Sarco/Endoplasmic Reticulum Ca²⁺ ATPase Activation Shifts Cardiac Fuel Preference Without Impairing Cardiac Function in Obese Mice

Abstract

Background

Activity of SERCA (sarco/endoplasmic reticulum Ca²⁺ ATPase) affects cardiac metabolism and function and is under investigation in clinical trials. SERCA function also regulates glucose metabolism in multiple tissues, but it is unknown how these effects extend to the heart. Pressure overload‐induced cardiac hypertrophy corresponds to increased cardiac glucose oxidation in obesity. We hypothesized that SERCA activation by the compound CDN1163 would increase ATP demand due to Ca²⁺ cycling, leading to increased fat oxidation to maintain cardiac energy homeostasis.

Methods

In vivo [U‐¹³C₃]lactate tracer experiments were performed to assess fasting cardiac metabolic fluxes in Mc4r ^−/− mice fed a Western diet and treated with CDN1163. Fluxes were estimated by fitting a mathematical model of cardiac metabolism to ¹³C enrichment measurements of plasma and tissue metabolites taken at the end of the isotope infusion. Metabolic flux measurements were combined with echocardiography, gene expression, and enzymatic assays to assess the effects of SERCA activation on heart function and metabolism following 8 weeks of CDN1163 treatment.

Results

CDN1163 increased cardiac ATPase activity, decreased cytosolic Ca²⁺ signaling, and decreased cardiac glucose uptake and glycolysis in obese mice. A greater fraction of mitochondrial acetyl‐coenzyme A was obtained from nonglycolytic sources, such as fat, to sustain citric acid cycle flux, corresponding to an increase in mitochondrial activity. CDN1163 treatment upregulated gene expression of enzymes in β‐oxidation and lipid handling. No changes in basal cardiac function or compensatory left ventricular remodeling were observed.

Conclusions

SERCA activation promotes flux from nonglucose substrates to fuel cardiac mitochondrial metabolism in obese mice.

Nonstandard Abbreviations and Acronyms

AcCoA

acetyl coenzyme A

CAC

citric acid cycle

CaMKII

Ca²⁺/calmodulin‐dependent kinase II

MASLD

metabolic dysfunction‐associated steatotic liver disease

MID

mass isotopomer distribution

PGC1

peroxisome proliferator‐activated gamma coactivator 1

R_g

rate of glucose uptake

SERCA

sarco/endoplasmic reticulum Ca²⁺ ATPase

WD

Western diet

Research Perspective.

What Is New?

• Activation of calcium reuptake by SERCA (sarco/endoplasmic reticulum Ca²⁺ ATPase) promotes a shift in cardiac metabolic substrate use in obese mice away from glucose, without affecting cardiac function or morphology.

What Question Should Be Addressed Next?

• Future research should determine if this modulation of cardiac metabolism by SERCA activation has cardioprotective effects in the setting of cardiovascular disease.

• Future research should also determine the contributions of nonglucose carbon (ketones, fats, branched chain amino acids, etc) that support the citric acid cycle for cardiac energy metabolism in the setting of SERCA activation.

The prevalence of obesity is a serious threat to human health as it increases risk for numerous adverse outcomes, including death from a cardiovascular event. Energy deficits are an underlying culprit in the cause of heart failure (HF), representing an inability of metabolism to supply the heart with sufficient energy to pump blood against pressure to the body. There is evidence to suggest that energy deficits in HF may result from a shift in metabolic substrate use in cardiomyocytes. ¹ , ² , ³ , ⁴ A clear understanding of how metabolic pathway activity responds to obesity and other perturbations may help identify targets for preventing, treating, and reversing cardiac disease. The effectiveness of this strategy has been highlighted in the context of cardiac repair following acute hypoxia. For example, increased anaplerotic flux entering the citric acid cycle (CAC) improved cardiac contractility during recovery from hypoxia‐reoxygenation injury when perfused rat hearts were supplemented with fumarate. ⁵

ATP hydrolysis powers the heart for contractions, and healthy cardiomyocytes—in perfused rat hearts ⁶ or isolated from mouse and human hearts ⁷ —can use a variety of substrates to fuel mitochondrial metabolism, including glucose, lactate, fatty acids, and ketones. ⁸ , ⁹ In obesity, hearts increase uptake and oxidation of substrates, including fatty acids. This is likely, in part, a function of increased delivery of free fatty acids via the circulation due to peripheral insulin resistance. ¹⁰ , ¹¹ This substrate overload is hypothesized to also contribute to the development of cardiac disease by increasing mitochondrial stress. ¹² Cardiomyocytes from failing hearts do not exhibit the same flexibility in substrate uptake, which impairs contractile fiber function. ⁷ Furthermore, mutations in enzymes that catalyze reactions of fatty acid oxidation result in cardiomyopathies, ¹³ indicating that cardiac substrate metabolism is a key factor in maintaining and preserving cardiac function.

Calcium is a major regulator of muscle contraction as well as mitochondrial metabolism, where it can allosterically activate many of the dehydrogenases of the CAC. Dysregulated intracellular Ca²⁺ homeostasis is a hallmark of many diseases associated with obesity, including HF ¹⁴ and metabolic‐associated steatotic liver disease (MASLD). ¹⁵ SERCA (sarco/endoplasmic reticulum Ca²⁺‐ATPase) is the major transporter that actively pumps cytosolic Ca²⁺ into the sarcoplasmic reticulum of myocytes or the endoplasmic reticulum of liver hepatocytes, thereby regulating cytosolic and sarcoplasmic reticulum/endoplasmic reticulum Ca²⁺ levels. We previously tested the effects of pharmacological SERCA activation to restore hepatic Ca²⁺ homeostasis and limit MASLD severity in a mouse model of obesity using the small‐molecule SERCA activator CDN1163. ¹⁶ , ¹⁷ Studies from our group and others show that CDN1163 improves liver health in obese mice, ¹⁷ , ¹⁸ but it may also act on multiple SERCA isoforms expressed in additional tissues outside the liver. ¹⁹

Because obesity is also strongly associated with cardiovascular disease, it is important to determine the impacts of SERCA activation on cardiac metabolism and function. Notably, SERCA function is diminished in HF. ¹⁴ SERCA‐based gene therapies are currently being explored to increase SERCA2a expression and activity in HF (MUSIC‐HFrEF1 [Modulation of SERCA2a of Intra‐Myocytic Calcium Trafficking in Heart Failure With Reduced Ejection Fraction], NCT04703842; MUSIC‐HFpEF [Modulation of SERCA2a of Intra‐Myocytic Calcium Trafficking in Heart Failure With Preserved Ejection Fraction], NCT06061549). Here, we demonstrate that the pharmacological activation of SERCA with CDN1163 causes shifts to cardiac metabolism in obese mice, without impairing function, paving the way for designing therapeutics that can prevent the metabolic defects that precede a cardiac event.

