Deletion of PTP1B in cardiomyocytes alters cardiac metabolic signaling to protect against cardiomyopathy induced by a high-fat diet

Abstract

Cardiomyocytes (CMs) normally use fatty acid oxidation (FAO) as their primary energy source. In response to pathological stress, the substrate preference of CMs switches from FAO to glucose metabolism, leading to the development of heart failure. Obesity increases this pathological risk of cardiovascular disease. We focused on protein tyrosine phosphatase 1B (PTP1B), an inhibitor of insulin signaling, the abundance and activity of which are increased in brain, muscle and adipose tissues in obese and/or diabetic animals and in obese human patients. We generated mice with cardiomyocyte-specific deficiency in PTP1B (PTP1B^fl/fl::αMHC^Cre/+) to investigate the cardiomyocyte-specific role of PTP1B in response to cardiac dysfunction induced by high-fat diet (HFD) feeding. Although no physiological or functional cardiac differences were observed at baseline, PTP1B^fl/fl::αMHC^Cre/+ mice were protected against development of cardiac hypertrophy, mitochondrial dysfunction, and cardiac steatosis induced by HFD feeding. Metabolomics data revealed that hearts with cardiomyocyte-specific deletion of PTP1B had increased FAO and lipolysis but reduced glucose metabolism. Furthermore, phosphoproteomic analyses and mechanistic studies identified an axis involving the kinases PKM2 and AMPK downstream of PTP1B in the heart, which collectively acted to promote FAO and suppress lipogenesis. Together, these results suggest that cardiomyocyte-specific deletion of PTP1B prevents a substrate switch from FAO to glucose metabolism, protecting the heart against the development of HFD-induced cardiac hypertrophy and dysfunction.

Introduction

Obesity has reached epidemic proportions worldwide and continues to rise at an alarming rate. Projections from the World Health Organization suggest that 50% of the US will be classically defined as obese by the year 2030, with the most jeopardized demographic being children¹. Obesity also increases the risk for developing multiple diseases, including type 2 diabetes, certain cancers, and cardiovascular disease (CVD)². With regards to the latter, obesity mediates adverse effects on glucose and lipid levels, increases arterial blood pressure, induces inflammation, and reduces pulmonary function, characteristics that can lead to cardiac hypertrophy, dysfunction and/or heart failure if left untreated^{3, 4, 5, 6}. Unfortunately, despite knowing the contributing factors that lead to HFD-related CVD, the molecular mechanisms for how obesity and the consequent associated physiological changes directly lead to CVD remain poorly understood.

Cardiac metabolism is complex and involves hormonal regulation and activation of multiple downstream signaling pathways that control the balance between glucose and fatty acid utilization⁷. Under normal physiological conditions, 70–80% of the adenosine triphosphate (ATP) needed for metabolic function in the heart is derived from fatty acid oxidation (FAO), whereas the remainder is generated by carbohydrate metabolism^8–10. However, there is a shift from FAO to glucose utilization during cardiac hypertrophic growth and pathological remodeling^{8, 9}.

In contrast, cardiac substrate metabolism switches to favor glucose metabolism over FAO in response to hemodynamic stress and pathological stimuli¹¹. Specifically, in response to obesity and a high-fat diet (HFD), the heart undergoes alterations in energy metabolism, increasing fatty acid uptake and enhancing FAO as a result of increased supply of fatty acids^{12,13, 14}. This change in energy metabolism contributes to lipotoxic heart disease, which may increase the risk of heart failure. However, uncovering the mechanisms governing excess lipid accumulation and adverse sequelae in cardiomyocytes in response to HFD entails more comprehensive investigation.

Systemic metabolic substrate preference is largely under the control of the circulating hormone insulin. Indeed, insulin signaling is a critical modulator of the body’s ability to metabolize carbohydrates, lipids and proteins¹⁵ and is integral in regulating homeostatic processes that control cellular proliferation¹⁶, differentiation¹⁷, and apoptosis¹⁸ . In the presence of insulin, the insulin receptor (IR) phosphorylates insulin receptor substrate (IRS) proteins, inducing the activation of critical downstream pathways¹⁷. These include modulation of the phosphatidylinositol 3-kinase (PI3K)–protein kinase B/AKT (AKT) pathway, which mediates the metabolic actions of insulin, and the Ras–mitogen-activated protein kinase (MAPK) pathway, which regulates gene transcription and cooperates with the PI3K pathway to regulate cell growth and differentiation^19–21.

PTPN1 encodes the non-transmembrane protein tyrosine phosphatase non-receptor type 1 (PTP1B), a ubiquitously expressed protein that inhibits insulin signaling by directly dephosphorylating IRS-1.^22–24,25. PTP1B is therefore considered an emerging potential therapeutic target against the development of obesity, insulin resistance, and diabetes. Indeed, obesity or diabetes is associated with increased PTP1B levels and activity in the brain, muscle and adipose tissues in animals^26–30 and human obese patients^{26, 31–33}. Conversely, mice with germline deletion of PTP1B are resistant to obesity and have increased insulin sensitivity induced by a HFD³⁴. Although liver-specific or adipose-specific deletion of PTP1B does not affect weight gain in HFD-fed mice, they protect against the induction of obesity-induced ER stress^{23, 35}, suggesting a non-autonomous regulation of metabolism and/or obesity by PTP1B. Similarly, mice with neuronal-specific or pro-opiomelanocortin (POMC)-specific PTP1B deletion exhibit decreased body weight gain in response to HFD feeding, effects associated with increased leptin sensitivity and improved glucose homeostasis^{36, 37}.

In addition to IRS-1, PTP1B dephosphorylates and activates pyruvate kinase muscle isozyme 2 (PKM2) in pancreatic cancer cells and cultured adipocytes, inducing glycolysis by promoting the conversion of phosphoenolpyruvate to pyruvate^{35, 38}. Activated PKM2 also affects the adenosine monophosphate (AMP)/ATP ratio by inhibiting AMP-activated protein kinase (AMPK) activity^{38, 39}, thereby potentially controlling metabolic and cellular energy homeostasis by blocking FAO^{40, 41} and increasing lipogenesis⁴². Additionally, AMPK controls the expression of nicotinamide phosphoribosyl-transferase (NAMPT)⁴³, the rate-limiting enzyme in the nicotinamide adenine dinucleotide (NAD) salvage biosynthesis pathway that is responsible for converting nicotinamide (NAM) to nicotinamide ribonucleotide (NMN)⁴⁴. This PKM2/AMPK axis may be regulated by PTP1B, but how and whether this is modulated specifically in the heart remains unknown.

Given the complexities of PTP1B in both insulin/AKT and PKM2/AMPK signaling, it is critical to determine the direct role of PTP1B in the heart. Increased PTP1B activity is associated with increased incidence of heart failure in both rats and humans⁴⁵. Genome wide expression analysis studies demonstrate that increased pressure overload in the heart increases the expression of PTP1B⁴⁵. Both cardiac contractile and intracellular Ca²⁺ signaling dysfunction are also associated with elevated expression levels of PTP1B⁴⁶. Conversely, endothelial cell-specific deletion of PTP1B protects against pressure overload-induced heart failure by driving the activation of vascular endothelial growth factor (VEGF) signaling and angiogenesis, inducing migration and proliferation of microvascular endothelial cells, reducing hypoxia, and preventing fibrosis^{47, 48}.

We believe PTP1B is an integral signaling protein involved in metabolism and a nodal enzyme critical for the regulation of cardiac function. However, a direct role for PTP1B in cardiac insulin resistance in cardiomyocytes (CMs) and its effect on HFD-induced pathological cardiac remodeling remains unknown. Here, we leveraged a global and integrated metabolomics and phosphoproteomics approach to directly investigate the role PTP1B in CMs in response to a HFD, to assess whether deletion of this enzyme in the heart may protect against the development of HFD-induced cardiomyopathy.

Results

Generation of cardiomyocyte-specific PTP1B knock-out mice.

To investigate the role of PTP1B in HFD-induced cardiomyopathy, we generated mice with CM-specific deletion of PTP1B. We crossed PTP1B^fl/fl mice³⁶ with mice expressing Cre recombinase under the control of the α-MHC promoter to generate PTP1B^fl/fl::αMHC^Cre/+ mice, as well as their background controls, PTP1B^+/+::αMHC^Cre/+ mice (Figure S1A, B). PTP1B protein was undetectable in CMs isolated from PTP1B^fl/fl: αMHC^Cre/+ mice but was detected in CMs from male and female PTP1B^+/+::αMHC^Cre/+ mice. PTP1B expression in other cardiac cell types and/or in other tissues, including the spleen, remained unchanged (Figure S1, C–E). Male and female PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl:: αMHC^Cre/+ mice were born at expected mendelian ratios (Table S11).

Cardiomyocyte-specific deletion of PTP1B prevents HFD-induced cardiomyopathy.

In PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl:: αMHC^Cre/+ mice fed a normal diet (ND), hearts were of similar size and weight for both males and females and for either genotype (Figure 1A, 1B, S2A, S2B), suggesting that PTP1B has minimal physiological effects on the heart at baseline. After 10 weeks of high fat diet (HFD) feeding in male control PTP1B^+/+::αMHC^Cre/+ mice, we found that PTP1B abundance increased in response to HFD (Figure S1F). Consequently, control hearts developed hypertrophy with a 1.2 fold change in heart weight to tibia length, whereas hearts from PTP1B^fl/fl::αMHC^Cre/+ mice remained normal, suggesting that deletion of PTP1B in CMs protects against HFD-induced cardiac hypertrophy in males (Figure 1A, 1B).

Figure 1. CM-specific deletion of PTP1B prevents HFD-induced cardiomyopathy.

