Enhanced NCLX-dependent mitochondrial Ca²⁺ efflux attenuates pathological remodeling in heart failure

Abstract

Mitochondrial calcium (_mCa²⁺) uptake couples changes in cardiomyocyte energetic demand to mitochondrial ATP production. However, excessive _mCa²⁺ uptake triggers permeability transition and necrosis. Despite these established roles during acute stress, the involvement of _mCa²⁺ signaling in cardiac adaptations to chronic stress remains poorly defined. Changes in NCLX expression are reported in human heart failure (HF) and models of cardiac hypertrophy. Therefore, we hypothesized that altered _mCa²⁺ homeostasis contributes to the hypertrophic remodeling of the myocardium that occurs upon a sustained increase in cardiac workload. The impact of _mCa²⁺ flux on cardiac function and remodeling was examined by subjecting mice with cardiomyocyte-specific overexpression (OE) of the mitochondrial Na⁺/Ca²⁺ exchanger (NCLX), the primary mediator of cardiac _mCa²⁺ efflux, to several well-established models of hypertrophic and non-ischemic HF. Cardiomyocyte NCLX-OE preserved contractile function, prevented hypertrophy and fibrosis, and attenuated maladaptive gene programs in mice subjected to chronic pressure overload. Hypertrophy was attenuated in NCLX-OE mice, prior to any decline in cardiac contractility. NCLX-OE similarly attenuated deleterious cardiac remodeling in mice subjected to chronic neurohormonal stimulation. However, cardiomyocyte NCLX-OE unexpectedly reduced overall survival in mice subjected to severe neurohormonal stress with angiotensin II + phenylephrine. Adenoviral NCLX expression limited _mCa²⁺ accumulation, oxidative metabolism, and de novo protein synthesis during hypertrophic stimulation of cardiomyocytes in vitro. Our findings provide genetic evidence for the contribution of _mCa²⁺ to early pathological remodeling in non-ischemic heart disease, but also highlight a deleterious consequence of increasing _mCa²⁺ efflux when the heart is subjected to extreme, sustained neurohormonal stress.

Graphical abstract

1. INTRODUCTION

Cardiomyocytes continuously fine-tune their metabolism to ensure an adequate ATP supply to fuel myofilament cross-bridge cycling and maintain ionic homeostasis in the face of constantly changing workloads. Calcium (Ca²⁺) released from the sarcoplasmic reticulum (SR) following each cardiac action potential activates myofilament cross-bridge cycling to enable contraction of the heart. Simultaneously, Ca²⁺ is transmitted to mitochondria to stimulate TCA cycle dehydrogenases and ATP synthesis. Thus, Ca²⁺ is proposed as the key second messenger that coordinates the parallel activation of myocardial energy consumption and production [1, 2].

Ca²⁺ released from the SR rapidly enters the mitochondrial matrix via the mitochondrial calcium uniporter channel complex (mtCU), a multi-protein channel that spans the inner mitochondrial membrane [3, 4]. The mitochondrial calcium uniporter protein (MCU) forms the Ca²⁺-selective pore of the mtCU and enables the efficient uptake of Ca²⁺ near SR release sites when the concentration rises above ~400nM [5, 6]. The mtCU is regulated by a number of subunits including EMRE, MICU1, MICU2, MICU3, MCUR1, and MCUB, which together control the permeation and Ca²⁺-sensitive gating of the mtCU [7–19]. Recent investigations employing genetic manipulation of MCU or the regulatory subunits of the mtCU have revealed that acute Ca²⁺ uptake through the uniporter is necessary to increase mitochondrial metabolism to support a sudden increase in cardiac work rate [20–25]. This phenomenon is critical for situations that require a rapid but transient increase in cardiac output, such as sympathetic fight-or-flight responses. In contrast, excessive _mCa²⁺ uptake through the mtCU is detrimental in settings of acute pathological cellular stress, such as occurs during cardiac ischemia-reperfusion (IR). Here, excessive _mCa²⁺ uptake and matrix Ca²⁺ accumulation triggers mitochondrial permeability transition and necrotic cell death [26] (reviewed in [27, 28]). Blockade of mtCU-dependent _mCa²⁺ uptake in the adult heart via inducible genetic disruption of MCU [21, 22] or inducible overexpression of the inhibitory mtCU subunit, MCUB [25], is sufficient to protect the heart from IR injury and cardiomyocyte death.

Precisely how altered _mCa²⁺ homeostasis affects the heart’s adaption to chronic rather than acute stress remains controversial. Altered cytosolic Ca²⁺ cycling including increased cellular Ca²⁺ influx; increased SR Ca²⁺ leak leading to diminished SR Ca²⁺ load and decreased cytosolic Ca²⁺ transient amplitude; slowed Ca²⁺ transient decay; and increased diastolic Ca²⁺ levels are reported in settings of cardiac pressure overload caused by aortic stenosis or chronic hypertension [29–39]. These same features of aberrant Ca²⁺ handling are observed in chronic heart diseases such as hypertrophic cardiomyopathy [40–48]. These reports show a clear association between altered cardiomyocyte calcium signaling, cardiac hypertrophy, and impaired contractile function. However, much less is known regarding whether such long-term alterations in cytosolic Ca²⁺ handling lead to coordinate changes in mitochondrial Ca²⁺ homeostasis, and what impact any chronic changes in _mCa²⁺ may have on the development or progression of non-ischemic heart disease.

Several studies suggest that mtCU activity is impaired in the end-stage failing heart [49, 50], and that moderately increasing _mCa²⁺ uptake in this context can improve contractile function [51]. In contrast, other reports indicate that cardiac MCU protein expression increases in experimental pressure overload-induced cardiac hypertrophy and in patients with aortic stenosis [52, 53]. These observations raise the possibility that _mCa²⁺ uptake and mitochondrial matrix Ca²⁺ are elevated when cardiac workload and cytosolic Ca²⁺ levels are chronically elevated, just as they are when SR Ca²⁺ leak is increased in the post-infarction failing heart [54]. The findings that cardiac _mCa²⁺ content is enhanced due to increased MCU expression and _mCa²⁺ uptake along with reduced NCLX expression and _mCa²⁺ efflux, and that heart size is increased in a mouse model of mitochondrial cardiomyopathy due to Tfam deletion [55], provide additional support for an association between _mCa²⁺ signaling and cardiac hypertrophy. Indeed, increased _mCa²⁺ content may help to fuel cardiac growth because it can stimulate TCA cycle flux [56–62] and respiratory complex activity [63, 64] to increase cellular ATP production needed to power anabolic reactions including protein synthesis. Chronic elevation of _mCa²⁺ in the face of sustained cellular Ca²⁺ stress can also contribute to deleterious myocardial remodeling and cardiac dysfunction by predisposing to _mCa²⁺ overload, cardiomyocyte death, and subsequent replacement fibrosis [28, 65–67]. Further complicating the question of how altered _mCa²⁺ homeostasis impacts the heart’s response to chronic pathological stress is evidence that diminished _mCa²⁺ content resulting from increased Na⁺-driven mitochondrial Na⁺/Ca²⁺ exchange in end-stage heart failure (HF) limits mitochondrial energetics, perturbs redox state, and further impairs cardiac contractility [68, 69].

Early experiments aiming to clarify the role of _mCa²⁺ signaling in the heart’s adaptation to chronic stress attempted to limit _mCa²⁺ accumulation by targeting rapid _mCa²⁺ uptake through the mtCU. However, neither constitutive nor inducible genetic Mcu deletion showed any effect to prevent cardiac hypertrophy or preserve contractile function in mice subjected to chronic pressure overload with transverse aortic constriction [22, 70]. These studies also yielded inconsistent results on the efficacy of genetic Mcu disruption in altering steady-state _mCa²⁺ content over the long-term, and so left the question of _mCa²⁺’s role in the heart’s response to chronic stress unresolved. Chronic pharmacological inhibition of MCU with ruthenium red was recently reported to attenuate cardiac hypertrophy and improve contractile function in response to transverse aortic constriction [53]. Interpretation of these findings is confounded, though, by the fact that ruthenium red can affect the activity of other cellular Ca²⁺ channels such as the L-type calcium channel and ryanodine receptor [71–74]. Given the challenges in specifically manipulating _mCa²⁺ homeostasis over chronic time scales via genetic or pharmacological targeting of MCU, we sought an alternative approach to perturb _mCa²⁺ signaling within the heart.

