Cardiac-specific Conditional Knockout of the 18-kDa Mitochondrial Translocator Protein Protects from Pressure Overload Induced Heart Failure

Heart failure (HF) is characterized by abnormal mitochondrial calcium (Ca2+) handling, energy failure and impaired mitophagy resulting in contractile dysfunction and myocyte death. We have previously shown that the 18-kDa mitochondrial translocator protein of the outer mitochondrial membrane (TSPO) can modulate mitochondrial Ca2+ uptake. Experiments were designed to test the role of the TSPO in a murine pressure-overload model of HF induced by transverse aortic constriction (TAC). Conditional, cardiac-specific TSPO knockout (KO) mice were generated using the Cre-loxP system. TSPO-KO and wild-type (WT) mice underwent TAC for 8 weeks. TAC-induced HF significantly increased TSPO expression in WT mice, associated with a marked reduction in systolic function, mitochondrial Ca2+ uptake, complex I activity and energetics. In contrast, TSPO-KO mice undergoing TAC had preserved ejection fraction, and exhibited fewer clinical signs of HF and fibrosis. Mitochondrial Ca2+ uptake and energetics were restored in TSPO KO mice, associated with decreased ROS, improved complex I activity and preserved mitophagy. Thus, HF increases TSPO expression, while preventing this increase limits the progression of HF, preserves ATP production and decreases oxidative stress, thereby preventing metabolic failure. These findings suggest that pharmacological interventions directed at TSPO may provide novel therapeutics to prevent or treat HF.

mitochondrial Ca 2+ serves as a regulator of energy production 6,14,15 . However, during HF, mitochondrial Ca 2+ uptake is typically reduced 3,14,16 , which can impact the production of energy and contribute to the pathogenesis of HF. The 18-kDa translocator protein (TSPO) is a five transmembrane domain protein that is primarily localized in the OMM and ubiquitously expressed throughout the body with elevated expression in steroid-synthesizing tissues, such as adrenal glands, gonads and the brain, as well as in the heart 17 . In addition to regulating steroidogenesis 18 , there is evidence that TSPO is involved in modulating oxidative stress 19 , and, critically, mitochondrial physiology and metabolism 20 . These latter effects include modulation of Ca 2+ transport 21 , regulation of transcription of genes involved in mitochondrial electron transport chain, and thus ATP production 22 , and mitochondrial quality control 23 .
The above effects of TSPO could be explained via modulation of VDAC 23 , a ubiquitous channel located on the OMM which regulates the energy balance of mitochondria, serving as a common pathway for metabolite exchange between mitochondria and the cytoplasm 24 . Alterations in VDAC function, expressed as open-probability (P o ), can impact oxidative phosphorylation by controlling the rate of adenine nucleotide exchange across the OMM 25 . Recent evidence has shown that expression of the TSPO is increased during stress conditions [26][27][28][29][30] , with the degree of expression reflecting the severity of the condition 31 . Thus, conditions which increase TSPO expression, such as HF, can modulate metabolism and affect cell survival through its effect on VDAC P o ; however, no data exist on the role played by TSPO during HF development. Our previous work showed that altering TSPO activity with a pharmacologic ligand protoporphyrin IX (hemin) decreased Ca 2+ uptake in mitochondria isolated from cardiac tissue by decreasing the open probability of VDAC 21 , which raises the possibility that TSPO may play a role in mitochondrial Ca 2+ handling.
The objectives of the present study were to test the hypotheses that: a) HF increases expression of TSPO, with a concomitant decrease in inward mitochondrial Ca 2+ entry and systolic function; and thus b) genetic modulation of the TSPO could normalize mitochondrial Ca 2+ uptake and prevent the abnormalities in cardiac structure and function evident in HF, specifically by normalizing cellular energetics and redox balance inside cardiomyocytes. These hypotheses were tested using a cardiomyocyte-specific, conditional knockout of the TSPO in a murine model of pressure-overload HF. Our studies show for the first time that TSPO-KO animals were protected from the progression of HF when compared to wild-type (WT) controls. In concordance with this protection, TSPO-KO mice had elevated activity of the mitochondrial complex I which led to the reduction of the oxidative stress in TSPO-KO TAC mice, shifting redox balance toward a more reduced state and preserving the mitochondrial membrane potential (ΔΨ m ). Accordingly, the TAC-induced decline in mitochondrial Ca 2+ uptake was prevented in TSPO-KO vs. WT mice, with correspondent preservation of ATP generation. Furthermore, the signaling cascade activating the removal of damaged mitochondria via mitophagy was significantly impaired in WT TAC and restored in TSPO KO cardiomyocytes, thus contributing to improved mitochondrial function. Together, these data suggest that TSPO can be a potential pharmacological target to correct the abnormalities in mitochondrial Ca 2+ transport and mitochondrial bioenergetics associated with HF.

Results
TSPO expression is upregulated in pressure overload-induced heart failure. To examine the role of the TSPO, we generated cardiac-specific, conditional TSPO-KO mice using the Cre-loxP system as described in Methods and illustrated in Fig. 1A 32 . DNA was taken from tails, amplified using polymerase chain reaction, processed through DNA electrophoresis, and visualized under UV lighting (Fig. 1B). WT animals showed bands at 935 base pairs (bp), while KO animals displayed bands at 201 bp confirming that the homozygous line for Tspo containing loxP sites was established.