A. Representative photographs of hearts from male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice maintained on a normal diet (ND) or high-fat diet (HFD) for 10 weeks. Scale bar, 50 mm. B. Heart weight to tibia length ratios from male control or PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD. N= 8–12 mice/group. *p<0.05 and **p<0.01 by two-way ANOVA with Bonferroni post-hoc test. C. Representative H&E whole heart cross-sections from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ male mice maintained on ND or HFD for 10 weeks. Scale bars, 1mm (upper) or 50μm (lower). N=5 mice/group. D. Wheat germ agglutinin (WGA) staining (red) of heart cross-sections from control and PTP1B^fl/fl::αMHC^Cre/+ male mice fed a ND or HFD for 10 weeks. Scale bar, 50μm. E. Frequency distribution of CM area from control or PTP1B^fl/fl::αMHC^Cre/+ hearts from male mice fed a ND or HFD for 10 weeks. N=4 mice/group, with at least 1×10³ cell counts per heart. Data are presented as means ± SEM. * p<0.05, by Student’s t-test. F. Representative photomicrograph of ventricular myocytes isolated from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ male mice fed a ND or HFD for 10 weeks. Scale bar= 50 μm. N=4–6 mice/group. 200 cells were quantified per mouse.

Female mice have reduced metabolic and inflammatory effects, particularly as mediated by HFD-induced stress^49–52, resulting in milder cardiovascular remodeling phenotypes as compared to males at similar time points^53–56. Therefore, to better modulate the potential sex-specific effects of cardiac-specific PTP1B deletion in these mice, we extended our timeline and conducted experiments with female mice after 20 weeks of HFD feeding, as compared to just 10 weeks of HFD feeding in male mice. However, even at this later time point, we did not observe any overt phenotypic changes in heart size or weight in female mice in response to HFD in either genotype (Figure S2A, S2B), suggesting that female mice may be more physiologically cardio-protected against HFD-induced hypertrophy as has been previously suggested^{53, 57}.

Nonetheless, we found that HFD feeding increased overall CM cell size, induced myocardial disarray and resulted in enlarged nuclei in both male and female control PTP1B^+/+::αMHC^Cre/+ mice (Figure 1C–E, S2C–E). In contrast, deletion of PTP1B in hearts isolated from both male and female HFD-fed mice showed near normal myocardial size and structure (Figure 1C–E, S2C–E) as compared to control mice, suggesting that deletion of PTP1B protects hearts against development of HFD-induced hypertrophy. Cardiac fibrosis or collagen deposition did not differ between PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl:: αMHC^Cre/+ mice from either sex, whether they were fed a ND or HFD (Figure S3A–B). HFD feeding increased the mRNA expression of the inflammatory cytokines IL-1β and IL6 in PTP1B^+/+::αMHC^Cre/+ hearts but not in PTP1B^fl/fl::αMHC^Cre/+ hearts (Figure S3C). To directly ascertain the effects of PTP1B on cardiac hypertrophy, we isolated individual CMs and found that HFD significantly enlarged the length, width, and overall area of CMs from HFD-fed PTP1B^+/+::αMHC^Cre/+ control mice, as compared to ND-fed mice (Figure 1F, S2F, Table S1, and Table S12). In contrast, CMs from HFD-fed mice with cardiac-specific PTP1B deletion had normal length, width, and overall area, similar to those of CMs from mice fed a ND.

Echocardiographic analysis software that measures multilayer global longitudinal strain (GLS) can identify early abnormalities of myocardial dysfunction^{58, 59}. Therefore, to begin to assess the functional effects of HFD on cardiac-specific PTP1B deletion, we measured left ventricular wall peak longitudinal strain, strain rate, wall velocity and cardiac speckle tracking displacement in PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl:: αMHC^Cre/+ male and female mice after 8 weeks of HFD feeding (Table S2, Table S13). Even at this early stage, both endocardial and epicardial speckle tracking parameters were higher in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mouse hearts, indicating preservation of cardiac function. Specifically, male HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice exhibited a normal range for endocardial GLS, whereas HFD-fed PTP1B^+/+::αMHC^Cre/+ hearts had a lower value (Table S2). Similarly, epicardial GLS was also preserved in PTP1B^fl/fl::αMHC^Cre/+ mice as compared to PTP1B^+/+::αMHC^Cre/+ mice (Table S2). In addition, cardiac deformation, as reflected by strain and peak velocity rates, was higher in male HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice than in HFD-fed PTP1B^+/+::αMHC^Cre/+ control mice (Table S2). Female mice exhibited similar but less pronounced trends, such that HFD-fed PTP1B deleted hearts had greater endocardial and epicardial GLS as compared to HFD-fed control mice (Table S13). HFD-fed PTP1B^fl/fl::αMHC^Cre/+ female mice also showed significantly preserved endocardial GLS rates, but no notable differences in epicardial GLS rates, peak velocity, or displacement measurements (Table S13). Together, these data suggest that deletion of PTP1B may be functionally cardioprotective against the cardiac stress associated with a HFD.

To evaluate the functional and structural changes over time in response to HFD, we utilized conventional echocardiography and measured cardiac functional parameters following 8, 10, or 12 weeks of ND or HFD in PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice. We did not observe any significant differences in M-mode at 8 weeks of HFD feeding, despite the changes in GLS measurements. This was not unexpected because early HFD feeding does not typically result in overt functional changes using M-mode echocardiography in mice^{60, 61}. However, we began to see significant cardioprotective effects of PTP1B deletion in hearts starting as early as 10 weeks on HFD, including normalization of left ventricular mass, end-diastolic internal dimensions of the left ventricle (LVIDd), and posterior wall thickness in both systole (LVPWs) and diastole (LVPWd), results that were further confirmed after 12 weeks on HFD (Figure 2A–E, Table S3, Data File S1).

Figure 2. CM-specific PTP1B deletion preserves cardiac function in response to HFD.

A. Representative echocardiography of PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ male mice fed a ND or HFD for up to 12 weeks. Two-headed arrows indicate left ventricular chamber size or left ventricular wall thickness. LVIDd, left ventricular internal diameter in end diastole. LVPWd, left ventricular posterior wall thickness in diastole. B and C. Quantification of LVPWd (B) and LVIDd (C) (N=10–15 mice/group). D. Real-time q-PCR analysis of the expression of hypertrophy-related genes (MYH6, MYH7, and ANP) in CMs from PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ male mice maintained on a ND or HFD for 10 weeks. Gene expression was normalized to 18S and Eef1 (which encodes eukaryotic elongation factor-1) mRNAs. N=6–7 mouse hearts/group and each sample was assessed in technical triplicate. Data in graphs are presented as means ± SEM. *p<0.05, ** p<0.01, ***p<0.001 by ANOVA with Bonferroni post-hoc test.

With respect to female mice, we similarly observed significant cardioprotective effects in PTP1B^fl/fl::αMHC^Cre/+ mice after 20 weeks of HFD feeding, as compared to control HFD-fed mice, including normalization of LVID in both systole (LVIDs) and diastole (LVIDd), normalization of left ventricular anterior wall thicknesses in both systole (LVPWs) and diastole (LVAWd), and normalization of LVPW in systole (LVPWs) (Figure S4A–F, Table S14, Data File S1).

To validate the echocardiographic data at the molecular level, we next assessed changes in fetal gene mRNA expression profiles in mice fed a HFD. We observed a significant increase in the expression of ANP (which encodes atrial natriuretic factor) and a switch from MYH6 (which encodes αMHC) to MYH7 (which encodes β-myosin heavy chain) in CMs from male PTP1B^+/+::αMHC^Cre/+, but not in those from PTP1B^fl/fl::αMHC^Cre/+ mice fed HFD for 10 weeks (Figure 2F). However, no significant changes in the fetal gene expression program were detected in CMs isolated from either control or PTP1B^fl/fl::αMHC^Cre/+ female mice, even after 20 weeks of HFD feeding (Figure S4G).

In this regard, female mice may be more protected against development of HFD-associated cardiomyopathy due to increased levels of estrogen^53–56. To determine if aged female mice progressed and developed more severe HFD-associated heart pathologies later, mirroring the effects observed in male mice at an earlier time point, we continued HFD feeding in PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice for 50 weeks. In this cohort, we found that the LVAW and LVPW thicknesses were all elevated, whereas LVID chamber dimensions were significantly decreased in both systole and diastole in the HFD-fed PTP1B^+/+::αMHC^Cre/+ control mice, as compared to ND-fed control female aged mice. Moreover, PTP1B^fl/fl::αMHC^Cre/+ aged female mice were protected against HFD-induced cardiac dysfunction, showing normalized cardiac functional parameters that were similar to those in ND-fed control female aged mice (Table S15). Together, these data suggest that deletion of PTP1B in CMs attenuates and protects both male and female mice against the development of HFD-induced cardiomyopathy.

Cardiomyocyte-specific deletion of PTP1B mitigates HFD-induced mitochondrial dysfunction.

Mitochondrial dysfunction is a pathogenic hallmark of HFD- and obesity-induced cardiomyopathy^{62, 63}. TEM showed no differences in mitochondrial ultrastructure between the PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ hearts under conditions of ND (Figure 3A). In contrast, we observed disrupted mitochondrial ultrastructure, in which a subset of interfibrillar mitochondria was significantly swollen with disorganized and reduced cristae density in HFD-fed PTP1B^+/+::αMHC^Cre/+ hearts (Figure 3A–B). These changes were not observed in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ hearts, where mitochondrial integrity appeared normal and preserved (Figure 3A–B). Additionally, the mitochondrial number did not vary between HFD-fed PTP1B^+/+::αMHC^Cre/+and PTP1B^fl/fl::αMHC^Cre/+ hearts (Figure S5A).

Figure 3. CM-specific deletion of PTP1B preserves mitochondrial function in the presence of HFD.