Our laboratory recently demonstrated that the mitochondrial Na⁺/Ca²⁺ exchanger, NCLX, is the primary physiological route for mitochondrial Ca²⁺ efflux in cardiomyocytes [75]. We also reported that NCLX expression modestly increases in end-stage ischemic hearts at the time of transplantation [75]. NCLX extrudes 1 Ca²⁺ ion from the mitochondrial matrix in exchange for the import of 3 Na⁺ ions, and constitutes the rate-limiting step of mitochondrial Ca²⁺ exchange [76–79]. Modulation of net mitochondrial NCLX activity therefore can exert a substantial influence over mitochondrial Ca²⁺ homeostasis, and so offers an alternative strategy to MCU modulation for investigating the role of _mCa²⁺ in cardiac responses to chronic stress. Here, we used mice with cardiomyocyte-specific overexpression of NCLX to test the hypothesis that _mCa²⁺ signaling contributes to pathological cardiac remodeling and HF in response to sustained elevations in cardiac workload and cytosolic Ca²⁺ cycling. We employed three surgical models of chronically increased hemodynamic or neurohormonal load on the heart and assessed the impact of cardiomyocyte NCLX overexpression on contractile function, cardiac hypertrophy, and fibrotic remodeling of the myocardium. We then utilized an in vitro model of cardiomyocyte hypertrophy to examine the mechanism by which NCLX overexpression limits pathological cardiac growth. Our findings support a critical role for _mCa²⁺ in driving early hypertrophic remodeling and subsequent contractile dysfunction in non-ischemic heart disease, but also indicate that limiting _mCa²⁺ accumulation can be fatal under conditions of extreme neurohormonal stress.

2. RESULTS

2.1. Transgenic expression of NCLX in cardiomyocytes protects against pressure overload-induced cardiac remodeling and failure.

We recently developed and characterized mice with cardiomyocyte-specific, doxycycline-controlled overexpression of NCLX [75] (Fig. 1A). This model exhibits a ~2-fold increase in cardiac NCLX protein expression in adult mice (Fig. 1B–C; full-length blots shown in Extended Fig. 1) and results in a ~38% increase in the _mCa²⁺ efflux rate during electrical pacing in intact, isolated adult cardiomyocytes, with no significant effect on the rate of _mCa²⁺ influx [75]. To test how increased capacity for _mCa²⁺ efflux affects the heart’s response to chronic hemodynamic stress, following the removal of doxycycline to allow for transgene activation we subjected mice with cardiomyocyte-specific NCLX overexpression (TRE-NCLX x αMHC-tTA, NCLX-Tg) and control mice (αMHC-tTA) to 12 weeks of pressure overload induced by transverse aortic constriction (TAC) (Fig. 1D). No significant differences in survival were observed among experimental groups (Supplemental Fig. S1A). αMHC-tTA mice exhibited a gradual increase in left ventricular end systolic dimension (LVESD) and a corresponding decrease in left ventricular fractional shortening (%FS) during the 12 weeks of TAC (Fig. 1 E–G). Both the increase in LVESD and decrease in %FS were attenuated in NCLX-Tg mice (Fig. 1 E–G). These findings suggest that enhancing cardiomyocyte _mCa²⁺ efflux is sufficient to prevent the decline in contractile function that typically occurs in response to a sustained increase in cardiac afterload. We further examined the impact of NCLX overexpression on pathological cardiac remodeling during pressure overload. The TAC-mediated increase in the heart weight-to-tibia length (HW/TL) ratio (Fig. 1H) and increase in cardiomyocyte cross-sectional area (CSA) (Fig. 1 I–J) observed in αMHC-tTA control mice, were both attenuated with NCLX overexpression (36% increase in αMHC-tTA HW/TL and 16% increase in NCLX-Tg HW/TL vs. sham; 63% increase in αMHC-tTA CSA and 26% increase in NCLX-Tg CSA vs. sham). Likewise, the extent of myocardial fibrosis resulting from 12 weeks of TAC was diminished with NCLX overexpression (Fig. 1 K–L). Together, these data indicate that increasing cardiomyocyte NCLX activity protects against pressure overload-induced pathological cardiac remodeling.

Figure 1: Cardiomyocyte-specific NCLX overexpression protects against pressure overload-induced HF and pathological remodeling.

A) Genetic approach for doxycycline-controlled, cardiomyocyte-specific expression of a human NCLX transgene in mice. TRE-NCLX mice expressing human NCLX cDNA under the control of a tetracycline-responsive promoter were crossed to αMHC-tTA (cardiomyocyte-specific, doxycycline-off) transgenic mice [75]. All mice were fed doxycycline-containing chow during development and through three weeks of age to allow for normal cardiac development. NCLX transgene expression was induced by removal of doxycycline at the time of weaning at 3 weeks of age. B-C) Western blots for NCLX protein expression in heart lysates from adult αMHC-tTA and TRE-NCLX x αMHC-tTA mice, and quantification of NCLX expression normalized to total OXPHOS complexes I-V as a mitochondrial loading control. Data analyzed by t-test. *P<0.05. (n=4 mice/group). D) Timeline of transverse aortic constriction (TAC) experimental protocol. DOX, doxycycline administration. Left ventricular end diastolic dimension (LVEDD) (E), end systolic dimension (LVESD) (F), and percent fractional shortening (%FS) (G) over 12 weeks after TAC or sham surgery. Data analyzed by 2-way ANOVA with Tukey’s post-hoc test. *P <0.05, **P<0.01, ***P<0.001 αMHC-tTA TAC vs. TRE-NCLX x αMHC-tTA TAC. ^#P <0.05, ^####P<0.0001 αMHC-tTA sham vs. TAC. (n=7–18 mice / group). H) Heart weight-to-tibia length ratio (HW/TL) 12 weeks post sham or TAC surgery. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05 vs. sham. (n=7–18 mice / group). I) Left ventricular tissue at 12 weeks post sham or TAC surgery, stained with wheat germ agglutinin (WGA, red) to delineate sarcolemma and with DAPI (blue). Scale bars = 50 µm. J) Quantification of cardiomyocyte cross-sectional area (CSA). Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^###P<0.001 vs. sham; *P < 0.05 αMHC-tTA vs. TRE-NCLX x αMHC-tTA. (n=3–5 mice / group). K) Masson’s trichrome stain for myocardial collagen deposition (blue) at 12 weeks post-surgery. Scale bars = 100 µm. L) Quantification of fibrotic area as percent of tissue area. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01 vs. sham; *P < 0.05 αMHC-tTA vs. TRE-NCLX x αMHC-tTA. (n=3–5 mice / group). qPCR quantification of mRNA expression of fetal (M-N) and pro-fibrotic (O-P) genes in hearts of mice 12 weeks after sham or TAC surgeries. Nppa, natriuretic peptide type A; Acta1, α-skeletal muscle actin; Postn, periostin; Spp1, osteopontin. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05, ^##P<0.01 vs. sham; *P < 0.05, **P < 0.01 αMHC-tTA vs. TRE-NCLX x αMHC-tTA. (n=3–7 mice / group).

Examination of cardiac gene expression revealed that 12 weeks of TAC increased the expression of fetal genes such as atrial natriuretic peptide (Nppa or ANP), α-skeletal muscle actin (Acta1), and brain natriuretic peptide (Nppb or BNP) (Fig. 1 M–N; Supplemental Fig. S1B) in control αMHC-tTA mice. The ratio of α-myosin heavy chain (Myh6) to β-myosin heavy chain (Myh7) gene expression in αMHC-tTA hearts was also decreased by 12 weeks of TAC (Supplemental Fig. S1D). The expression of such developmental genes is consistent with the re-activation of the fetal gene program that is classically observed in pathological cardiac hypertrophy [80]. The induction of Nppa, Acta1, and Nppb by 12 weeks of TAC was attenuated in NCLX overexpressor mice (Fig. 1 M–N; Supplemental Fig. S1B). This result is in agreement with our observation that NCLX overexpression limited the extent of TAC-induced cardiac hypertrophy (Fig. 1 H–J). Consistent with our observation that NCLX overexpression limited TAC-induced cardiac fibrosis (Fig. 1 K–L), the expression of pro-fibrotic genes such as periostin (Postn) and osteopontin (Spp1) with 12 weeks of TAC was reduced in NCLX-Tg mice (Fig. 1 O–P).

2.2. Cardiomyocyte NCLX overexpression attenuates early pressure overload-induced cardiac hypertrophy

To better understand the temporal dynamics of how enhanced cardiomyocyte NCLX expression and _mCa²⁺ efflux capacity alters the heart’s response to chronic pressure overload, we also evaluated cardiac pathology at an earlier time point, 2 weeks after TAC-induced pressure overload (Fig. 2A). No significant contractile dysfunction or changes in left ventricular dimensions were yet apparent at this early stage in the hearts of either control αMHC-tTA mice or TRE-NCLX x αMHC-tTA NCLX overexpressor mice subjected to TAC (Fig. 1 E–G). However, the hearts of αMHC-tTA mice did exhibit hypertrophy, indicated by a significant increase in heart weight-to-tibia length ratio (53% increase in αMHC-tTA HW/TL) and cardiomyocyte cross-sectional area (49% increase in αMHC-tTA CSA) as compared to sham animals (Fig. 2 B–D). In contrast, no significant increase in either the heart weight-to-tibia length ratio or cardiomyocyte cross-sectional area was detected in NCLX-Tg mice (Fig. 2B–D).