Recent technological advancements in flux analysis have allowed the simultaneous assessment of metabolic fluxes in multiple tissues of the same animal using a single isotope tracer experiment. Our previous study of wild‐type (WT) and genetically obese Mc4r ^−/− mice revealed that flux dysregulation in the liver, heart, and skeletal muscle varied by tissue type and extent of disease severity. ²⁰ Cardiac fatty acid oxidation was impaired and glucose oxidation was elevated in Western diet (WD)‐fed Mc4r ^−/− mice compared with lean controls. This metabolic adaptation predisposes obese mice to developing impairments in left ventricular (LV) structure and function. ²¹ WD‐fed Mc4r ^−/− mice also develop a severe form of MASLD with fibrosis, ²² , ²³ and patients with MASLD are at a greater risk of developing cardiovascular disease. ²⁴ Mutations in MC4R are the most common monogenic form of obesity in humans, ²⁵ and variants in MC4R are associated with polygenic obesity in the general population. ²⁶ Therefore, Mc4r ^−/− mice provide a relevant model for basic research into the metabolic drivers of HF in patients with obesity.

Here we extend the fluxomics approach of Rahim et al ²⁰ to understand the effects of pharmacological SERCA activation in the heart during obesity. We hypothesized that SERCA activation would increase ATP demand due to Ca²⁺ cycling in the heart, leading to increased fat oxidation to maintain cardiac energy homeostasis. In addition, we assessed cardiac function by echocardiography to determine whether 8 weeks of CDN1163 treatment affects cardiac function. Overall, we found that CDN1163 treatment decreased cardiac glucose uptake and glycolysis while preserving cardiac function. To sustain CAC flux, increased supply of mitochondrial acetyl‐coenzyme A (AcCoA) was obtained from nonglycolytic sources, such as fat, ketones, and branched chain amino acids, thereby shifting cardiac fuel use from glucose to alternative oxidative substrates.

METHODS

Data Availability

The authors declare that all supporting data are available within the article and its supplemental material.

Animal Models and Treatments

All protocols and procedures were approved by the Vanderbilt Institutional Animal Care and Use Committee. All studies were performed on 16‐ to 17‐week‐old, Mc4r ^−/− (knockout [KO]) and Mc4r ^+/+ (WT) male mice from Mc4r ^+/− breeders backcrossed to C57Bl/6J background for at least 10 generations. Mice were maintained on a 12‐hour light–dark cycle with ad libitum access to water and a standard rodent chow diet (5L0D, 29% protein, 58% carbohydrates, 13% fat by caloric contribution; LabDiet, St. Louis, MO) until 8 weeks of age. At 8 weeks of age, Mc4r ^−/− mice were switched to WD (D12079B, 17% protein, 43% carbohydrates, and 40% fat by caloric contribution; Research Diets Inc., New Brunswick, NJ), a regimen that induces MASLD with hepatic fibrosis. ²³ Half of the WD‐fed mice were randomly assigned to be given intraperitoneal injections of CDN1163 (50 mg/kg body weight) thrice weekly for the course of the diet period. All remaining animals were administered vehicle injections consisting of 10% DMSO (Sigma Aldrich D8418), 10% Tween 80 (Sigma Aldrich P1754) in 0.9% saline (Baxter 2F7123). CDN1163 was synthesized by the Vanderbilt Molecular Design and Synthesis Center, and fresh solutions were prepared weekly. WT littermates maintained on chow diet and receiving vehicle injections for the entire treatment period served as a lean control group.

In Vivo Tracer Infusions and Sample Collection

One week before isotope infusion studies, jugular vein and carotid artery catheters were surgically implanted in 16‐week‐old mice for infusion and sampling of blood, respectively. ²⁷ In vivo studies were performed by infusing a cocktail of tracers into mice through a jugular vein catheter over a total time course of 120 minutes, following an overnight fast (16 hours). A primed (200 μmol/kg), continuous (50 μmol/kg per min) infusion of sodium [¹³C₃]lactate was administered for 120 minutes. A 12 μCi bolus of [1‐¹⁴C]2‐deoxyglucose was given 25 minutes before the end of each infusion. Blood samples were collected from a carotid artery catheter just before [¹³C₃]lactate infusion, and then at 2, 5, 10, 15, and 25 minutes after the bolus of [1‐¹⁴C]2‐deoxyglucose. Hematocrit was maintained through an infusion of donor erythrocytes (suspended in 10 U/mL heparinized saline with tetrahydrolipstatin); blood glucose concentrations were monitored with a glucometer. After the final blood sample, mice were rapidly euthanized by cervical dislocation, and excised tissues were snap‐frozen in liquid nitrogen for further analysis. Separate cohorts of age, sex, and treatment‐matched mice from each group were maintained similarly—except without surgery or isotope tracer infusions—for in vivo cardiac function tests and collection of samples for gene expression analysis and enzymatic assays, following a short fast (5 hours). For the latter, mice were placed in a restrainer before obtaining plasma (isolated from blood from the cut tail), and tissues were rapidly harvested and snap‐frozen in liquid nitrogen post euthanasia. All plasma and tissue samples were stored at −80 °C until processed for further analysis.

In Vivo Cardiac Function Tests

In vivo cardiac functional parameters were evaluated with transthoracic echocardiography in conscious mice with the Vevo2100 Imaging System (VisualSonics, Inc.) at the end of the diet period. Following chest fur removal, prewarmed echo transmission gel was applied to the chest wall before the acquisition of echo images, and parasternal long‐ and short‐axis view at the papillary muscle level and 2‐D guided M‐mode images were recorded. LV diameter at end‐systole and end‐diastole, end‐systolic and end‐diastolic posterior and anterior wall thickness, ejection fraction, and fractional shortening were measured in 3 consecutive beats according to the guidelines and standards of the American Society of Echocardiography leading edge method. ²⁸ Qualitative and quantitative measurements were calculated by a reviewer masked to groups using the VisualSonics VEVO2100 Imaging System software.