A. Representative TEM images of cardiac mitochondria from male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. Scale bars, 2 μm (upper) or 500 μm (lower). N=5 sections from 3 mice per group. B. Quantitive assessment of mitochondrial swelling from male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10 weeks, based on the number of swollen mitochondria as a percentage of total mitochondria per field (N=3 hearts/group, with quantification from at least 5 TEM sections per heart). C. Real-time qPCR analysis of SOD (which encodes superoxide dismutase) and CAT (which encodes catalase) in CMs. Gene expression was normalized to 18S and Eef1 mRNAs. N=7–9 mouse hearts/group and each sample was assessed in technical triplicate. Data in graphs are presented as means ± SEM. *p<0.05, **p<0.01, ***p<0.001 by 2-way ANOVA with Bonferroni post-hoc test. D.Oxygen consumption rates (OCR) measured in mitochondria from the hearts of PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. OCR was measured after the sequential addition of the complex V inhibitor oligomycin (Oligo), the protonophore FCCP, and the complex III inhibitor antimycin A (AntiA) to analyze ATP-linked respiration, proton leak respiration, maximal respiratory capacity, mitochondrial reserve capacity, and non-mitochondrial respiration. Data are presented as means ± SEM. *p<0.05 by two-way ANOVA with Bonferroni post-hoc correction. E and F. Representative Western blots of hearts from male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10 weeks using an antibody cocktail that recognizes the five mitochondrial oxidative phosphorylation complexes (E). Quantification of the abundance of OXPHOS mitochondrial complexes (F). *p<0.05, one-way ANOVA with Bonferroni post-hoc test. G. Representative image of JC-1 fluorescence in adult CMs from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10 weeks. Red fluorescence indicates the JC-I mitochondrial aggregate, whereas green fluorescence indicates monomeric JC-1. Scale bar, 50 μm. H. Quantificative assessment of the red to green fluorescence intensity ratio indicating changes in mitochondrial membrane potential (N = 6 mice/group). *p<0.05 by one-way ANOVA with Bonferroni post-hoc test.

A potential mechanism by which HFD is thought to mediate cardiac dysfunction is through increased generation of reactive oxygen species (ROS) and therefore oxidative stress⁶⁴. Inhibiting PTP1B reduces oxidative stress in response to HFD feeding^{65, 66}. To determine if CM-specific deletion of PTP1B provides a similar antioxidant effect, we measured the mRNA expression of SOD (which encodes superoxide dismutase) and CAT (which encodes catalase), two enzymes that protect cells against ROS. We observed a significant increase in both SOD and CAT in CMs isolated from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice as compared to CMs from PTP1B^+/+::αMHC^Cre/+ control mice (Figure 3C). These data suggest that CM-specific deletion of PTP1B could mediate an antioxidant effect against development of HFD-induced cardiomyopathy. Moreover, the upregulation of SOD and CAT expression may also be an adaptive response to maintaining cellular homeostasis in response to HFD in the PTP1B^fl/fl::αMHC ^Cre/+ mouse hearts.

Next, we examined the effect of CM-specific PTP1B deletion on the respiratory activity of mitochondria. Basal mitochondrial oxygen consumption rates (OCR) were similar between PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mitochondria fed a ND from both male and female mouse hearts (Figure 3D). Moreover, OCR was similar between ND-fed PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mitochondria in the presence of the ATP synthase inhibitor oligomycin, the mitochondrial oxidative phosphorylation uncoupler FCCP, and after treatment with rotenone and antimycin (Figure 3D). In contrast, OCR was significantly impaired in HFD-fed PTP1B^+/+::αMHC^Cre/+ mitochondria, with reduced overall baseline levels of respiration, but was not affected in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mitochondria. These results suggest that deletion of PTP1B in CMs protects against HFD-induced mitochondrial dysfunction in both male and female mouse hearts (Figure 3D).

Next, we examined the protein levels of electron transport chain (ETC) complexes. We observed comparable levels of all OXPHOS complexes in both male and female ND-fed PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ hearts (Figure 3E,F, S5B–C). However, in response to HFD, the abundance of NADH:ubiquinone oxidoreductase subunit B8 (NDUFB8; complex I) and succinate dehydrogenase subunit b (SDHB; complex II) were significantly increased in PTP1B^fl/fl::αMHC^Cre/+ mouse hearts, raising the possibility of increased flux through complex I (Figure 3E–F, S5B–C).

To evaluate the mitochondrial membrane potential, we used JC-1 to stain CMs isolated from HFD-fed PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice. We found that PTP1B^fl/fl::αMHC^Cre/+ CMs had active and hyperpolarized mitochondria as compared to those from PTP1B^+/+::αMHC^Cre/+ CMs, confirming that the depletion of PTP1B in CMs prevents HFD-induced mitochondrial dysfunction (Figure 3G–H). Together, these results indicate that CM-specific deletion of PTP1B preserves mitochondrial function and structural integrity following HFD.

Cardiac-specific deletion of PTP1B shifts cardiac metabolism from glycolysis to fatty acid oxidation.

To better characterize the metabolic phenotype of CM-specific deletion of PTP1B, we analyzed the cardiac tissue metabolome of HFD-fed PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice. Our resulting dataset identified a total of 741 metabolites (Table S16), 86 of which significantly differed between control and PTP1B^fl/fl::αMHC^Cre/+ hearts. Indeed, many of these metabolites affect critical metabolic signaling pathways, as analyzed with MetaboAnalyst⁶⁷ (Figure 4A, Tables S4–S8). Specifically, levels of metabolites central to glycolysis, such as glucose 6-phosphate, fructose 6-phosphate, fructose 1,6-bisphosphate, were decreased in both male and female PTP1B^fl/fl::αMHC^Cre/+ mouse hearts (Table S4). Moreover, HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mouse hearts exhibited decreased TCA cycle intermediates (Table S5). Given these results, we measured the activity of pyruvate dehydrogenase (PDH), an enzyme that catalyzes pyruvate into acetyl-CoA, the rate-limiting step for glucose oxidation^68–70. In HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mouse hearts, PDH phosphorylation was increased, indicative of decreased PDH activity (Figure 4B–C, S6A–B).

Figure 4. Metabolomics analysis reveals increased triglyceride mobilization and fatty acid oxidation in hearts from male PTP1B^fl/fl::αMHC^Cre/+ mice.

A. Enrichment analysis of metabolomics data from whole hearts from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD was performed using MetaboAnalyst. N=4 mice/group. B-C. Representative Western blots (B) and quantification of PDH phosphorylated at Ser²⁹³ (C) in heart lysates. n=6–7 mice/group. D. Oil Red O staining for lipid droplets in heart sections from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. Scale bar, 50 μm. N=4 mice/group. E and F. Real-time PCR analysis of SREBP (E) and HSL (F) in CMs from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD. Gene expression was normalized to 18S and Eef1 mRNAs. N=7–9 mice/group and each sample was assessed in technical triplicate. G-H. Representative Western blots (G) and quantification of HSL phosphorylation (H) in heart lysates from PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10-weeks. n=6–7 mice/group. I-L. Gene expression of PPARα (I), CPT1b (J), ACADVL (K), and HADHB (L) in CMs from PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice. Gene expression was normalized to 18S and Eef1 mRNAs. N=7–9 mice/group and each sample was assessed in technical triplicate. Data in graphs are presented as means ± SEM. *p<0.05, ** p<0.01, ***p<0.001 by 2-way ANOVA with Bonferroni post-hoc test.

Next, we asked if fat utilization was altered in PTP1B^fl/fl::αMHC^Cre/+ hearts. Oil red O staining showed that ND feeding did not lead to lipid accumulation in PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mouse hearts. In contrast, HFD feeding promoted lipid accumulation in PTP1B^+/+::αMHC^Cre/+ hearts, but not in PTP1B^fl/fl::αMHC^Cre/+ hearts (Figure 4D, S6C). To evaluate the effects of PTP1B on myocardial lipid metabolism, we measured the expression of lipogenic genes by quantitative real-time polymerase chain reaction (qRT-PCR) analysis. The expression of SREBP1c (which encodes sterol regulatory element binding protein c) was significantly decreased in PTP1B^fl/fl::αMHC^Cre/+ hearts, as compared to PTP1B^+/+::αMHC^Cre/+ controls, irrespective of whether the mice were fed a normal diet or HFD (Figure 4E, S6D). SREBP modulates cardiac lipid metabolism and its dysregulation is implicated in cardiovascular diseases. Specifically, inhibiting SREBP in the heart reduces cardiac lipid accumulation, as well as decreases mitochondrial and endoplasmic reticulum stress⁷¹. Moreover, elevated SREBP1 levels are observed in heart tissues from diabetic patients, which is positively correlated with their myocardial dysfunction⁷². Therefore, the decrease in SREBP levels in the heart resulting from CM-specific deletion of PTP1B suggests decreased lipid accumulation and improved myocardial function in response to HFD in these mice.

In addition to changes in SREBP levels, we also observed significantly reduced phospholipid and lysophospholipid metabolic intermediates in both male and female PTP1B^fl/fl::αMHC^Cre/+ hearts (Table S6, S7). Moreover, the mRNA abundance and phosphorylation of the key enzyme hormone-sensitive lipase (HSL), which mediates triglyceride hydrolysis, was increased in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ hearts (Figure 4F–H, S6E–G), which paralleled the reduced lipid droplet accumulation in the cardiac tissues in these mice. FAO metabolites, such as acylcarnitines (AC), were also significantly increased in both male and female HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mouse hearts (Table S8). ACs arise from the conjugation of acyl-coenzyme A with carnitine for the transport of long-chain fatty acids across the inner mitochondrial membrane for β-oxidation.

To further validate these findings, we measured for changes in fatty oxidation in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice. We measured the expression of key FAO-related transcripts by qRT-PCR analysis, which showed increased expression of PPARα, which encodes peroxisome proliferator activated receptor α (a nuclear hormone receptor that regulates the oxidation and transport of fatty acids^{73, 74}) in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ hearts as compared to HFD-control hearts (Figure 4I). Further, CM-specific deletion of PTP1B increased the expression of CPT1b (which encodes carnitine palmitoyl-transferase 1b, a mitochondrial long chain fatty acyl importer), in response to both ND and HFD feeding (Figure 4J, S6H). ACADVL, which encodes acyl-CoA dehydrogenase very long chain (the protein that catalyzes the first step of the mitochondrial beta oxidation pathway), was also elevated in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice (Figure 4K, S6I). Similarly, the expression of HADHB (which encodes acetyl-CoA acyltransferase) was increased in PTP1B^fl/fl::αMHC^Cre/+ mouse hearts in response to both ND and HFD feeding (Figure 4L, S6J). Further, primary CMs isolated from control HFD-fed male mice showed significantly decreased dependency on glucose and an increased dependency on FAO when in the presence of the PTP1B inhibitor DPM1001 (Figure S7A–B). Together, these data suggest that PTP1B acts as a molecular switch such that CM-specific PTP1B deletion increases cardiac FAO and decreases both lipogenesis and glucose utilization.

