 (Thermo Fisher Scientific, #14347-1-AP, 1:2000), anti-rabbit IgG-peroxidase antibody produced in goat (Sigma-Aldrich, #A6154, 1:5000) and anti-mouse IgG-peroxidase antibody produced in goat (Sigma-Aldrich, #A4416, 1:10,000).

The α-MTP activity was calculated as previously reported (Ljubkovic et al, 2019). About 20 mg of frozen heart powder was lysed in lysis buffer (5 mM Hepes, 1 mM EGTA, 1 mM DTT, 0.1% TX100, pH 8.7) and clarified by centrifugation 1 min at 1×g at 4 °C. Samples were quantified spectrophotometrically and 40 to 400 μg were aliquoted in Eppendorf. Samples were added with 1 ml of reaction buffer (100 mM Na-pyrophosphate, 1 mM EDTA, 0.15 mM NADH, 50 μM acetoacetyl-CoA) and 300 μl were aliquoted in a 96-well plate. The plate was read at λ = 340 nm for 160 cycles, and the V_max was calculated by linear regression on the exponential portion of the curve.

Palmitate oxidation capacity was evaluated in mitochondria isolated from cardiac tissues and cells using [1−¹⁴C]-palmitic acid (PerkinElmer). Hearts were homogenized in isolation buffer (Tris-HCl 50 mM, KCl 100 mM, MgCl₂ 5 mM, EDTA 1 mM, ATP 1.8 mM, pH 7.2) added with proteases and phosphatases inhibitors. For cells, the medium was removed, and cells were washed with PBS and gently detached by scraping in buffer added with proteases and phosphatase inhibitors. Samples were homogenized in a glass potter tissue grinder and centrifuged at 400 × g for 1 min at 4 °C to remove nuclei and other large membranous organelles. Supernatants were further centrifuged at 16,000 × g for 5 min at 4 °C. The pellet enriched in mitochondria was resuspended in 250 μl resuspension buffer (Sucrose 250 mM, K₂HPO₄ 15 mM, MgCl₂ 2 mM, EDTA 0.5 mM, pH 7.2) for the analysis. A small aliquot was used for protein quantification by BCA assay (Millipore). About 50 μl of samples was incubated with 100 μl of assay buffer (20 mM Hepes, 0.24 mM fatty acid-free BSA, 0.5 mM l-carnitine) and 2 μCi (380 μM) [1−¹⁴C]-palmitic acid (PerkinElmer) for 1 h in mild agitation at RT. About 100 μl of a solution 1:1 v/v phenylethylenamine 100 mM/methanol was added and incubated for 1 h further at room temperature. The reaction was stopped with 100 μl of HClO₄ 0.8 N and centrifuged at 16,000×g for 10 min. Supernatants (containing ¹⁴CO_2, used as internal negative control) and pellets (containing ¹⁴C-acid soluble metabolites (ASM)) were mixed with 4 ml of scintillation liquid (UltimaGold, PerkinElmer) and read by a β-counter. In each experimental condition, the amount of ¹⁴C-CO₂ was always ≤5% of ¹⁴C-ASM. Results were expressed as picomoles of ¹⁴C-ASM per hour normalized on mg of proteins.

Lipid composition analysis was performed extracting lipids in chloroform/methanol mixture and partitioned with water. The organic phase was dried and then suspended in chloroform/methanol for the analysis of total fatty acid and cholesterol contents. Total fatty acids were determined as methyl esters by gas chromatography (Agilent Technologies). The cholesterol analysis was achieved by high-pressure liquid chromatography HPLC (Jasco).