To determine if the TSPO has a role in HF, we induced pressure-overload by surgical transverse aortic constriction (TAC) in [11][12] week-old mice. Representative images in Fig. 1C depict an aorta from a sham mouse (left) and an aorta from a mouse after one week of TAC (right). The degree of constriction was verified by color and pulse wave measurements of flow velocity in both carotid arteries one week after surgery (Fig. 1D,E), with values of 6.3 ± 0.4 mm/sec and 6.2 ± 0.4 mm/sec in wild-type TAC (WT TAC) and TSPO knockout TAC (KO TAC) mice, respectively (Fig. 1E). These measurements demonstrated identical increases in flow velocity following a week of TAC, indicating similar degrees of aortic constriction in the WT and KO animals. After 8 weeks, hearts were harvested and mitochondria were isolated from the whole hearts. As seen in Fig. 1F,G, 8 weeks of pressure-overload increased TSPO protein expression in the mitochondrial fraction of hearts from WT animals by 164% (p < 0.01, n = 4 hearts) over sham controls (WT sham, n = 4 hearts), an increase that was abrogated in the KO animals. As demonstrated by immunohistochemistry ( Supplementary Fig. 1), very low TSPO levels are likely due mainly to TSPO in cells types other than cardiac myocytes where TSPO upregulation by inflammation has been reported previously 33,34 . These results provide the first evidence that cardiac TSPO is upregulated in failing hearts following TAC.
TSPO ablation slows progression toward heart failure following TAC surgery. To determine if TSPO has an adaptive or maladaptive effect on the development of pressure-overload HF, the anatomic and physiologic responses to pressure-overload were assessed weekly by echocardiography monitoring in unconscious mice. Sham animals, both WT and KO, did not exhibit morphologic or performance changes in the 8-week observation period after sham surgery. After TAC surgery, the increases in wall thickness after one week were similar in the WT and KO animals ( Fig. 2A), which further supports a similar hemodynamic load in the two TAC groups. Following the initial hypertrophic compensation to the increase in afterload with TAC, the progression toward cardiac dilatation was significantly slower in KO TAC than WT TAC as manifest in both left ventricular end-diastolic dimension (LVEDD) and left ventricular end-systolic dimension (LVESD) (Fig. 2B,C). Similarly, WT TAC mice had more cardiac dilation than KO TAC mice, illustrated in the short axis m-mode images (Fig. 2D) and significantly higher ventricular cavity area (Fig. 2E, p < 0.01). These changes demonstrate the development of a dilated cardiomyopathy in the WT TAC animals, with relative preservation of chamber size in the KO TAC animals.
In parallel with the differences in wall thickness and ventricular dimensions, WT TAC animals developed reduced systolic function. As seen in Fig. 2F, WT animals subjected to TAC had a significant reduction in ejection fraction (EF) from 63.9 ± 1.0% at baseline to 34.5 ± 2.4% after 8 weeks (p < 0.01). In contrast, EF was minimally reduced in the KO TAC animals, from 65.5 ± 1.5% at baseline to 55.8 ± 2.0% after 8 weeks of TAC, resulting in a significant difference in final ejection fraction between the groups (p < 0.001, Fig. 2F, Table 1). Representative images of m-mode echocardiography demonstrate these changes in ventricular dimensions and ejection fraction (Fig. 2G). Together, these in vivo data indicate that knocking out TSPO limited HF development and preserved cardiac function with pressure overload.
TSPO deletion protects the heart against pressure overload-induced hypertrophy and pathological cardiac remodeling. To further characterize the severity of HF development, anatomical and histological analyses were conducted. Consistent with the echocardiographic findings, measurement of heart weight-to-tibial length (HW:TL) ratios revealed greater TAC-induced cardiac mass (due to greater dilation) in WT compared to KO animals (16.4 ± 1.0 vs 14.0 ± 1.2 mg/mm, p < 0.05, n = 14-18 hearts, Fig. 3A Abnormalities in histology paralleled the gross anatomic changes, including higher collagen formation in WT TAC mice than KO TAC mice (9.2 ± 0.8% and 3.9 ± 0.4%, respectively; n = 3-4 hearts p < 0.01, Fig. 3C,D). These data show that KO TAC animals were relatively protected from TAC-induced cardiac remodeling. In addition to the anatomical, functional, and structural changes, WT TAC mice had more clinical evidence of pulmonary congestion confirmed by a higher lung weight-to-tibial length (LW:TL) ratio (13.8 ± 1.9 mg/mm,   Fig. 3E). In contrast, KO TAC mice did not experience a significant increase in the LW:TL ratio compared to sham controls. WT TAC mice exhibited significantly lower body weight (28.5 ± 0.7 g) compared to sham controls (31.0 ± 0.3 g, p < 0.01), while KO TAC mice did not (31.8 ± 0.8 g, Fig. 3F). Consistent with clinical signs of HF, there was a significant increase in atrial natriuretic peptide hormone (ANP) mRNA levels in both TAC groups when compared with their corresponding sham controls (Fig. 3G), although the WT TAC mice experienced a greater increase in ANP than did the KO TAC mice (p < 0.05). Furthermore, the size of cardiac myocytes, determined by confocal fluorescent microscopy imaging, was significantly increased in WT TAC (n = 80, P < 0.001) animals compared to WT shams (n = 56) (Fig. 3H). Cardiomyocyte cell size was also significantly smaller in KO TAC (p < 0.005, n = 66) compared to WT TAC, indicating that TSPO-KO also limited pathological myocyte hypertrophy. Taken together, TSPO KO mice develop compensated hypertrophy in response to TAC, whereas WT mice progress fully to HF in the same 8 week period (based on EF, chamber size, cardiac fibrosis, lung weight and ANP expression).