To confirm the metabolic impact of HFD treatment, we assessed levels of circulating insulin and triglycerides in ND- and HFD-fed male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl:: αMHC^Cre/+ mice. Although no changes were observed in response to ND, HFD-fed PTP1B^+/+::αMHC^Cre/+ mice had increased levels of both insulin and triglyceride, as compared to HFD-fed PTP1B^fl/fl:: αMHC^Cre/+ mice (Figure S7C–D), suggesting that CM-specific deletion of PTP1B is protective against the development of HFD-mediated metabolic dysregulation.

Phosphoproteomic analyses show PTP1B deletion in CMs protects against pathological changes in multiple downstream cardiac signaling pathways in response to HFD feeding.

Tyrosine phosphorylation specifically controls multiple, critical cellular processes including growth, differentiation, survival, and metabolism. It is a reversible and dynamic process, whereby the phosphorylation state is governed by opposing actions of protein tyrosine kinases (PTKs) and protein tyrosine phosphatases (PTPs)⁷⁵. Given that PTP1B modulates multiple receptor tyrosine kinase signaling pathways, including VEGF, PI3K/AKT, and Ras/MAPK signaling (Figure S8)^{25, 35, 38, 76}, we next sought to determine the downstream signaling effects of PTP1B deletion in hearts from HFD-fed mice. We conducted a phosphotyrosyl (pY)-specific proteomics screen in hearts from PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed HFD for 10 weeks and identified a total of 1970 unique pY peptides. Of these, 267 were differentially phosphorylated proteins (DPPs) that showed either significantly increased (97.35%) or decreased phosphorylation (2.65%) (Figure 5A). Specifically, the phosphorylation of PDH was increased 2.4 fold in HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mouse hearts, validating our previous molecular metabolomics profile data on this enzyme (Fig 4B–C). We also performed a functional enrichment analysis for the differentially phosphorylated proteins (DPPs) through KEGG pathway analysis⁷⁷, which identified a total of 11 significantly enriched pY pathways in PTP1B-deleted hearts, including VEGF, hypertrophic cardiomyopathy (HCM), cAMP, autophagy, diabetic cardiomyopathy, and PI3K-AKT signaling pathways (Figure 5B).

Figure 5. Loss of PTP1B in CMs induces the phosphorylation of ERK, AKT, PKM2 and AMPK.

A. Volcano plot of the average pY phosphoproteomics data for male PTP1B^fl/fl::αMHC^Cre/+ and PTP1B^+/+::αMHC^Cre/+ hearts (n=5 mice/group). Volcano plots are depicted as the fold change of each phophosite. The grey dotted lines indicate the p<0.05 cutoff (calculated from Welch’s t-test). B. The heat map of differentially phosphorylated proteins in key KEGG pathways. C-D. Representative Western blots (C) and quantification of. acetyl CoA carboxylase 1 (ACC1) phosphorylation (D) in cardiac lysates from male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10 weeks. N=6–8 biological replicates per group. E-G. Representative Western blots (E) and quantification of the phosphorylation of IR (F) or VEGFR (G) in cardiac lysates from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD and injected with saline or insulin (10 mU/g i.p.) for 10 minutes. H-I. Representative Western blot (H) and quantification of phosphorylated NFAT (I) in cardiac lysates from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a HFD for 10 weeks. J-L. Representative Western blot (J) and quantification of the phosphorylation of ERK (K) or AKT (L) in cardiac lysates from male PTP1B^+/+::αMHC ^Cre/+ or PTP1B^fl/fl::αMHC ^Cre/+ mice fed a ND or HFD for 10 weeks. M-O. Gene expression of BAX (M) and BCL2 (N) and ratio of BCL2 to BAX (O) in CMs from PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ fed a ND or HFD for 10 weeks. Gene expression was normalized to 18S and Eef1 mRNAs. N=7–9 mice per group and each sample was assessed in technical triplicate. P-R. Representative Western blots (P) and quantification of mTOR phosphorylated at Ser²⁴⁸ (Q) and the ratio of LC3II to LC3I (R) in cardiac lysates from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. S-U. Representative Western blots (S) and quantification of PKM2 phosphorylated at Tyr¹⁰⁵ (T) and AMPK phosphorylated at Thr¹⁷² (U) in cardiac lysates from male PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. For (C) to (U), N=6–8 hearts/group. Data are presented as means ± SEM. *p<0.05, ** p<0.01, ***p<0.001 by 2-way ANOVA with Bonferroni post-hoc test.

We next sought to validate the phosphoproteomic analysis to better understand the molecular mechanisms underlying the metabolic shift in HFD-fed PTP1B-deleted hearts. KEGG analysis of showed an enrichment for cyclic AMP (cAMP), a pivotal second messenger enzyme involved in the regulation of glycogen, sugar, and lipid metabolism, in PTP1B^fl/fl::αMHC^Cre/+ hearts (Table S17). The cAMP-dependent protein kinase A (PKA) was also hyperphosphorylated in PTP1B^fl/fl::αMHC^Cre/+ hearts, along with the protein phosphatase 1 (PP1) regulatory subunits 12A and 12B (PPP1R12A and PPP1R12B) and its catalytic subunit, PP1β (PPP1CB) (Table S17). Similarly, we found that phosphorylation of acetyl CoA carboxylase 1 (ACC1), the enzyme that catalyzes the conversion of acetyl coenzyme A to malonyl coenzyme A, was increased in PTP1B^fl/fl::αMHC^Cre/+ hearts. ACC1 is the rate-limiting enzyme in lipogenesis⁷⁸; as a downstream regulator of PKA, ACC1 hyper-phosphorylation in PTP1B deleted hearts suggested attenuation of fatty acid synthesis (Table S18). Increased ACC1 phosphorylation was further validated by Western blotting analysis of hearts from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice and from control mice (Figure 5C, D). Together with the reduced SREBP mRNA expression levels and the decreased amounts of phospholipids and lysophospholipid metabolites in PTP1B^fl/fl::αMHC ^Cre/+ hearts (Fig 4E, Tables S6, S7), these data suggest that CM-specific PTP1B deletion prevents lipogenesis.

Because PTP1B inhibits IR and VEGFR signaling^{79, 80, 81}, we measured the effect of intraperitoneally injected insulin on IR or VEGFR phosphorylation in PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice. Although we did not observe significant differences at baseline, hearts isolated from PTP1B^fl/fl::αMHC^Cre/+ mice fed either a ND or HFD exhibited increased insulin-induced activation of both IR and VEGFR as compared to insulin-stimulated control mouse hearts (Figure 5E–G).

Vascular endothelial growth factor (VEGF) and insulin activate multiple integral downstream signaling pathways, including Ras/MAPK, AKT, PKC, and Ca²⁺-calcineurin signaling^{82, 83}. To assess how PTP1B regulation of IR and VEGFR signaling modulates downstream pathway activation, we conducted KEGG analysis of hearts from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice and PTP1B^+/+::αMHC^Cre/+ control mice. We found that PTP1B^fl/fl::αMHC^Cre/+ hearts showed hyperphosphorylation of nuclear factor of activated T cells (NFAT) (Figure 5H–I), a downstream effector of both VEGF and cardiac hypertrophy signaling pathways (Figure S7, Table S18, S19). Normally found in a hyper-phosphorylated inactive state in the cytosol, NFAT dephosphorylation is required for nuclear translocation and for the induction of genes that modulate cardiac hypertrophy⁸⁴. Therefore, increased phosphorylation of NFAT in PTP1B-deleted CMs suggested that these mice were protected against cardiac hypertrophy in response to HFD, validating our phenotypic, echocardiographic, and fetal gene expression data (Figure 1A–F, 2A–F; Table S2, S3).

To further understand the effects of cardiac-specific PTP1B deletion on these pathways, we measured changes in the phosphorylation of ERK, a downstream effector of Ras/MAPK signaling through VEGFR. Hearts isolated from both male and female PTP1B^fl/fl::αMHC^Cre/+ mice fed either a ND or HFD showed significant increases in phosphorylated ERK (Figure 5J–K, S9A–B). We next assessed the effects of PTP1B deletion on the PI3K/AKT signaling pathway. Similarly, we found that both male and female PTP1B^fl/fl::αMHC^Cre/+ mouse hearts had significantly increased levels of AKT phosphorylation in response to both ND and HFD feeding (Figure 5J, 5L, S9A, S9C).

AKT can modulate apoptosis and autophagy. Moreover, HFD increases apoptosis and reduces autophagy, although the underlying mechanisms remain unclear⁸⁵. In this regard, the B-cell leukemia/lymphoma 2 (BCL-2) protein is an important regulator of cell death and apoptosis⁸⁶. To maintain cardiac homeostasis, BCL-2-associated athanogene 3 protein (BAG3) binds to BCL-2 to prevent its degradation, thereby inhibiting apoptosis⁸⁷. Our proteomics data showed that the phosphorylation of BAG3 was significantly increased in PTP1B^fl/fl::αMHC^Cre/+ hearts (Table S20), suggesting PTP1B may prevent apoptosis by preserving BCL2 degradation^88–90. To validate this notion, we measured the expression of mRNA encoding various apoptosis regulators, including of B-cell lymphoma 2 (BCL2) and BCL associated X (BAX)⁹¹. Although we did not see any overt effects on BAX expression (Figure 5M), BCL2 expression was significantly increased (Figure 5N). Moreover, we found that the ratio of BCL2 to BAX (BCL2/BAX), a measurement of apoptosis, was increased in CMs from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice (Figure 5O). Together, these data suggest that CM-specific deletion of PTP1B protects against apoptosis in response to HFD.

Because AKT-dependent enhancement of protein synthesis is also mediated in part by its downstream activation of mammalian target of rapamycin (mTOR), we next sought to determine the effects of CM-specific PTP1B deletion on autophagy. mTOR phosphorylation was increased in hearts from control HFD-fed PTP1B^+/+::αMHC^Cre/+ mice, but not in those from PTP1B^fl/fl::αMHC^Cre/+ mice (Figure 5P–Q, S9A, S9D), despite the upstream increase in AKT in these mice (Figure 5J, 5L, S9A, S9C). This decrease in mTOR activity was validated by measuring autophagy, which was enhanced in hearts from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice, as evidenced by an increase in LC3B-II-to LC3B-I ratios (Figure 5P, 5R, S9A, S9E).