Cell Culture

Human induced pluripotent stem cell derived cardiac myocytes (CMM‐100‐012‐000.5, Cellular Dynamics, Madison, WI) were cultured as per the manufacturer’s instructions in proprietary manufacturer‐provided cardiac myocyte maintenance medium in 35‐mm tissue culture dishes (Cellvis D35‐14‐1.5P). Cells were maintained at 37 °C and 5% CO₂.

Calcium Imaging

Cells were treated with 100 μM CDN1163 for 4 hours before loading with 2 μM Fura‐2 AM (Invitrogen) for 25 minutes. This CDN1163 concentration was chosen based on a previous study of skeletal muscle myotubes. ²⁹ Cells were then washed, incubated (for 15 minutes), and perfused with Krebs‐Ringer HEPES buffer containing (mM) 119.0 NaCl, 4.7 KCl, 1.2 MgSO₄, 1.2 KH₂PO₄, and 10.0 HEPES (pH=7.35, adjusted by NaOH) supplemented with 2 mM glucose. ³⁰ Fura‐2 AM Ca²⁺ fluorescence (emission at 488 nm) was measured in response to 340 and 380 nm excitation (F340/F380) every 5 seconds as an indicator of intracellular Ca²⁺ using a Ti2 microscope (Nikon) and a back‐illuminated sCMOS Prime 95B camera (Teledyne Photometrics).

Metabolite Extraction and Derivatization

Plasma and tissue metabolites were extracted and derivatized as described elsewhere. ³¹ Briefly, plasma glucose was extracted using cold acetone to precipitate protein. Samples were air dried followed by immediate conversion into 3 separate glucose derivatives (di‐O‐isopropylidene, methyloxime pentapropionate or aldonitrile pentapropionate) according to protocols described elsewhere. ³² Polar metabolites were isolated from 40 μL of plasma or 30 to 50 mg of heart or liver tissue using a biphasic methanol/water/chloroform extraction. An internal standard of 75 μM norvaline was added for metabolite quantification. The polar and nonpolar layers of the extract were isolated using a fine‐tipped pipette and air‐dried overnight for storage at −80 °C before derivatization.

Metabolites from the polar layer were converted to their methyloxime tert‐butyldimethylsilyl derivatives using MtBSTFA+1% TBDMCS (Regis Technologies 1‐270144‐200). Derivatized samples were analyzed by gas chromatography mass spectrometry. Sample volumes of 1 μL were injected using a 5:1 split into an Agilent 7890A gas chromatography system equipped with 2 HP‐5 ms (15 m×0.25 mm×0.25 μm; Agilent J&W Scientific) capillary columns and interfaced with an Agilent 5977C mass spectrometer. Previously defined temperature programs for methyloxime tert‐butyldimethylsilyl ³³ and glucose derivatives ³² were used for data collection. Derivative peaks were integrated using a custom MATLAB function ³⁴ to obtain mass isotopomer distributions (MIDs) for the metabolite fragment ions shown in Table S1.

Metabolic Flux Analysis

Metabolic flux analysis was performed by minimizing the sum of squared residuals between the model‐simulated and experimentally measured metabolite MIDs summarized in Table S1. The full MIDs of all modeled metabolites (after correction for natural isotope abundance) are presented in Figure S1. The Isotopomer Network Compartmental Analysis software package ³⁵ was used to develop metabolic atom‐mapping models and to determine all fluxes by least‐squares regression. The reaction network used for modeling fluxes consists of glycolytic and CAC reactions in the heart, as well as gluconeogenic reactions in the liver that clear circulating lactate and supply glucose to the heart under fasting conditions. Unlabeled carbon sources entering glycolysis from glycogen breakdown and supplying AcCoA from fat oxidation were included in the model, as well as carbon sinks for alanine, lactate, and glutamate, based on physiological considerations described in prior studies. ³⁶ , ³⁷ , ³⁸ The complete reaction network is provided in Table S2.

Plasma, liver, and heart metabolite MIDs were provided as measurements for Isotopomer Network Compartmental Analysis. Measurement uncertainty was assessed by calculating the root‐mean‐square deviation between the MID of unlabeled samples from control mice (not infused with tracers) and the theoretical MID computed from the known abundances of naturally occurring isotopes. Best‐fit flux solutions were determined for each animal by fitting all the experimental MID measurements (Table S1) to the multicompartment isotopomer network model (Table S2). To ensure a global solution was obtained, flux estimations were repeated a minimum of 100 times from randomized initial guesses. A chi‐square test was used to assess goodness of fit, and sensitivity analysis was performed to determine 95% CIs associated with the calculated flux values. Initially, fluxes in the hepatic and extrahepatic compartments were estimated relative to liver citrate synthase flux by constraining V_CS.l to an arbitrary value of 100, and relative cardiac fluxes were estimated by setting the hexokinase flux (V_HK.h) to 100. The absolute V_HK.h flux was assumed to equal cardiac R_g, an index of tissue‐specific glucose uptake determined from [1‐¹⁴C]2‐deoxyglucose administration. ³⁹

Stoichiometric Analysis of ATP Production

Best‐fit metabolic flux analysis solutions were used to estimate net ATP production rate within the central metabolic network of cardiomyocytes. Fluxes of nucleotide triphosphate‐consuming reactions (hexokinase + phosphofructokinase + pyruvate carboxylase) were subtracted from those of nucleotide triphosphate‐producing reactions (phosphoglycerate kinase + pyruvate kinase + succinyl‐CoA synthetase). Similarly, the fluxes of reactions involved in the interconversion of NAD⁺ and NADH (glycerol 3‐phosphate dehydrogenase + glyceraldehyde 3‐phosphate dehydrogenase + lactate dehydrogenase + pyruvate dehydrogenase + isocitrate dehydrogenase + α‐ketoglutarate dehydrogenase + malate dehydrogenase) were summed to estimate the net NADH production rate. One reaction in the CAC produces FADH₂ (succinate dehydrogenase). Finally, using P/O ratios of 2.7 for NADH and 1.6 for FADH₂ ⁴⁰ and assuming that all nucleotide triphosphates are energetically equivalent to ATP, the net ATP production rate was calculated, and ATP yield was determined after normalization to glucose uptake.