Cardiac TSPO-KO prevented TAC-induced impairment of mitochondrial Ca 2+ uptake. HF has been characterized by an alteration in mitochondrial Ca 2+ uptake 14,16 . Our previous work using hemin-induced TSPO activation suggested that TSPO can suppress mitochondrial Ca 2+ uptake 21 . The current study used the WT vs. TSPO-KO to test whether the dramatic rise in TSPO in TAC might contribute to reduced mitochondrial Ca 2+ in TAC. Mitochondrial Ca 2+ uptake amplitude (assessed using X-Rhod-1), was similar in myocytes from WT and TSPO-KO mice under sham conditions ( Fig. 4A-C). However, TAC caused a dramatic 86-93% reduction in WT mouse mitochondrial Ca 2+ uptake at the three [Ca 2+ ] studied ( Fig. 4B,C). In sharp contrast, TSPO-KO mice exhibited nearly unaltered mitochondrial Ca 2+ uptake after TAC (<20% reduction at 1.35 µM Ca 2+ , no change at 2 µM Ca 2+ and slight increase at 10 µM Ca 2+ ). Moreover, after TAC, mitochondrial Ca 2+ uptake in KO myocytes was 7.7-9.5-fold higher than that in WT (black vs. green bars in Fig. 4C). We conclude that normal physiological levels of TSPO expression do not limit acute mitochondrial Ca 2+ uptake, but that the high levels of TSPO in TAC-WT may contribute to reduced mitochondrial Ca 2+ uptake. Thus, preventing the TAC-induced increase in TSPO expression in HF by conditional cardio-specific silencing of TSPO preserved mitochondrial Ca 2+ uptake. Since opening of the mitochondrial permeability transition pore (mPTP) has been associated with Ca 2+ overload 35 , increased ROS 36 and resultant myocyte death, the effect of TSPO-KO on mPTP was examined using calcein release from mitochondria [37][38][39] . Figure 4D shows that both WT and TSPO-KO sham had a similar low mPTP activity that increased progressively (and similarly) as [Ca 2+ ] was raised. Myocytes from TAC hearts (both WT and TSPO-KO) exhibited higher mPTP opening at any given [Ca 2+ ], but no differences in mPTP activity were seen between TSPO-KO and WT TAC myocytes (Fig. 4D). These data indicate that TAC caused increased mPTP activation, but TSPO ablation did not prevent this effect. Of note, the high mPTP rate in WT TAC mitochondria occurred at a lower free mitochondrial [Ca 2+ ] (Fig. 4C), so this might reflect a higher ROS level in the HF myocytes that would promote mPTP at lower free [Ca 2+ ]. We also measured basal levels of total mitochondrial Ca 2+ content measured in isolated mitochondria lysed in 0Ca 2+ /0Na + conditions. Baseline total mitochondrial Ca 2+ was elevated in both WT TAC and KO TAC compared to their corresponding sham controls (Fig. 4E); as higher total mitochondrial Ca 2+ content also promotes mPTP 40 , this might explain the similar mPTP in WT TAC vs. KO TAC mitochondria.
ATP generation was reduced in WT TAC mice, but preserved in TSPO-KO TAC mice. Ca 2+ entry into the mitochondria can regulate energy production via activation of several key dehydrogenases 3,14,16 , and enhancing mitochondrial Ca 2+ in the context of HF can maintain NADH oxidation 3 and, hence ATP production 6   compared to WT sham cardiomyocytes in the presence of pyruvate/glutamate (p < 0.01), but not in the presence of methyl-succinate. This finding suggests that complex I mediated respiration drove significantly more ATP production in KO sham cardiomyocytes vs. WT. With methyl-succinate to drive complex II mediated respiration, there was no difference between WT and KO shams (Fig. 5C,D). In WT mouse myocytes, TAC nearly abolished the ATP rise induced by either pyruvate and glutamate or methyl-succinate. However, in TSPO-KO TAC myocytes, ATP levels achieved were higher than in the WT animals after TAC for either substrate (albeit still lower than sham).
Corresponding to the improved ATP levels in the TSPO-KO TAC animals (vs. WT TAC), TSPO ablation resulted in restoration of the mitochondrial reserve capacity in TSPO-KO TAC cardiomyocytes utilizing substrate for either mitochondrial complex I or II (Fig. 5B,D, right panels). Furthermore, the reserve capacity was higher in sham TSPO-KO vs. WT mice, but only for pyruvate/glutamate as substrates for mitochondrial complex I.
Together, these data confirm the adverse effect of TAC-induced heart failure on ATP production 37,41 and suggest that the effects of the TSPO-KO cardioprotection after TAC may be mediated by enhancement of energy production via complex I.
