Article ATPase inhibitor factor-1 disrupts mitochondrial Ca2+ handling and promotes pathological cardiac hypertrophy through CaMKIIδ

Background : ATPase inhibitor factor-1 (IF1) preserves cellular ATP under conditions of respiratory collapse, yet the function of IF1 under normal respiring conditions is unresolved. We tested the hypothesis that IF1 promotes mitochondrial dysfunction and pathological cardiomyocyte hypertrophy in the context of heart failure (HF). Methods and results Cardiac expression of IF1 was increased in mice and in humans with HF, downstream of neurohumoral signaling pathways and in patterns that resembled the fetal-like gene program. Adenoviral expression of wild type IF1 in primary cardiomyocytes resulted in pathological hypertrophy and metabolic remodeling as evidenced by enhanced mitochondrial oxidative stress, reduced mitochondrial respiratory capacity, and the augmentation of extra-mitochondrial glycolysis. Similar perturbations were observed with an IF1 mutant incapable of binding to ATP-synthase (E55A mutation), indication that these effects occurred independent of binding to ATP synthase. Instead, IF1 promoted mitochondrial fragmentation and compromised mitochondrial Ca2+ handling, which resulted in sarcoplasmic reticulum Ca2+ overloading. The effects of IF1 on Ca2+ handling were associated with the cytosolic activation of CaMKII and inhibition of CaMKII or co-expression of catalytically dead CaMKIIδC was sufficient to prevent IF-1 induced pathological hypertrophy. Conclusions IF1 represents a novel member of the fetal-like gene program that contributes to mitochondrial dysfunction and pathological cardiac remodeling in HF. Furthermore, we present evidence for a novel, ATP-synthase independent, role for IF1 in mitochondrial Ca2+ handling and mitochondrialto nuclear crosstalk involving CaMKII.


Introduction
The heart requires tremendous amounts of ATP to sustain the systemic circulation, most of which is generated by mitochondria. In patients with HF, metabolic roadblocks and structural damage to mitochondria occur that induce