AC16 cells empty and mel were plated in cellview cell culture slide, glass bottom, (Greiner) at concentration of 15,000 cells/well. For hiPSC-CMs, cellview cell culture slide was coated with poly-l lysine and ES-qualified Geltrex (0.17 mg/ml). Mel null hiPSC-CMs were plated at a concentration of 30,000 cells/well and infected with 30,000 MOI AAV6-empty and AAV6-MycMel as described above. hiPSC-CMs were used 7 days after the infection. Cells were fixed with 4% PFA for 10 min at RT, permeabilized with 0.1% TX100 in PBS, and washed in TBS + 10% Tween 20 (TBST). Cells were blocked with PBS + 3% BSA + 1% goat serum for 1 h and incubated 1 h at RT in a humidity chamber with a solution of rabbit anti-HADHA (α-MTP) (Abcam, ab203114) diluted 1 μg/ml or mouse anti-Myc tag (Thermo Fisher Scientific, MA1-21316) diluted 5 μg/ml in PBS + 3% BSA + 1% goat serum. Anti-myc tag was used since no antibodies for wild-type Melusin work in IF in our hands. Cells were washed three times for 5 min with PBS and incubated at RT in a humidity chamber for 1 h with a dilution of goat anti-rabbit secondary antibody Alexa Fluor™ 633 (Thermo Fisher Scientific, A21070) 4 μg/ml in PBS + 1% BSA or goat anti-mouse secondary antibody Alexa Fluor™ 488 (Thermo Fisher Scientific, A11001) 1 μg/ml in PBS + 1% BSA. Negative controls were performed by staining cells with no primary antibody and secondary antibody only. Cells were washed three times for 5 min with PBS and incubated in a humidity chamber for 1 min at RT with a dilution of 0.1 μg/ml of DAPI (Thermo Fisher Scientific) in deionized water for nuclear staining. Cells were washed three times for 5 min with PBS, and cover slips were mounted using the Prolong Antifade Mounting Medium (Thermo Fisher Scientific). Images were acquired using the 63x objective of Leica SP8 confocal microscope at ex:488 nm/em:520 nm for the anti-mouse secondary antibody Alexa Fluor™ 488, at ex:631 nm/em:650 nm for the anti-rabbit secondary antibody Alexa Fluor™ 633.

Proximity ligation assay between Melusin and α-MTP was performed using the Duolink® In Situ Red Starter Kit Mouse/Rabbit (Sigma-Aldrich), following the manufacturer’s recommendations. AC16 cells empty and mel were plated in cellview cell culture slide, glass bottom, (Greiner) at concentration of 15,000 cells/well. For hiPSC-CMs, cellview cell culture slide was coated with poly-l lysine and ES-qualified Geltrex (0.17 mg/ml). Mel null hiPSC-CMs were plated at a concentration of 30,000 cells/well and infected with 30,000 MOI AAV6-empty and AAV6-MycMel as described above. hiPSC-CMs were used 7 days after the infection. Cells were fixed with 4% PFA for 10 min at RT, permeabilized with 0.1% TX100 in PBS, and washed in TBS + 10% Tween 20 (TBST). Cells were blocked with Duolink® Blocking Solution and incubated for 1 h at 37 °C in a preheated humidity chamber with a solution of rabbit anti-HADHA (α-MTP) (Abcam, ab203114) and mouse anti-Myc tag (Thermo Fisher Scientific, MA1-21316) diluted respectively 1 μg/ml and 5 μg/ml in Duolink® Antibody Diluent. Cells were washed twice for 5 min with Wash Buffer A at RT and incubated in a preheated humidity chamber for 1 h at 37 °C with a dilution of PLA PLUS and MINUS probes 1:5 in Duolink® Antibody Diluent. After washes, cells were incubated in a preheated humidity chamber for 30 min at 37 °C with a solution 1:40 of ligase in Duolink® Ligation Buffer. Cells were washed and incubated in a preheated humidity chamber for 100 min at 37 °C with a solution 1:80 of polymerase in Amplification Buffer. Cells were washed twice for 10 min with Wash Buffer B at RT and the chambers and silicone around the wells were removed from the slide completely. Cover slips were mounted with the Duolink® In Situ Mounting Medium with DAPI and the edges of the cover slip were sealed with nail polish. PLA images were acquired at least 20 min after using the 63x objective of Leica SP8 confocal microscope at ex:594 nm/em:624 nm.

The glucose flux through the tricarboxylic acid cycle was analyzed on living cells using 6[¹⁴C]-glucose (PerkinElmer). The cell medium was removed, and cells were washed in PBS and detached by gently scraping in PBS. Cells were centrifuged at 100 × g for 5 min and resuspended at 2 × 10⁵ cells in 1 ml of Hepes-Ca⁺⁺ buffer (145 mM NaCl, 5 mM KCl, 1 mM MgSO₄, 10 mM Hepes sodium salt, 10 mM glucose, 1 mM CaCl₂, pH 7.4). A small aliquot was used for protein quantification by BCA assay. Cells were transferred in a vial and added with 2 μCi (80 μM) of 6[¹⁴C]-glucose. The vial, connected with a tight connector closed by a rubber cap with another vial containing 0.5 ml KOH 0.8 N, was incubated for 1 h at 37 °C in agitation. Later, 0.4 ml of HClO₄ 3.7 N was added to the cell suspension by a syringe to stop the reaction. The system was adjusted to have sample vials at 37 °C and the KOH vials in ice, in order to maximize the reaction of ¹⁴CO₂ produced by the TCA cycle with the KOH, and maintained 1 h in agitation. The so generated [¹⁴C]-K₂CO₃, index of the TCA cycle activity, was mixed with 4 ml scintillation liquid (UltimaGold, PerkinElmer) and read by a β-counter. The marked ¹⁴CO₂ produced via the tricarboxylic acid cycle was expressed as picomoles CO₂ produced per hour normalized on the mg of cell proteins.