HFD induces mTOR activity in both the liver and skeletal muscle, leading to impaired insulin signaling⁹². The reduction in HFD-induced mTOR activity by deletion of PTP1B in CMs suggests a cardio-protective role against HFD-induced insulin resistance. However, the mechanism for how this occurs remains unclear, particularly given the increase in the phosphorylation of the upstream regulator of mTOR, AKT. It is possible that a parallel signaling pathway differentially modulates mTOR and autophagy in response to HFD feeding in mice with a CM-specific deletion of PTP1B. In this regard, PKM2, a rate-limiting glycolytic enzyme, is dephosphorylated and activated by PTP1B in both pancreatic cancer cells and cultured adipocytes^{35, 38}（Figure S8). To assess whether CM-specific PTP1B affects this pathway, we examined the phosphorylation status of PKM2, which was significantly elevated only in hearts from HFD-fed PTP1B^fl/fl::αMHC^Cre/+ mice (Figure 5S–T, S9F–G), suggesting decreased PKM2 activity in response to PTP1B deletion. PKM2 normally inhibits AMPK activity⁹³, and concomitantly, we found that AMPK phosphorylation was significantly increased in PTP1B^fl/fl::αMHC^Cre/+ hearts, even in those from mice fed a ND (Figure 5S, 5U, S9F, S9H). Together, these results suggest that the absence of PTP1B in CMs improves insulin resistance and HFD-associated cardiomyopathy by increasing AMPK activity in the heart.

PTP1B differentially regulates NAD⁺ to modulate cardiac metabolic functions in response to HFD.

Our data suggested that HFD feeding altered metabolic signaling profiles in PTP1B^fl/fl::αMHC^Cre/+ mice. AMPK can mediate NAD⁺ metabolism by activating nicotinamide phosphoribosyltransferase (NAMPT), a key energy sensing enzyme that regulates NAD⁺ synthesis, thereby increasing cellular NAD⁺ levels and influencing the activity of NAD⁺-dependent enzymes⁹⁴. We hypothesized that CM-specific deletion of PTP1B protects against impaired HFD-induced cardiomyopathy by activating the NAD⁺ biosynthesis pathway downstream of the PKM2/AMPK signaling axis. We first measured changes in NAD⁺ metabolic intermediates, which showed increased levels of nicotinamide ribonucleotide (NMN) only in male PTP1B^fl/fl::αMHC^Cre/+ hearts (Table S9), suggesting that NAM is converting to NMN, and that deletion of PTP1B in CMs is activating NAD+ biosynthesis through this pathway. However, hearts from both male and female PTP1B^fl/fl::αMHC^Cre/+ mice had diminished amounts of nicotinamide (NAM) metabolites (Table S9). To validate the role for NAM metabolites in the regulation of PTP1B in the heart, we measured the levels of NAD⁺, NADH, NADP⁺, and NADPH directly in the heart. We found that all NAD⁺ intermediates were increased in CM-specific deleted PTP1B hearts (Figure 6A). Next, we measured the expression of genes related to NAD⁺ synthesis. The expression of NAMPT, NMNAT, NAPRT, and NADSYN were all significantly higher in hearts from both male and female PTP1B^fl/fl::αMHC^Cre/+ mice in response to ND or HFD feeding (Figure 6B, S10). Moreover, NAMPT protein abundance was also elevated in hearts isolated from PTP1B^fl/fl::αMHC^Cre/+ mice (Figure 6C–D). Together, these results suggest that CM-specific deletion of PTP1B promotes NAD⁺ biosynthesis by regulating metabolic processes in cardiac mitochondria and AMPK-modulated metabolic signaling pathways.

Figure 6. NAD⁺ production is increased in male PTP1B^fl/fl::αMHC^Cre/+ mice.

A. Levels of cardiac NAD⁺, NADH, NADP⁺ and NADPH were evaluated in male PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. N=7–9 mice/group, and each sample was assessed in technical triplicate. B. NAMPT, NMNAT, NAPRT, and NADSYN were measured by quantitative real-time PCR in CMs from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. Gene expression was normalized to 18S and Eef1 mRNAs. N=7–9 mice/group and each sample was assessed in technical triplicate. C-D. Representative Western blots (C) and quantification of NAMPT normalized to the loading control GAPDH (D) in hearts from PTP1B^+/+::αMHC^Cre/+ and PTP1B^fl/fl::αMHC^Cre/+ mice fed a ND or HFD for 10 weeks. N=6–8 hearts/group. E-G. Representative Western blots (E) and quantification of the phosphorylation of PKM2 (F) or AMPK (G) in CMs from HFD-fed male PTP1B^+/+::αMHC^Cre/+ control mice treated with DMSO, DPM1001 or Compound 3000 (Comp.3K). N=3 mouse hearts/group. H. Real-time qPCR analysis of the expression of MYH7, ANP, BNP, and CPT1b in CMs from male HFD-fed PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice treated with DMSO or AMPK inhibitor (Comp.C). Gene expression was normalized to 18S and Eef1 mRNAs. N=6–7 mice/group and each sample was assessed in technical triplicate. I-J. Representative Western blot (I) and quantification of HSL phosphorylation (J) in Compound C-treated CMs from HFD-fed PTP1B^+/+::αMHC^Cre/+ or PTP1B^fl/fl::αMHC^Cre/+ mice. N=4 mice/group. Data in graphs are presented as means ± SEM. *p<0.05, ** p<0.01, ***p<0.001 by 2-way ANOVA with Bonferroni post-hoc test.

PTP1B deletion exerts cardioprotective effects through increased activation of the PKM2-AMPK signaling axis.

To confirm the role of PTP1B regulation on PKM2 in response to HFD, we performed pharmacological analysis on adult CMs from HFD-fed PTP1B^+/+::αMHC^Cre/+ control mice. Treatment with either the PTP1B inhibitor DPM1001 or the PKM2 inhibitor Compound 3000 resulted in an increase in the phosphorylation of both PKM2 and AMPK (Figure 6E–G). To further validate the cardioprotective and cardiometabolic effects of PTP1B deletion in CMs, we sought to determine if inhibition of AMPK would reverse the cardioprotective effects of PTP1B deletion. Indeed, PTP1B^fl/fl::αMHC^Cre/+ CMs treated with the AMPK inhibitor Compound C had fetal gene expression profiles similar to those from control HFD-fed PTP1B^+/+::αMHC^Cre/+ hearts (Figure 6H). Moreover, treatment of PTP1B^fl/fl::αMHC^Cre/+ with Compound C also decreased the mRNA expression of CPT1b (Figure 6H). Finally, Compound C also attenuated the increase in the phosphorylation of HSL in PTP1B^fl/fl::αMHC^Cre/+ CMs (Figure 6I–J). Together, these data confirm that mice with CM-specific deletion of PTP1B are cardioprotected from HFD feeding through aberrant regulation of a PTP1B-PKM2-AMPK signaling axis, the effects of which mediate decreased lipogenesis and increased lipolysis, autophagy, FAO, and NAD⁺ production (Figure 7).

Figure 7. PTP1B in the heart functions as a metabolic switch.

PTP1B deletion increases cardiac FAO and decreases both lipogenesis and glucose utilization. Mechanistically, PTP1B deletion inactivates PKM2, thereby increasing AMPK and autophagy and leading to activation of NAD⁺ biosynthesis and cardioprotection from HFD feeding. Created in BioRender. Science, M. (2025) https://BioRender.com/79zn7z8.

Discussion

Using a HFD-induced cardiomyopathy mouse model, we observed that CM-specific deletion of PTP1B in mice ameliorated HFD-induced cardiomyopathy (specifically, hypertrophy) and diminished cardiac steatosis. CM-specific deletion of PTP1B activated VEGF/ERK and AKT signaling but decreased mTOR phosphorylation. The suppression of mTOR and subsequent increase in autophagic activity was regulated by a PTP1B-PKM2-AMPK axis in the heart. In addition, deletion of PTP1B in CMs altered lipid metabolism and mitochondrial function in response to HFD-induced cardiomyopathy, promoting increased FAO and decreased glycolysis in response to HFD. CM-specific deletion of PTP1B also promoted NAD⁺ biosynthesis through AMPK-mediated metabolic signaling pathways (Figure 7, S8). Together, our data indicate that CM-specific deletion of PTP1B protects the heart against the development of HFD-induced cardiomyopathy by directly regulating cardiac metabolic signaling.

Patients with systolic dysfunction show increased PTP1B activity⁴⁵. Conversely, germline deletion of PTP1B in mice improves cardiac output without affecting infarct size⁹⁵. Moreover, miR-206, which directly inhibits PTP1B expression, can reduce CM apoptosis and myocardial infarct size in rats⁹⁶. Although these data suggest a protective role for PTP1B in the heart, the underlying mechanisms have remained unclear. Several previous studies focused only on the molecular effects of PTP1B in cardiac endothelial cells. For example, overexpression of PTP1B in bovine aortic endothelial cells inhibits VEGF-induced AKT phosphorylation⁹⁷. Conversely, deletion of PTP1B in endothelial cells protects against cardiac hypertrophy induced by transverse aortic constriction and improves cardiac VEGF signaling and angiogenesis^{47, 98}. Moreover, deletion of PTP1B in endothelial cells also promotes VEGF-induced ERK activation, increasing cell proliferation and migration⁴⁸. Here, we showed that CM-specific deletion of PTP1B increased the phosphorylation of both IR and VEGFR, validating that these receptors are critical substrates for PTP1B in the heart^{79, 80, 81}. Consequently, CM-specific deletion of PTP1B also led to increased downstream activity of ERK and AKT.