Figure 2: Cardiomyocyte NCLX overexpression attenuates early pressure overload-induced cardiac hypertrophy.

A) Timeline of 2-week transverse aortic constriction experimental protocol. DOX, doxycycline administration. B) Heart weight-to-tibia length ratio (HW/TL) of αMHC-tTA and TRE-NCLX x αMHC-tTA mice at 2 weeks post sham or TAC surgery. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05 vs. sham. (n=4–8 mice / group). C) Left ventricular tissue at 2 weeks post sham or TAC surgery, stained with wheat germ agglutinin (WGA, red) to delineate sarcolemma and with DAPI (blue). Scale bars = 50 µm. D) Quantification of cardiomyocyte cross-sectional area (CSA). Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01 vs. sham. (n=4–8 mice / group). E-F) qPCR quantification of mRNA expression of fetal genes in hearts of mice 2 weeks after sham or TAC surgeries. Nppa, natriuretic peptide type A; Acta1, α-skeletal muscle actin. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05 vs. sham. (n=4–8 mice / group).

We observed a significant induction of the expression of the fetal gene Acta1 and a trend towards increased expression of Nppa, in the hearts of αMHC-tTA mice following 2-week TAC (Fig. 2 E–F). In contrast, the expression of Nppa and Acta1 was not increased in NCLX overexpressor mice (Fig. 2 E–F). Consistent with this 2-week timepoint falling within the early stages of cardiac stress responses and remodeling, the expression of Npr1 was not yet changed after 2-weeks of TAC (Supplemental Fig. S2A); however, in agreement with our results for the 12-week TAC study (Supplemental Fig. S1D), the ratio of Myh6/Myh7 expression was significantly reduced by 2-week TAC in animals of both genotypes (Supplemental Fig. S2B). Nppb expression was likewise increased after 2-weeks of TAC in NCLX-Tg mice (Supplemental Fig. S2C). These results collectively suggest that NCLX overexpression and enhanced capacity for cardiomyocyte _mCa²⁺ efflux attenuates the early hypertrophic response to chronic pressure overload, and abrogates or delays the activation of certain specific stress-responsive gene programs within the heart. Protein expression of the mtCU components MCU, EMRE, and MICU1 did not differ between sham αMHC-tTA and NCLX-Tg hearts under basal conditions, although MCU expression tended to increase in NCLX-Tg hearts with 2 weeks of TAC (Supplemental Fig. S3A–B). This may represent a compensatory response that develops over time during hypertrophic stimulation in the NCLX-Tg hearts in order to increase net _mCa²⁺ accumulation in the face of increased NCLX-dependent _mCa²⁺ extrusion. Interestingly, EMRE protein expression decreased significantly with 2 weeks of TAC only in the hearts of αMHC-tTA mice (Supplemental Fig. S3C), consistent with the notion that these hearts are experiencing some degree of _mCa²⁺ overload that then elicits a compensatory downregulation of _mCa²⁺ uptake. That we did not detect a significant decrease in EMRE expression in NCLX-Tg hearts with 2 weeks of TAC further supports the idea that NCLX overexpression is sufficient to limit net _mCa²⁺ accumulation in response to pressure overload.

We previously demonstrated that NCLX overexpression is a powerful approach to minimize _mCa²⁺ overload and cell death during acute cellular Ca²⁺ stress that occurs in cardiac IR injury and myocardial infarction [75]. Like IR injury, chronic pressure overload increases cytosolic Ca²⁺ levels [33, 38] and can culminate in cardiomyocyte dropout [65–67]. Therefore, we hypothesized that augmenting NCLX function in this context would be cardioprotective. In contrast to the appreciable fibrosis observed in αMHC-tTA hearts at 12 weeks post-TAC (Fig. 1 K–L), when examined at 2 weeks post-TAC neither genotype yet showed an increase in cardiac fibrosis (Supplemental Fig. S2D). The expression of Postn, a marker of activated fibroblasts [81], Spp1, Tgfb1, and α-smooth muscle actin (Acta2) were not yet significantly increased by this short-term stress (Supplemental Fig. S2 E–H). This result supports the notion that this 2-week TAC timepoint is prior to any substantial increase in myofibroblast activation or differentiation [82]. This suggests that NCLX overexpression limits pressure overload-induced cardiomyocyte hypertrophy, prior to any distinct role in preventing cardiomyocyte death and replacement fibrosis.

2.3. Cardiomyocyte NCLX overexpression attenuates neurohormonal-induced cardiac hypertrophy.

Next, to validate the effects of NCLX overexpression on cardiac hypertrophy in a second, independent in vivo model of sustained cellular Ca²⁺ stress, we implanted NCLX-Tg and control mice with osmotic minipumps to deliver a low dose of angiotensin II (AngII) for 2 weeks (1.44mg/kg/day, Fig. 3A). This treatment paradigm causes an increase in blood pressure and induces cardiac hypertrophy in mice [83, 84]. Angiotensin II also directly enhances cytosolic Ca²⁺ cycling in cardiomyocytes via activation of angiotensin II type 1 receptors [85, 86], which contributes to an increase in _mCa²⁺ content [87].

Figure 3: Cardiomyocyte NCLX overexpression attenuates hypertrophy in mice infused with angiotensin II.

A) Timeline of angiotensin II (AngII) infusion experimental protocol. DOX, doxycycline administration. B) Heart weight-to-tibia length ratio (HW/TL) of αMHC-tTA and TRE-NCLX x αMHC-tTA mice 2 weeks after sham or angiotensin II minipump implantation surgery. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05, ^###P<0.001 vs. sham. (n=5–10 mice / group). C) Left ventricular tissue 2 weeks after sham or angiotensin II minipump implantation surgery, stained with wheat germ agglutinin (WGA, red) to delineate sarcolemma and with DAPI (blue). Scale bars = 50 µm. D) Quantification of cardiomyocyte cross-sectional area (CSA). Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01 vs. sham. (n=5–10 mice / group). E-G) qPCR quantification of mRNA expression of fetal genes in hearts of αMHC-tTA and TRE-NCLX x αMHC-tTA mice 2 weeks after sham or angiotensin II minipump implantation surgery. Nppa, natriuretic peptide type A; Nppb, natriuretic peptide type b; Myh6, α-myosin heavy chain; Myh7, β-myosin heavy chain. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01, ^###P<0.001, ^####P<0.0001 vs. sham. (n=5–10 mice / group).

Two weeks of low-dose AngII infusion had no detrimental effect on survival or cardiac function (Supplemental Fig. S4 A–D). AngII treatment was sufficient to cause cardiac hypertrophy as indicated by an increase in the heart weight-to-tibia length ratio (46% increase in αMHC-tTA vs. sham; compared to 29% increase in NCLX-Tg vs. sham); and by an increase in cardiomyocyte cross-sectional area (21% increase in αMHC-tTA vs. sham) in control αMHC-tTA mice (Fig. 3 B–D). We again noted a trend towards increased MCU protein expression in the hearts of NCLX-Tg mice subjected to AngII infusion (Supplemental Figure S5). Consistent with our findings in pressure overload (Fig. 1–2), cardiomyocyte cross-sectional area was not significantly increased by 2-week AngII infusion in NCLX overexpressor mice (Fig. 3 C–D). Interestingly, despite this attenuation of AngII-induced cardiomyocyte hypertrophy in NCLX overexpressor mice, we still observed a significant induction of fetal genes including Nppa, Nppb, and Acta1, and a significant reduction in the Myh6/7 ratio, in the hearts of both αMCH-tTA and NCLX-Tg mice with 2 weeks of AngII infusion (Fig. 3 E–G; Supplemental Fig. S4 E–F). These findings indicate that NCLX overexpressor hearts still sense and respond to angiotensin II stimulation with expected changes in gene expression. This observation raises the intriguing possibility that enhanced _mCa²⁺ efflux may slow stress-induced cardiomyocyte growth by limiting anabolic processes needed for an increase in cell mass, despite preserved activation of mitochondrial-independent signaling pathways classically associated with hypertrophy.

2.4. Cardiomyocyte NCLX overexpression attenuates pathological remodeling but reduces survival in mice infused with high-dose angiotensin II + phenylephrine.

Apart from attenuating cardiac hypertrophy, our findings indicated that the other major protective effect of cardiomyocyte NCLX overexpression on cardiac remodeling following TAC was to minimize the extent of fibrosis throughout the heart (Fig. 1 K–L), likely by minimizing _mCa²⁺ overload-induced cardiomyocyte dropout and replacement fibrosis. To further explore this notion, as well as the proposed effect of Na⁺-driven _mCa²⁺ extrusion through NCLX to starve the mitochondria of Ca²⁺ and compromise cellular energetics in settings of severe cardiac stress such as end-stage HF [50, 68, 69], we next evaluated cardiac function and remodeling using a more extreme chronic cytosolic Ca²⁺ stress model. Here, we subjected αMHC-tTA and TRE-NCLX x αMHC-tTA mice to 4-week osmotic minipump infusion of combined high doses of the Gq-coupled receptor agonists angiotensin II (10mg/kg/day) and phenylephrine (50mg/kg/day) (AngII/PE) (Fig 4A).