Mitochondrial FAD redox potential remained more reduced in TSPO-KO TAC animals. It has been previously reported that restoring mitochondrial Ca 2+ uptake in the failing hearts plays an essential role  However, there was no significant difference in mPTP activity between the two TAC groups at all Ca 2+ levels. n = 3 hearts, 10-25 cells. Data expressed as mean ± SEM. *p < 0.05 vs corresponding sham group, **p < 0.01 vs corresponding sham group, † p < 0.05 vs WT sham. (E) Measurement of basal free Ca 2+ in isolated mitochondria from 4 experimental groups revealed that there was a small increase (P < 0.05) in TAC groups compared to the corresponding sham controls (n = 6 in each group), however no difference was observed between WT (n = 10) and KO TAC (n = 11, P = 0.3). in maintaining the NADH/FAD redox balance 3 . We therefore hypothesized that the inability of cells to maintain normal mitochondrial Ca 2+ uptake in WT TAC mice could contribute to dysregulation of the redox potential and electron flux. To test that hypothesis, we monitored mitochondrial redox potential by FAD/FADH 2 autofluorescence under conditions of increased cell work induced by electrical field stimulation at 4 Hz and subsequent isoproterenol (ISO) stimulation (Fig. 6A). Quantitative analysis in WT TAC cardiomyocytes revealed nearly maximal FAD/FADH 2 ratio (oxidized) at basal conditions, significantly more oxidized than WT sham (0.94 ± 0.01, n = 10 vs 0.56 ± 0.04, n = 14, p < 0.001, Fig. 6B). Most importantly, we found WT TAC myocytes were not able to maintain a redox state when cells were exposed to high work load induced by 4 Hz stimulation (FS) with isoproterenol. That is, the degree of FAD reduction, which inversely correlates with NADH oxidation, was nearly ten-fold higher in WT TAC cardiomyocytes vs. WT sham cardiomyocytes (Fig. 6C). In TSPO-KO TAC cardiomyocytes, there was still significant reduction in FAD/FADH 2 with stimulation plus isoproterenol, but this was significantly smaller when compared to WT TAC myocytes (−0.16 ± 0.1, n = 10 vs −0.29 ± 0.2, n = 10, p < 0.001). In both WT and KO sham cardiomyocytes, FAD/FADH 2 levels were well maintained during 4 Hz FS and ISO stimulation (Fig. 6C). These data suggest a possible impairment in the electron flow at the level of the respiratory chain in WT TAC mice which was partially restored in KO TAC mice.
Complex I activity was impaired in WT TAC and associated with decreased mitochondrial membrane potential and enhanced ROS generation. We 41 and others 42 previously reported that electron transfer during HF is impaired at the level of the mitochondrial complex I, which can increase ROS accumulation. We found that WT TAC mice had reduced complex I activity compared to WT sham mice (7.4 ± 0.3 vs 8.5 ± 0.2 mOD/min, p < 0.01), whereas KO TAC mice had preservation of complex I activity (similar to KO sham). Importantly, we found that complex I activity after TAC was much higher in TSPO-KO vs. WT mice (9.5 ± 0.3 vs 7.4 ± 0.3 mOD/min, p < 0.01, Fig. 6D). Together, these data show that complex I activity is impaired in WT TAC mice, but preserved in KO TAC mice. To examine whether KO TAC mice have less ROS generation compared to WT TAC mice, superoxide generation was measured in cardiomyocytes using MitoSOX Red fluorescent dye 38,39 . Figure 6E shows that WT cardiomyocytes had significantly higher superoxide production both under sham conditions and after TAC, following electrical field stimulation at 0.5 Hz or beta-adrenergic stimulation with 100 nM isoproterenol. Notably, superoxide generation in cardiomyocytes from KO TAC animals was similar to that of sham animals (1.02 ± 0.01, n = 10 vs 1.04 ± 0.01, n = 19, p = 0.1). Together, these data show that ROS generation was increased by HF in WT TAC mice, and limited to control levels in TSPO-KO animals. Furthermore, we found that mitochondrial membrane potential (ΔΨ m ) was significantly decreased in WT TAC cardiomyocytes, but not in TSPO-KO TAC (Fig. 6F). Application of 10 µM rotenone, a mitochondrial complex I inhibitor, led to a significant decrease in ΔΨ m in WT sham from 1516 ± 89 to 858 ± 43 a.u., n = 39, p < 0.001 but not in WT TAC myocytes where basal level was already at 911 ± 24 a.u., n = 17 (not shown), confirming that the inability of mitochondria to maintain ΔΨ m in HF was related to impairment of the mitochondrial complex I. In TSPO-KO TAC mice, basal levels of ΔΨ m were significantly closer (1404 ± 72 a.u., n = 24) to that observed in sham controls (1567 ± 81 a.u., n = 39). We conclude that TSPO ablation improved mitochondrial oxidative phosphorylation via restoration of the electron flux flow at the level of the mitochondrial complex I.