Enzymatic Assays

Frozen ventricular tissues were homogenized in CelLytic MT buffer (Sigma Aldrich C3228) with 1 mM phenylmethylsulfonyl fluoride (Research Products International P20270), Halt Protease and Phosphatase Inhibitor, and 5 mM EDTA (Thermo Fisher 78442) in a 1:20 ratio of tissue mass to lysis buffer. The extract was then centrifuged at 16 000×g for 10 minutes at 4 °C. The supernatant was transferred to a clean tube, and protein concentration was measured with the bicinchoninic acid method. ATPase activity was quantified using an enzymatic assay (Abcam ab43055) on lysate diluted to working range with assay buffer according to the manufacturer’s protocol. Citrate synthase activity of cell lysates was quantified using a reagent kit (Sigma Aldrich CS0720) according to the manufacturer’s protocol.

Western Blotting

Protein lysates from enzymatic assays were also used for Western blotting. Samples were added at 15 μg protein/lane for separation with NuPAGE 4% to 12% Bis Tris gels and MOPS‐SDS at 120 V and transferred to nitrocellulose membranes. Primary antibodies were directed against total CaMKII (panCaMKII [pan‐Ca²⁺/calmodulin‐dependent kinase II]; made at Vanderbilt University by the laboratory of Roger Colbran ⁴¹ ), CaMKII autophosphorylated at T286/7 (pCaMKII pT287; Santa Cruz Biotechnology sc‐12 886‐R), and PGC1 (peroxisome proliferator‐activated gamma coactivator 1; ABclonal A19674). Sample loading was assessed by GAPDH (Cell Signaling Technology 2118) or β‐actin (Cell Signaling Technology 4970) and Ponceau staining (Thermo Scientific A40000279). Secondary antibody (anti‐rabbit HRP [horseradish peroxidase], Cell Signaling Technology 7074; anti‐goat HRP, Santa Cruz Biotechnology sc‐2354) was incubated for 1 hour, and visualization was performed using enhanced chemiluminescence substrate (Thermo Scientific 34580) and an Odyssey FC (LiCor) gel imaging system. Intensities of pCaMKII were normalized to panCaMKII within each lane and to vehicle treatment within each membrane. Intensities of PGC1 were normalized to β‐actin within each lane and to vehicle treatment within each membrane.

Gene Expression Analysis

Total RNA was isolated from 20 to 30 mg ventricular tissue with the use of TRIzol reagent (Invitrogen 15596026). RNA yield was quantified with the NanoDrop 2000 spectrophotometer (Thermo Scientific). From isolated RNA, cDNA was synthesized with iScript cDNA Synthesis Kit (BioRad 1708891), diluted to 5 ng/μL, mixed with custom primer sets (Integrated DNA Technologies) and iQ SYBR Green Supermix (BioRad 1 708 880), then analyzed on a CFX96 Real‐Time PCR System (BioRad). Relative gene expression was normalized to peptidylprolyl isomerase A (Ppia) using the 2^−ΔΔCt method. Primer sets are provided in Table S3.

Statistical Analysis

All statistical analysis between experimental groups was performed in GraphPad Prism 10. Statistical tests performed are indicated in figure legends.

RESULTS

Pharmacological SERCA Activation Reduces Uptake and Oxidation of Glucose by the Heart

Mc4r ^−/− mice were fed WD and treated with intraperitoneal injections of either CDN1163 or vehicle for 8 weeks. We confirmed that ATPase activity was elevated in the hearts of CDN1163‐treated mice, indicative of in vivo SERCA activation (Figure 1A). Additionally, hearts from CDN1163‐treated mice had decreased T287 autophosphorylation of CaMKII (Figure 1B and 1C), a sensor of cytosolic Ca²⁺. The combination of increased ATPase activity and decreased cytosolic Ca²⁺ signaling demonstrates in vivo cardiac SERCA activation with CDN1163. To further confirm that CDN1163 alters intracellular Ca²⁺ transport in cardiomyocytes, we treated human induced pluripotent stem cell‐derived cardiac myocytes with CDN1163 and measured cytosolic Ca²⁺ levels using fluorescence microscopy. Cells treated with CDN1163 had lower basal levels of cytosolic Ca²⁺, consistent with increased sarcoplasmic reticulum uptake of cytosolic Ca²⁺ due to SERCA activation in cultured cardiomyocytes (Figure S2).

Figure 1. CDN1163 exerts direct effects on cardiac metabolism.

A, ATPase activity measured in ventricular lysates of KO WD mice after 8 wks of vehicle or CDN1163 injections. B, Representative blots of phospho‐CaMKII T287, pan‐CaMKII, and GAPDH. C, Quantification of pCaMKII expression relative to pan‐CaMKII. D, Overview of isotope infusion protocol: a primed (0.200 mmol/kg), continuous (0.050 mmol/kg per min) infusion of [U‐¹³C₃]lactate was administered for 120 min, and a 12 μCi bolus of [1‐¹⁴C]2DG was given 25 min before the end of the infusion. E, Cardiac glucose uptake from 2DG radiotracer measurements. F, Tail blood glucose levels of mice following an overnight fast. G, Plasma free fatty acids of mice following an overnight fast. H, Plasma insulin levels of mice following an overnight fast. (I) Atom percent enrichment of plasma lactate, (J) plasma glucose, and (K) heart lactate. Data represent means±SEM, with unpaired t test (A and C) or 1‐way ANOVA with Tukey’s post hoc test (E–K), where *P<0.05, **P<0.01, and ****P<0.0001. 2DG indicates 2‐deoxyglucose; CaMKII, Ca²⁺/calmodulin‐dependent kinase II; KO, knockout; WD, Western diet; and WT, wild type.

At the end of the 8‐week period of WD feeding (with vehicle or CDN1163 treatment), jugular vein and carotid artery catheters were surgically implanted to enable intravenous infusions and blood sampling, respectively, and the mice were allowed to recover for 1 week. A separate cohort of lean control mice (age‐matched WT littermates maintained on chow diet and administered vehicle treatments for 8 weeks) was similarly equipped with catheters. Following an overnight fast, conscious and unrestrained mice received concurrent infusions of [¹³C₃]lactate and [1‐¹⁴C]2‐deoxyglucose over a 2‐hour time course (Figure 1D).