Myocardial expression of IF-1 is increased in heart failure and induces metabolic reprogramming in NRVM
To corroborate previous reports that indicated that the expression of IF1 is increased in failing hearts 14 , IF1 mRNA expression was determined in the LV lysates obtained from several models of murine HF: mice with cardiomyocyte autonomous expression of activated Gαq 15 , mice subjected to transverse aortic constriction (TAC) 20 , or mice subjected to a myocardial infarction (MI) 16 . In all 3 models, IF1 expression levels were increased by 3-fold compared to control mice ( Figure 1A). The increased mRNA expression of IF1 corresponded with a similar increase in IF1 protein levels ( Figure   S1A). IF-1 mRNA levels were also higher in cardiac lysates obtained from patients with end stage HF compared to control hearts ( Figure 1B). In Neonatal Rat Ventricular Cardiomyocytes (NRVM), overexpression of Gαq ( Figure S1B) and treatment with isoproterenol ( Figure 1C) which both stimulate pathological hypertrophy resulted in a significant increase in the expression of IF-1. In contrast, treatment with insulin like growth factor-1 (IGF-1) which stimulates physiological cardiac growth did not affect IF-1 expression levels ( Figure 1C). The expression of IF1 increased during consecutive stages of embryonic development, to reduce after birth ( Figure S1C). IF-1 expression increased again in mice with HF.
Together these findings suggest that IF1 is regulated downstream of maladaptive neurohumoral signaling pathways in patterns mimicking the fetal-like gene program. (LDH) and Pyruvate kinase (PRK) assayed with RT-qPCR (n=4). Data are presented as mean ± SEM. *p < 0.05, **p < 0.01 and ***p < 0.001 vs ad-CTL by nonparametric Mann-Whitney test (C, G and H). Data are presented as mean ± SEM. * p < 0.05, ** p < 0.01 and *** p < 0.001 vs SHAM/Ctrl/ad-CTL using the ad-CTL using the Mann-Whitney U test or T-test where appropriate.
To determine the consequences of increased expression of IF1 on mitochondrial function in cardiomyocytes, NRVM's were infected with an adenoviral vector expressing human IF1 (ad-IF1-WT), or a control virus expressing GFP (ad-CTL) 22 .
Infection with ad-IF1-WT resulted in a 7-fold increase in the protein levels of IF1 ( Figure S2A). As expected, the drop in mitochondrial membrane potential following respiratory collapse was increased by ad-IF1-WT ( Figure S2B-C) and preserved intracellular ATP levels, consistent with stronger inhibition of the reverse mode of ATPsynthase ( Figure S2D).
Infection with IF1 did not affect cellular viability for up to 2 days after infection ( Figure S2E).
To determine the consequences of increased IF1 expression under respiring conditions, mitochondrial oxygen consumption rate (OCR) was measured using a SeaHorse metabolic flux analyzer. Overexpression of IF-1 did not influence basal or ATP-linked respiration, nor did it affect the proton leak ( Figure 1D, E and S2F). IF1 did, however, result in a dose dependent decrease in the maximal mitochondrial respiration and a consequent reduction in mitochondrial spare capacity ( Figure 1E and S2F). Reducing IF1 expression levels with an adenovirus expressing siRNA-IF1 did not influence mitochondrial respiration, arguing against a direct effect of IF1 on respiratory complexes ( Figure S2G and S2H). Next, we determined the consequences of IF1 on extra-mitochondrial metabolism by measuring extracellular acidification rates as a proxy of anaerobic glycolysis 23 . Infection with Ad-WT-IF1 resulted in a significant increase in anaerobic glycolysis ( Figure   1F, G), accompanied by increased expression of the key glycolytic enzymes lactate dehydrogenase and pyruvate kinase ( Figure 1H). Together these findings suggest that IF1 compromises maximal mitochondrial respiration and promotes extramitochondrial glycolysis without affecting ATP-linked respiration.
IF1 induces mitochondrial oxidative stress and stimulates cardiomyocyte hypertrophy IF1 has been shown to both increases-or decrease mitochondrial ROS emissions depending on the cell type and experimental context 5,9,24 . We therefore hypothesized that the reductions in maximal mitochondrial respiration could be explained by mitochondrial oxidative stress and/or downregulation of mitochondrial respiratory complexes. Compared to Ad-CTL infected cells, infection with Ad-IF1-WT resulted in a 40% increase in fluorescence intensity of the mitochondrial-specific ROS indicator MitoSox (Figure 2A). Similar increases in the expression of the ROS-responsive genes NADPH oxidase 2 (NOX2), Nuclear receptor factor-2 (NRF2) and Heat shock protein-60 (HSP60) were observed ( Figure 2B). ROS-mediated mitochondrial DNA damage was also increased by 50% in ad-IF1-WT infected cells ( Figure   2C). The protein expression of four out of five respiratory chain complexes were significantly reduced in ad-IF1-WT infected cells ( Figure 2D). The reductions in maximal mitochondrial respiration induced by IF1 are thus accompanied by mitochondrial oxidative stress and reductions in electron chain complexes.
Overexpression of IF1 also resulted in a marked increase in cardiomyocyte size, ( Figure 2E) accompanied by increases in mRNA levels atrial natriuretic peptide (ANP) and B-type natriuretic peptide (BNP) ( Figure 2F). The increases in cell size and natriuretic peptides were equivalent to those of 50 µM of phenylephrine ( Figure S3A and S3B). The combination of phenylephrine and ad-WT-IF1 did not result in further increase in cell size or ANP. (Figure S3A and S3B). IF1 expression thus appears to be sufficient to promote pathological cardiomyocyte hypertrophy.  Infection with ad-IF1-E55A resulted in similar increases in IF1 expression compared to ad-IF1-WT ( Figure S4A). As expected, the drop in mitochondrial membrane potential following respiratory collapse was reduced by ad-IF1-E55A, consistent with diminished inhibition of the reverse mode of ATP-synthase ( Figure S4B). Despite its inability to bind to ATP-synthase, infection with ad-IF1-E55A resulted in similar reductions in maximal mitochondrial respiration and mitochondrial spare capacity as did WT-IF1 ( Figure 3A and 3B). Infection with ad-IF1-E55A also resulted in a marked increase in glycolysis ( Figure 3C and 3D) as well as increased expression of LDH and PDK ( Figure 3E). Mitochondrial ROS emissions were also increased after infection with ad-IF1-E55A ( Figure 3F), as were the expression levels of NOX2, NRF2 and HSP60 ( Figure 3G). Moreover, downregulation of the respiratory chain complexes II, IV and V was also observed with ad-IF1-E55A ( Figure S4C). Finally, infection with ad-IF1-E55A induced cardiomyocyte hypertrophy that was proportional to what had been observed with Ad-IF1-WT ( Figure 3H) and ANP and BNP mRNA expression were also significantly increased ( Figure S4D). Together these data demonstrate that the effects of IF1 on mitochondrial function and cardiomyocyte hypertrophy are independent of binding of IF1 to ATP-synthase.  NRF-2 and HSP-60. (H) Bar graph depicting differences in cardiomyocyte crossectional area (n=5). Data are presented as mean ± SEM. *p < 0.05 and **p < 0.01 vs ad-CTL using the Mann-Whitney U test or T-test where appropriate.