Mitophagy was impaired in WT TAC, but preserved in KO TAC cardiomyocytes. Efficient degradation and replacement of dysfunctional mitochondria is essential for cardiac myocyte survival. Damaged mitochondria are eliminated via mitochondrial autophagy or mitophagy. Upon mitochondrial depolarization, the serine/threonine kinase PTEN-inducible kinase 1 (PINK1) accumulates on the mitochondrial surface, recruiting the E3 ubiquitin ligase (Parkin) to the outer mitochondrial membrane from the cytosol. Parkin then polyubiquitinates mitochondrial membrane proteins, marking them for degradation and recognition by the adaptor protein SQSTM1/p62. SQSTM1/p62 binds to the microtubule-associated protein 1A/1B-light chain 3 (LC3) on the autophagosome and to the ubiquitinated mitochondrial proteins, tethering the autophagosome to the mitochondrion 7 . The autophagosome then engulfs the mitochondrion and fuses with lysosomes, allowing mitochondrial proteins to be degraded and recycled.
We evaluated mitophagy by monitoring mitochondrial Parkin accumulation and LC3-mediated autophagosome formation in cardiac myocytes using confocal microscopy. As shown in Fig. 7A-C, WT TAC myocytes exhibited increased mCherry-Parkin puncta accumulation (marking damaged mitochondria) which was not accompanied by an increase in LC3-mediated autophagosome formation. This indicates impairment of mitophagy flux in TAC-induced HF. In contrast, mitophagy flux was enhanced in TSPO KO TAC mice (i.e. lower accumulation of Parkin puncta, but more LC3-autophagosomes). While lower levels of both mCherry-Parkin and LC3-GFP puncta were observed in WT sham myocytes, both were increased in TSPO KO sham myocytes, consistent with potential TSPO-induced suppression of mitophagy even in control myocytes. These results were further confirmed by blots assessing key proteins involved in mitophagic signaling. There were no significant changes in PINK-1 expression among the groups (Fig. 7D,E); however, WT TAC had reduced expression of TOMM22, which imports PINK1 into the mitochondrial matrix for cleavage by mitochondrial processing proteinase (MPP) and presenilin-associated rhomboid-like (PARL; Fig. 7D,G). Conversely, expression of the adaptor protein SQSTM1/p62 was greatest in WT TAC, but this did not correlate with LC3-II levels (Fig. 7D,G). In TSPO KO TAC myocytes, TOMM22 levels were partially increased while p62 levels were lower compared to WT TAC samples; however, the conversion of LC3-I to LC3-II was improved (Fig. 7D-G). As LC3-II is a lipidated form of LC3-I that localizes on phagophores and autophagosomes, the LC3-II/LC3-I ratio is indicative of mitophagic flux and was highest in KO TAC myocytes. Taken together, these measurements demonstrate that there was an increase in the removal of damaged mitochondria via mitophagy in KO TAC cardiomyocytes, but that this mitophagic flux was impaired in WT TAC cardiomyocytes.

Discussion
Our results yield several key findings with respect TSPO function in pressure-overload induced heart failure (TAC). First, we demonstrated that TAC resulted in a 3-fold increase in expression of the TSPO in WT animals (Fig. 1F,G), an upregulation of TSPO that parallels that seen in cardiac ischemia and reperfusion 29 and under stress conditions in other tissues 27,28,30 . Pressure overload for 8 weeks in WT animals resulted in the expected changes in cardiac morphology, function, and metabolism associated with HF (see Figs 2 and 3, and Table 1). These included an increase in cardiac chamber size, a 46% decrease in ejection fraction, an increase in myocyte hypertrophy and fibrosis, and clinical signs of HF such as pulmonary congestion and an increase in atrial natriuretic peptide.
Second, this study demonstrated that cardiac-specific conditional KO of TSPO substantially abrogated the deleterious effects of TAC. By preventing TSPO upregulation during TAC, there was well-preserved left ventricular function (only 15 vs. 46% reduction in ejection fraction in TSPO-KO vs. WT animals). In addition, TSPO-KO decreased associated HF markers, with lower levels of ventricular dilation, pulmonary congestion, ANP levels and extents of cardiac fibrosis and hypertrophy.
Third, TSPO-KO had several salutary effects on mitochondrial function in TAC mice, including normalization of mitochondrial Ca 2+ uptake (Fig. 4) and ATP production (Fig. 5), lower production of ROS, maintenance of mitochondrial membrane potential, redox state (Fig. 6) and preservation of mitophagy (Fig. 7). Remarkably, the TSPO-KO at baseline (sham) was not appreciably different than the WT sham in any of the parameters that we assessed. We conclude from this observation that physiological levels of TSPO expression do not have overt negative effects, but that the dramatic upregulation of TSPO in TAC (and other stress conditions) has impactful negative effects that worsen HF progression and function. It is possible that the TSPO upregulation exceeds the stoichiometry of its normal physiological targets or partners, allowing it to interact preferentially with new targets that mediate pathological effects.
As a modulator of VDAC 43 , TSPO can affect inward mitochondrial Ca 2+ flux. We have previously shown that the TSPO ligand protoporphyrin IX (hemin) reduced Ca 2+ uptake in isolated mitochondria 21 , as was shown with another TSPO ligand PK11195 44,45 . Upregulation of TSPO expression during HF coincided with a decrease in mitochondrial Ca 2+ uptake, which was normalized in TSPO-KO mice (Fig. 4). Together, these data are consistent with excess TSPO being a negative modulator of mitochondrial Ca 2+ entry, potentially by interacting with and modulating VDAC open probability 21 . VDAC operates at the OMM, and controls Ca 2+ entry into the intermembrane space in series with the mitochondrial Ca 2+ uniporter (MCU) on the IMM 46,47 . A recent study 13 demonstrated that reducing Ca 2+ flux via VDAC severely affected heart contractility in zebra fish, while interventions to increase Ca 2+ flux through the OMM restored normal contraction of the heart. These data highlight the possible impact of OMM proteins such as VDAC and TSPO in the regulation of mitochondrial Ca 2+ uptake and cardiac function.