Activated ERK1/2 induces physiological hypertrophy, increases contractile force and reduces fibrosis^{99, 100}. Similarly, increased AKT activity protects against ischemia-reperfusion injury¹⁰¹. However, overt cardiac-specific overexpression of constitutively active AKT can also induce pathological cardiac hypertrophy in mice^{102, 103}, suggesting that dosage of AKT is critical to the modulation of cardiac hypertrophy. Together, ERK and/or AKT activation in the heart may potentially have positive therapeutic effects, by maintaining cardiac performance and preventing the transition to maladaptive hypertrophy and heart failure. However, the beneficial effects may depend on various factors, including whether the effects are acute or chronic and require further research. AKT also inhibits autophagy by activating mTOR, thereby leading to accumulation of damaged organelles, which can contribute to the development of cardiac dysfunction. Indeed, defects in autophagy and mitophagy are implicated in HFD-associated mitochondrial dysfunction in the heart^{104, 105}. In response to short-term HFD feeding, genetic ablation of Atg7, which encodes a critical autophagy factor, impairs mitophagy and leads to cardiac dysfunction¹⁰⁵. In response to deletion of PTP1B, we observed increased AKT activity but decreased mTOR activity and increased autophagy, which may partially explain how PTP1B deletion may protect against HFD-induced cardiac hypertrophy and mitochondrial dysfunction. However, the regulation of mTOR and autophagy by PTP1B does not appear to be through the expected canonical pathways, but rather through a PTP1B-PKM2-AMPK axis. We showed that deletion of PTP1B decreased PKM2 activity, thereby protecting against pathological cardiac hypertrophy. Direct inhibition of PKM2 is also protective against development of right ventricular dysfunction in mice subjected to pulmonary artery banding¹⁰⁶ or with pulmonary hypertension¹⁰⁷.

Mitochondrial dysfunction and energy imbalance are critical components of HFD-induced cardiomyopathy, which leads to increased reliance on glucose and decreased FAO^108–110. With regards to glycolysis, we observed that deletion of PTP1B in CMs led to reduced steady-state levels of glycolytic intermediates and TCA intermediates, suggesting reduced activity or potentially increased metabolic flux. With regards to FAO, the directionality of the change in FAO in the heart is variable and model dependent. Thus, it might not be apparent if decreased FAO levels are causal to or a consequence of cardiac dysfunction in response to HFD feeding. For example, patients with heart failure with preserved ejection fraction (HFpEF) have reduced FAO^{61, 111, 112}. In addition, multiomics analysis also found that patients with HCM had reduced levels of myocardial FAO intermediates, including decreased levels of acylcarnitine (AC) and reduced expression of CPT1, which encodes a critical regulator of mitochondrial long-chain FAO^113–115. These studies suggest that increasing FAO could be beneficial in certain contexts. Conversely, increased FAO is associated with increased production of reactive oxygen species (ROS) in hearts from obese mice and rats^{116, 117, 118}. In a physiological context, increased ROS induced by FAO could be self-limiting¹¹⁹ and a short-term negative feedback mechanism as an adaptation to pathological stress imposed on the heart. As a result, the negative effects of ROS might not outweigh the positive benefits of enhanced FAO. However, considerations for therapeutic modulation of FAO in hypertrophy and heart failure remain to be determined.

Excessive accumulation of lipids is a cause of HFD-induced cardiomyopathy in the absence of underlying vascular disease^{61, 120}. Moreover, HFD induces an imbalance between lipid uptake and oxidation by affecting either increased lipogenesis or decreased FAO^121–123. Here, we found that phospholipids and lysophopholipids, which are CVD risk factors in patients^{124, 125}, were reduced in response to CM-specific deletion of PTP1B. Phospholipid metabolism plays a critical role in cellular adaptation to changes in growth and hypertrophy¹²⁶. Thus, reduced phospholipid levels in hearts from HFD-fed CM-specific deleted PTP1B mice may be a consequence of reduced hypertrophy (Figure 2, A–F, Figure S4, A–G, and Table S6). Moreover, although it remains unclear whether HFD-associated cardiomyopathy is a direct result of abnormal FAO or accumulation of toxic lipids or both, our finding that deletion of PTP1B in CMs increased cardiac FAO and lipolysis while decreasing lipogenesis indicates that PTP1B may be critical for regulating lipid metabolism in the heart. Increased FAO has been thought to contribute to HFD-associated cardiomyopathy but is now thought not to contribute to the development of cardiac dysfunction^{61, 111, 112}. Specifically, increasing FAO by deletion of the gene encoding acetyl coenzyme A carboxylase 2 (ACC2) in the heart does not cause cardiac dysfunction in mice¹¹¹. This likely means that cardiac lipotoxicity is not due to increased FAO per se, but rather to an imbalance of fatty acid supply, storage, and use. Similarly, in cardiometabolic HFpEF, cardiac dysfunction occurs due to accumulation of cardiac lipids and reduced FAO¹¹².

Our findings suggested that increased FAO was critical to preserve cardiac function in response to HFD feeding. CM-specific deletion of PTP1B led to increases in FAO, levels of acylcarnitine, and mitochondrial respiration and decreased accumulation of toxic lipids in the heart. Moreover, deletion of PTP1B in CMs restored the OXPHOS complex by increasing the abundance mitochondrial complexes I and II of the ETC, thereby increasing ATP synthesis and decreasing free radical production^{127, 128}. A lower membrane potential (Δψ) may correlate with an increase in ROS production^{129, 130}, and we showed that Δψ was increased in PTP1B^fl/fl::αMHC^Cre/+ CMs from mice fed a HFD, suggesting decreased ROS production and preserved mitochondrial integrity. Indeed, our gene expression profiling suggested that CM-specific deletion of PTP1B potentiated cardioprotective effects at the molecular level, even in the absence of a pathological stimulus such as HFD (Fig. 4I–L). Therefore, the cardiometabolic protective effects in our mice could help protect against and even mitigate future cardiac pathologies in response to stress. Along these lines, increased FAO does not always induce mitochondrial or cardiac dysfunction in non-obese mice. Specifically, under conditions of pressure overload, mice with a CM-specific deletion of ACC2 have a substrate utilization profile similar to that of sham animals, with increased FAO and decreased glycolysis, thereby indicating a protective effect development of cardiac hypertrophy and fibrosis in response to pathological stress¹³¹.

There are several lines of evidence that CM-specific deletion of PTP1B leads to the activation of lipolysis and a higher utilization of FAO for CM energetics in response to HFD, including fewer neutral lipids, increased PKA signaling, increased phosphorylation of HSL, increased phosphorylation (and therefore deactivation) of ACC, higher PPARα and Cpt1β expression (Figure 4, G–L, 6, H–I, Table S5, S6, S19, S20). Moreover, the decreased ACC activity as mediated by PTP1B inhibition led to increased Cpt1 activity and therefore to induction of FA synthesis (Figure 4J). Together, these data identify PTP1B as a mediator of an integral molecular switch in cardiac energetics. Consequently, it may be possible that the effects of PTP1B in the heart modulate more than just HFD-induced cardiac dysfunction; it may also be a nodal enzyme critical for the regulation of other cardiac-associated stress responses as well, including heart failure and metabolic cardiac diseases. CM-specific deletion of PTP1B maintained lower levels of circulating insulin and triglyceride, suggesting protection against HFD-mediated metabolic dysfunction (Figure S7C–D). Additional studies to prove this possibility for a therapeutic potential for PTP1B inhibition in the heart are needed to prove this point and are currently ongoing in the lab.

Nevertheless, due to its ability to promote FAO and lipolysis, PTP1B appears to be a promising therapeutic target to treat human obesity and type 2 diabetes. Indeed, the PTP1B gene is located within the chromosomal region of 20q13.1–13.2, a locus that is linked to the development of both obesity and type 2 diabetes^132–134. Six single nucleotide polymorphisms for PTP1B have been identified in the French-Canadian population^{132, 135}. Further, genome-wide association studies have indicated that rare single nucleotide polymorphisms (SNPs) within the PTP1B gene correlate with the development of type 2 diabetes in Danish¹³⁶ and Canadian¹³⁷ populations and are also early predictors of insulin resistance in individuals of Italian descent¹³⁸. As well, the SNP rs3787348 in PTP1B is associated with the effects of weight reduction therapy on BMI and waist circumference among obese Japanese patients¹³⁹. Finally, four PTP1B SNPs in an obese population are predictors of dyslipidemia¹⁴⁰.

CM-specific deletion of PTP1B promoted NAD⁺ synthesis through its ability to enhance AMPK-mediated metabolic signaling and increase NMN levels (the NAD⁺ precursor). NAD⁺ is a central metabolite in the salvage pathway and is involved in energy and redox homeostasis. Stimulating NAD⁺ synthesis protects against the development of both diabetic cardiomyopathy¹⁴¹ and HFpEF¹⁴². Moreover, deletion of PTP1B led to increased abundance of NAMPT, which encodes the rate-limiting enzyme in the salvage pathway. CM-specific overexpression of NAMPT protects against the development of HFD-induced cardiac hypertrophy and diastolic dysfunction in mice¹⁴¹. Furthermore, NAMPT overexpression restores levels of NAD⁺ and NADP⁺, protecting mice deficient in the mitochondrial complex I subunit NDUFS4 from developing diabetic cardiomyopathy¹⁴³. Finally, systemic NAMPT overexpression protects mice against the development of angiotensin II-induced hypertension¹⁴⁴. Here, we showed that increased NAD⁺ signaling in PTP1B^fl/fl::αMHC ^Cre/+ hearts was mediated by the effects of deletion of PTP1B on PKM2 and AMPK, leading to elevated mitochondrial FAO^{94, 145} and increased phosphorylation of NAMPT¹⁴⁶. In line with our findings, intraperitoneal injection of NMN stimulates NAD⁺ biosynthesis in obese mice¹⁴⁷, preventing the development of cardiac hypertrophy¹⁴⁸. We observed increased levels of NMN in male PTP1B^fl/fl::αMHC^Cre/+ hearts without concomitant increase in NR (Table S9). Instead, we observed increased NAMPT levels, suggesting that the mechanism of regulation by PTP1B deletion in CMs is mediated by increases in the NAM-->NMN-->NAD arm of the salvage pathway. However, it may also be possible that this can be achieved by blockade of the other arm of this pathway, which is NR-->NMN-->NAD. Future research will be needed to determine the precise mechanism and the translational application of NAD⁺, NMN, and its precursors in the prevention of diabetic cardiomyopathy.