Figure 4: Cardiomyocyte NCLX overexpression attenuates remodeling but reduces survival in mice infused with chronic high-dose angiotensin II + phenylephrine.

A) Timeline of angiotensin II + phenylephrine (AngII/PE) infusion experimental protocol. DOX, doxycycline administration. B) Kaplan-Meier survival curve of αMHC-tTA and TRE-NCLX x αMHC-tTA mice for 4 weeks after sham or AngII/PE osmotic minipump implantation surgeries. The number of animals in each group at the start of the study is indicated in parentheses. Data analyzed by log-rank (Mantel-Cox) test. *P<0.05 αMHC-tTA + AngII/PE vs. TRE-NCLX x αMHC-tTA + AngII/PE. C) Heart weight-to-tibia length ratio (HW/TL) of αMHC-tTA and TRE-NCLX x αMHC-tTA mice 4 weeks after sham or AngII/PE minipump implantation surgery. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05 vs. sham; *P < 0.05 αMHC-tTA vs. TRE-NCLX x αMHC-tTA. (n=5–7 mice / group). D) Masson’s trichrome stain for myocardial collagen deposition (blue) in αMHC-tTA and TRE-NCLX x αMHC-tTA hearts 4 weeks after sham or AngII/PE minipump implantation surgery. Black scale bars for whole-heart cross sections = 2mm. White scale bars for higher magnification micrographs (bottom row) = 100 µm. E) Quantification of fibrotic area as percent of tissue area. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01 vs. sham; *P < 0.05 αMHC-tTA vs. TRE-NCLX x αMHC-tTA. (n=3–4 mice / group). F-H) qPCR quantification of mRNA expression of pro-fibrotic genes in hearts of mice after 4 weeks of AngII/PE infusion. Postn, periostin; Tfgb1, transforming growth factor β−1; Acta2, α-smooth muscle actin. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05, ^##P<0.01 vs. sham. (n=5–7 mice / group).

In stark contrast to the other models examined above, high-dose AngII/PE infusion caused a significant increase in mortality, specifically among NCLX overexpressor mice, with most death occurring during the first 2 weeks of infusion (Fig. 4B). Echocardiography over this time span revealed a transient decrease in left ventricular end diastolic and end systolic dimensions, with a corresponding increase in fractional shortening, in TRE-NCLX x αMHC-tTA mice after 1 week of AngII/PE infusion (Supplemental Fig. S6 A–C). The functional difference subsided by 2 weeks and little death was observed through the remainder of the study (Supplemental Fig. S6C; Fig. 4B). We were able to harvest cardiac tissue from a small number of the TRE-NCLX x αMHC-tTA mice that died spontaneously during the first two weeks of AngII/PE infusion. von Kossa staining of these hearts revealed substantial Ca²⁺ deposition throughout the myocardium (Supplemental Fig. S6D), consistent with cellular Ca²⁺ overload and necrosis as has been reported in reported previously in other mouse models of pathological cardiomyocyte Ca²⁺ loading [88]. We did not detect any positive von Kossa staining in control TRE-NCLX x tTA hearts collected up to 24 hours post-euthanasia, indicating that such Ca²⁺ deposition was not merely a post-mortem artifact. We also found no evidence for aortic aneurysm among animals that died spontaneously. These incidental findings suggest that when paired with extreme, sustained cytosolic Ca²⁺ stress, cardiac NCLX overexpression can be detrimental, perhaps because it compromises mitochondrial Ca²⁺ buffering capacity and/or mitochondrial bioenergetics within the heart.

In control αMHC-tTA mice that survived for the duration of the 4-week study, the combined high dose of AngII/PE elicited significant cardiac hypertrophy (16% increase in αMHC-tTA HW/TL vs. sham) (Fig. 4C). Consistent with our previous models, cardiac hypertrophy was attenuated in NCLX-Tg mice that survived to study termination (11% increase in NCLX HW/TL vs. sham) (Fig. 4C). MCU protein expression was increased in the hearts of NCLX-Tg mice subjected to AngII/PE infusion (Supplemental Figure S7 A–B), once again suggestive of an eventual compensatory upregulation of _mCa²⁺ uptake in response to hypertrophic stimulation. Such upregulation of MCU may counteract _mCa²⁺ depletion or exaggerated cytosolic Ca²⁺ overload resulting from diminished _mCa²⁺ buffering in the NCLX-Tg hearts. Given enough time, it may also allow sufficient _mCa²⁺ accumulation to allow for some pathological cardiac growth in NCLX-Tg hearts, albeit at slower rate as compared to the control genotype.

Chronic AngII/PE infusion caused cardiac injury as indicated by patchy fibrotic scarring throughout the myocardium of αMHC-tTA mice that survived to study termination, and this fibrotic effect was attenuated in in surviving NCLX-Tg mice (Fig. 4 D–E). Chronic AngII/PE infusion significantly increased the expression of Postn and Tgfb1 in αMHC-tTA mice (Fig. 4 F–H). The finding that Acta2 expression did not increase with AngII/PE infusion (Fig. 4H), despite increased cardiac fibrosis, is in agreement with our findings in the 12-week TAC model (Supplemental Fig. S1F). This observation may reflect that hearts subjected to 12-week TAC or 4-week AngII/PE treatment are in a more chronic phase of the fibrotic response, where the transient expression of some fibrotic genes such as α-smooth muscle actin subsides as myocardial scar tissue matures [81, 89]. Consistent with our finding of reduced fibrotic deposition in NCLX-Tg hearts after 4 weeks of AngII/PE infusion (Fig. 4 D–E), Postn expression in the hearts of TRE-NCLX x αMHC-tTA mice was not significantly increased by AngII/PE treatment (Fig. 4F). Together, these findings support the notion that increasing _mCa²⁺ efflux is sufficient to limit myocardial fibrosis in settings of chronic Ca²⁺ overload, likely by minimizing cardiomyocyte death and subsequent replacement scarring of the myocardium. The results of this high-dose AngII/PE model further reveal that although increased cardiomyocyte NCLX activity consistently mitigates pathological cardiac remodeling over chronic time scales, it also has deleterious effects that negatively impact overall survival in the context of extreme neurohormonal stress that may cause severe cytosolic Ca²⁺ overload.

2.5. NCLX expression attenuates _mCa²⁺ accumulation, oxidative metabolism, and cardiomyocyte hypertrophy in vitro.

Finally, we employed a reductionist in vitro model of cardiac hypertrophy using neonatal rat ventricular myocytes (NRVMs) transduced with adenovirus encoding human NCLX (Ad-NCLX) or control adenovirus encoding β-galactosidase (Ad-LacZ) to investigate the cellular mechanism by which NCLX overexpression limits pathological cardiomyocyte growth. 48hrs of phenylephrine stimulation increased steady-state mitochondrial Ca²⁺ content as indicated by an increase in resting mito-R-GECO1 fluorescence in NRVMs transduced with Ad-LacZ (Fig. 5A and Supplemental Fig. S8A). Adenoviral NCLX expression attenuated the phenylephrine-induced increase in _mCa²⁺ (Fig. 5A and Supplemental Fig. S8A), consistent with an increased capacity for _mCa²⁺ efflux. Phenylephrine treatment increased cytosolic Ca²⁺ concentration, as measured by the ratiometric Ca²⁺-sensitive dye Fura-2, AM, in NRVMs transduced with either Ad-LacZ or Ad-NCLX (Fig. 5B and Supplemental Fig. S8B). Notably, cytosolic Ca²⁺ was significantly higher in PE-treated NRVMs transduced with Ad-NCLX than those transduced with Ad-LacZ, indicating that the reduction in _mCa²⁺ accumulation with exogenous NCLX expression is not attributable to diminished cytosolic Ca²⁺ loading. This finding further supports the notion that during hypertrophic stimulation, NCLX overexpression limits or delays mitochondrial Ca²⁺ accumulation. Phenylephrine stimulation likewise increased mitochondrial superoxide production in NRVMs transduced with Ad-LacZ, and this effect was significantly reduced in NRVMs transduced with Ad-NCLX (Fig. 5C and Supplemental Fig. S8C).

Figure 5: NCLX expression limits oxidative capacity and biosynthetic potential of cardiomyocytes during hypertrophic stimulation in vitro.

