mTORC1 Regulates the Metabolic Switch of Postnatal Cardiomyocytes During Regeneration

Abstract

The metabolic switch from glycolysis to fatty acid oxidation in postnatal cardiomyocytes contributes to the loss of the cardiac regenerative potential of the mammalian heart. However, the mechanisms that regulate this metabolic switch remain unclear. The protein kinase complex mechanistic target of rapamycin complex 1 (mTORC1) is a central signaling hub that regulates cellular metabolism and protein synthesis, yet its role during mammalian heart regeneration and postnatal metabolic maturation is undefined. Here, we use immunoblotting, rapamycin treatment, myocardial infarction, and global proteomics to define the role of mTORC1 in postnatal heart development and regeneration. Our results demonstrate that the activity of mTORC1 is dynamically regulated between the regenerating and the non-regenerating hearts. Acute inhibition of mTORC1 by rapamycin or everolimus reduces cardiomyocyte proliferation and inhibits neonatal heart regeneration following injury. Our quantitative proteomic analysis demonstrates that transient inhibition of mTORC1 during neonatal heart injury did not reduce protein synthesis, but rather shifts the cardiac proteome of the neonatal injured heart from glycolysis towards fatty acid oxidation. This indicates that mTORC1 inhibition following injury accelerates the postnatal metabolic switch, which promotes metabolic maturation and impedes cardiomyocyte proliferation and heart regeneration. Taken together, our results define an important role for mTORC1 in regulating postnatal cardiac metabolism and may represent a novel target to modulate cardiac metabolism and promote heart regeneration.

Graphical Abstract

1. INTRODUCTION

The inability of the adult mammalian heart to regenerate damaged tissue after injury has resulted in a significant health and economic burden from widespread heart failure with reduced ejection fraction [1]. In contrast, the neonatal mouse heart possesses a remarkable ability to regenerate after injury, in stark contrast to the adult mouse heart [2, 3]. Shortly after birth, cardiomyocytes transition from hyperplastic to hypertrophic growth, resulting in loss of the cardiac regenerative potential [4–6]. One key biological process of postnatal cardiomyocyte maturation is a switch in metabolic substrate utilization. Highly proliferative embryonic and neonatal cardiomyocytes rely upon glycolysis, but exposure to an oxygen-rich environment after birth stimulates a metabolic switch towards the more energy efficient mitochondrial oxidative phosphorylation and fatty acid utilization as the primary energy source [5–7]. The cardiac metabolic state can have important implications in modulating cardiac maturation and regeneration. However, the mechanisms that control this metabolic switch in postnatal cardiomyocytes is not fully understood.

mTORC1 has been implicated as a master regulator of protein synthesis and cellular metabolism in response to nutrient and growth signals, where mTORC1 function regulates the levels of key metabolic enzymes involved in glycolysis, oxidative phosphorylation, and fatty acid synthesis [8]. mTORC1’s regulation of protein synthesis and metabolism is mediated via direct phosphorylation of its primary downstream targets eukaryotic translation initiation Factor 4E binding protein 1 (4EBP1) and ribosomal protein S6 kinase (S6K). mTORC1 promotes glycolysis through phosphorylation of 4EBP1 and S6K, which increases levels of master regulators of glycolysis such as hypoxia-inducible factor 1α (HIF1α) and MYC that can induce the expression of glycolytic enzymes [9–11]. In addition, mTORC1 controls oxidative phosphorylation through multiple regulatory mechanisms at the transcriptional and translational level [12, 13]. Additionally, mTORC1 negatively regulates autophagy by phosphorylating the autophagy inducer unc-51 like kinase 1 (ULK1). High levels of ULK1 phosphorylation at Serine 757 (ULK1-S757) inhibit ULK1 activation [14]. However, whether mTORC1 regulates the postnatal metabolic switch in cardiomyocytes remains undetermined.

mTOR plays an essential role during cardiac development and homeostasis. mTOR inhibition through cardiomyocyte-specific deletion of Mtor leads to lethality during both embryonic and postnatal development due to impaired cardiomyocyte survival and proliferation [15, 16]. Furthermore, inhibition of mTORC1 disrupts adult cardiac function, where inducible cardiomyocyte-specific knockout of the mTORC1 subunit; Raptor; is lethal within 8 weeks of due to increased apoptosis and disruption of metabolic substrate use [17]. Interestingly, recent studies demonstrate that modulating mTORC1 function through modifications to the upstream regulator tuberous sclerosis complex (TSC2) can regulate cardiac metabolism following ischemia reperfusion injury [18, 19]. These results demonstrate an important role for mTORC1 in postnatal heart function and metabolism.

mTORC1 is a critical node in regulating key metabolic pathways; however, the role of mTORC1 in the postnatal metabolic switch, cardiomyocyte cell cycle activity, and cardiac regeneration remains unclear. In this study, we define the distinct function of mTORC1 in the postnatal heart as well as following injury. We demonstrate that transient inhibition of mTORC1 activity blocks cardiomyocyte proliferation and neonatal mouse heart regeneration. Furthermore, our quantitative proteomic analysis demonstrates that inhibition of mTORC1 in the postnatal heart induces an accelerated metabolic switch from glycolysis to fatty acid oxidation. Our results reveal a novel role for mTORC1 function in regulating postnatal cardiomyocyte metabolism and mammalian cardiac regenerative potential.

2. METHODS

2.1. Animals

Untimed pregnant CD1 females were obtained from Charles River Laboratories, and neonatal mice were then used for all experiments. All experimental procedures involving animals were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin-Madison. All animal experiments were performed with age-matched mice.