The oxygen consumption rate (OCR) of cells expressing or not Melusin was evaluated using the Resipher (Lucid Scientific) technology.

hiPSC-CMs wild-type and null for Melusin were plated overnight at near confluence (90,000 cells per well) in a 96-well plate in limited medium (DMEM glucose - glutamine - pyruvate -, NaHCO₃ 3.7 g/L, 0.5 mM glucose, 2 mM glutamax, 1% FBS, 1% penicillin/streptomycin) added with 1 mM l-carnitine. As negative controls, some wells were added with 40 μM of the carnitine palmitoyltransferase-1 (CPT-1) inhibitor Etomoxir.

AC16 cells empty and mel were plated overnight at the confluence (30,000 cells per well) in a 96-well plate in a limited medium added with 1 mM l-carnitine. As negative controls, wells were added with 40 μM Etomoxir.

The morning after, the plate was attached to the Resipher in the incubator and left to equilibrate for ̴5 h. The OCR at this time point was considered as time = 0. The plate was then detached from the machine, and a palmitate-BSA conjugate was added to the wells (150 μM for AC16 cells, 1 mM for hiPSC-CMs). The plate was attached again to the Resipher, and the OCR was followed for 25–30 h. The maximum pick of OCR was considered as time = 1 to measure the maximal OCR increase (t = 1 – t = 0).

Total RNA from frozen heart powder or cell cultures was isolated using Trizol Reagent (Thermo Fisher Scientific), following the manufacturer’s recommendations. RNA was reverse transcribed by using Applied Biosystem high-capacity cDNA reverse transcription kit. Gene expression analysis was performed using TaqMan Gene Expression Assays (Applied Biosystems) on an ABI Prism 7900HT sequence detection system (Applied Biosystems). 18S (Thermo Fisher Scientific) was used as endogenous control and gene-specific primers were used for other genes (Roche Applied Sciences). Analysis was performed using the ^ΔΔCt method to determine fold changes. The specific primers used were: ITGB1BP2 forward 5′- CCTTCCTGATTCCTGTTGCCA-3′, reverse 5′-TCTACAGTTCGCTTTCGGCA-3′; α-MTP forward 5′-TTCTTAAAGACACCACAGTGACG-3′, reverse 5′-CTTCTTCACTTTGTCGTTCAGC-3′; β-MTP forward 5′-CAAGCAATGTGGCTAGAGAGG-3′, reverse 5′-AGAGATACAAGCCATGGTGACA-3′; NONO forward 5′-AGATGGCTATGGGAGGTGCTA-3′, reverse 5′-AGCCTGACCAAAGCGTTCAG-3′; HPRT forward 5′-TGACACTGGCAAAACAATGCA-3′, reverse 5′-GGTCCTTTTCACCAGCAAGCT-3′; TNNT2 forward 5′-TTCACCAAAGATCTGCTCCT CGCT-3′, reverse 5′-TTATTACTGGTGTGGAGTG GGTGTGG-3′; GLUT1 forward 5′-CCTGTCTCTTCCTACCCAACC-3′, reverse 5′-GCAGGAGTGTCCGTGTCTTC-3′; GLUT4 forward 5′-CTCATGGGCCTAGCCAATG-3′, reverse 5′-GGGCGATTTCTCCCACATAC-3′;

Malonyl-CoA level was measured using the Mouse Malonyl Coenzyme A ELISA kit (CUSABIO), following the manufacturer’s recommendations. Briefly, 100 mg of cardiac tissue from wild-type, Mel null and Mel over mice were homogenized in 1 ml of PBS and stored overnight at −20 °C. The day after, samples were centrifuged at 5000×g for 5 min at 4 °C, and the supernatant was collected for the assay. Supernatants were diluted 1:3000 in dilution buffer and 100 μl was added in the wells of a coated assay plate and incubated 2 h at 37 °C. A standard solution was used as calibrator at 10, 5, and 1 ng/ml, while dilution buffer was used as 0. Samples were removed and 100 μl of biotin-antibody 1x was added and incubated 1 h at 37 °C. Wells were washed with wash buffer and incubated with 100 μl of HRP-avidin 1 × 1 h at 37 °C. Wells were washed again and filled with 90 μl of TMB substrate for 15–30 min at 37 °C. About 50 μl of Stop solution was added to the wells before reading the optical density at 450 nm.