A) Basal, steady-state _mCa²⁺ content in neonatal rat ventricular myocytes (NRVMs) transduced with adenovirus encoding β-galactosidase (Ad-LacZ) or human NCLX (Ad-NCLX) and treated with vehicle or phenylephrine (PE), as indicated by fluorescence of the genetically-encoded mitochondrial Ca²⁺ reporter, mito-R-GECO1. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^####P<0.0001 vs. vehicle; ****P<0.0001 Ad-LacZ vs. Ad-NCLX. Cell number per experimental group (n) is indicated within the corresponding bar of the graph. Individual data points are depicted in Supplemental Fig. S8A. B) Basal, steady-state cytosolic Ca²⁺ level in NRVMs, as indicated by fluorescence of the ratiometric Ca²⁺ reporter dye, Fura-2, AM. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^####P<0.0001 vs. vehicle; ****P<0.0001 Ad-LacZ vs. Ad-NCLX. Cell number per experimental group (n) is indicated within the corresponding bar of the graph. Individual data points are depicted in Supplemental Fig. S8B. C) Mitochondrial superoxide production in NRVMs, as indicated by fluorescence of the dye MitoSOX Red. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01 vs. vehicle; *P<0.05 Ad-LacZ vs. Ad-NCLX. Cell number per experimental group (n) is indicated within the corresponding bar of the graph. Individual data points are depicted in Supplemental Fig. S8C. D) Basal oxygen consumption rate per unit cellular protein in NRVMs after 24 hours of treatment with vehicle or phenylephrine. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^####P<0.0001 vs. vehicle; ****P<0.0001 Ad-LacZ vs. Ad-NCLX. (n=8–15 replicates / group). E) Western blots for puromycin incorporation, and mitochondrial loading control ATP5A, in NRVMs treated with puromycin (puro) for 1–4 hours starting at 24 hours of phenylephrine stimulation. Corresponding full-length western blots are shown in Extended Fig. 2. F) Quantification of cellular puromycin incorporation normalized to loading control, ATP5A. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^#P<0.05 vs. vehicle; **P<0.01 Ad-LacZ vs. Ad-NCLX. (n=3 replicates / group). G) Protein/DNA ratio in NRVMs after 48 hours of vehicle or PE treatment. Data analyzed by 2-way ANOVA with Sidak’s post-hoc test. ^##P<0.01, ^####P<0.001 vs. vehicle; *P<0.05 Ad-LacZ vs. Ad-NCLX. (n=15 replicates / group).

We next examined NRVMs at 24 hours of phenylephrine stimulation, a timepoint at which we could not yet detect cardiomyocyte hypertrophy, to investigate how limiting _mCa²⁺ accumulation affects cellular metabolism that may be required for cardiomyocyte growth. Whereas phenylephrine treatment increased oxygen consumption per unit cellular protein in NRVMs transduced with Ad-LacZ, this effect was abrogated in NRVMs transduced with Ad-NCLX (Fig. 5D). This finding is consistent with the notion that increased _mCa²⁺ stimulates TCA cycle flux and oxidative phosphorylation, as well as with our finding that adenoviral NCLX expression attenuates the phenylephrine-induced increase in _mCa²⁺ (Fig. 5A and Supplemental Fig. S8A). Together, these results suggest that increased NCLX expression limits the oxidative capacity of cardiomyocytes during hypertrophic stimulation by suppressing _mCa²⁺-dependent stimulation of metabolism. Measurement of cellular puromycin incorporation showed that 24 hours of PE stimulation increased the rate of de novo protein synthesis in NRVMs transduced with Ad-LacZ, and that this PE-dependent effect was abolished in NRVMs transduced with Ad-NCLX (Fig. 5E–F; full-length blots shown in Extended Fig. 2). This observation indicates that by limiting the ability for cardiomyocytes to increase oxidative metabolism in response to hypertrophic stimulation, NCLX overexpression consequently limits the cells’ capacity for anabolic growth. Indeed, fitting with this model and in agreement with our chronic in vivo studies, the increase in the protein/DNA ratio of NRVMs after 48 hours of phenylephrine stimulation was significantly reduced in cells transduced with Ad-NCLX (Fig. 5G).

3. DISCUSSION AND CONCLUSIONS

The net increase in _mCa²⁺ concentration is a key mechanism that increases cardiac ATP production to augment an increase in cardiac workload and ATP consumption. At the same time, excessive _mCa²⁺ accumulation has long been recognized as a driver of cardiomyocyte death and the resulting contractile dysfunction following myocardial infarction. A key unanswered question in the HF field is what causative role, if any, altered _mCa²⁺ homeostasis plays in the development of non-ischemic heart disease and failure. Here, we report that cardiomyocyte-specific, transgenic overexpression of the mitochondrial sodium-calcium exchanger, NCLX, attenuates cardiac hypertrophy and fibrosis and improves contractile function in a mouse model of chronic pressure overload via 12-week transverse aortic constriction (Fig. 1). We demonstrate that cardiomyocyte-specific NCLX overexpression mitigates early hypertrophic remodeling that occurs in response to 2-week TAC or 2-week angiotensin II stimulation, prior to the manifestation of contractile dysfunction (Fig. 2–3). NCLX overexpression also reduces cardiac hypertrophy myocardial fibrosis in mice subjected to chronic neurohormonal overload with infusion of a combined high dose of angiotensin II and phenylephrine (Fig. 4). This provides genetic evidence that enhancing _mCa²⁺ efflux mitigates maladaptive cardiac remodeling and the development of contractile failure when hearts are subjected to sustained hemodynamic and neurohormonal load, and supports the hypothesis that _mCa²⁺ accumulation contributes to early pathological remodeling and subsequent contractile dysfunction in non-ischemic heart disease. We further demonstrate that NCLX limits _mCa²⁺ accumulation, oxidative metabolism, and protein biosynthesis during hypertrophic stimulation in vitro (Fig. 5), suggesting that an increase in _mCa²⁺ content drives increased mitochondrial metabolism that is required to fuel cardiomyocyte growth. However, our in vivo findings also suggest the existence of a lower threshold or “floor” for matrix Ca²⁺ content, below which interventions that increase cardiomyocyte _mCa²⁺ efflux switch from being adaptive to being maladaptive and increase the likelihood of death. This finding emphasizes that therapies targeting _mCa²⁺ exchange to either increase or decrease net _mCa²⁺ accumulation in chronic heart disease may need to be matched to the specific nature of Ca²⁺ dysregulation and the patient’s stage of disease.

Several prior studies have attempted to explore the relationship between altered cardiomyocyte _mCa²⁺ homeostasis and the development of maladaptive remodeling and contractile dysfunction in response to chronic hemodynamic stress. Mcu deletion has proven ineffective at limiting cardiac hypertrophy or contractile dysfunction with chronic pressure overload [22, 70], possibly because alternative, slower routes of _mCa²⁺ uptake are sufficient to increase _mCa²⁺ content upon chronic elevations in cytosolic Ca²⁺ [21, 22, 90]. The NCLX overexpression models used in the current study overcome this technical limitation and provide novel insight into the contribution of _mCa²⁺ signaling to pathological cardiac remodeling and the development of non-ischemic HF.

Our findings indicate that cardiomyocyte NCLX overexpression mitigates early pathological cardiac remodeling and the development of contractile dysfunction via two distinct mechanisms. First, the observation that cardiomyocyte NCLX overexpression minimizes myocardial fibrosis in response to 12-week TAC and 4-week infusion with angiotensin II + phenylephrine is consistent with our earlier report that cardiomyocyte NCLX overexpression protects against cell death in response to acute cellular Ca²⁺ overload [75]. This result suggests that increased capacity for _mCa²⁺ efflux similarly lowers the overall risk of _mCa²⁺ overload and necrotic cardiomyocyte death during chronic stress. Our finding of reduced fibrotic deposition in the hearts of NCLX-Tg mice subjected to 4-week AngII/PE infusion, despite similar activation of the expression of some pro-fibrotic genes such as transforming growth factor β−1, also fits with a model of a net delay in fibrotic remodeling with cardiomyocyte NCLX overexpression. That is, with a delay in _mCa²⁺ overload and cardiomyocyte death in NCLX-Tg mice, less total time would have passed between the initial stimulus for the induction of pro-fibrotic gene expression, subsequent synthesis of extracellular matrix proteins, and study termination.

Second, cardiomyocyte NCLX overexpression attenuates cardiomyocyte hypertrophy. This result was consistent across all in vivo models examined (Fig. 1–4) and was recapitulated in NRVMs (Fig. 5). The observation that cardiomyocyte hypertrophy was still attenuated in NCLX-Tg mice even in instances where there was similar activation of the fetal gene program suggests that the rate of cardiomyocyte growth in response to the same initial stressor was slowed by increased _mCa²⁺ efflux activity. That NCLX overexpression alone does not affect heart weight, cardiomyocyte cross sectional area, protein synthesis rates, protein/DNA ratio, nor mitochondrial or cytosolic Ca²⁺ content under homeostatic conditions is consistent with the view that NCLX overexpression specifically mitigates stress-induced cardiac hypertrophy by limiting net _mCa²⁺ accumulation in response to elevations in cytosolic Ca²⁺ concentration. This suggests that an increase in net _mCa²⁺ accumulation during chronic stress directly contributes to the process of pathological cardiac hypertrophy. We acknowledge that we consistently observed trends towards upregulation of MCU in NCLX-Tg hearts when subjected to chronic stress (Supplemental Figs. 3, 5, and 7). This indicates that given enough time, hearts have some capacity to compensate for NCLX overexpression, particularly when subjected to perturbations that increase cardiac workload. Such compensatory mechanisms to augment _mCa²⁺ accumulation may help to explain the somewhat modest effects of NCLX overexpression on cardiac remodeling in our chronic in vivo studies.

