β-hydroxybutyrate administered at reperfusion reduces infarct size and preserves cardiac function by improving mitochondrial function through autophagy in male mice

Abstract

Ischemia/reperfusion (I/R) injury after revascularization contributes ~50% of infarct size and causes heart failure, for which no established clinical treatment exists. β-hydroxybutyrate (β-OHB), which serves as both an energy source and a signaling molecule, has recently been reported to be cardioprotective when administered immediately before I/R and continuously after reperfusion. This study aims to determine whether administering β-OHB at the time of reperfusion with a single dose can alleviate I/R injury and, if so, to define the mechanisms involved. We found plasma β-OHB levels were elevated during ischemia in STEMI patients, albeit not to myocardial protection level, and decreased after revascularization. In mice, compared with normal saline, β-OHB administrated at reperfusion reduced infarct size (by 50%) and preserved cardiac function, as well as activated autophagy and preserved mtDNA levels in the border zone. Our treatment with one dose β-OHB reached a level achievable with fasting and strenuous physical activity. In neonatal rat ventricular myocytes (NRVMs) subjected to I/R, β-OHB at physiologic level reduced cell death, increased autophagy, preserved mitochondrial mass, function, and membrane potential, in addition to attenuating reactive oxygen species (ROS) levels. ATG7 knockdown/knockout abolished the protective effects of β-OHB observed both in vitro and in vivo. Mechanistically, β-OHB’s cardioprotective effects were associated with inhibition of mTOR signaling. In conclusion, β-OHB, when administered at reperfusion, reduces infarct size and maintains mitochondrial homeostasis by increasing autophagic flux (potentially through mTOR inhibition). Since β-OHB has been safely tested in heart failure patients, it may be a viable therapeutic to reduce infarct size in STEMI patients.

1. Introduction

Heart failure is a global health problem. ST-elevation myocardial infarction (STEMI) is the leading cause of heart failure, with high morbidity and mortality [1]. It is extremely important to expeditiously restore the blood flow and oxygen supplementation to the ischemic myocardium to lower mortality [2]. Emergent percutaneous coronary intervention (PCI) is the most effective way to achieve timely and effective reperfusion, resulting in improved clinical outcomes [2]. However, the reperfusion is usually associated with additional damage to the ischemic myocardium up to 50% of the final infarct size, which is called ischemia/reperfusion (I/R) injury [3]. The I/R injury has been recognized for many years. However, there is still no standard therapy to reduce I/R injury after prompt reperfusion therapy to further reduce the infarct size in clinical practice [4]. Therefore, investigating how to alleviate I/R damage after establishing blood flow with PCI treatment is greatly needed.

Acute myocardial infarction (AMI) is closely related to metabolism derangements [5]. The heart is largely dependent on the oxidation phosphorylation in mitochondria to generate adequate ATP to maintain its energy-intensive electrical and mechanical activities. Upon AMI, the availability of energy substrates, such as fatty acids and glucose and oxygen, is greatly limited, and the mitochondria are seriously damaged with extended ischemia [6,7]. As each mitochondrion has its own copy of DNA, the levels of mitochondrial DNA (mtDNA) will represent the mitochondrial abundance, and the intact mtDNA levels can indicate functional mitochondrial genomes [8]. During reperfusion, reactive oxygen species (ROS) generated from damaged mitochondria and other sources will mediate mitochondrial permeability transition pore opening, inflammation, and intracellular calcium overload, which will contribute to further cell damage [9], thus initiating cell death and compromising myocardium contract function [10]. Mitochondria are a source of ROS, and thus play a central role in this process [11]. Excessive ROS from damaged electron transported chain will cause further mitochondrial dysfunction, creating a vicious circle. Consequently, to mitigate I/R injury, it is crucial to maintain cardiac mitochondrial homeostasis in response to acute heart stress by enhancing mitochondrial dynamics and quality control [12,13].

Beta-hydroxybutyrate (β-OHB) caught attention in the field of heart failure treatment because during heart failure, utilization of β-OHB significantly increases as an alternative fuel source [6]. In addition to providing energy, β-OHB serves as a signaling molecule in the regulation of multiple pathways, including autophagy, inflammation, and apoptosis [14]. Increasing β-OHB levels before I/R injury reduces infarct size in rodents [15,16], which is impractical to achieve prior to an un-predictable heart attack in humans. Continuously given low dose β-OHB at reperfusion protected the heart from I/R injury in mice [17]. Pre-treatment with continuous IV β-OHB also reduced infarct size in pig similarly to the effect of SGLT2 inhibitor, empagliflozin [18]. These studies were done in whole animals and the direct effect of β-OHB on cardiomyocytes during I/R were not investigated. Furthermore, the exact mechanisms of β-OHB’s cardioprotective effects haven’t been thoroughly defined. It has been noted that β-OHB is an endogenous histone deacetylase (HDAC) inhibitor and reduces ROS abundance at physiological levels [19]. The cardioprotective effects of HDAC inhibitors have been widely investigated, which include induction of cell-protective autophagy, suppression of cell apoptosis, and reduction of mitochondrial damage [12,20,21]. Whether β-OHB affects the pathways that are regulated by HDAC inhibitors, such as autophagy and whether its protective effects are dependent on autophagy, needs to be further explored.