Although our results suggested a protective cardiac effect in response to deletion of PTP1B in the heart, another paper suggested that CM-specific deletion of PTP1B in mice may be pathological, inducing a hypertrophic phenotype that is exacerbated by pressure overload¹⁴⁹. Specifically, argonaute 2 (AGO2), a critical component of the RNA-induced silencing complex, is inactivated in response to CM-specific deletion of PTP1B, thereby preventing miR-208b-mediated inhibition mediator complex subunit 13 (MED13) and leading to thyroid hormone-mediated pathological cardiac hypertrophy. There could be several reasons why we and others^{36, 150, 151}might have seemingly opposing results to this paper. First, inhibition of miR-208b improves rather than exacerbates cardiac dysfunction in titin-induced dilated cardiomyopathy¹⁵². Second, the upregulation of MED13 in the heart confers resistance to obesity, regulating systemic energy homeostasis and influencing cardiac metabolic processes that increase energy expenditure and improve insulin sensitivity¹⁵³. Third, Coulis et al. used PTP1B^fl/fl mice as their control group instead of αMHC^Cre/+ mice, and the expression of the αMHC-Cre transgene by itself in some mouse strains can induce a cardiac phenotype, including hypertrophy¹⁵⁴. Finally, it is possible that different pathological stimuli can lead to different outcomes. For example, overexpression of NAMPT can induce heart failure by activating Sirt1¹⁵⁵ in response to pressure overload or can reduce cardiac diastolic dysfunction, apoptosis and proinflammatory signaling by regulating NAD⁺, NADP⁺ and NADPH production in response to HFD¹⁴¹.

Our data suggest potential sex-driven differences in the cardiometabolic alterations associated with HFD-related cardiomyopathy. It took longer for female mice to develop signs of cardiomyopathy in response to HFD feeding (20 weeks compared to 10 weeks for the males), irrespective of genetic background. In addition, hypertrophy in response to HFD was less pronounced in female mice. However, gene expression of the key metabolic enzymes involved in NAD⁺ synthesis were elevated in both male and female PTP1B^fl/fl::αMHC^Cre/+ mouse hearts, suggesting that the effects of pathological stress could have similar causes in both males and females.

Together, our results suggest that CM-specific deletion of PTP1B is protective against development of HFD-induced cardiomyopathy. Deletion of PTP1B in CMs mediates a substrate switch from glucose metabolism to FAO, protecting hearts against the development of HFD-induced cardiac hypertrophy and dysfunction. These protective effects are mediated by a PTP1B/PKM2/AMPK axis that is critical for the regulation of NAMPT and NAD⁺ biosynthesis.

Materials and methods

Mice

To generate mice with cardiomyocyte-specific PTP1B knock-out (PTP1B^fl/fl::αMHC^Cre/+), C57Bl6J mice with loxP-flanked PTP1B floxed alleles (PTP1B^fl/fl) (courtesy of Benjamin G. Neel, NYU Grossman School of Medicine, Laura and Isaac Perlmutter Cancer Center, New York¹ and purchased from the Mutant Mouse Resource and Research Center (MMRRC) at The Jackson Laboratory, an NIH-funded strain repository [B6;129S4-Ptpn1^tm2Bbk/Mmjax, RRID:MMRRC_032243-JAX]³⁶ were mated with transgenic mice that express Cre driven by the α myosin heavy chain (MYH6) promoter (αMHC^Cre/+)¹⁵⁶. Mice were weaned at Day 30 (4 weeks of age) and immediately placed on either normal diet (ND) (PicoLab Rodent Diet20#5053) or HFD (Research Diet #12492, 60 kcal% fat) at post-natal day 30, for a period of 10 weeks (males) or 20 weeks (females). Unless otherwise noted, endpoint analyses were conducted in 14- (males) or 24- (females) week old mice. Unlike males, which showed a significant cardiac phenotype by 10 weeks on HFD, female mice did show any overt phenotype in this same time course. Therefore, female mouse studies were conducted over a period of 20 weeks to reveal their phenotypes. All other experimental groups were age- and sex-matched mice. All mice were maintained on the C57Bl6J background and αMHC^Cre/+ mice were used as our control groups. All animal procedures were approved by the Masonic Medical Research Institute Animal Care and Use Committee. Our PHS assurance number is D16–00144(A3228–01) and we are an AAALAC (#001865) accredited institution.

RNA Extraction and Real-time PCR analysis

RNA was isolated using Trizol (Invitrogen) and purified with RNeasy kits (Qiagen). A total of 1 μg RNA was reverse-transcribed with Iscript Supermix (Bio-rad). The resulting cDNA was used to amplify the target genes by SYBR Green PCR Master Mix (Thermo Fisher Scientific). 18S ribosomal RNA (18S), eukaryotic translation elongation factor 1 (EEF1A1) and ribosomal protein L4 (RPL4) were used as control housekeeping genes. Data were quantified using the comparative C_T method (ΔΔCT). For primer sequences and PCR conditions, see Table S10.

Biochemical Studies

Tissue lysates were prepared by homogenizing the tissue in radioimmunoprecipitation (RIPA) buffer (25 mmol/l Tris-HCl [pH 7.4], 150mmol/l NaCl, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 5mmol/l EDTA), 1mmol/l sodium fluoride, 1mmol/l sodium orthovanadate, and a protease cocktail at 4°C, followed by sonication. For immunoblots, proteins were resolved by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. Immunoblots were performed with anti-PTP1B (Abcam), anti-AKT, anti-phospho-AKT (Ser⁴⁷³), anti-extracellular signal-regulated kinase (ERK) 1/2, anti-phospho-ERK 1/2 (Thr²⁰²/Tyr²⁰⁴), anti-AMPK, anti-phospho-AMPK (Thr¹⁷²), anti- microtubule-associated protein 1A/1B light chain 3B (LC3B), anti-PKM2, anti-phospho-PKM2 (Tyr¹⁰⁵), anti-ACC1, anti-phospho-ACC1 (Ser⁷⁹) (Cell Signaling Technology) or anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Santa Cruz Biotechnology, Inc).

Histology

Hearts were isolated and fixed in 4% paraformaldehyde for 24 hours and paraffin embedded. Sections were stained with hematoxylin/eosin (H&E) or Masson’s trichrome and images were taken using a Keyence BXZ microscope. Cryosections (8 μm) of hearts were permeabilized with 0.1% Triton X in PBS for 5 mins at room temperature, hearts were blocked with normal horse serum for 30 mins. Coverslips were incubated with wheat germ agglutinin (WGA)-Alexa Fluor 647 (1:100) (Thermo Fisher Scientific) antibodies in a humidity chamber. Slides were imaged by a Zeiss Confocal microscope. The quantitative analysis of CM area size included 5 different fields from four mice per group and 3 sections per heart. The total number of myocytes counted was 1–2×10³ cells per mouse.

Echocardiography and Speckle Tracking Analysis

Transthoracic echocardiography was conducted on non-anesthetized animals as described previously², using a Visual Sonics Vevo 3100^® high-frequency ultrasound rodent imaging system. Hearts were imaged in the two-dimensional parasternal short-axis view, and an M-mode echocardiogram of the mid ventricle was recorded at the level of papillary muscles. Heart rate, posterior wall thickness (LVPW), and end-diastolic and end-systolic internal dimensions of the left ventricle (LVIDd and LVIDs, respectively) were measured from the M-mode image.

Early-stage heart function was assessed using 2D speckle-tracking and quantitative myocardial deformation analysis using the Vevo Strain software (Vevo LAB v5.6.1). Three consecutive cardiac cycles from the long-axis B mode videos were chosen for analysis using the M mode setting, to limit respiratory artifacts in LV wall motion. Utilizing the Free Curve software tool, the borders of the endocardium and epicardium were determined and segmented into 6 regions. An initial execution of speckle tracking resulted in estimations of velocity, displacement, and strain rate for each segment. Wall motion tracking was also manually adjusted frame by frame. Peak systolic strain values were measured in each segment of the LV for the endocardial layer, mid-myocardial layer and epicardial layer.

Mitochondrial Respiration

Mitochondria were isolated from hearts as previously described¹⁵⁷. Briefly, hearts were dissected and washed in ice-cold mitochondrial isolation buffer. Tissues were cut into small pieces and homogenized with a Potter-Elvehjem tissue grinder. Tissue pieces were disrupted with at least 30 strokes and then centrifuged at gradient speed to obtain a pellet. The pellet was resuspended in mitochondrial assay solution containing 1 mM pyruvate, 2 mM glutamine, 10 mM glucose, and 5 mM malate. Mitochondrial respiration was determined using an XF96 Extracellular Flux Analyzer (Seahorse Bioscience). For those measurements, 50μg mitochondria were plated and centrifuged at 2,000 g for 20 mins to promote adherence. Oligomycin (1μg/mL) and carbonyl cyanide-p-trifluoromethoxyphenyl-hydrazon (FCCP) (20mM) were used to inhibit ATP synthase. Rotenone and Antimycin (40mM) were used to inhibit complex I and complex III-dependent respiration. All readings were normalized to protein content.

The MitoFuel Seahorse assay was conducted using an XF96 Extracellular Flux Analyzer, as directed following the manufacturer’s protocol. Briefly, adult CMs from HFD-fed mice were isolated, plated onto 96-well Seahorse XF96 plates at 5,000 cells per well, and cultured for 6 hours with or without the PTP1B inhibitor DPM1001 (100nM). Prior to the assay, the culture media was replaced with DMEM base medium supplemented with 1 mM pyruvate, 2 mM glutamine, 10 mM glucose, and 200 μM BSA-palmitate. To inhibit the glucose oxidation pathway, the mitochondrial pyruvate carrier inhibitor UK5099 (2 μM) was used. The long-chain FAO pathway was inhibited using the carnitine palmitoyl transferase 1A inhibitor etomoxir (4 μM). Additionally, the glutamine oxidation pathway was blocked with the glutaminase inhibitor BPTES (3 μM).