Our in vivo results raised the question of how precisely _mCa²⁺ contributes to cardiomyocyte hypertrophy. Adenoviral-mediated NCLX expression in neonatal rat ventricular cardiomyocytes in vitro revealed that limiting _mCa²⁺ accumulation during pro-hypertrophic stimulation is associated with an impaired ability to increase mitochondrial oxidative metabolism. These observations fit with the notion that increased _mCa²⁺ levels enhance metabolic flux by stimulating the activity of pyruvate dehydrogenase and TCA cycle dehydrogenases [57, 62, 91, 92]. Such increased mitochondrial metabolism is likely required to provide the energy and to generate the macromolecules that are required for cell growth. Fitting with this view, elevated _mCa²⁺ concentration in mPTP-deficient, cyclophilinD-null hearts is associated with an increase in net glucose oxidation [93]. The Tian laboratory recently described a mechanism by which a shift from fatty acid oxidation to glucose oxidation increases anapleurotic input of pyruvate into the TCA cycle and thereby drives an increase in de novo aspartate biosynthesis [94]. Such increased aspartate production supports nucleotide and protein synthesis to enable anabolic growth of the heart in response to hypertrophic stimulation. These reports establish a model in which increased _mCa²⁺ signaling drives increased glucose oxidation, to increase biosynthetic pathways that fuel biomass deposition within the heart. Our current results indicate that cardiomyocyte NCLX overexpression limits _mCa²⁺ accumulation when the heart is subjected to sustained stress, and so limits the increase in oxidative metabolism that is needed to provide the energy and molecular building blocks required for cardiomyocyte growth. Indeed, we find that exogenous NCLX expression in vitro mitigates the phenylephrine-induced increase in the rate of de novo protein synthesis and cardiomyocyte growth, processes that ultimately depend upon the integration of cellular bioenergetic status (availability of ATP to invest in energy-consuming anabolic reactions such as protein translation) and the availability of macromolecules such as amino acids, nucleotides, and lipids, whose synthesis is coupled to TCA cycle flux. We therefore propose that NCLX overexpression limits _mCa²⁺ accumulation when the heart is subjected to chronic stress, thereby limiting the oxidative capacity of the cardiomyocytes and so limiting the cells’ potential for anabolic growth.

The fact that NCLX mediates mitochondrial Na⁺ influx in exchange for _mCa²⁺ efflux can help to reconcile the deleterious effect of cardiomyocyte NCLX overexpression on overall animal survival we observed in the model of extreme neurohormonal overload with combined high-doses of angiotensin II + phenylephrine (Fig. 4B) with the cardioprotective phenotype we observed in all other stress conditions. Cytosolic Na⁺ concentration is increased in the failing heart [95, 96]. Elevated cytosolic Na⁺ levels drive increased mitochondrial Na⁺/Ca²⁺ exchange through NCLX, such that the mitochondrial matrix may become deficient in Ca²⁺ [50, 68]. Under these conditions, ATP production can be impaired, and mitochondrial redox balance can be disrupted, resulting in increased ROS stress [97]. These combined consequences of _mCa²⁺ depletion in the failing heart can limit cardiac contractility and increase the susceptibility to fatal cardiac arrhythmia [51, 69]. We found that a significant number of NCLX overexpressor animals receiving AngII/PE died within the first two weeks of treatment. The abrupt onset of this mortality, along with the absence of overt systolic dysfunction, are in line with possibility that NCLX-Tg mice exhibit a greater propensity for arrhythmia under extreme adrenergic stress, leading to sudden cardiac death. These findings could be explained by increased intracellular Na⁺ accumulation due to increased plasma membrane Na⁺/Ca²⁺ and Na⁺/H⁺ exchange in the mice treated with AngII/PE [98, 99], as compared to the less severe models we tested. Such increased cytosolic Na⁺ would favor mitochondrial Na⁺/Ca²⁺ exchange, limit _mCa²⁺ accumulation, and augment increases in cytosolic Ca²⁺ levels, effects that would have been exacerbated in NCLX-Tg hearts. Our anecdotal observations of pronounced Ca²⁺ deposition in the hearts of NCLX-Tg mice that died spontaneously following AngII/PE infusion (Supplemental Fig. 6D) agree with this interpretation and suggest that impaired _mCa²⁺ accumulation compromises mitochondrial Ca²⁺ buffering capacity, possibly leading to energetic impairment and cytosolic Ca²⁺ overload that culminate in necrosis. Such tissue-level damage likely disrupts the electrical syncytium of the myocardium, resulting in aberrant propagation of electrical activity and increasing the chance for fatal cardiac arrhythmia. Liu et al’s finding that NCLX inhibition protects against sudden cardiac death in a guinea pig HF model [69] lends further support to the view that excessive NCLX activity in the failing or extremely-stressed heart increases the overall risk for arrhythmia, separable from the protective effects of increased NCLX activity against deleterious cardiac remodeling that develops over chronic time scales.

In conclusion, our data provide evidence for enhancing NCLX efflux as an effective strategy to mitigate pathological _mCa²⁺ overload and minimize detrimental hypertrophic remodeling as chronically-stressed hearts are beginning to progress towards failure. Thus, increasing NCLX function within the heart may be a suitable therapeutic strategy to slow or delay the onset of cardiac pathology and contractile dysfunction that develops in non-ischemic heart disease and inherited cardiomyopathies. Our findings further demonstrate that although increasing _mCa²⁺ efflux is protective early in the disease process, if applied in settings of extreme stress—perhaps less clinically appropriate—it can become deleterious. Our results combined with earlier reports [51, 69] suggest a threshold or switch for net _mCa²⁺ exchange in chronic heart disease, where increasing _mCa²⁺ content is maladaptive and contributes to pathological remodeling and dysfunction early in the heart’s response to increased workloads, but later on in the disease process or with more severe perturbation of cellular ion homeostasis—where the mitochondria could become starved of Ca²⁺—increasing _mCa²⁺ content instead becomes adaptive in order to augment mitochondrial energetics and correct mitochondrial redox balance. Our study thus adds nuance and complexity to our understanding of the multifarious roles of _mCa²⁺ signaling as the heart remodels in response to chronic stress and progresses towards failure.

Future research to fully elucidate the specific molecular mechanisms linking increased _mCa²⁺ to cardiomyocyte growth will offer insight into novel therapeutic strategies to mitigate pathological hypertrophy in chronic heart disease. Further work will also be necessary to distinguish any specific effects of altered _mNa⁺ content, as opposed to altered _mCa²⁺ homeostasis, in mediating pathological remodeling and contractile dysfunction in HF, as well as the anti-hypertrophic effects of NCLX overexpression. Finally, research into the molecular mechanisms regulating NCLX activity is needed to identify novel, tractable ways to enhance or inhibit NCLX-dependent _mCa²⁺ exchange as appropriate for the treatment of HF and other human disease.

4. MATERIALS AND METHODS

Mice

The generation of mice with doxycycline-controlled, cardiomyocyte-restricted overexpression of NCLX was described previously [75]. In brief, TRE-NCLX mice expressing human SLC8B1 cDNA under the control of a P_tight Tet-responsive promoter were crossed to α-myosin heavy chain tetracycline transactivator (αMHC-tTA) transgenic mice. This strategy allows inducible expression of human NCLX within the cardiomyocytes upon removal of doxycycline (DOX) from the animals’ diet. Breeder cages were provided with chow containing DOX (TestDiet, #MRMH1500/625PPM DOXY 1/2 IRR) and litters were maintained on DOX until weaning at 3 weeks of age in order to prevent transgenic NCLX overexpression during development. Pups were genotyped for the TRE-NCLX transgene using the forward primer: 5’-TGTCGGCCATTTCTACCACACTGA-3’ and the reverse primer: 5’-ACACATACAAGCCCAGGTAACCCA-3’. PCR reaction mixture contained 1uL tail DNA in DirectPCR Lysis reagent (Viagen Biotech #102-T), 1x Taq buffer (Syd Labs #MB042-EUT), 80µM each dNTPs (New England Biolabs #N0447L), 800nM each forward and reverse primers, 200µM betaine (Affymetrix #77507), and 1.25 U Taq polymerase (Syd Labs #MB042-EUT). The PCR conditions were: denaturation at 95°C for 3 minutes, followed by 35 cycles (95°C for 30 seconds, 62°C for 30 seconds, 72°C for 40 seconds), followed by 10 minutes at 72°C. Pups were genotyped for the αMHC-tTA transgene using the forward primer: 5’-AGCGCATTAGAGCTGCTTAATGAGGTC-3’ and the reverse primer: 5’-GTCGTAATAATGGCGGCATACTATC-3’. PCR reaction mixture contained 1uL tail DNA in DirectPCR Lysis reagent, 1x Taq buffer, 80µM each dNTPs, 800nM each forward and reverse primers, and 1.25U Taq polymerase. The PCR conditions were: denaturation at 95°C for 3 minutes, followed by 35 cycles (95°C for 30 seconds, 66°C for 30 seconds, 72°C for 1min), followed by 10 minutes at 72°C.