IF-1 induces mitochondrial fission through DRP1
Altered mitochondrial dynamics are thought to contribute to mitochondrial dysfunction in heart failure [25][26][27] . Because IF1 has been implicated in the regulation of mitochondrial dynamics 6, 7, 34 , we determined mitochondrial volume and morphology using confocal Z/y scanning followed by image deconvolution and 3D reconstruction. Typical examples of confocal images and multislice 3D reconstructions are depicted in figure 4A and 4B. Total mitochondrial volume and mitochondrial DNA copy numbers were not affected by IF1( Figure S5A and S5B). Infection with Ad-IF1-WT did, however, increase the number of mitochondria per cell and reduced the average mitochondrial volume and increased the fission index ( Figure 4C), indicating that IF1 promotes mitochondrial fission in NRVC. A key step in the activation of mitochondrial fission is the phosphorylation and translocation of the GTPase protein called Dynamin-related protein 1 (DRP-1) to the outer mitochondrial membrane 28 . Infection with IF1 resulted in a marked increase in the mitochondrial translocation of DRP1 ( Figure 4D and E). The myocardial expression of mitofusin-2 was decreased by IF1 ( Figure 4F

IF-1 reduces mitochondrial Ca 2+ and induces sarcoplasmic reticulum Ca 2+ overload
It has been well described that mitochondrial dysfunction in the context of cardiomyocyte hypertrophy is associated with reductions in mitochondrial Ca Ca +2 26, 29 . Moreover, it has been proposed that this is related to excessive mitochondrial fission which diminished the mitochondrial capacity to store Ca Ca 2+ 30 . To determine whether IF1 affects mitochondrial Ca 2+ levels, we first measured mitochondrial Ca2+ content with the mitochondrial Ca 2+ specific dye Rhod2-AM. Cells infected with Ad-IF1-WT demonstrated a significant reduction in Rhod2-AM fluorescence, consistent with reduced mitochondrial Ca 2+ content ( Figure 5A). To confirm these observations, we also quantified FCCP-induced mitochondrial Ca2+ release using the cytosolic Ca 2+ -sensitive dye Fluo4-AM. Indeed, FCCP-induced mitochondrial Ca2+ release was also significantly lower in NRVM infected with Ad-IF1-WT than in those infected with Ad-CTL ( Figure 5B and 5C). To determine how the reductions in mitochondrial Ca 2+ content affected Ca 2+ concentrations in other cellular compartments, we also determined Ca 2+ concentrations in the cytosol and the sarcoplasmic reticulum using the ratiometric high affinity Ca 2+ selective fluorescence indicator Fura 2-AM. Overexpression of ad-IF1-WT did not affect basal cytosolic Ca 2+ levels ( Figure 5D). However, potassium chloride-induced sarcoplasmic Ca 2+ release was significantly increased in cells infected with ad-IF1-WT compared to ad-Ctrl. (Figure 5E and 5F). Next, we assessed the mRNA expression of a panel of genes involved in the regulation of mitochondrial Ca 2+ handling. Interestingly, IF1 resulted in a significant increase in the expression of the dominant-negative pore-forming subunit of the mitochondrial Ca 2+ uniporter ß (MCUB, Figure 5G). In summary, IF1-induced mitochondrial fragmentation and upregulation of MCUB compromises mitochondrial Ca 2+ handling and promotes SR Ca 2+ overloading.