2.2. Myocardial Infarction (MI) Model

Neonatal mice at postnatal day (P) 1 underwent a surgically induced MI as described previously [3]. Neonates were anesthetized via hypothermia by placing them on ice for up to 5 min. Once anesthetized, a lateral incision was made into the skin to view the muscle above the fourth intercostal space. A lateral thoracotomy was then performed by blunt dissection of the fourth intercostal space. A C-1 tapered needle connected to 6–0 prolene suture (Ethicon Inc, Bridgewater, NJ) was then passed through the ventricle around the left anterior descending coronary artery, which was then tied off to induce a clinically significant MI. The lateral thoracotomy was then closed by using the prolene suture to tie the ribs together, whereas the skin incision was closed by using surgical adhesive glue (3M). Mice were then warmed in-hand until they began to awaken, mice were then placed onto a heating pad until fully recovered. Sham-operated mice underwent the same procedure, including anesthetization, however their coronary artery was not ligated.

2.3. Drug Administration

Neonatal mice were weighed daily and injected with saline, 1.0 mg/kg Rapamycin, or 1.0 mg/kg Everolimus. Stock solutions of both Rapamycin and Everolimus were diluted 100 mg/ml in dimethyl sulfoxide (Sigma) and then further diluted in saline prior to injection. Saline was used as the vehicle control for all experiments. All treated mice were given intraperitoneal injections on days 4, 5, and 6 post-MI. In mice treated with puromycin for measuring protein translation, mice were weighed 1 h prior to harvest and injected with 0.04 μmol/g puromycin (VWR, #75844–852).

2.4. Immunoblotting Assay

Myocardial ventricular lysates were obtained from ventricular tissue below the ligature that was flash frozen in liquid nitrogen and prepared in RIPA buffer (Thermo Fisher, #89900) with 1X HALT protease and phosphatase inhibitor cocktail (Thermo Fisher, #78440). Protein concentration for each sample was then determined by BCA assay (Thermo Fisher, #23225). Samples were then prepared in SDS-PAGE loading buffer (LI-COR, #928–40004) and ran on either 10% or 15% polyacrylamide containing gels with transfer blotting on to a polyvinylidene fluoride membrane. Proteins of interest were targeted with the following antibodies, phospho-p70 S6K T389 (CST, #9205 Lot-16, 1:1000), p70 S6K (CST, #9202 Lot-21, 1:1000), phospho-4EBP1 T37/T46 (CST, #2855 Lot-30, 1:1000), phospho-4EBP1 S65 (CST, #9451 Lot 17, 1:1000), 4EBP1 (CST, #9452 Lot-12, 1:1000), phospho-ULK1 S757 (CST, #14202 Lot-7, 1:1000), Raptor (CST, #2280 Lot-13), puromycin (Millipore Sigma, #MABE343, 1:5000), and GAPDH (Proteintech, 60004–1, 1:1000). Total protein was examined using the Revert total protein stain kit following the manufacturer’s protocol (LI-COR, #926–11015). The antibody binding was then visualized using the Odyssey FC (LI-COR) imaging system and Image Studio Software version 4.0. Staining quantification was then conducted using the Image Studio Software version 4.0 with comparisons to the appropriate total form or loading control.

2.5. Immunofluorescent Staining

Paraffin embedded hearts were sectioned and placed onto glass slides, slides were then deparaffinized, rehydrated, and boiled in IHC antigen retrieval solution (Invitrogen, #00-4955-58). Sections were then blocked with 5% goat serum (Vector Laboratories, #S-1000) and incubated overnight at 4°C with primary antibodies. Primary antibodies used in this study were phospho-histone H3 Ser10 (Millipore, catalog # 06–570, 1:100), aurora B Kinase (Sigma, catalog # A5102, 1:100), and cardiac troponin T (Abcam, catalog # ab829, 1:100). Sections were washed and incubated with a corresponding secondary antibody conjugated to Rabbit-488 (Invitrogen, #A-11008, 1:400) or Mouse-555 (Invitrogen, #A28180, 1:400) for 1 h at room temperature, with nuclei then stained using DAPI (Sigma, #D9542). Slides were mounted using Fluoromount-G (Thermo Fisher, #00-4958-02) and imaged for quantification on a Keyence BZ-X800 microscope, with high magnification images taken on a Nikon A1RS confocal microscope.

2.6. TdT-mediated dUTP Nick-End Labeling Assay

TdT-mediated dUTP nick-end labeling (TUNEL) staining was carried out according to the manufacturer’s protocol using the Fluorescein In Situ Cell Death Detection Kit (Sigma, catalog # 11684795910). Staining was performed on 6 hearts per group, and three sections per heart.

2.7. Trichrome Staining

Hearts were harvested and fixed using 4% paraformaldehyde (PFA) in PBS overnight at 4 °C. The tissue was then washed with PBS and incubated overnight in 70% ethanol for dehydration before processing and paraffin embedding. Cardiac fibrosis and scar formation was then assayed using Masson’s trichrome staining with Newcomer Supply’s “Trichrome, Masson, Aniline Blue Stain Kit” (Newcomer Supply, #9179B) following the kit protocol. At least 4 biological replicates were analyzed with at least 3 independent sections from each biological replicate being stained for analysis.