Adult TRE-NCLX x αMHC-tTA mice and age-matched αMHC-tTA controls were used in surgical studies between 8–25 weeks of age. Both male and female mice were included. Experiments were performed using a numbered ear-tagging system in order to blind the experimenter to mouse genotype and experimental group. All animal experiments followed AAALAC guidelines and were approved by Temple University’s IACUC.

Western blot analysis

Hearts were excised from mice, rinsed in ice-cold phosphate buffered saline, immediately snap frozen using liquid nitrogen-cooled metal tongs, and stored at −80°C until use. Neonatal rat ventricular cardiomyocytes (NRVMs) were trypsinized off of tissue culture plates, washed once in phosphate buffered saline, pelleted, and snap frozen in liquid nitrogen. Heart tissue was homogenized in ice-cold 1X RIPA buffer (Millipore #20–188) supplemented with 1X SigmaFast protease inhibitor cocktail (Sigma-Aldrich #S8830–20TAB) and 1X PhosStop phosphatase inhibitor (Roche #04 906 845 001) using a bead mill homogenizer (VWR, #75840–022). NRVM pellets were lysed in this same RIPA buffer. Crude homogenates were sonicated for 10 seconds and then centrifuged for 5 minutes at 5,000 x g to pellet insoluble material. Lysate supernatants were collected and protein concentration was determined using a bicinchoninic acid protein assay (bioWORLD #20831001–1). Samples were mixed 4:1 with 5x SDS sample buffer (250µM Tris-HCl, pH 7.0; 40% (volume/volume) glycerol; 8% (weight/volume) sodium dodecyl sulfate; 20% (volume/volume) β-mercaptoethanol; 0.1% (weight/volume) bromophenol blue) and boiled for 5 minutes. For western blots for NCLX, 50µg of protein lysate per well was loaded on 10% (weight/volume) polyacrylamide Tris-glycine SDS gels. 20µg NRVM protein lysate per well was used for western blotting for puromycin. For large-format western blots of mouse heart tissue from surgical cohorts, 25µg protein per well was loaded on NuPAGE 4–12% Bis-Tris gels (Thermo Fisher, # WG1403BOX). Proteins were separated by electrophoresis and then transferred to polyvinyldienefluoride membranes (Millipore #IPFL00010), which were then blocked for 1 hour with blocking buffer (Rockland #MB-070). Membranes were incubated overnight at 4°C with primary antibodies diluted in 50% blocking buffer / 50% Tris-buffered saline (bioWORLD #42020056–3) + 0.1% TWEEN (Sigma-Aldrich #P9416) (TBS-T). Membranes were washed 3 times in TBS-T, incubated for 1.5 hours at room temperature with secondary antibodies diluted in 50% blocking buffer / 50% TBS-T, and washed 3 times in TBS-T before imaging on a LI-COR Odyssey infrared imager system. Primary antibodies and dilutions used in this study included: rabbit polyclonal against NCLX (GeneTex #GTX87452), 1:1000; mouse monoclonal total OXPHOS rodent antibody cocktail (Abcam #ab110413), 1:1000; rabbit monoclonal against MCU (Cell Signaling Technology #14997), 1:1000; rabbit antibody against EMRE (Bethyl Laboratories #A300-BL19208), 1:1000; custom-made rabbit polyclonal against MICU1[18] 1:500; mouse monoclonal against ATP5A (Abcam #ab14748), 1:2000; and mouse monoclonal against puromycin (EMD Millipore #MABE343), 1:5000. Secondary antibodies and dilutions included: IRDye 680RD goat anti-rabbit (LI-COR #926–68071), 1:10,000; IRDye 800CW goat anti-mouse (LI-COR #926–32210), 1:10,000; IRDye 800CW goat anti-rabbit (LI-COR #925–32210), 1,10,000; IRDye 680RD goat anti-mouse (LI-COR #925–68070), 1:10,000. All full-length western blots are shown in Extended Figs. 1–2, Supplemental Fig. S3, Supplemental Fig. S5, and Supplemental Fig. S7. Western blot densitometry was measured using LI-COR Image Studio software (LI-COR, version 2.0.38).

Transverse aortic constriction

Transverse aortic constriction was performed as described previously [100]. Briefly, mice were anesthetized with 3% isoflurane. Anesthetized mice were intubated and the chest cavity entered via partial thoracotomy to the second rib. The aortic arch was exposed and the transverse aorta was constricted by tying a 6–0 suture against a 27½ -gauge needle, which was then promptly removed. For sham mice, the chest was entered and the aorta exposed, but no suture was secured around the transverse aorta.

Left ventricular echocardiography

Transthoracic echocardiography of the left ventricle was performed on a Vevo 2100 imaging platform (VisualSonics) as described elsewhere [21]. In brief, mice were anesthetized with 1.5% isoflurane in 100% oxygen during imaging and M-mode images were collected in the short-axis view. Recordings were analyzed using VisualSonics Vevo LAB software (VisualSonics version 3.1.1). For TAC studies, the efficacy of aortic constriction was confirmed using pulsed wave Doppler echocardiography to measure blood flow velocity through the site of constriction. The aortic pressure gradient was calculated according to the formula: pressure gradient = 4 X V _max² (Li 2016) [101]. For mice subjected to TAC surgery, animals with an aortic pressure gradient <30mmHg were excluded from subsequent analysis [102].

Tissue gravimetrics and histology

Hearts were collected at the indicated time points and massed. Mouse tibia length was measured for normalization of heart mass. The atria were removed from the ventricles, and the ventricles were divided into thirds for further analysis. Ventricle base samples were rinsed in ice-cold phosphate buffered saline, then immediately snap frozen using liquid nitrogen-cooled metal tongs and stored at −80°C until use. Ventricle apex samples were rinsed in ice-cold phosphate buffered saline, incubated in RNAlater (Thermo Fisher, #AM7020) for 24 hours at 4°C, and then stored at −80°C until use. Mid-ventricle samples were rinsed in ice-cold phosphate buffered saline and then fixed in 10% buffered formalin (EKI, #4498–1GL) before dehydration and embedding in paraffin, as described previously [75]. Paraffin-embedded ventricle samples were cut to 5-μm sections using a microtome and placed on glass slides. For measurement of cardiomyocyte cross-sectional area, heart sections were stained with TRITC-conjugated wheat germ-agglutinin (Sigma Aldrich #L-5266) at 100μg/mL to label the sarcolemma. Coverslips were mounted using ProLong Gold Antifade Mountant with DAPI (Invitrogen #P36935). The left ventricle was imaged on an Axio Observer Z1 fluorescence microscope (Zeiss) and cardiomyocyte cross sectional area quantified using ImageJ (National Institutes of Health). A minimum of 80 cardiomyocytes were measured per mouse. For measurement of myocardial fibrosis, heart sections were stained with Masson’s trichrome (Sigma-Aldrich #HT15). Images of the myocardium were captured at 20x magnification using an Eclipse Ni-E light microscope (Nikon). The percent fibrosis was quantified in ImageJ by dividing the area of tissue staining blue (collagen) by the total tissue area. For assessment of cellular calcium overload, heart sections were stained with von Kossa stain (Abcam #ab150687) according to the manufacturer’s instructions. All images of whole heart cross sections were captured on a SMZ1000 stereomicroscope (Nikon).

qPCR gene expression analysis

mRNA was isolated from ventricle apex samples preserved in RNAlater using a RNeasy Fibrous Tissue Kit (Qiagen, #74704) according to the manufacturer’s instructions. cDNA was prepared from mRNA using the High-Capacity cDNA Reverse Transcription kit (Thermo Fisher #4368813) following the manufacturer’s instructions. qPCR was performed on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad) using PowerUp SYBR Green Master Mix (Applied Biosciences #100029283). The real-time PCR conditions were: UDG activation at 50°C for 2 minutes, initial denaturation at 95°C for 10 minutes, followed by 40 cycles (95°C for 15 seconds, 60°C for 30 seconds, 72°C for 30 seconds). qPCR primers against mouse transcripts are listed in Table 1.

Table 1:

Sequences of qPCR primers for mouse genes.