We hypothesized that β-OHB exerts cardioprotective effects by enhancing autophagy and maintaining mitochondrial homeostasis during I/R injury. To test this hypothesis, two approaches were used: i) mice were subjected to I/R surgery and ii) neonatal rat ventricular myocytes (NRVMs) were cultured under normoxia and simulated I/R conditions to gain insight into the mechanisms by which β-OHB alleviates myocardial I/R injury. Our studies demonstrated that β-OHB reduced infarct size by ~50% and preserved mitochondrial functions through increasing the autophagic flux when given at the time of reperfusion. Loss of ATG7, an essential autophagy gene, abolished β-OHB’s cardioprotective effects. Additionally, β-OHB induced autophagy by inhibiting mTOR signaling pathway.

2. Materials and methods

2.1. Animal care

All experiments involving animals were approved by the Institutional Animal Care and Use Committee of Heersink School of Medicine, the University of Alabama at Birmingham, and performed under the Guidelines for the Care and Use of Laboratory Animals published by the US National Institutes of Health (2011) [22]. All mice were bred at a constant 25 °C pathogen-free environment under a 12-h-day/night light cycle and had free access to laboratory chow and water.

2.2. Mouse cardiac I/R surgery and tissue harvest

The surgical procedure was detailed in previous publications [12,23]. Briefly, 8- to 12-week-old male C57BL/6 J wild-type mice were anesthetized with isoflurane and then placed on an electronic heating pad at 37 °C with the ventilation of mixed oxygen and isoflurane (1.5–2%). After the heart was exposed, the left anterior descending artery was ligated with a reversible knot to achieve ischemia. After 45 min ischemia, control (normal saline) or 10 mmol/kg β-OHB (dissolved in water at the concentration of 1 mol/L, given at the dose of 10 μL stock/g of body weight) was given by intraperitoneal injection. Then the knot was released immediately to allow reperfusion, followed by chest closing. The survival rate is >90% after 24 h. After 24 h of reperfusion, the heart was harvested and the left ventricle was divided into 3 parts: ischemia zone, border zone, and remote zone according to the color difference under stereomicroscope. The heart tissue was snap frozen with liquid nitrogen and lysed with Triton X-100 lysis buffer. Then protein, DNA, and RNA were isolated separately to evaluate the autophagic flux, mtDNA level, and related gene expression levels.

2.3. 2,3,5-Triphenyltetrazolium chloride staining and infarct size determination

The 2,3,5-triphenyltetrazolium chloride staining was done following the detailed protocol as previously published [23]. Briefly, the mice were given 1000 U/kg heparin prior to sacrifice. The heart was excised and retrogradely perfused with PBS through an aortic cannula. The heart was perfused with 1% 2,3,5-triphenyltetrazolium chloride for 6–7 min at 37 °C. The left anterior descending artery was tied using the suture left in place at the site of the previous occlusion, and then perfused with 5% phthalo blue dye in normal saline to delineate the area at risk. After this procedure, the infarct area, area at risk, and non-infarct area were clearly marked by different colors of staining. After fixation with 10% zinc formalin for 1 day at 4 °C, the left ventricle was cut into 5–6 slices parallel to the short axis of the heart, weighed, and images were taken. The non-infarct zone was stained into blue, area at risk was stained into light red, and infarct zone was shown as light yellow. Then the areas of three zones were quantified with Image J software [24]. The percentage of area at risk and infarct zone were determined based on the weight of each slice.

2.4. Echocardiography

The heart function was measured using a standard protocol before and 24 h after the I/R surgery using the VEVO3100 imaging system. Left ventricular ejection fraction (LVEF) was measured by tracing the LV 2D images in parasternal long axis view. The M mode and strain images were taken to show the wall motion.

2.5. Measurement of circulating β-OHB levels in mouse

Mouse circulating β-OHB levels were measured at baseline, right before reperfusion, and 15 min / 30 min / 45 min / 1 h / 2 h / 3 h / 4 h / 6 h / 8 h / 24 h after reperfusion using over-the-counter Precision Xtra blood glucose and ketone monitoring system (Abbott). At different time points, fresh blood samples were obtained by cutting mouse tails and applied to the test strip immediately. The circulating β-OHB levels were recorded to make the concentration-time curve.