IF1-induced cardiomyocyte hypertrophy is dependent upon CaMKII activation
To study the role of CaMKII activation in the pathological effects of IF1 described above, NRVMs were co-infected with an adenovirus expressing a catalytically dead mutant of the predominant cardiac isoform of CaMKII (CaMKIIδ; ad-dnCAMKII).
As expected, infection with ad-dnCAMKII reduced CaMKII-dependent phosphorylation of PLN ( Figure 6C). Co-infection with ad-dnCAMKII did not alter IF1-induced changes in maximal mitochondrial respiration ( Figure 6D), nor did it affect mitochondrial ROS production ( Figure 6E). Ad-dnCAMKII did, however, block IF1-induced cardiomyocyte hypertrophy ( Figure 6E) and IF1-induced expression of ANP ( Figure 6F). Together these findings indicate that the effects of IF1 on cardiac hypertrophy are dependent on the activation of CaMKIIδ (Figure 7). The primary function of IF1 is to block the reverse mode of ATP-synthase, which can occur upon dissipation of the mitochondrial membrane potential under ischemic conditions or in critically damaged mitochondria [14]. By blocking ATP-synthase reversal, IF1 prevents mitochondria from becoming ATP consumers rather than generators. Under these conditions, IF1 also suppresses programmed cell death, and stimulates parkin-dependent elimination of dysfunctional mitochondria [8,9,15]. It has long been thought that the biological function of IF1 was restricted to conditions of respiratory compromise. However, several recent publications have suggested that IF1 can also inhibit the forward direction of ATP-synthase and thereby controls the rate of ATP synthesis under normal respiring conditions [6,16] Conversely, different research groups have rather demonstrated that IF1 promotes ATP synthase activity by stimulating mitochondrial cristae formation [8,9]. These conflicting results may reflect differences in cell types and experimental conditions, as most studies so far have been performed in immortalized cell lines and cancer models. In our hands, IF1 did not influence ATP-linked respiration and the structural and biochemical consequences of overexpression of IF1 also occurred with an IF1 mutant that is unable to bind to ATP-synthase. Our results therefore strongly suggest that the biological effects of IF1 under normal respiring conditions are independent of the canonical pathway that requires binding to ATP-synthase. Overexpression of IF1 reduced maximal mitochondrial respiration and stimulated glycolysis. The reductions in mitochondrial respiratory capacity were associated with mitochondrial oxidative stress, significant damage to mitochondrial DNA and downregulation of mitochondrial respiratory chain complexes. In addition, IF1 stimulated mitochondrial fission and fragmentation, which by itself is sufficient to reduce mitochondrial respiration and promote oxidative stress [17]. IF1 has consistently been implicated in mitochondrial oxidative stress, yet the exact mechanism responsible remains enigmatic. Some authors have suggested that IF1-induced oxidative stress is caused by increased in the mitochondrial membrane potential, secondary to ATP-synthase inhibition [5]. However, our results clearly indicate that the mitochondrial oxidative stress induced by IF1 is independent on its capacity to bind to ATP-synthase. Further research is therefore required to determine the mechanisms of IF1-induced mitochondrial ROS.
The profound effects of IF1 on mitochondrial Ca 2+ handling were, arguably, the most intriguing finding of our study.
Overexpression of IF1 reduced the mitochondrial capacity to store Ca 2+ and resulted in SR Ca 2+ overloading. Aberrant SR Ca 2+ handling is a central mechanism responsible for various pathophysiological changes in failing myocytes including but not restricted to cardiac arrythmia's, transcriptional activation of hypertrophy, mitochondrial dysfunction, and cell death [18]. IF1 could thus reflect a novel mechanistic link between mitochondrial dysfunction and dysregulated Ca 2+ handling in HF. Our findings confirm and extend upon a recent study in which exogenous treatment with IF1 corresponded with increases in cytosolic Ca 2+ levels in skeletal muscle cells [11]. The mechanisms responsible for IF1-induced reductions in mitochondrial Ca 2+ are unknown, yet it is tempting to speculate that they rely on intrinsic biochemical properties of the IF1 molecule. For instance, it was recently discovered that Ca 2+ influences the self-association and structure of IF1 and it is possible that IF1 influences the Ca 2+ buffering in the mitochondrial matrix [19]. IF1 has also been shown to regulate critical Ca 2+ handling proteins such as the MCU [7]. While we did not detect changes in the expression of MCU, IF1 did promote the expression of the negative regulator of MCU, MCUB. Finally, we cannot exclude the possibility that the reductions in mitochondrial Ca 2+ were the consequence of mitochondrial oxidative stress or the mitochondrial fragmentation observed in our model. Nevertheless, our study contributes to the growing body of evidence that IF1 regulates mitochondrial Ca 2+ . protected from pressure overload-induced cardiomyocyte hypertrophy [14]. We confirm and extend upon these observations by providing a mechanistic link between SR Ca 2+ overloading and the activation of CaMKII. Multifunctional