Specific activity measurements of ¹⁴C in heart tissue and plasma samples were used to determine cardiac R_g. CDN1163 treatment caused a substantial reduction in cardiac R_g in obese WD‐fed mice compared with vehicle‐treated animals (Figure 1E). Cardiac R_g of lean control animals was at an intermediate level between these 2 groups. However, the difference in R_g between vehicle‐ and CDN1163‐treated obese mice was not due to changes in glucose or free fatty acid availability, as circulating concentrations of both sources were unchanged by CDN1163 treatment (Figure 1F and 1G). CDN1163 treatment decreased circulating fasting plasma insulin levels (Figure 1H), which may contribute to the decrease in cardiac glucose uptake. However, both obese groups were hyperinsulinemic and hyperglycemic compared with lean controls (Figure 1F and 1H).

To further investigate how changes in glucose uptake correspond to flux rerouting in downstream pathways of intermediary metabolism, we measured the ¹³C enrichments of metabolites extracted from plasma and heart tissues (Table S1, Figure S1) of animals infused with [¹³C₃]lactate. Both circulating lactate (Figure 1I) and circulating glucose (Figure 1J) were enriched with ¹³C, the latter derived from gluconeogenesis with ¹³C‐lactate as substrate. Importantly, the infusion of [¹³C₃]lactate did not change circulating levels of lactate or glucose in any of the 3 groups studied, as pre‐ and postinfusion concentrations of each metabolite remained the same (Figure S3). Uptake and metabolism of circulating ¹³C‐labeled carbon sources enriched tissue lactate within the heart (Figure 1K). Note that differences in the atom percent enrichment of plasma glucose between groups mirrors the differences in the atom percent enrichment of plasma lactate. Because all mouse groups received the same infusion rate of [¹³C₃]lactate, the increase in circulating lactate atom percent enrichment in the KO WD CDN1163 group indicates a decrease in the endogenous lactate turnover rate of CDN1163‐treated animals.

We normalized the atom percent enrichment of measured heart metabolites to that of heart lactate to determine the fractional contribution of lactate carbon to each downstream analyte (Figure 2A and 2B). ⁴² Heart lactate acquires ¹³C from both the uptake of enriched plasma lactate and from the conversion of labeled glucose to pyruvate, because lactate rapidly equilibrates with pyruvate due to lactate dehydrogenase activity. Therefore, fractional contributions of CAC intermediates from lactate reflect the oxidation of glycolytic intermediates that enter the CAC via pyruvate. In CDN1163‐treated mice, there was a decrease in the fractional contributions of lactate to the formation of citrate and glutamate (Figure 2A and 2B), indicating a reduction in the glycolytic contribution to the CAC. Although the hearts of KO WD mice fueled their CAC almost entirely from glycolytic sources, CDN1163 treatment restored an intermediate balance of glycolytic and nonglycolytic sources (~50% each) that closely matched results obtained in lean control mice (Figure 2A and 2B). Consistent with these fractional contributions, the ratio of M+2 Citrate/M+3 phosphoenolpyruvate also decreased in response to CDN1163 treatment (Figure 2C), indicating a reduction in the relative contribution from glycolysis to the oxidative arm of the CAC. ⁴³

Figure 2. Isotope infusions reveal differences in metabolite labeling in CDN1163‐treated obese mice.

Fractional contribution of (A) cardiac lactate to cardiac citrate and (B) cardiac lactate to cardiac glutamate. C, Ratio of cardiac M+2 Cit to cardiac M+3 phosphoenolpyruvate. D, Fractional contribution of cardiac citrate to cardiac succinate and (E) plasma glucose to cardiac PEP. Fractional contributions were calculated as the APE of the product metabolite normalized to the APE of its indicated upstream substrate. Data represent means±SEM, with 1‐way ANOVA with Tukey’s post hoc test, where *P<0.05, ***P<0.001, and ****P<0.0001. APE indicates atom percent enrichment; Cit, cardiac citrate; Glc, plasma glucose; Glu, cardiac glutamate; KO, knockout; Lac, cardiac lactate; PEP, phosphoenolpyruvate; Suc, cardiac succinate; WD, Western diet; and WT, wild type

In contrast to the shift in lactate fractional contributions, there was no difference in the fractional contribution of citrate to succinate (Figure 2D), suggesting that anaplerotic and cataplerotic routes of carbon flux to/from the CAC were not impacted by CDN1163 treatment. Furthermore, there was no difference in the fractional contribution of plasma glucose to phosphoenolpyruvate among all 3 groups (Figure 2E). This finding indicates that although lactate fractional contributions and cardiac R_g were reduced by CDN1163 treatment, the molar oxidation rate of glucose to phosphoenolpyruvate remained unchanged, suggesting that the contribution from nonglucose sources to the glycolytic pathway was consistent across all groups. Therefore, the major effect of SERCA activation on cardiac metabolism was to reduce the oxidation of pyruvate derived from ¹³C‐enriched glucose and lactate.

Metabolic Flux Analysis Reveals Substrate Switching in Response to SERCA Activation With CDN1163

To further explore the metabolic differences between obese vehicle‐ and CDN1163‐treated mice, we performed metabolic flux analysis by fitting a comprehensive atom‐mapping model (Figure 3A, Table S2) to the ¹³C‐labeling measurements of Table S1. ²⁰ , ³⁷ In a previous study of WD‐fed Mc4r ^−/− mice, we found that fatty acid oxidation was impaired, and glucose oxidation was elevated compared with chow‐fed WT mice. ²⁰ Here, treatment with CDN1163 lowered glucose uptake (represented by V_HK.h in the reaction network of Figure 3A) even below the rate observed in lean control mice (Figure 1E). The decrease in V_HK.h corresponded to a relative increase in the flux of AcCoA from nonglucose sources (eg, fatty acids, ketones), represented in the model as a single β‐oxidation flux (V_βOx.h) for simplicity (Figure 3B). This shift in substrate use results in a repartitioning of fluxes that contribute to AcCoA production, indicated by a switch in the ratio of β‐oxidation flux to pyruvate dehydrogenase flux (V_βOx.h/V_PDH.h) from <1 in vehicle‐treated mice to >1 in CDN1163‐treated mice (Figure 3C). The relative increase in β‐oxidation flux partially compensated for reduced glucose uptake in the heart by redirecting carbon from alternative sources into the CAC for oxidation, reflected by an increase in citrate synthase flux relative to glucose uptake (V_CS.h/V_HK.h) in CDN1163‐treated mice compared with their vehicle‐treated littermates (Figure 3D). Additionally, there was no evidence of increased cataplerosis in the hearts of CDN1163‐treated mice, as flux through succinate dehydrogenase mirrored the change in citrate synthase flux (Figure 3E).