2.8. Cardiomyocyte Cross-Sectional Area Quantification

To quantify cardiomyocyte cross sectional area, we stained paraffin embedded heart sections with a wheat germ agglutinin (WGA) (Thermo Fisher, #W11261) antibody that was pre-conjugated to an Alexa Fluor 488 fluorophore for visualization. Slides were incubated with the WGA antibody for 1 h at room temperature, rinsed with PBS-Tween (0.5%) and mounted in Fluoromount-G (Thermo Fisher, #00-4958-02) mounting medium, after which slides were then imaged. Fiji was used to quantify cardiomyocytes by measuring the cross-sectional area of 150 cardiomyocytes across 5 biological replicates for MI and 4 biological replicates for sham (SH) with 3 sections measured per biological replicate.

2.9. Global Quantitative Proteomic Analysis

Hearts were excised from the mice, the atria were dissected from the heart, and the hearts were quickly washed with PBS. Ventricular tissue below the ligature was then snap-frozen with liquid nitrogen and cryopulverized. The global ventricular proteome was extracted as previously described with Azo lysis buffer with minor modification [20, 21]. In a cold room (4 °C), ventricular tissue was thoroughly homogenized in 0.2% Azo lysis buffer (0.2% w/v Azo, 25 mM ammonium bicarbonate (ABC), 10 mM L-methionine, 1 mM DTT, and 1X HALT protease and phosphatase inhibitor) with a Teflon pestle. Samples were then sonicated in a sonicating water bath for 10 min at 4 °C prior to centrifugation at 21,100 g for 30 min at 4 °C. The supernatant was transferred to a new tube, and sample aliquots were diluted 1:50 in water before Bradford protein assay (Bio-Rad, Hercules, CA, USA, Cat# 5000006). Samples were normalized to 1 mg/mL in 0.1% Azo, reduced with 30 mM DTT at 37 °C for 1 h and alkylated with 30 mM chloroacetamide for 45 min. Trypsin Gold (Promega, Madison, WI, USA, Cat# V5820) was added to each sample in a 1:50 ratio (wt/wt) of protease:protein and incubated on 1000 rpm shaker at 37 °C overnight. The reaction was halted with 1% FA (final concentration). Azo was degraded at 305 nm (UVN-57 Handheld UV Lamp; Analytik Jena, Jena, TH, DEU) for 5 min, samples were cleared at 21,100 g at 4 °C for 30 min, and the peptides were desalted with 100 μL Pierce C18 tips (ThermoFisher Scientific, Cat# 87784) using the manufacturer’s protocol. Peptides were then dried in a vacuum centrifuge prior to reconstitution in 0.1% FA. Peptide concentration was estimated with A205 readings on a NanoDrop.

200 ng of peptide digest was separated using a nanoElute nano-flow ultra-high pressure LC system (Bruker Daltonics) coupled online to a trapped ion mobility quadruple-time-of-flight mass spectrometer (timsTOF Pro, Bruker Daltonics) using a CaptiveSpray nano-electrospray ion source. After a 10-min wash at 2% mobile phase B (mobile phase A (MPA): 0.1% FA; mobile phase B (MPB): 0.1% FA in acetonitrile) on a C18 trap column, peptides were separated on a capillary C18 column (25 cm length, 75 μm inner diameter, 1.6 μm particle size, 120 Å pore size; IonOpticks, Fitzroy, VIC, AUS) at a flow rate of 400 nL/min using a stepwise gradient of 2–17% MPB from 10–70 min, 17–25% MPB from 70–100 min, 25–37% MPB from 100–110 min, and 37–85% MPB from 110–120 min, with a 10 min wash at 85% MPB for a total runtime of 130 min. Samples were measured in data-independent analysis - parallel accumulation-serial fragmentation (diaPASEF) mode using 32 windows ranging from 0.6 to 1.421/K₀ and 400 to 1200 m/z [22].

Raw mass spectrometry (MS) files were processed with DIA-NN in double-pass mode using an in silico generated spectral library from the mouse Uniprot reviewed proteome (UP000000589, accessed 20 April 2022) [23]. Peptides were required to be between 200 and 1700 m/z with no more than 2 missed cleavages. Cysteine carbamidomethylation was set as a fixed modification, while methionine oxidation and N-terminal acetylation were set as variable modifications. Mass accuracy was set to 15 ppm. Match-between runs was enabled at 1% false-discovery rate (FDR). Heuristic protein grouping was turned off. Protein and peptides identifications were set at a 1% FDR. Protein and peptide output is included in Table S1.

Global protein abundance analysis was performed using “DAPAR” [24], “DEP” [25], and “IHW” [26] for R (version 4.1.0) as previously described [27]. Proteins were filtered to remove contaminants and proteins that were not quantified in at least three of five biological replicates within one sample group. Label-free quantification values were Log₂-transformed and normalized to the median of the total data set. Missing values were imputed via ssla for partially observed values with a sample group or set to 2.5% quantile for values missing across an entire sample group. Limma tests were performed to evaluate statistical significance, p-values were adjusted via independent hypothesis weighting based on the number of peptides observed per protein group [26], and statistical significance required a Log₂-fold change of 0.25 or greater in either direction and an FDR-adjusted p-value of 0.05 or less. A list of differentially expressed proteins is included in Table S2. Data were further visualized in R. Gene ontology analysis was performed using STRING [28], and the significance threshold was set at an FDR-adjusted p-value of 0.05.

2.10. Statistical Analysis

Graphs and statistical analysis were performed with Prism v9.4.0 (GraphPad). Two-way ANOVAs were performed to compare mTORC1 function during development and injury response, followed by Sidak’s multiple comparison tests to detect significant differences between analysis conditions. One-way ANOVAs were performed to compare mTORC1 treatment conditions, followed by Tukey’s multiple comparison tests to detect significant differences between treatment and control conditions. A p-value < 0.05 was determined as significant, with all error bars shown as standard error of the mean unless specified.