The remarkable beneficial effects of the TSPO-KO (at the mitochondrial, cardiac structure and functional level) are consistent with TSPO upregulation being a central amplifier of the progression from compensated hypertrophy (seen in TSPO-KO mice after TAC) to more full-blown HF (seen in WT TAC). A primary effect of TSPO overexpression to inhibit mitochondrial Ca 2+ uptake could lead to most of the mitochondrial sequelae (on ATP and ROS production redox state and membrane potential), but we cannot rule out additional primary TSPO effects. Better maintenance of mitochondrial and energetic function by TSPO-KO, as seen here during TAC, could maintain cardiac function and slow the progression to HF 48 . Notably, the TSPO-KO TAC mice (vs. WT TAC) had higher respiratory reserve values and higher activity of the mitochondrial complex I, indicating preservation of mitochondrial bioenergetics. These findings agree with those of Gatliff et al. 49 who showed that CF35 cells with siRNA-induced TSPO knockdown showed greater mitochondrial ATP production compared to control cells, while an overexpression of TSPO decreased ATP. Taken together, our data show that TSPO-KO mice were resistant to the progression of HF, and this protective effect was most likely due to a combination of enhanced mitochondrial Ca 2+ uptake and energy production.
This situation contrasts with the known deleterious effect of mitochondrial Ca 2+ overload under acute ischemia/reperfusion 39,50 . With ischemia/reperfusion, significant Ca 2+ overload during reperfusion, mediated by sarcolemmal Na + /Ca 2+ exchanger 50,51 , can result in opening of the mitochondrial permeability transition (mPTP), leading to mitochondrial swelling, membrane rupture and myocyte death. In contrast to the detrimental effect of Ca 2+ overload under this acute pathologic challenge, the better maintained physiological mitochondrial Ca 2+ uptake during chronic TAC in TSPO-KO promoted improved mitochondrial energetics and cardiac function. Notably, mPTP, which is a key turning point in above ischemia/reperfusion damage, was not influenced by TSPO-KO (Fig. 4D). Consistent with these findings, cyclosporine A treatment, a known mPTP inhibitor, while limiting deterioration of mitochondrial function in HF animals, was not associated with a commensurate improvement in cardiomyocyte Ca 2+ handling or contractility 52 .
HF has been shown to result in increased oxidative stress 38,53 and a higher FAD/FADH 2 ratio 54 . These abnormalities have been shown to result in greater myocyte loss and dysfunction 55 , observations consistent with those observed with progressive HF in the WT TAC animals in the current study. In contrast, TSPO-KO mice demonstrated preserved function and less fibrosis, coupled with lower oxidative stress. The lower oxidative stress in the KO animals is likely due to the preservation of complex I activity 38,42 , as well as the higher ATP generation resulting in activity of key Ca 2+ -dependent dehydrogenases 56 .
It is increasingly clear that cardiac mitochondrial ROS production is dynamically regulated by Ca 2+ , ADP and the redox state of mitochondrial pyridine nucleotides 57 . One aspect is that ROS production can be increased by mitochondrial [Ca 2+ ] that is either too high or too low (involving NADH, NADPH and FADH 2 ). This "Redox-optimized ROS balance" hypothesis 58,59 postulates that the physiological steady state in cardiac mitochondria is tuned to an intermediate redox state that prevents excessive ROS formation at the ETC under highly reduced conditions [60][61][62] , but also maintains anti-oxidative capacity under highly oxidized conditions 58,59,63 . Our data fit with this concept, as basal FAD/FADH 2 ratio was 0.56 under sham conditions, with minimal ROS generation. During HF, however, Ca 2+ uptake was depressed and FAD/FADH 2 was severely oxidized, with a concordant increase in ROS generation. TSPO ablation restored normal Ca 2+ flux to mitochondria, improved complex I activity, shifting redox index toward an "intermediate state" and reduction in ROS generation (see Fig. 6), which ultimately resulted in the improvement of cardiac function in vivo.
Our results are also consistent with results of Gatliff et al. 23 , in mouse embryonic fibroblasts and canine mammary gland epithelia. They concluded that TSPO, via its interaction with VDAC1, causes mitochondrial uncoupling, increased ROS production and inhibition of mitophagy by limiting the Parkin-dependent ubiquination. This fits with our data from KO TAC mitochondria, myocytes and hearts, where mitochondrial coupling in complex I was enhanced, ROS levels were lower and mitophagy was preserved vs. WT TAC. Moreover, this enhanced mitochondrial health and Ca 2+ handling in TSPO KO hearts was associated with dramatically improved cardiac functional response to pressure overload, preventing transition from compensated hypertrophy to HF. These results should also further stimulate studies aimed at clarifying the detailed molecular mechanisms by which elevated levels of the OMM protein TSPO in HF mediate multiple pathological changes in mitochondrial function and turnover.