IF1-induced SR Ca
CaMKII has been implicated in a myriad of pathogenic cellular responses in HF, which include mitochondrial reprograming, mitochondrial oxidative stress and mitochondrial fragmentation [20][21][22]. Nevertheless, IF1-induced reductions in mitochondrial respiration and mitochondrial ROS were not affected by catalytically dead CaMKII. The activation of CaMKII thus appears to be a consequence of IF1-induced mitochondrial dysfunction rather than a cause. Of note, CaMKII can also be activated by ROS and we cannot determine whether the activation of CaMKII in our model was dependent upon Ca 2+ or ROS.
While our study provides compelling evidence for a role of IF1 in controlling mitochondrial oxidative stress and mitochondrial Ca 2+ homeostasis in cardiomyocytes, several limitations need to be acknowledged. One of the main limitations of our experimental model is that neonatal cardiomyocytes display a relatively immature metabolic phenotype that is highly glycolytic. Therefore, the role of IF1 on mitochondrial metabolism may be different in adult cardiomyocytes or in vivo. Moreover, ATP-linked respiration is modest in NRVMs, which may have prevented us from detecting subtle changes in ATP-lined respirations induced by IF1. Nevertheless, studies with the ad-IF1-E55A mutant clearly demonstrated that neither the metabolic, nor the structural changes induced by IF1 were dependent on binding of IF1 to ATP-synthase. Another limitation of our study is that we do not provide mechanisms responsible for IF1-induced mitochondrial oxidative stress and mitochondrial fission since, to our opinion, this is beyond the scope of the current investigation.
Based on our results we suggest that the increase in the expression of IF1 in HF is maladaptive and contributes to mitochondrial fragmentation, aberrant Ca2+ handling and CAMKII-dependent pathological remodeling. Our study reinforces the concept that mitochondrial dysfunction has profound effects in failing cardiomyocytes that extend beyond bioenergetic insufficiency. Finally, the present work underscores a central role of mitochondria in pathological growth responses in the heart and supports current efforts to design mitochondrial targeted therapies for HF [23,24]. IF1 represents a novel member of the fetal-like gene program that contributes to mitochondrial dysfunction and pathological cardiac remodeling in HF. Furthermore, we present evidence for a novel, ATP-synthase independent, role for IF1 in mitochondrial Ca 2+ handling and mitochondrial-to nuclear crosstalk involving CaMKII.

Material and Methods
The use of animals for these studies was in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The studies were submitted to and approved by the Institutional Animal Care and Use Committee of the University of Groningen or the University of California San Diego.
Heart failure samples mRNA was isolated from viable left ventricular tissue from 3 different murine models for chronic heart failure; first, Gαq-40 transgenic mice has been described previously [25]. Second, transverse aortic constriction (TAC) was performed with a 7-0 nylon suture between the carotid arteries around a 27G needle, as described [26]. Third, post-myocardial infarction (MI) heart failure was achieved through a permanent ligation of the left coronary artery with a Premilene 6-0 suture [26]. Human myocardial tissue was obtained from patients with end-stage ischemic heart failure (n=10) and from donor hearts rejected due to technical reasons (n=19) within the heart transplant program at the Heart-und Diabetes Center NRW. The study was approved by the local ethical committee and conducted in accordance with the guidelines in the Declaration of Helsinki.