Figure 3. CDN1163 treatment leads to a distinct cardiac metabolic flux profile in the hearts of obese mice.

A, Overview of reaction network used for modeling (see Table S2 for more details). Mass isotopomer distributions of underlined metabolites were used for modeling (see Table S1 and Figure S1 for more details). B, Flux of βOx relative to glucose uptake (ie, HK). C, Flux of βOx relative to PDH. D, Flux of CS relative to HK. E, Flux of SDH relative to HK. Data represent means±SEM, with unpaired t test, where *P<0.05. βOx indicates β‐oxidation; CS, citrate synthase; HK, hexokinase; PDH, pyruvate dehydrogenase; SDH, succinate dehydrogenase; and V, flux.

Eight Weeks of Treatment With CDN1163 Does Not Affect Cardiac Function or Morphology

Given that alterations in the heart’s ability to use metabolic fuel substrates contributes to cardiomyopathy, ¹³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ we asked if the differences in metabolism caused by CDN1163 treatment would correspond to alterations in cardiac function. Cardiac function was assessed in conscious mice with transthoracic echocardiography and revealed no changes in cardiac output (Figure 4A), stroke volume (Figure 4B), ejection fraction (Figure 4C), and fractional shortening (Figure 4D). There was no compensation by LV hypertrophy, as evidenced by no change in LV mass (Figure 4E). Body composition analysis of the mice also did not indicate a change in lean mass in CDN1163‐treated mice, so the LV mass normalized to lean mass was also unchanged (Figure 4F). Additionally, end‐diastolic and end‐systolic LV anterior wall thickness (Figure 4G and 4H), as well as end‐diastolic and end‐systolic LV posterior wall thickness (Figure 4I and 4J), were the same between vehicle‐ and CDN1163‐treated mice. With no changes to wall thicknesses, the end‐diastolic and end‐systolic LV inner diameter (Figure 4K and 4L) also remained unaffected by CDN1163 treatment. In all, despite substantial changes to cardiac glucose uptake and energy metabolism, CDN1163 treatment did not impair cardiac function in obese Mc4r ^−/− mice fed WD for 8 weeks. However, we cannot rule out the possibility that a longer treatment time course is needed to reveal functional differences between these groups.

Figure 4. SERCA activation by CDN1163 has no effect on cardiac function.

Echocardiographic measurements of (A) cardiac output, (B) stroke volume, (C) ejection fraction, (D) fractional shortening, and (E) corrected LV mass. F, Corrected LV mass normalized to lean body mass. G, LVAWd thickness. H, LVAWs thickness. I, LVPWd thickness. J, LVPWs thickness. K, LVIDd. L, LVIDs. Data represent means±SEM, with unpaired t test, where *P<0.05. LV indicates left ventricular; LVAWd, end‐diastolic LV anterior wall; LVAWs, end‐systolic LV anterior wall; LVIDd, end‐diastolic LV inner diameter; LVIDs, end‐systolic LV inner diameter; LVPWd, end‐diastolic LV posterior wall; LVPWs, end‐systolic LV posterior wall; and SERCA, sarco/endoplasmic reticulum Ca²⁺ ATPase.

SERCA Activation Upregulates Genes Involved in Energy Production From Fat Oxidation in the Obese Mouse Heart

To sustain cardiac function, hearts of CDN1163‐treated mice increased energy production from oxidation of nonglucose substrates. To assess whether the shift from glucose to nonglucose substrates correlated with an upregulation of β‐oxidation and ketone oxidation capacity, we profiled changes in gene expression within these pathways (Figure 5A). Gene expression of LCAD (long chain acyl CoA dehydrogenase; gene: Acadl) was increased >2‐fold (Figure 5B), whereas MCAD (medium chain acyl CoA dehydrogenase; gene: Acadm) was unchanged by CDN1163 treatment (Figure 5C). Although there was an increase in the gene expression of Acadl, we did not see a similar increase in expression for the genes catalyzing the downstream reactions of β‐oxidation (Figure 5D through 5F). However, the reaction catalyzed by acyl CoA dehydrogenases is considered the major rate‐limiting step of β‐oxidation. ⁴⁸ Although we did not observe a change in circulating free fatty acids (Figure 1F), we saw a decrease in expression of Cd36, a cytosolic fatty acid importer (Figure 5G) along with an increase in the expression of fatty acid binding protein 3 (Fabp3) (Figure 5H), a cytosolic chaperone of lipids, in hearts from CDN1163‐treated mice, possibly indicating an increased demand for fatty acid trafficking within cardiomyocytes. There was no change in the expression of genes within the ketone oxidation pathway (Figure 5I through 5K). Altogether, CDN1163 treatment upregulated key genes involved in fatty acid oxidation and lipid handling, providing further evidence that cardiac function was preserved by switching from glucose to fat oxidation.

Figure 5. SERCA activation induces expression of β‐oxidation genes.

A, Overview of β‐oxidation and ketone oxidation pathways. B, Gene expression of long chain acyl CoA dehydrogenase (Acadl). C, Gene expression of medium chain acyl CoA dehydrogenase (Acadm). D, Gene expression of enoyl CoA hydratase (Echs1). E, Gene expression of 3‐hydoxyacyl CoA dehydrogenase (Hadh). F, Gene expression of β ketothiolase (Hadhb). G, Gene expression of fatty acid translocase (Cd36). H, Gene expression of fatty acid binding protein 3 (Fabp3). I, Gene expression of β‐hydroxybutyrate dehydrogenase (Bdh). J, Gene expression of succinyl CoA:3‐oxoacid CoA transferase (Oxct1). K, Gene expression of acetyl CoA acetyltransferase (Acat1). Data represent means±SEM, with unpaired t test, where *P<0.05 and ****P<0.0001. CoA indicates coenzyme A; and SERCA, sarco/endoplasmic reticulum Ca²⁺ ATPase.