Limitations
As with any murine animal model, one can neither extrapolate these findings to other HF causes (e.g. myocardial infarction, anthracycline toxicity), nor to human or large animal models. TSPO has numerous reported effects on steroid transport and lipid metabolism 64 , porphyrin metabolism 65 , and fatty acid oxidation 66 . While beyond our present scope, one or more of these pathways could also be involved in the protective effect of TSPO-KO in HF besides the mechanisms reported in this study. are taken as 100%. n = 3-5 animals, 8-15 cells used for each group in confocal microscopy experiments. Data expressed as mean ± SEM. *p < 0.05 vs corresponding sham group, **p < 0.01 vs corresponding sham group, † p < 0.05 vs WT sham, ‡ p < 0.01 vs WT sham, # p < 0.05 vs WT TAC, ## p < 0.01 vs WT TAC.

Conclusion
Taken together, these experiments have shown that HF induced by pressure overload resulted in increased expression of TSPO. Abrogating this increase by genetic ablation resulted in cardioprotection reflected by preserved ventricular function, cellular histology and clinical signs of heart failure. This cardioprotective effect was associated with preserved mitochondrial Ca 2+ uptake and complex I-mediated respiration leading to the normalization of ATP production oxidative stress and mitochondrial quality control. These findings are the first observations that modulation of the mitochondrial outer membrane translocator protein could be a target for the prevention or treatment of HF.

Methods
All protocols were in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication NO. 85-23, revised 1996) and approved by the University of California Davis Institutional Animal Care and Use Committee.
Model of pressure overload hypertrophy/heart failure. C57Bl/6J mice were obtained from The Jackson Laboratory (Sacramento, CA). The aorta was surgically constricted as previously described 67 . A 5-mm thoracotomy was made and the transverse aorta was isolated and a suture looped around a blunt 27-gauge needle, yielding a transverse aortic diameter of approximately 0.4-mm. The sham procedure was identical except that the aorta was not ligated. Echocardiography. Ventricular function was monitored using 2D echocardiography on a weekly basis using a Visualsonics Vevo 2100 echocardiography machine and a MS 550D probe . Animals were anesthetized using isoflurane (0.75-1%). 2D echocardiography was performed at weekly intervals until the mice were ready to be sacrificed after 8 weeks of sham or TAC operation. Ejection fraction was determined from 2D images by manual determination of endocardium and epicardium from end-diastolic and end-systolic images.

Generation of cardiac-specific conditional TSPO
Morphometric Measurements. Eight weeks after surgery, mice were anaesthetized with isoflurane and sacrificed by cervical dislocation. Hearts were excised, weighed and immediately fixed by perfusion in a retrograde manner with 4% paraformaldehyde in phosphate-buffered saline (PBS) followed by immersion fixation in buffered 4% paraformaldehyde. Hearts were kept at 4 °C before the tissue was processed. Lungs were extracted, blotted dry, and weighed. Increased cardiac mass was assessed by comparing the ratios between heart weight-to-tibial length. Pulmonary congestion was assessed by comparing lung weight-to-tibial length ratios.
Fibrosis Measurement. Extracted hearts were rinsed with PBS, placed in 10% Neutral Buffered Formalin, and paraffin-embedded. Hearts were sliced into 5 µm sections and collagen fibers were stained with Picro-Sirius Red stain. Images were taken with an Olympus BH-2 Bright-field Microscope using the 20x objective lens. Quantification of fibrosis was performed using ImageJ software.
Cell Isolation. Left ventricular myocytes were isolated using a Langendorff perfusion system 38  the same intracellular solution but without saponin, and extramitochondrial Ca 2+ ([Ca 2 + ] em ) was elevated to initiate mitochondrial Ca 2+ uptake. X-Rhod-1 was excited with the 543-nm line of a green HeNe laser, and emitted fluorescence was measured at 552-617 nm. Mitochondrial X-Rhod-1 fluorescence intensity (F) in each experiment was normalized to the level of fluorescence recorded before stimulation (F 0 ) but after cell permeabilization. Changes in [Ca 2+ ] m are expressed as ΔF/F0, where ΔF = F − F 0 .