Neonatal rat cardiomyocytes isolation
The use of animals for these studies was in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study was submitted to and approved by the Committee for Animal Experiments of the University of Groningen. Euthanasia was performed by quick decapitation. Primary neonatal rat ventricular cardiomyocytes (NRVMs) were isolated from neonatal rats of 1-2 days old, as previously described [27]. NRVMs were grown in MEM (Sigma M1018) supplemented with 5% fetal bovine serum (FBS) (Thermo Fisher SV30160) and penicillin-streptomycin (100 U/ml-100 μg/ml) (Thermo Fisher 15070063). IF1 wild-type sequence (IF1-WT) was cloned into adenovirus pSF-Ad5-WT OG617 using the E1A promoter (Oxford Genetics). Inactive form of IF1 (E55A) sequence was cloned previously by Prof. David Sabatini (Addgene #85404) [28]. (PerkinElmer NEL112001EA). The specific primary and secondary antibodies that were used are depicted in table S1.

Real time PCR
To analyze gene expression, total RNA was isolated using TRI reagent according to the protocol provided (Sigma, T9424).
RNA concentrations have been determined with a Nanodrop 2000 (Thermo Scientific) and cDNA was synthetized by reverse transcription using QuantiTect Reverse Transcription Kit (Qiagen 205313) and real time qPCR were performed with IQ SYBR green (BioRad 170-8885) using specific primers (see table 1). Relative expression levels were calculated using 2 (-ΔΔCT) .
To assess mitochondrial DNA to nuclear DNA ratio and DNA damage, Total DNA including mtDNA was extracted from the non-infarcted left ventricle using Nucleospin Tissue XS (Macherey-Nagel 740.901.50). mtDNA-to-nDNA ratio was determined by quantitative real-time polymerase chain reaction (qRT-PCR), as described previously [29]. Mitochondrial DNA copies were corrected for nuclear DNA values, and the calculated values were expressed relative to the control group Immediately after mitochondrial stain cells were washed 3 times in PBS 1X and fixed with PFA 4% and processed as described in the cell size section. FITC-labelled α-actinin (Sigma A7811) was employed as a marker for cardiomyocytes.
Image acquisition was performed using Leica Sp8 Lightning confocal microscope. Z-stacks was obtained from two independent channels and imaging deconvolution were performed using Huygens Pro. Processed image were analyzed by Imaris software. Cell surface was measured, and 3D reconstruction was performed by surface area detailed level (0.2 µM) from red (mitochondria) and green channel (α-actinin) respectively. Number of particle and total mitochondrial volume was quantified per cell. Krebs buffer was added. The ratio of 340/380 wavelength fluorescence analysis was performed using ImageJ.

Cell size
For cell-size measurement, cells were cultured on laminin (Millipore CC095)-coated coverslips for 48h, and then transfected with the specific adenoviruses. The cardiomyocytes were fixed with 4% paraformaldehyde (Merck 4005) in phosphate buffer for 5 minutes at room temperature. After, the cells were washed in PBS 1x and followed by permeabilization with PBS + 0.3% Triton-X100 (Sigma-Aldrich, T9284) on ice for 5 minutes. For images acquisition we used Leica SP8 epifluorescence microscopy (Leica, Germany) and for the analysis we used ImageJ software (NIH, Bethesda, MD, USA) for the determination of cellular area. Five observation fields were selected randomly on each cover slip, and 5-10 cells within each observation field were selected for the determination of the mean cardiomyocyte surface area according to the image analysis system. For the measurements, we used at least 5 different fields from 5 independent cultures in each condition (>50 cells).

Statistical analysis
All data are presented as mean ± SEM. Comparisons between groups were performed using the Student t test, the Mann-Whitney U test, Kruskal-Wallis test, or 1-way ANOVA, followed by the Tukey post hoc test, where appropriate. A P value <0.05 was considered statistically significant.