3. RESULTS

3.1. mTORC1 Signaling is Dynamically Regulated During Regenerative and Non-Regenerative States

The role of mTORC1 activity during postnatal heart maturation and regeneration remains undefined. To analyze mTORC1 activity during regenerative and non-regenerative cardiac injury responses, we utilized the neonatal mouse myocardial infarction (MI) models. MI occurring on postnatal day 1 (P1) is associated with cardiomyocyte proliferation and a robust regenerative response. In contrast, MI at postnatal day 7 (P7) results in a loss of this proliferative and regenerative capability, leading to scar formation. These models offer a valuable platform to examine the role of mTORC1 signaling in different injury responses as well as during postnatal maturation in uninjured controls. This approach is instrumental in establishing the role of mTORC1 in postnatal maturation, disease, and regeneration.

To define mTORC1 activity during endogenous cardiac regeneration, we measured the phosphorylation levels of primary downstream targets of mTORC1, S6K, 4EBP1, and ULK1, after P1 MI compared to uninjured controls. Hearts were harvested for protein analysis at key timepoints of the cardiac injury response: 1-day post-surgery (DPS), 7-DPS, and 28-DPS (Figure 1A). First, we quantified a significant increase in the phosphorylation of S6K-T389 at 7 DPS during the regenerative injury response compared to the uninjured controls (Figure 1B). mTORC1 phosphorylation of 4EBP1 occurs through sequential phosphorylation of multiple 4EBP1 phosphorylation residues, starting with phosphorylation of the rapamycin insensitive Threonine residues 37 and 46 [29]. We analyzed both rapamycin insensitive and sensitive residues of 4EBP1 by measuring the phosphorylation of 4EBP1-T37/46 and 4EBP1-S65, respectively. We detected no difference in 4EBP1-T37/46 phosphorylation during the regenerative injury response at all timepoints compared to uninjured controls (Figure S1A). Interestingly, phosphorylation of the rapamycin sensitive 4EBP1-S65 residue was significantly increased at 7 DPS after MI compared to uninjured controls (Figure 1C). This was accompanied by a significant increase in ULK1-S757 phosphorylation at 7 DPS following MI compared to uninjured controls, further demonstrating the increase in mTORC1 activity (Figure 1D). Additionally, Raptor levels remained unchanged at all timepoints in both regenerating hearts and uninjured controls (Figure 1E), indicating that mTORC1 target phosphorylation changes during regeneration are due to increased mTORC1 function rather than mTORC1 expression levels. These results demonstrate that mTORC1 activity is upregulated during the cardiac regenerative response.

Figure 1. mTORC1 signaling is dynamically regulated during postnatal cardiac development and following injury.

A. Schematic timeline of MI and collection timepoints for analyzing mTORC1 regulation during regenerative injury response. B. Representative immunoblots and quantification of S6K phosphorylation (pS6K) in uninjured and regenerative mice, demonstrating a significant increase in pS6K levels at 7DPS after P1 MI compared to uninjured controls (n=4 per group). C. Representative immunoblots and quantification of 4EBP1 rapamycin sensitive phosphorylation site S65 (p4EBP1-S65) in uninjured and regenerative mice, demonstrating a significant increase in p4EBP1 levels at 7DPS after P1 MI compared to uninjured controls (n=4 per group). D. Representative immunoblots and quantification of ULK1 phosphorylation (pULK1) in uninjured and regenerative mice, demonstrating a significant increase in pULK1 levels at 7DPS after P1 MI compared to uninjured controls (n=4 per group). E. Representative immunoblots and quantification of Raptor in uninjured and regenerative mice, demonstrating no change in Raptor levels between MI and uninjured controls at any timepoint (n=4 per group). F. Schematic timeline of MI and collection timepoints for analyzing mTORC1 regulation during non-regenerative injury response. G. Representative immunoblots and quantification of pS6K in uninjured and non-regenerative mice, demonstrating a significant increase in pS6K levels 7DPS after P7 MI compared to uninjured controls (n=4 per group). H. Representative immunoblots and quantification of p4EBP1–65 in uninjured and non-regenerative mice, demonstrating no change in p4EBP1 levels between P7 MI and uninjured controls at any timepoint (n=4 per group). I. Representative immunoblots and quantification of pULK1 in uninjured and non-regenerative mice, demonstrating an increase in pULK1 levels at 1DPS after P7 MI compared to uninjured controls (n=4 per group). J. Representative immunoblots and quantification of Raptor in uninjured and non-regenerative mice, demonstrating no change in Raptor levels between P7 MI and uninjured controls (n=4 per group). Error bars indicate mean ±S.E.M. Two-way ANOVA with Sidak’s multiple comparisons test, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

To determine mTORC1 activity during the cardiac non-regenerative injury response, we induced MI at P7, which naturally does not regenerate. Similarly, hearts were harvested for protein analysis at 1-, 7-, and 28-DPS together with uninjured controls (Figure 1F). Interestingly, we quantified a significant increase in S6K-T389 phosphorylation at 7DPS in non-regenerating hearts compared to the uninjured controls, similar to the regenerative injury response (Figure 1G). However, at all examined timepoints, there was no difference in 4EBP1-T37/46 or 4EBP1-S65 phosphorylation between the non-regenerating hearts and uninjured controls (Figures 1H, S1B). A notable increase in ULK1-S757 phosphorylation was only detected at 1 DPS in the non-regenerating hearts (Figure 1I). Furthermore, Raptor levels showed no significant variation at any timepoint when comparing non-regenerating hearts to uninjured controls (Figure 1J).