Basal Mitochondrial Ca 2+ Measurement. Mitochondria were isolated from left ventricles using a standard differential centrifugation method. Briefly, minced hearts were suspended in mitochondrial isolation buffer (IB 1 ) in 67 mM sucrose, 50 mM Tris/HCl, 50 mM KCl, 10 mM EDTA, and 0.2% BSA with a pH = 7.2, homogenized with a glass Teflon pestle, and subsequently centrifuged at 700 g for 10 min at 4 °C. The supernatant was collected and then spun at 8000 g for 10 min at 4 °C. The pellet was then resuspended in ice-cold buffer (IB 2 ) containing 250 mM sucrose, 3 mM EGTA/Tris, and 10 mM Tris/HCl with a pH = 7.2, followed by centrifugation at 8000 g for 10 min at 4 °C. Mitochondria were then resuspended in IB 2 . Protein concentration was measured using the Bradford Assay method. Basal mitochondrial calcium was measured in isolated mitochondria using Molecular Devices Spectra Max M5 fluorometer in combination with Ca 2+ indicator Rhod-2. A calibration curve was generated in a 96-well plate with Ca 2+ suspended in deionized water, ranging from 0 µM to 3 µM Ca 2+ . 10 µg of isolated mitochondria were spun down at 8000 g at 4 °C for 10 minutes. The IB 2 was then removed from the pelleted mitochondria, and replaced with water to lyse the mitochondria. All wells were incubated with Rhod-2 for 30 minutes at 37 °C and then measured using a fluorescent plate reader at Ex/Em 552/581 nm. ATP Measurements. ATP measurements were performed indirectly via the free magnesium (Mg 2+ ) concentration using the fluorescent dye mag-fluo-4 (Invitrogen, ThermoFisher Scientific, Grand Island NY) 37,68 . Since free [Mg 2+ ] i is kept constant within a rather narrow range, ATP hydrolysis leads to concomitant increase in free [Mg 2+ ] i as measured with fluorescent Mg 2+ indicators such as mag-fluo-4. Therefore, an increase in mag-fluo-4 fluorescence indicates a decrease in ATP concentration. For ATP measurements myocytes were loaded with 10 µM mag-fluo-4 (λ ex = 488 nm, λ em = 565-605 nm) for 30 min at 37 °C. All data from these measurements are expressed as R = 1 − F/F 0 . mPTP activity. mPTP activity was monitored in permeabilized myocytes loaded with 5 μM calcein/AM (λ ex = 488 nm, λ em = 510 nm) for 40 min at 37 °C 37,38 . Opening of mPTP resulted in the loss of mitochondriatrapped calcein (620 Da) and a decrease of fluorescence. At the end of each recording 10 μg/ml of the pore-forming antibiotic alamethicin was applied to provide a control measure for maximum calcein release from mitochondria. Loss of mitochondrial calcein induced by elevating Ca 2+ ] em was quantified as the rate of decline of fluorescence (d(ΔF)/dt) calculated from the linear fit to the initial decrease of calcein fluorescence. The rate of decline was normalized (%) to the basal decline of calcein fluorescence addition (taken as 100%) before [Ca 2+ ] em elevation. ROS Generation. ROS production was measured in intact myocytes loaded with 0.5 μM MitoSox Red (λ ex = 543 nm, λ em = 555-617 nm) for 30 min at 37 °C 37,68 . Changes in MitoSox Red fluorescence intensity (F) were normalized to the level of fluorescence recorded prior to stimulation (F 0 ), and expressed as ΔF/F 0 .
Mitochondrial Redox State. Flavin adenine dinucleotide (FAD)-linked protein autofluorescence (λ ex = 488 nm, λ em = 510 nm) was measured to evaluate mitochondrial redox state 37,38,68 . Data are presented as the ratio of oxidized FAD to reduced FADH 2 (FAD/FADH 2 ) calculated as (F − F min )/(F max − F min ) where F is the fluorescence intensity, and F min is the fluorescence obtained after addition of 4 mM NaCN (inhibits respiration and promotes maximal FAD reduction, i.e. FADH 2 formation), taken as 0%. F max is the fluorescence obtained after addition of 10 μM FCCP (stimulates maximal respiration, completely oxidizing the mitochondrial FADH 2 pool), taken as 100%.
Mitochondrial Complex I activity. We used colorimetric assays (Abcam, Cambridge MA) kits for these two complexes. Mitochondria were isolated using the standard procedure stated above. After isolation, proteins were extracted from the mitochondria using the provided detergent solution. Twenty micrograms of proteins were then loaded into wells pre-coated with antibodies specific for Complex I. Assay time, reagents and buffers added, were followed exactly as stated in the protocol booklets for each complex.
Mitophagy measurements. Adenoviral gene transfer in isolated myocytes that were kept in short-term culture (24-48 h; multiplicity of infection of 500) was used to express mCherry-Parkin and GFP-LC3 to monitor Parkin accumulation in mitochondria and LC3-mediated autophagosome formation 69 . Freshly isolated myocytes were plated on sterile, laminin-coated glass coverslips in PC-1 medium, supplemented with penicillin and streptomycin (50 μg/ml). PINK1, Parkin, SQSTM1/p62, TOMM22 and accumulation of LC3-II were also monitored by Western blot. diluted to a concentration of 1:1000, and secondary antibodies were conjugated with Horseradish Peroxidase was diluted to a concentration of 1:2000. Signals were detected using an Alpha FluorChem Imaging Systems 8900 and bands were quantified relative to housekeeping protein ( Supplementary Fig. 2).

SDS-PAGE and Western
Immunohistochemistry. Hearts were fixed in 4% PFA and then embedded in paraffin. 5 micrometer sections were cut and these sections were deparaffined in a series of xylene and alcohol concentrations 70 . Antigen retrieval was performed using heat in sodium citrate buffer. Blocking was done at room temperature for 2 hours using TBS with 10% goat serum and 1% BSA. Primary antibodies were incubated at 1:300 overnight. TSPO (red, ABCAM ab109497) and α-actinin (green, Sigma A7811) were the primary antibodies. Images were taken using a Zeiss LSM 700 Confocal Laser Scanning Microscope.
Statistical analysis. All data are expressed as mean ± standard error. Statistical significance of differences between experimental groups was determined using Student paired t test or two-way ANOVA, followed by Tukey post-test when appropriate. A value of p < 0.05 was considered statistically significant. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.