Detection of Mitochondrial Membrane Potential

JC-1 (Abcam, USA) was used to assess the mitochondrial membrane potential. Isolated cardiomyocytes were incubated with 5μM JC-1 staining solution at 37 °C protected from the light for 10 min according to the manufacturer’s instruction. Adult cardiomyocytes were washed by culture medium and imaged by a Zeiss Confocal microscope. JC1 monomers and aggregates were both excited at 488 nm. Detection of fluorescence for JC1 monomers and aggregates were performed respectively at 530 and 590 nm. Cells with high mitochondrial membrane potential (Δψ) promote the formation of red fluorescent JC-1 aggregates, whereas cells with low Δψ exhibit green fluorescence¹⁵⁸. Ratio F(aggregate)/F(monomer) was subsequently evaluated using Image J³.

Transmission Electron Microscopy (TEM)

TEM studies were performed by the SUNY Upstate University Transmission Electron Microscopy Center. Briefly, mice were sacrificed and their hearts were perfused with 2.5% (vol/vol) glutaraldehyde in 0.1M Na-cacodylate buffer (pH 7.4). The hearts were immediately cut into pieces < 1mm³ and fixed in 2.5% (vol/vol) glutaraldehyde in 0.1M Na-cacodylate buffer (pH 7.4) for 2 hours. After fixation with 1% osmium tetroxide in 0.1M cacodylate buffer, the tissue was dehydrated in a graded series of ethanol washes and then embedded in Spurr’s resin. Semithin (0.5μm) and ultrathin (90nm) sections were cut, mounted on copper grids, and stained with uranyl acetate and lead citrate. Sections were viewed with a JEM 2100F transmission electron microscope.

Adult Primary Cardiomyocyte Isolation

Mice were injected with intraperitoneal heparin (40 units/mice), and their hearts were collected, isolated and perfused through the aorta. The perfusion buffer consisted of KCl (14.7mM), NaCl (120.4mM), KH₂PO₄ (0.6mM), Na₂HPO₄ (0.6mM), MgSO₄7H₂O (1.2mM), 2,3 butanedione monoxime (10mM), taurine (30mM), HEPES (10mM), and glucose (5.5mM). The heart was digested with collagenase II digestion buffer (2 mg/ml) for approximately 8–10 min. The heart was cut from the cannula and placed in the dish with digestion buffer and stopping buffer (12.5μM CaCl₂ and 10% exosome depleted FBS in perfusion buffer). Isolated cardiomyocytes were cultured in a six well plate (treated with laminin), with myocyte culture medium (ScienCell #6201), 5% exosome depleted FBS, and 1% penicillin/streptomycin.

Phospho-proteomics Analysis

Whole heart pieces from PTP1B^fl/fl::αMHC^Cre/+ and control mice were homogenized in 9M Urea with 0.3% Triton X-100 and protease and phosphatase inhibitor cocktails (Thermo Fisher Scientific, Halt). Samples were centrifuged at 17,000xg for 5 minutes and the pellet of insoluble proteins was discarded. Chloroform/methanol precipitation was performed with 2 mg of lysate to remove Triton. Samples were then reduced, alkylated, digested with trypsin, and desalted using Pierce Peptide Desalting Columns following the manufacturer’s protocol. Phosphopeptides were enriched with a High Select Fe-NTA Phosphopeptide Enrichment Kit (Thermo Scientific) according to the manufacturer’s protocol. The enriched peptides were dried with a vacuum centrifuge. Dried peptides were resuspended in 0.15% formic acid in HPLC Water and loaded onto a Vanquish Neo UHPLC system (Thermo Fisher Scientific) with a heated trap and elute workflow containing a c18 PrepMap, 5mm, trap column (cat #160454) with a forward-flush configuration connected to a 25cm Easyspray analytical column (cat #ES802A rev2), 2u,100A, 75um × 25 of 100% Buffer A (0.1% formic acid in water) and a column oven set to 40 °C. Peptides were eluted over a 150 min gradient using 80% acetonitrile and 0.1% formic acid (buffer B), going from 4% to 5% (5 min), to 35% (125 min), and then to 99% (20 min), after which all peptides were eluted. Spectra were acquired with an Orbitrap Eclipse Tribrid mass spectrometer with FAIMS Pro interface (Thermo Fisher Scientific) running Tune 3.5 and Xcalibur 4.5. For all acquisition methods, spray voltage was set to 1600V, and ion transfer tube temperature was set to 300°C. FAIMS switched between CVs of −45 V, – 55 V, and −65 V with cycle times of 1.5sec. MS1 spectra were acquired at 120,000 resolutions, with a scan range from 375 to 1600 m/z, normalized AGC target of 300%, and maximum injection time of 50ms. S-lens RF level was set to 30. Precursors were filtered using monoisotopic peak determination set to peptide. DDMS2 scan was used in isolation mode Quadrupole, Isolation Window (m/z): 1.6; activation type set to HCD with 30% Collision Energy (CE), orbitrap was set as a detector with resolution of 30K, AGC target was set to 50,000; maximum injection time was set to 54ms, micro scans: 1 and datatype was set to Centroid.

Phosphoproteomic data were analyzed using Proteome Discoverer 2.5 (Thermo Fisher Scientific) using Sequest HT search engines. Mouse Uniprot protein sequence database (uniprot-proteome_UP000000589) was used to generate search parameters that included precursor mass tolerance of 10 ppm and 0.02 Da for fragments, allowing two missed trypsin cleavages, oxidation (Met) and acetylation (protein N-terminus), and phosphorylation (Ser, Thr, and Tyr) as variable modifications, and carbamidomethylation (Cys) as a static modification. Percolator PSM validation was used with the following parameters: strict false discover rate (FDR) of 0.01, relaxed FDR of 0.05, maximum ΔCn of 0.05, and validation based on q-value. Precursor Ions Quantifier settings were to use Unique + Razor for peptides. Precursor abundance was based on intensity, normalization was based on total peptide amount, protein abundance was calculated by the summed intensity of connected peptides, and protein ratios were calculated based on protein abundance.

Metabolomics Studies

Metabolomic analysis was performed by Metabolon, Inc. Briefly, heart samples were immediately snap-frozen in liquid nitrogen and shipped overnight on dry ice to the Metabolon, Inc. Samples were prepared using the automated MicroLab STAR^® system from Hamilton Company. Several recovery standards were added prior to the first step in the extraction process for QC purposes. To isolate proteins, small molecules bound to proteins or trapped in the precipitated protein matrix were dissociated. To recover chemically diverse metabolites on these complexes, proteins were precipitated with methanol under vigorous shaking for 2 min (Glen Mills GenoGrinder 2000), followed by centrifugation. The resulting extract was divided into five fractions: two for analysis by two separate reverse phases/UPLC-MS/MS methods with positive ion mode electrospray ionization (ESI), one sample for analysis by RP/UPLC-MS/MS with negative ion mode ESI, one fraction for analysis by HILIC/UPLC-MS/MS with negative ion mode ESI, and one sample was reserved for backup. Samples were placed briefly on a TurboVap^® (Zymark) to remove the organic solvent. The sample extracts were stored overnight under nitrogen before preparation for analysis.

Several types of controls were analyzed in concert with the experimental samples: a pooled matrix sample generated by taking a small volume of each experimental sample as a technical replicate of the data set; extracted water samples as process blanks; and a cocktail of QC standards, chosen so as to not interfere with the measurement of endogenous compounds, were spiked into every analyzed sample. These controls allowed for optimal instrument performance monitoring and aided the chromatographic alignment. Instrument variability was determined by calculating the median relative standard deviation (RSD) for the standards that were added to each sample prior to injection into the mass spectrometer. Overall process variability was determined by calculating the median RSD for all endogenous metabolites (non-instrument standards) present in 100% of the pooled matrix samples. Experimental samples were randomized across the platform run with QC samples spaced evenly among the injections.

Peaks were quantified using the area-under-the-curve. For studies spanning multiple days, a data normalization step was performed to correct variation resulting from instrument inter-day tuning differences. Briefly, each compound was corrected in run-day blocks by registering the medians to equal one (1.00) and normalizing each data point proportionately (termed the “block correction”). For studies that did not require more than one day of analysis, no normalization was necessary, other than for purposes of data visualization. In certain instances, biochemical data may have been normalized to an additional factor (such as cell counts, total protein as determined by Bradford assay, osmolality, and so on) to account for differences in metabolite levels due to differences in the amount of material present in each sample.

NAD⁺ and NADH Measurements

NAD⁺ and NADH were measured using the EnzyChrom^™ NAD⁺/NADH Assay Kit according to the manufacturer’s protocol (ECND-100, Bioassay Systems, Hayward, CA). Briefly, 20mg of fresh heart tissue was weighed out for each sample and washed with cold PBS. Samples were homogenized in a 1.5mL Eppendorf tube with either 100 μL NAD extraction buffer for NAD determination or 100 μL NADH extraction buffer for NAD determination. Extracts were heated at 60°C for 5 min and neutralized by addition of 20 μL Assay Buffer and 100 μL of Neutralizing Buffer. Samples were vortexed and spun down at 14,000 rpm for 5 min. Supernatants were then collected and used for the NAD/NADH assay. Here, 40 mL of the supernatant was added to 80 mL of the Working Reagent and optical density at 565nm was read at time zero and after a 15-min incubation at room temperature.

Statistical Analysis

All values in graphs are expressed as means ± SEM. Simple group comparisons were performed with the student’s t-test. Longitudinally measured variables were analyzed using two-way repeated measures ANOVA, followed by post-hoc pairwise comparisons performed in conjunction with a Bonferroni correction (GraphPad Prism 9). Diagnostic tools were used to assess model assumptions. A nominal significance level of 0.05 was used throughout. For metabolomics data, to account for the multiple testing, the False Discovery Rate (FDR) method was used with the threshold for computed q-values being set at 0.05¹⁵⁹.

Supplementary Material

Figs. S1–S10.

Tables S1–S20.

Data File S1.

Data and Materials Availability

The mass spectrometry proteomics data was uploaded to the MassIVE repository (ftp://massive-ftp.ucsd.edu/v09/MSV000097436/). The metabolomics data have been deposited to the Metabolights repository with the dataset identifier MTBLS977. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. The α-MHC-Cre mice are available from E.D.A. under a material transfer agreement with the University of Utah.