Collectively, our data demonstrate that downstream targets of mTORC1 are dynamically regulated during regeneration and the non-regenerative injury response, suggesting that mTORC1 may play an important role during postnatal cardiac regeneration and maturation.

3.2. mTORC1 Inhibition Prevents Neonatal Cardiac Regeneration

To determine whether mTORC1 is required for neonatal cardiomyocyte cell cycle activity and neonatal cardiac regeneration, neonatal mice were treated with the mTOR inhibitors rapamycin (1 mg/kg) or everolimus (1 mg/kg) daily for 3 days at 4, 5, and 6 DPS after a P1 MI or sham surgery (Figure 2A). Since mTORC1 plays a critical role during postnatal development and long-term inhibition in this stage can be detrimental, the dosing strategy was determined by optimizing for the dose concentration and treatment duration that resulted in no mortality at baseline (Figure S2). Dosages of 1.0 mg/kg, 2.5 mg/kg, and 10 mg/kg were tested. The 1.0 mg/kg dosage was chosen for our examination, as it was the only one that did not result in lethality or disrupt postnatal development in uninjured animals treated with rapamycin when administered on days 4, 5, and 6 DPS. Additionally, we utilized this acute treatment protocol as acute rapamycin or everolimus treatments have been demonstrated to be mTORC1 specific with minimal impacts to mTORC2 [30].

Figure 2. mTORC1 inhibition prevents neonatal cardiac regeneration.

A. Schematic timeline of MI strategy and rapamycin (Rapa) or everolimus (Evero) treatments. B. Immunostaining and quantification of pH3 positive cardiomyocytes demonstrating a significant decrease in cardiomyocyte mitosis following mTORC1 inhibition in both sham and MI groups (Sham: n=6 per group, MI: n=5 for Saline, and n=6 for Rapa and Evero). C. Immunostaining and quantification of Aurora B Kinase positive cardiomyocytes demonstrating a significant decrease in cardiomyocyte cytokinesis following mTORC1 inhibition post-MI but not sham groups (Sham: n=6 per group, MI: n=5 for Saline, and n=6 for Rapa and Evero). D. Quantification of TUNEL positive nuclei demonstrating no significant change in apoptosis in sham group, but a significant increase in apoptosis was quantified at 7dps in everolimus treated mice (n=6 per group). E. Masson’s Trichrome histological staining and fibrotic area quantification of heart sections from saline, rapamycin, or everolimus treated mice at 28 days post-MI showing a reduced regenerative capacity after mTORC1 inhibition (n=5 per group). F. Heart weight to body weight ratio demonstrating a significant increase in heart weight to body weight ratio following mTORC1 inhibition post-MI compared to controls (n=6 for Saline, n=8 for Rapa, n=7 for Evero). G. Representative images of WGA immunofluorescent staining and cardiomyocyte cross-sectional area quantification showing a significant increase in cardiomyocyte cross-sectional area at 28 days post-MI following mTORC1 inhibition. Quantification of cross-sectional area of 450 cardiomyocytes (150 cardiomyocytes per section for 3 sections) per biological replicate (n=5 per group). (Scale bars: B and C, 20 μm; E, 1 mm; G, 50 μm). Error bars indicate mean ± S.E.M. One-way ANOVA with Tukey’s multiple comparison tests, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

To examine whether mTORC1 inhibition alters cardiomyocyte cell cycle activity, we analyzed cardiomyocyte mitosis and cytokinesis at 7 DPS by immunostaining for phospho-Histone H3 and Aurora B Kinase, respectively, in both sham control and MI treated mice. We quantified a significant decrease in the number of cardiomyocytes undergoing mitosis in both sham and MI groups following mTORC1 inhibition (Figure 2B). We also quantified a significant decrease in cardiomyocyte cytokinesis following mTORC1 inhibition only in the MI group but not in sham controls (Figure 2C). To assess if acute mTORC1 inhibition in the postnatal heart influences apoptosis, we conducted TdT-mediated dUTP nick-end labeling (TUNEL) staining on ventricular tissue. While no difference was observed in the number of TUNEL positive nuclei in sham controls, a significant increase was noted in the everolimus-treated group post-MI, a change that was not seen in the rapamycin-treated group (Figure 2D). These results demonstrate that mTORC1 inhibition blocks cardiomyocyte cell cycle activity, which is the main source of the newly formed cardiomyocytes during neonatal cardiac regeneration.

To determine whether mTORC1 inhibition prevents neonatal cardiac regeneration, we performed Masson’s trichrome staining to assess scar formation and myocardial regeneration. Rapamycin and everolimus treated mice exhibited significantly larger scar size compared to saline treated mice along with left ventricular dilation, both hallmarks of loss of cardiac regeneration (Figure 2E and Figure S3). However, no significant histological changes were observed between saline, rapamycin, and everolimus treated mice under sham conditions (Figure S4A). We quantified a significant increase in heart weight to body weight ratios in rapamycin and everolimus treated mice post-MI compared to sham control and saline treated MI groups (Figure 2F). Furthermore, we measured cardiomyocyte cross-sectional area by wheat germ agglutinin (WGA) staining. We quantified a significant increase in cardiomyocyte cross sectional area in rapamycin and everolimus treated mice (Figure 2G). Interestingly, mTORC1 inhibition did not alter cardiomyocyte cross-sectional area following sham surgery, indicating that the impact of mTORC1 inhibition on cardiomyocyte hypertrophy is an injury-specific response (Figure S4B). This demonstrates that transient inhibition of mTORC1 by rapamycin or everolimus does not disrupt postnatal development, but promotes cardiac hypertrophy and pathological remodeling post-MI. Collectively, our results demonstrate that mTORC1 function is critical for neonatal cardiomyocyte proliferation and cardiac regeneration.