Gene	Forward Primer (5’→3’)	Reverse Primer (5’→3’)
Nppa	GGGTAGGATTGACAGGATTGG	CTCCTTGGCTGTTATCTTCGG
Nppb	GCACAAGATAGACCGGATCG	CCCAGGCAGAGTCAGAAAC
Npr1	CACAGTCAACACAGCTTCAAG	TTCATTTCCACGTCACCTCG
Acta1	TGAACCCCAAAGCTAACCG	CCCCAGAATCCAACACGATG
Myh6	GCAGAACAGTAAAATTGAGGACG	CGCAGCTTCTCCACCTTAG
Myh7	GACGACGTCACCTCCAACA	TGCTCCGGTGCTCATTCATC
Postn	AAGAGATGGTCACTTCACGC	GCACTGGAGGGTATTTAGGATG
Spp1	GCTTGGCTTATGGACTGAGGTC	CCTTAGACTCACCGCTCTTCATG
Acta2	GTGAAGAGGAAGACAGCACAG	GCCCATTCCAACCATTACTCC
Tgfb1	CCTGAGTGGCTGTCTTTTGA	CGTGGAGTTTGTTATCTTTGCTG
Rps13	GCACCTTGAGAGGAACAGAA	GAGCACCCGCTTAGTCTTATAG

Osmotic minipump infusion studies

Osmotic minipump implantation surgeries were performed as described previously [103]. Mice were anesthetized with 3% isoflurane and a small midline incision was made in the back. After subcutaneous insertion of the minipump, the incision was closed with 5–0 absorbable suture. For sham mice, an incision was made in the back and then closed with 5–0 absorbable suture. All mice were treated preemptively for infection via subcutaneous injection of the antibiotic cefazolin (Sandoz, #007813450) at a dose of 40mg/kg. Osmotic minipumps (Alzet Model 1004, #000992) were set to deliver the Gq-coupled receptor agonist angiotensin II (Sigma-Aldrich #A9525) dissolved in sterile saline at a dose of 1.44mg/kg/day for 2 weeks [83, 84]. In a separate study, osmotic minipumps (Alzet Model 2004, #0000298) were set to deliver a combination of high doses of the Gq-coupled receptor agonists angiotensin II and phenylephrine hydrochloride (Sigma-Aldrich #P6126) dissolved in sterile saline at a dose of 10mg/kg/day angiotensin II + 50mg/kg/day phenylephrine hydrochloride [104] for 4 weeks.

Neonatal rat ventricular myocyte hypertrophy in vitro

Neonatal rat ventricular myocytes (NRVMs) were isolated from 1 day-old Sprague-Dawley rats as described previously [105]. NRVMs were plated at 100,000 cells per 35mm collagen-coated glass-bottomed dish (MatTek #P35GCOL-1.5–10-C) for imaging assays; at 200,000 cells/well of 6-well tissue culture plates for puromycin incorporation assay and measurement of protein/DNA ratio; or 20,000 cells/well of XF96 cell culture microplates (Agilent #101085–004) for measurement of oxygen consumption. NRVMs were plated and maintained in DMEM (Corning, #10–013-CV) + 5% fetal bovine serum (FBS) (Peak Serum, #PS-FB3) + 1% penicillin/streptomycin (Sigma-Aldrich #P0781–100ML) + 0.1mM 5’-bromo-2’-deoxyuridine (BrdU) (Sigma-Aldrich #B5002–1G) + 2µg/mL vitamin B12 (Sigma-Aldrich #V2876–1G) for 4 days. 4 days after isolation, the media was changed to DMEM + 5% FBS 1% penicillin/streptomycin + 2µg/mL vitamin B-12. The following day, the media was changed to serum-free DMEM + 1% penicillin/streptomycin + 2µg/mL vitamin B12 containing control adenovirus encoding β-galactosidase (Ad-LacZ) or adenovirus encoding human NCLX (Ad-NCLX), either alone or in combination with adeno-associated virus 6 encoding the genetic mitochondrial Ca²⁺ reporter, mito-R-GECO1 (AAV6-mito-R-GECO1). After 24 hours of viral transduction, the media was changed to serum-free DMEM + 1% penicillin/streptomycin + 2µg/mL vitamin B12 supplemented with 100µM phenylephrine hydrochloride (PE) (Sigma-Aldrich #P6126–5G) or vehicle control. Subsequent experiments were conducted at 24 or 48 hours after the beginning of PE treatment as indicated.

Ca²⁺ and mitochondrial superoxide imaging

Cytosolic and mitochondrial Ca²⁺ content and mitochondrial superoxide production in NRVMs were assessed after 48 hours of treatment with phenylephrine or vehicle control. For cytosolic Ca²⁺ assessment, NRVMs were loaded with 1µM Fura-2, AM (Thermo Fisher Scientific #F1221) for 15 minutes at room temperature, washed twice in Tyrode’s buffer (150 mM NaCl, 5.4 mM KCl, 1.2mM MgCl₂, 10 mM glucose, 2 mM sodium pyruvate, 5 mM HEPES, 2 mM CaCl₂, pH 7.4), and then imaged at 340nm and 380nm excitation/510±40nm emission. For mitochondrial Ca²⁺ assessment, NRVMs transduced with AAV6-mito-R-GECO1 were washed twice in Tyrode’s buffer and imaged at 572±17.5nm excitation/632±30nm emission. For detection of mitochondrial superoxide production, NRVMs were loaded with 5µM MitoSOX Red (Thermo Fisher Scientific #M36008) for 15 minutes at 37°C, washed twice in Tyrode’s buffer, and then imaged at 490±10nm excitation and 585±20nm emission. All imaging was performed in a 37°C heated chamber in fresh Tyrode’s buffer on an Axio Observer Z1 fluorescence microscope (Zeiss).

Oxygen consumption measurements

After 24 hours of stimulation with phenylephrine or vehicle control, NRVMs were changed to basal DMEM (Corning # 90–113-PB), pH 7.4, supplemented with 25mM glucose, 4mM L-glutamine, 1mM sodium pyruvate, 0.2mM bovine serum albumin-conjugated sodium palmitate, 0.2mM L-carnitine hydrochloride, and 100µM phenylephrine or vehicle control. Basal oxygen consumption rate (OCR) was measured in NRVMs using a Seahorse Bioscience XF96 extracellular flux analyzer. The cells were then washed with phosphate buffered saline, and lysed by adding 1x RIPA buffer directly to the XF96 cell culture microplates. Protein concentration per well was determined using a bicinchoninic acid protein assay. The OCR measured for each well was normalized to the µg protein per well.

Puromycin incorporation assay for de novo protein synthesis

After 24 hours of stimulation with phenylephrine or vehicle control, NRVMs were changed to fresh media containing phenylephrine or vehicle and supplemented with 1µg/mL puromycin dihydrochloride (Sigma-Aldrich # P8833–25MG). Cells were collected 1, 2, or 4 hours later for western blotting for puromycin incorporation, as described above.

Assessment of NRVM hypertrophy

After 48 hours of stimulation with phenylephrine or vehicle control, NRVMs were trypsinized off of tissue culture plates, washed once in phosphate buffered saline, pelleted, snap frozen in liquid nitrogen, and stored at −80°C until analysis. Cell pellets were lysed using the CyQUANT Cell Proliferation Assay Kit (Thermo Fisher Scientific #C7026), and then sonicated. Protein concentration in the lysates was determined using a bicinchoninic acid protein assay. An aliquot of each lysate was treated with DNAse-free RNAse A (Thermo Fisher Scientific #EN0531) and the DNA concentration determined with the CyQUANT Cell Proliferation Assay Kit following the manufacturer’s instructions. NRVM hypertrophy was assessed by normalizing the protein concentration of each sample to its DNA concentration.

Statistical Analysis

All results are presented as mean ± S.E.M. Statistical analysis was performed using Prism 6.0 (GraphPad Software). Direct comparisons between two groups used a two-tailed t-test. The log-rank (Mantel-Cox) test was used for comparison of Kaplan-Meier survival curves. For analysis of longitudinal echocardiographic studies, 2-way ANOVA was used with Tukey post-hoc analysis. Grouped data were analyzed by 2-way ANOVA with Sidak’s post-hoc analysis. For all comparisons, P values less than 0.05 were considered significant.

Extended Data

Extended Figure 1: Full-length blots for Figure 1B.

Arrow indicates full-length NCLX.

Extended Figure 2: Full-length blots for Figure 5E.

Arrow indicates ATP5A band.

Highlights.

• The mitochondrial Na⁺/Ca²⁺ exchanger, NCLX, mediates mitochondrial calcium efflux.

• NCLX limits mitochondrial Ca²⁺ accumulation upon hypertrophic stimulation.

• NCLX limits oxidative metabolism and protein synthesis during hypertrophy.

• NCLX overexpression attenuates pathological cardiac hypertrophy in vivo.