We hypothesized that upregulation of β‐oxidation and increases in CAC flux relative to glycolytic flux were necessary to preserve ATP levels in the heart despite the dramatic reduction in glucose uptake caused by CDN1163 treatment. Stoichiometric analysis of the best‐fit flux solution enabled estimation of the net ATP yield per mole of glucose taken up by accounting for the net production of reduced cofactors, substrate level phosphorylation and dephosphorylation, and respiratory ATP generation within the metabolic network. This analysis revealed that the yield of ATP from all substrates, normalized to the glucose uptake rate, was increased in CDN1163‐treated mice compared with vehicle‐treated animals (Figure 6A). Furthermore, citrate synthase enzymatic activity, commonly used as an index of total mitochondrial mass, ⁴⁹ , ⁵⁰ , ⁵¹ was increased in hearts of CDN1163‐treated mice (Figure 6B). This was consistent with an increase in the protein expression of PGC1 (Figure 6C and 6D), a transcriptional coactivator of mitochondrial biogenesis and energy metabolism. ⁵² These changes occurred in combination with reduced gene expression of Ucp2 (Figure 6E), which encodes UCP2 (uncoupling protein 2). UCP2 is a mitochondrial proton pump that dissipates the proton gradient used for oxidative phosphorylation. ⁵³ Taken together, these measurements suggest an increase in mitochondrial efficiency for cardiac energy production in CDN1163‐treated mice through shifts in substrate use, increases in mitochondrial mass, and improved coupling of electron transport to ATP production.

Figure 6. CDN1163 treatment increases mitochondrial efficiency.

A, Total ATP yield per mole of glucose consumed, calculated from best‐fit flux values. B, CS activity measured in ventricular lysates as a proxy of mitochondrial mass and capacity. C, Western blot of PGC1 and (D) quantification of expression relative to β‐actin. E, Gene expression of Ucp2. Data represent means±SEM, with unpaired t test, where *P<0.05 and **P<0.01. CS indicates citrate synthase; and PGC1, peroxisome proliferator‐activated gamma coactivator 1.

DISCUSSION

The transition from obesity to cardiovascular disease to HF is multifactorial but is commonly associated with energetic defects stemming from metabolic adaptations that can no longer support function. ² , ⁴ , ⁷ , ⁵⁴ , ⁵⁵ , ⁵⁶ Therefore, therapies that prevent or shift these changes at an early stage may be helpful in preventing progression of the disease. ⁵⁷ These studies demonstrate that pharmacological targeting of SERCA may be a viable strategy to manipulate cardiac metabolism without compromising cardiac function in obese mice.

We were initially surprised by our finding of decreased cardiac glucose uptake in mice treated with CDN1163, as we expected that increasing ATPase activity would increase glucose uptake and oxidation. Molecular studies show that both insulin and free Ca²⁺ can independently stimulate GLUT4 translocation to the plasma membrane in cardiomyocytes, promoting glucose uptake. ⁵⁸ We found that plasma insulin and heart cytosolic Ca²⁺ were both decreased by CDN1163 treatment, corresponding to the reduced cardiac glucose uptake and glycolytic flux observed in our studies. Although the reduction in circulating insulin likely contributed to these effects, ²⁰ it is important to note that KO WD CDN1163 animals had insulin levels that were intermediate between WT chow and KO WD animals, whereas their cardiac R_g was below that of the WT chow group. These data suggest that additional factors, including a lowering of cytosolic Ca²⁺, may explain the disproportionate effects of CDN1163 on cardiac glucose metabolism.

Hearts subject to pressure overload and hypertension leading to LV hypertrophy exhibit greater reliance on glucose compared with fatty acids. ⁵⁹ , ⁶⁰ Initially, the greater reliance on glucose oxidation may be adaptive, as it demands less oxygen than fatty acid metabolism. ⁶¹ This effect also explains why cardiac glucose oxidation promotes repair in ischemia–reperfusion, but continued reliance on glycolysis for energy homeostasis may become pathological if prolonged over extended periods of time. ⁶² Mice that have increased rates of fatty acid oxidation due to deletion of ACC2 (AcCoA carboxylase 2) are protected from LV hypertrophy even when subjected to pressure overload, indicating that the metabolic protection afforded by increased fatty acid oxidation is important for preventing heart disease progression. ⁶³ Obese, CDN1163‐treated mice exhibit a metabolic phenotype that mirrors, in part, that of ACC2 deletion. It stands to reason that prophylactic CDN1163 treatment may also protect from pathological hypertrophy in prolonged pressure overload and ultimately HF with reduced ejection fraction, but this hypothesis would require testing in a translational mouse model.

In contrast, a decrease in glycolytic flux is consistent with the metabolic phenotype of diabetic cardiomyopathy. ⁶⁴ However, unlike some other hyperphagic mice, Mc4r ^−/− mice are not a model of diabetic cardiomyopathy, and they do not show the metabolic adaptations of diabetic cardiomyopathy on either a standard chow diet or WD. ²⁰ CDN1163 is effective at improving glucose tolerance, decreasing circulating insulin, and reducing the homeostatic model for the assessment of insulin resistance score in obese mice ¹⁷ , ¹⁸ and, by doing so, may prevent the development of diabetic cardiomyopathy. Future studies exploring the effects of CDN1163 in an appropriate disease model remain necessary to determine the effectiveness of CDN1163 in treating or preventing diabetic cardiomyopathy.