Since mTORC1 function is necessary for neonatal cardiac regeneration, we wanted to determine the downstream mechanism by which mTORC1 mediates its function. As mTORC1 regulates protein synthesis, we performed the SUnSET assay to measure protein translation in rapamycin and everolimus treated hearts [31, 32]. Following MI at P1, mice were treated with rapamycin or everolimus daily for 3 days at 4, 5, and 6 DPS. This was followed by a single injection with puromycin at 1 hr before harvest at 7 DPS, which can incorporate into newly synthesized proteins. No significant change in puromycin immunoblotting was detected during uninjured or regenerative conditions between saline, rapamycin, or everolimus treated mice, demonstrating that mTORC1 inhibition did not significantly impact protein synthesis (Figure S5A and S5B).

To define the impact of rapamycin and everolimus treatment on mTORC1 function following injury, we analyzed the phosphorylation of S6K, 4EBP1, and ULK1. We quantified a significant reduction in S6K phosphorylation in both rapamycin- and everolimus-treated mice post-MI compared to saline-treated controls (Figures S6A and S7A). We also quantified a significant decrease in phosphorylation of the rapamycin sensitive 4EBP1-S65 residue in everolimus-treated mice compared to saline and rapamycin treatments in both sham and MI groups, which may be due to the higher potency of everolimus compared to rapamycin (Figure S6B and S7B) [33]. No significant change was observed in ULK1-S757 phosphorylation in both sham and MI mice treated with rapamycin and everolimus compared to saline controls (Figures S6C and S7C). This is consistent with previous studies that have shown ULK1 phosphorylation to be largely resistant to rapamycin [34]. Taken together, our results demonstrate that mTORC1 is required for heart regeneration, and that this role is not mediated via regulation of protein synthesis but rather through downstream signaling of mTORC1.

3.3. Global Proteomic Analysis of mTORC1 Inhibition Defines its Role in Cardiac Metabolism

Since transient inhibition of mTORC1 following injury did not disrupt protein synthesis, we sought to investigate the distinct cardiac proteomic signature that drives the loss of regeneration by using a global bottom-up quantitative proteomic analysis. Mice were subjected to sham or MI surgeries at P1, treated intraperitoneally with either saline or rapamycin at 4, 5, and 6 DPS, and then hearts were collected at 7 DPS (Figure 3A). No significant histological changes were observed following mTORC1 inhibition within each surgical group (Figure S8). In our analysis, we identified over 6,300 protein groups in each sample group and found that nearly all these protein groups (6,322) were identified in 3 of 5 biological replicates in all four sample groups (Figure 3B), which confirms our earlier data that rapamycin treatment does not drastically alter gross proteome composition. Furthermore, the proteomic data showed a high degree of normality and median alignment (Figure S9A), low coefficients of variation among sample groups (Figure S9B), and high data completeness (Figure S9C), bolstering our confidence in downstream quantitative analyses. When examining our samples using a dimensional reduction analysis, we noticed the largest separation among sample groups was due to injury (separation along PC1) (Figure 3C). It is clear, however, that rapamycin treatment also affects proteome composition, as saline and rapamycin-treated samples cluster separately (Figure 3C). Therefore, we first examined how mTORC1 inhibition affects baseline cardiac proteome composition.

Figure 3. Global proteomic analysis of mTORC1 inhibition highlights its role in cardiac metabolism.

A. Outline of experimental design. Mice at P1 (n=5 per group) were given sham or myocardial infarction (MI) surgeries. Mice were treated 4, 5, 6 days post-surgery (DPS) with I.P. injection of saline (Sal) or rapamycin (Rapa). Hearts were harvested for global proteomic analysis at 7 DPS. B. Protein IDs were filtered to require n ≥ 2 per sample group for representation. Inlay: Bar chart demonstrating unique protein groups identified from each sample group (bar demonstrates mean, error bars demonstrate s.e.m.). C. Principal component analysis of per-sample Log₂ protein abundances demonstrates separation of sample groups based on treatment (Sal or Rapa) and surgery (sham or MI). D. Hierarchal heatmap demonstrating z-score normalized intensities of significantly differently expressed proteins between saline and rapamycin treated sham mice (adjusted p-value ≤ 0.05 and |Log₂ Fold Change| ≥ 0.25 required to be considered significant). The heatmap is separated into two clusters: proteins that decrease in expression with rapamycin treatment (grey) and proteins that increase with rapamycin treatment (black). E. and F. STRING Biological process gene ontology (GO) analysis of proteins that decrease in expression with rapamycin treatment (E) and proteins that increase in expression with rapamycin treatment (F). Dot size represents the number of identified proteins within a GO group. Gene ratio indicates the fraction of all proteins within the GO group that were identified. Color represents the FDR-adjusted p-value of the GO overrepresentation test. G. Boxplots displaying key proteins significantly dysregulated by rapamycin treatment in sham mice. Boxplots represent n=5 biological replicates per experimental group (* indicates adjusted p-value ≤ 0.05 and |Log₂ Fold Change| ≥ 0.25 as determined by limma testing). Error bars indicate mean ±S.E.M., dots represent individual data points.