Frozen hearts were lysed at 4 °C in 0.6 M perchloric acid (PCA) for NAD⁺ and NADP⁺ determination or in 0.1 M NaOH for NADH and NADPH determination. The lysate was clarified by centrifugation. NAD⁺/NADP⁺ contents were quantified as previously described (Bruzzone et al, 2009). NADH/NADPH content was measured by fluorescence, as previously reported (Pittelli et al, 2010).

Live hiPSC-CMs were resuspended in 200 μL of FACS buffer (PBS, 5% FBS, and 0.05% sodium azide) into a 96-well plate and centrifuged at 220×g for 4 min to remove the supernatant. Samples were incubated with 100 μL of fixable viability dye eFluor 450 (Invitrogen) diluted 1:1000 in PBS for 10 min at RT covered from light. Samples were washed in FACS buffer, fixed in PBS + 4% paraformaldehyde (Alfa Aesar) for 15 min at RT, and washed twice in FACS buffer. Cells were then permeabilized and blocked in PBS 0.75% saponin 5% FBS, and stained in 50 μL of PBS 0.75% saponin with anti-cTnT AF647 antibody (BD Pharmingen) diluted 1:100, incubating for 40 min at RT covered from light. An isotype control staining was performed in parallel. The stained samples were centrifuged and washed twice in PBS 0.75% saponin, and finally resuspended in FACS buffer for the analysis using a FACS Celesta. Data analysis was performed using FlowJo. Antibodies used were: Alexa Fluor 647 Mouse Anti-Cardiac Troponin T; clone 13-11 (BD Pharmingen); Alexa Fluor 647 Mouse IgG1, κ Isotype Control; clone MOPC-31 C (BD Pharmingen)

ROS amount was evaluated with two different approaches:

Mitochondria isolated as described above and resuspended in mitochondrial buffer were used to measure the electron flux from mitochondrial Complex I to Complex III. In a 96-well plate, 10 μl of mitochondria was incubated with 160 μl buffer A (5 mM KH₂PO₄, 5 mM MgCl₂, 5% BSA) and 100 μl buffer B (50 mM KH₂PO₄, 5 mM MgCl₂, 5% BSA, 25% saponin). Freshly prepared cytochrome c oxidized form 0.12 mM and NaN₃ 0.2 mM were added. The plate was left to equilibrate for 1 min at RT, then the reaction was started by adding NADH 0.15 mM. The kinetics of the reduction of cytochrome c was monitored at 550 nm for 6 min, using a Synergy HTX Multimode reader (Bio-Tek Instruments). The velocity rate was extrapolated from the curve, and results were calculated according to the Lambert–Beer equation and expressed as picomoles of cytochrome c reduced/minutes per mg of mitochondrial proteins.

The ATP content was measured in both isolated mitochondria and in cardiac total extracts using the ATP Bioluminescent Assay Kit (Sigma-Aldrich) following the manufacturer’s recommendations. Mitochondria were isolated as described above and resuspended in mitochondrial buffer. Hearts (15–20 mg) were homogenized in 350 μl of PBS added with proteases inhibitors, boiled 5 min at 95 °C to denature ATPases and centrifuged at 16000 × g for 5 min. Supernatants were used for the assay.

In a 96-well plate, 100 μl of ATP assay mix was added with 100 μl of sample or ATP standard as calibrator. Luminescence was immediately measured with a luminometer. Results were expressed as nanomoles of ATP per mg of mitochondrial or whole tissue proteins.

Lipid peroxidation in cardiac samples was evaluated by detecting malondialdehyde (MDA) and 4-hydroxyalkenals (HAE) using the Bioxytech LPO-586 (OxisResearch, USA), following the manufacturer’s recommendations. About 30–40 mg of cardiac tissue was homogenized in 350 μl of PBS added with 5 mM butylated hydroxytoluene (BHT) in acetonitrile and centrifuged at 3000×g for 10 min at 4 °C. Supernatant was saved for the assay, and a small aliquot was used for protein quantification. About 200 μl of the sample was mixed with 650 μl of N-methyl-2-phenylindole in acetonitrile (Reagent R1) diluted in methanol. About 150 μl of 15.4 M methanesulfonic acid (Reagent R2) was added and incubated for 1 h at 45 °C. Samples were finally centrifuged at 15,000 × g for 10 min, and supernatants were collected to measure the absorbance at 586 nm with a spectrophotometer. A blank with water and standards at different concentrations of MDA were used. Results were calculated as suggested and expressed as nanomoles of MDA + HAE per g of wet weight.