The reduction in glucose uptake and oxidation was compensated by increased reliance on nonglucose substrates, which sustained CAC flux and even increased its rate relative to glucose uptake. Furthermore, the source of AcCoA shifted away from glycolytic production through PDH (pyruvate dehydrogenase) and toward alternative sources, which is most likely the oxidation of fatty acids but could also include branched chain amino acids and ketone bodies. We found upregulation of Acadl, a gene encoding the rate‐limiting step of the fatty acid β‐oxidation pathway, ⁴⁸ which provides evidence for increased fat metabolism in CDN1163‐treated hearts. However, we did not find a similar upregulation of Oxct1, the rate‐limiting step of ketone oxidation. ⁶⁵ It is important to acknowledge that changes in enzyme expression do not necessarily indicate changes in flux, ⁶⁶ and so future studies with fatty acid or ketone tracers will be required to define the ultimate source of the nonglucose carbon supplying AcCoA in the hearts of CDN1163‐treated mice. This question is clinically significant because ketones become a major substrate for energy generation in failing hearts, ⁶⁵ and impairments in the ability of cardiomyocytes to oxidize ketones worsen the pathological remodeling that follows pressure overload. ⁶⁷

We also observed increases in cardiac citrate synthase activity and PGC1 expression in hearts of CDN1163‐treated mice, which are indicators of enhanced mitochondrial capacity. As the reaction catalyzed by LCAD occurs in mitochondria, increases in its expression and activity could necessitate increased mitochondrial mass. Additionally, by activating SERCA2 and increasing ATPase activity, the cellular ATP demand is expected to increase. Although SERCA is not the only cellular consumer of ATP, the energetic burden of sustaining increased ATPase activity may have increased mitochondrial flux and mass to preserve cardiac function. We also observed evidence for increased coupling of electron transport to oxidative phosphorylation, as indicated by decreases in Ucp2 expression in hearts of CDN1163‐treated mice. Together, these changes can increase mitochondrial capacity for ATP production, which was further investigated by calculating the total ATP yield per glucose consumed. The calculated ATP yield was found to be significantly elevated in response to CDN1163 treatment, stemming from an increase in the production of reduced cofactors from increased CAC flux. Although this calculation assumes the same coupling constants for each group of animals, the actual ATP yield would be even higher if respiratory coupling is increased by CDN1163. Furthermore, the calculated ATP yield did not consider the production of reduced cofactors by the oxidation of fatty acids, ketones, or branched chain amino acids, which produce NADH+H⁺ and FADH₂ accompanying the production of AcCoA. The specific contributions of these nonglucose sources supplying AcCoA can be determined in future studies with fatty acid and ketone tracers.

These metabolic changes did not lead to altered basal cardiac function at the end of the 8‐week treatment period. Mc4r ^−/− mice have been previously shown to develop dilated cardiomyopathy with diminished cardiac contractility by 26 weeks of age, but not prior. ⁶⁸ Mice used in this study were only 16 to 17 weeks of age and did not show signs of dilated cardiomyopathy, and CDN1163 did not accelerate progression of cardiomyopathy in these mice. SERCA can also be activated with an endogenous peptide, Dwarf Open Reading Frame. ⁶⁹ Overexpression of Dwarf Open Reading Frame improved contractility in a mouse model of dilated cardiomyopathy but did not affect cardiac function in WT mice. Similarly, we would not expect CDN1163 to negatively affect cardiac function in dilated cardiomyopathy, and we did not observe any negative effects on cardiac function in CDN1163‐treated mice.

Our experiments with CDN1163 and its effects on cardiac metabolism, function, and energetics were confined to an obese state with no overt cardiomyopathy or other cardiac diseases. The effects on cardiac metabolism were confined to an overnight‐fasted state, whereas cardiac function was measured after a short fast. An overnight fast was chosen to be consistent with prior in vivo metabolic tracer studies performed by our group ¹⁷ , ²⁰ , ³¹ , ⁷⁰ , ⁷¹ and others. ⁷² , ⁷³ This standard approach improves reproducibility by maintaining a well‐defined metabolic state, and it also maximizes ¹³C enrichment due to the tracer infusion without perturbing endogenous lactate levels. Genetic (Mc4r ^−/− ) and dietary (WD feeding) manipulations rendered the hearts of these obese mice more reliant on glucose for the production of AcCoA compared with lean (Mc4r ^+/+ ) mice. In contrast, CDN1163 treatment of obese mice restored relative AcCoA fluxes to a state representative of lean mice. ²⁰ Additionally, CDN1163‐treated obese mice repartitioned sources of AcCoA production such that CAC flux and maximal ATP yield both increased relative to glucose uptake. Obesity is a risk factor for developing hypertension, ⁷⁴ and prior studies show that this shift in substrate preference is protective against developing pressure overload induced hypertrophy. Although Mc4r ^−/− mice on WD are not known to be a model for hypertension, ⁷⁵ we hypothesize that the metabolic effects of CDN1163 would be protective in a HF model.

Conclusions

In summary, isotope tracers and metabolic flux analysis were used to determine the metabolic effects of SERCA activation by CDN1163 treatment on the hearts of obese mice. We found that CDN1163 had direct effects to increase ATPase activity and alter intracellular Ca²⁺ partitioning in hearts. This was associated with decreased cardiac glucose uptake and glycolysis but preserved cardiac function. To sustain CAC flux, a greater fraction of mitochondrial acetyl‐CoA was obtained from nonglycolytic sources, such as fat. This was consistent with enzymatic assays showing an increase in mitochondrial capacity and increased expression of genes involved in β‐oxidation and lipid handling. These cardiac adaptations occurred alongside positive changes to liver metabolism and improvement in MASLD severity. ¹⁷ We anticipate that the mechanism of action of CDN1163 may open avenues for modulating cardiac metabolism in preventative and therapeutic applications.

Sources of Funding

This research was supported by the National Institutes of Health (NIH) (R01 DK106348, U01 CA235508, and R21 DK137147 to J.D. Young). Deveena R. Banerjee was supported by the Vanderbilt Integrated Training in Engineering and Diabetes program (NIH grant T32 DK101003). C.M. Hasenour was supported by NIH grant K01 DK135924. A. Jacobson was supported by the NIH (R01DK129340 and R01DK115620) and the Leona M. and Harry B. Helmsley Charitable Trust (2306‐06066). Research was also supported by the Vanderbilt Mouse Metabolic Phenotyping Center (NIH grants U24 DK059637 and S10 OD025199), the Vanderbilt Diabetes Research and Training Center (NIH grant P30 DK020593), the Vanderbilt Digestive Disease Research Center (NIH grant P30 DK058404), and the Vanderbilt‐Ingram Cancer Center (NIH grant P30 CA068485).

Disclosures

M. Rahim is an employee of Cytokinetics.

Supporting information

Tables S1–S3

Figures S1–S3

Unedited Membranes