Using a limma-based differential expression analysis, we were able to identify 262 differentially expressed proteins (DEPs) between saline and rapamycin-treated sham P8 mouse hearts (Table S2). When plotting these DEPs on a z-score normalized heatmap, we noticed that sample groups clustered together and there were two distinct trends: a group of proteins that decreased with rapamycin treatment and a group of proteins that increased (Figure 3D). Interestingly, rapamycin treatment led to a larger cluster of proteins with both decreased and increased abundance when compared to saline-treated mice (Figure 3D). Using the group of downregulated proteins, we performed gene ontology (GO) analysis for biological processes with STRING and observed that rapamycin-treatment led to a decrease in abundance of proteins involved in cell cycle and translational regulation processes (Figure 3E, Table S3), including regulation of gene expression (GO: 0010468), nucleic acid metabolic processes (GO: 0090304), regulation of translation (GO: 0006417), and mitotic cell cycle (GO: 0000278). On the other hand, GO analysis of the upregulated proteins following rapamycin-treatment demonstrate an increase in level of proteins involved in metabolism including oxidative-reduction processes (GO:0055114) and lipid transport (GO:0006869) (Figure 3F, Table S3). This indicates that under baseline conditions, systemic administration of rapamycin leads to changes in protein levels related to cell cycle and metabolism.

Specifically, we noted a change in protein expression that indicates an increase in cardiomyocyte metabolic maturation after mTORC1 inhibition compared to control hearts. Of note, there was a significant reduction in the protein levels of glycolysis and branched chain amino acid (BCAA) catabolism enzymes such as ALDOC, FASN, and BCAT2 following mTORC1 inhibition (Figure 3G). Additionally, there was a significant increase in multiple mitochondrial metabolism proteins including ACAD10 and NDUFB3 (Figure 3G). Our proteomic analysis indicates that mTORC1 inhibition by rapamycin in the postnatal heart shifts the neonatal proteomic landscape from glycolysis and BCAA catabolism towards increased oxidative phosphorylation that may result in premature metabolic maturation. These results suggest that mTORC1 inhibition via rapamycin treatment shifts the cardiac proteomic landscape towards a more mature metabolic state without inducing hypertrophic remodeling in the uninjured heart (Figure 2E, Figure S4A).

3.4. mTORC1 Inhibition Prevents Cardiac Regeneration Through Metabolic Remodeling

Since we identified distinct separation between saline and rapamycin-treated mice post-MI at 7 DPS in our principal component analysis (Figure 3C), we next analyzed how rapamycin-treatment affects proteome composition after injury to glean insight into how mTORC1 inhibition prevents cardiac regeneration (Figure 3A). We identified 251 DEPs, with 117 proteins upregulated in the saline-treated regenerative mice and 134 proteins upregulated in the rapamycin-treated non-regenerative mice (Figure 4A). Using the proteins that were upregulated in the saline-treated regenerative mice, we identified GO processes related to nucleotide biosynthetic processes (GO: 0009165), positive regulation of cardiac muscle tissue growth (GO: 0055023), and mitotic cell cycle (GO: 0000278) (Figure 4B). As cardiac regeneration is a function of cardiac development [35], it is plausible that these developmental processes would be upregulated during a regenerative response to injury. Conversely, when we perform GO analysis on the proteins upregulated in the rapamycin-treated non-regenerative mice, we identified processes related to regulation of cell death (GO: 0010941) and Fibrinolysis (GO: 0042730) (Figure 4C, Table S4), indicating a pathological response to injury. Interestingly, we also detected an increase in expression of proteins associated with oxidative phosphorylation in the rapamycin-treated non-regenerative mice, including oxidative-reduction processes (GO: 0055114) and fatty acid metabolic processes (GO: 0006631) (Figure 4C, Table S4), further underlining the role of mTORC1 in cardiac metabolism and its influence over cardiac regeneration.

Figure 4. mTORC1 inhibition prevents cardiac regeneration through a metabolic remodeling.

A. Hierarchal heatmap demonstrating z-score normalized intensities of significantly differently expressed proteins between saline and rapamycin-treated MI mice (adjusted p-value ≤ 0.05 and |Log₂ Fold Change| ≥ 0.25 required to be considered significant). The heatmap is separated into two clusters: proteins that decrease in expression with rapamycin treatment (grey) and proteins that increase with rapamycin treatment (black). B and C. STRING Biological process gene ontology (GO) analysis of proteins that decrease in expression with rapamycin treatment after MI (B) and proteins that increase in expression with rapamycin treatment post-MI (C). Dot size represents the number of identified proteins within a GO group. Gene ratio indicates the fraction of all proteins within the GO group that were identified. Color represents the FDR-adjusted p-value of the GO overrepresentation test. D. Volcano plot displaying fold-change in protein expression between saline and rapamycin-treated MI mice. The number of significantly upregulated proteins per group is shown in the bottom corners of the comparison (n=5 per group, * indicates adjusted p-value ≤ 0.05 and |Log₂ Fold Change| ≥ 0.25). E. Boxplots displaying key proteins significantly dysregulated by rapamycin treatment after MI. Boxplots represent n=5 biological replicates per experimental group (* indicates adjusted p-value ≤ 0.05 and |Log₂ Fold Change| ≥ 0.25 as determined by limma testing). Error bars indicate mean ±S.E.M. Dots represent individual data points.