Small pieces of cardiac tissue were fixed with 2.5% glutaraldehyde (Sigma-Aldrich) + 100 mM cacodylate (Sigma-Aldrich) and postfixed with 1% osmium tetroxide (Electron Microscopy Science), 1.5% potassium ferrocyanide (Electron Microscopy Science) + 100 mM cacodylate. Samples were stained in 0.5% uranyl acetate (Electron Microscopy Science), dehydrated, and finally embedded in Embed-812 resin (Electron Microscopy Sciences). Blocks were cut in ultrathin sections (70 nm) and contrasted with uranyl acetate (Electron Microscopy Science) and Sato’s lead solutions (lead citrate tribasic, Sigma-Aldrich; lead nitrate, Merck; lead acetate trihydrate, Sigma-Aldrich; sodium citrate tribasic dehydrate, Sigma-Aldrich) and observed with Talos 120 C (FEI) electron microscope. Images were acquired with a 4k × 4 K Ceta CMOS camera (Thermo Fisher Scientific). Image J was used to process the images and count the cristae distance.

Mitochondrial membrane potential was evaluated using the JC-1 Mitochondrial Membrane Potential Flow Cytometry Assay Kit (Cayman). AC16 cells empty and mel were plated in a 12-well plate at the concentration of 120,000 cells per well. hiPSC-CMs wild-type and null for Melusin were plated in a 12-well plate at the concentration of 500,000 cells per well. Cells were treated with doxorubicin as described above. Cells were detached by trypsinization, resuspended in cell medium, and finally centrifuged at 200 × g for 5 min at RT. Cells were resuspended in 100 μl of assay buffer, excepting one well resuspended in assay buffer added with 50 μM FCCP, as a positive control. After 5 min of incubation at RT, cells were stained with 100 μl of JC-1 2X working solution diluted in culture media. A JC-1 unstained sample was used as a normalizer for doxorubicin auto-fluorescence. Data were immediately acquired at the Celesta FACS machine at ex:405 nm/em:605-308 (Perelman et al, 2012), analyzed using FlowJo and expressed as a percentage of mitochondria with a physiologic mitochondrial membrane potential compared to unstained controls.

The genomic DNA (gDNA) and mitochondrial DNA (mtDNA) were extracted by DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer’s recommendations and quantified using Nanodrop. The genomic and mitochondrial gene expression profiles were tested by Luna Universal qPCR sybr Master Mix (NEW ENGLAND BioLabs) using 10 ng of DNA for mtDNA and 40 ng of DNA for gDNA. The thermal cycling conditions at the QuantStudio^TM 6 Real-Time PCR were as follows: 1 cycle of 95 °C for 5 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min (Kumpunya et al, 2022). The relative content of mtDNA was normalized to genomic β-2-microglobulin (Δct = gDNA ct – mtDNA ct) and determined by the 2 × 2^Δct. Wildtype, Mel null and Mel over, untreated and treated, data were normalized on the media of the respective untreated wild-type control. The primer sequences used were: mtDNA: 5′-CGTACACCCTCTAACCTAGAGAAGG-3′, 5′- GGTTTTAAGTCTTACGCAATTTCC-3′ and β2m: 5′-TTCTGGTGCTTGTCTCACTGA -3′, 5′- CAGTATGTTCGGCTTCCCATTC-3′.

Statistical analyzes were performed using GraphPad Prism 9. The number of biological replicates is indicated in the Fig. legend, and data were presented as individual data points overlayed with means ± SD. The normal distribution of data was tested by Shapiro–Wilk tests, to determine whether parametric statistical tests could be employed. Depending on the variables involved, significance was tested using the unpaired two-tailed Student's t-test, one-way ANOVA, or two-way ANOVA with Tukey/Bonferroni post hoc correction, as detailed in the relative Fig. legend. For experiments in which we evaluated the amplitude of difference among one genotype and its control set to 100%, we performed a “one sample t-test” comparing the mean of the group to the hypothetical mean represented by 100% of the control group. A p value below 0.05 was considered to be statistically significant. The Mantel–Cox test was used in Kaplan–Meier survival curves.