Of specific interest, we identified differential expression of key cardiac and metabolic proteins between the non-regenerating rapamycin-treated mice and the saline-treated regenerative controls (Figure 4D and 4E). Regeneration is dependent on a glycolytic metabolic state, which is supported by a significant decrease in PDK4 levels, demonstrating an increase in glycolytic metabolism in saline-treated mice compared to rapamycin-treated mice. Rapamycin-treated hearts displayed reduced levels of PPM1K, ACSS1, and FASN compared to saline-treated controls, indicating a reduced rate of BCAA metabolism and lipid synthesis in the non-regenerative hearts. Furthermore, rapamycin treatment results in a significant upregulation of oxidative phosphorylation and fatty acid oxidation proteins post-MI, such as ACSL1, CPT1B, DECR1, and SLC25A20. This increase in oxidative phosphorylation proteins in the rapamycin-treated hearts is concomitant with a significant decrease in protein levels of the proliferation marker PCNA, supporting our results that rapamycin-treatment reduces cardiomyocyte proliferation and heart regeneration following MI (Figure 2). These results demonstrate that mTORC1 inhibition by rapamycin reduces the protein levels of glycolytic enzymes while increasing the protein levels of oxidative phosphorylation and fatty acid oxidation enzymes following injury, suggesting that mTORC1 inhibition accelerates the metabolic maturation of the postnatal heart that may contribute to the loss of cardiomyocyte proliferation and heart regeneration post-MI.

4. DISCUSSION

The postnatal metabolic switch of the mammalian heart from glycolysis to fatty acid oxidation is an essential part of cardiac maturation and maintenance of adult heart function; however, this switch contributes to loss of the endogenous regenerative potential of the mammalian heart. Recent studies demonstrate that inhibition of fatty acid oxidation and mitochondrial oxidative phosphorylation via pyruvate dehydrogenase kinase 4 (PDK4) deletion or inhibition of succinate dehydrogenase (SDH) can induce adult cardiomyocyte proliferation and heart regeneration [36, 37]. Thus, elucidating the mechanisms that control this metabolic switch can lead to the development of novel targeted approaches to promote cardiac regeneration.

mTORC1 represents a central node in regulating metabolic pathways and protein synthesis that influence cellular growth as well as cell fate [8]. However, the role of mTORC1 during postnatal cardiac maturation and its influence on heart regeneration is unclear. In this study, we demonstrate that mTORC1 signaling is dynamically regulated during postnatal development and regeneration in contrast to the non-regenerating heart. These results suggest that mTORC1 may play an important role during cardiomyocyte proliferation and regeneration. Interestingly, transient inhibition of mTORC1 at 4–6 days following MI at P1 through treatment with either rapamycin or the rapalog everolimus, results in impeded cardiomyocyte proliferation, increased scar size, and inhibition of neonatal mouse heart regeneration. A previous study demonstrated that early mTORC1 inhibition by rapamycin at P1 and P2 not only reduces cell proliferation and the activation of downstream targets, but also diminishes hypertrophic growth and sarcomere maturation [38]. Furthermore, the effect of mTORC1 inhibition, which increases scar size during regeneration, contrasts with the effect of mTOR inhibition after adult MI, where it leads to a reduction in scar size [39]. This highlights a distinct difference in the role of mTORC1 inhibition during neonatal heart regeneration compared to cardiac remodeling following adult injury. Although the overall levels of protein synthesis were not affected by transient inhibition of mTORC1, our proteomic analysis demonstrates that acute mTORC1 inhibition by rapamycin treatment results in a distinct proteomic signature, which exhibits reduced levels of key glycolytic enzymes and higher levels of oxidative phosphorylation and fatty acid oxidation proteins suggesting that mTORC1 inhibition promotes postnatal metabolic maturation. This is evident by the reduced levels of myocyte proliferation and inhibition of heart regeneration following injury. These results suggest that mTORC1 inhibition specifically accelerates metabolic maturation post-MI compared to sham controls, which reduces rates of cardiomyocyte proliferation and increases cardiomyocyte size following infarction. This is consistent with a recent study that demonstrates that mTORC1 inhibition can promote maturation of induced pluripotent stem cell derived cardiomyocytes in vitro [40].

Our results suggest that targeting mTORC1 could be a promising strategy for modulating the cardiomyocyte transition from a hyperplastic to a hypertrophic state to promote heart regeneration following injury. Furthermore, mTOR signaling may represent an evolutionarily conserved mechanism during regeneration, as a recent study demonstrates an important role for mTOR activity during zebrafish heart regeneration [41]. However, the downstream metabolic state following mTORC1 inhibition remains to be fully defined. Elucidating how mTORC1 impacts cardiomyocyte metabolism specifically with respect to oxidative phosphorylation and fatty acid synthesis is important to fully establish the role of mTORC1 on cardiomyocyte maturation and function. A limitation of our study is the use of pharmacological inhibitors of mTORC1, which can affect various cell types in the heart. Therefore, to better understand the role of mTORC1 in different cell types during heart regeneration, the development of cell-specific mTORC1 inhibition using targeted genetic approaches is warranted. In addition, rapamycin and everolimus are known immunosuppressants, thus whether this inhibition of myocyte proliferation and regeneration is partially mediated via targeting the immune response needs to be established. Lastly, it is crucial to understand the impact of mTORC1 pathway activation in the adult heart to establish whether targeting mTORC1 can effectively promote regeneration in adult hearts. Nevertheless, our current study highlights the importance of mTORC1 in regulating mammalian cardiomyocyte cell cycle activity and heart regeneration, suggesting that mTORC1 could be a promising therapeutic target for heart failure.

Highlights.

• mTORC1 activity is distinct between regenerating and non-regenerating hearts

• mTORC1 inhibition reduces cardiomyocyte proliferation and heart regeneration

• mTORC1 inhibition shifts the cardiac proteome towards metabolic maturation

• mTORC1 inhibition inhibits heart regeneration via metabolic remodeling

Data availability

Source data are available via the MassIVE repository with identifier MSV000092436 and the PRIDE repository via ProteomeXchange with identifier PXD043771. All data are available from the corresponding author upon request.