2.6. NRVMs isolation, culture, and simulated I/R

NRVMs were isolated from Sprague-Dawley rats with Neomyt kit (Cellutron, nc-6031) following the manufacturer’s protocol. Briefly, ventricles of 1–2-day neonatal rats were prewashed and digested for several rounds, then pelleted by centrifugation. Cells were plated on 24 well plates at a density of 2.8–3 × 10⁵ per well and cultured in DMEM: M199 3:1 medium containing 5% FBS supplemented with 100 μM bro-modeoxyuridine (Sigma Aldrich, St Louis, MO, USA) with daily medium change. Simulated I/R was performed on day 4 after isolation. Normal culture medium was replaced by ischemia buffer (in mM: 20 deoxy-glucose, 125 NaCl, 8 KCl, 1.2 KH₂PO₄, 1.25 MgSO₄, 1.2 CaCl₂, 6.25 NaHCO₃, 5 sodium lactate, 20 HEPES, pH 6.6) and cells were incubated in a 37 °C chamber with 95% nitrogen and 5% carbon dioxide [13]. Normoxia control cells were cultured with normal culture medium in an incubator with 95% room air and 5% carbon dioxide [23]. After ischemia, cells were incubated with a normal culture medium containing PBS or 2/4/8 mM β-OHB. Proteins were collected after 2 h of ischemia and 4 h of reperfusion. DNA and RNA were collected and extracted after 2 h of ischemia and 6 h of reperfusion.

2.7. Western blots analysis

For western blots, cells were lysed with SDS lysis buffer, and heart tissues were homogenized in lysis buffer based on weight, and then SDS buffer was added. After sonication and denaturation, proteins were separated on a 4–12% gradient SDS-PAGE gel and transferred to a PVDF membrane for blotting. The membrane was blocked for 60 min at room temperature in a mix of 5% milk in TBST (Tris-buffered saline +0.1% Tween 20) followed by incubating with primary antibodies overnight at 4°C. The next day, the membrane was washed with TBST before incubating in a secondary antibody diluted in a blocking buffer for 50 min at room temperature. After that, the membrane was washed three times with TBST and imaged with a Western HRP substrate kit. The images were quantified with Image Quant software (GE Healthcare Bio-Sciences AB, Sweden). Antibody used: GAPDH (1:5000, Fitzgerald 10R-2932), p62 (1:2000, Abcam ab91526), LC3 (1:1000, a gift from Dr. Joseph A. Hill’s laboratory), ATG7 (1:2500, Abcam ab133528), pmTOR (1:1000, Cell signaling technology 5536S), mTOR (1:1000, Cell signaling technology 2983S), pS6 (1:1000, Cell signaling technology 4858S), S6 (1:1000, Cell signaling technology 2217S), p4EBP1 (1:1000, Cell signaling technology 2855S), 4EBP1 (1:1000, Cell signaling technology 9644S), pAMPK (1:1000, Cell signaling technology 2535S), AMPK (1:1000, Cell signaling technology 2532S), pULK1 (1:1000, Cell signaling technology, 5869S), ULK1 (1:1000, Cell signaling technology, 8054S).

2.8. Cell death assay

The cell death was assessed using a cytotoxicity detection kit (LDH) (Roche 11644793001). On day 4 after isolation, NRVMs were subjected to ischemia for 5 h and reperfusion for 3 h as described above, then lysed with 0.9% Triton X-100. Cell culture mediums were collected right after ischemia, reperfusion, and lysis separately for detection of released LDH. After removing cell debris by centrifugation, the supernatant was mixed with LDH detection buffer and incubated at room temperature for 30 min. The absorbance was detected at 490 nm using SpectraMax i3x imaging cytometer (Molecular Devices, LLC., USA). The cell death was calculated by the ratio of LDH in ischemia and reperfusion media to the total LDH.

2.9. Cell viability assay

The cell viability was determined using MTT cell viability assay kit (Invitrogen V13154). NRVMs were treated with ischemia for 5 h and reperfusion for 3 h, as described before. Then cells were incubated with MTT diluted in no phenol-red culture medium overnight. HCl-SDS solution was added on the next day and incubated at 37 °C for 5 h. The absorbance at 570 nm was measured to evaluate the cell viability using SpectraMax i3x imaging cytometer.

2.10. qRT-PCR analysis

The DNA was isolated using the DNeasy Blood & Tissue Kits (Qiagen, 69506) following the manufacturer’s protocol. For mtDNA qPCR, primers specific for genes coded by mitochondrial DNA: ATP6, COXII, ND1, and ND6 were used (Supplement table 3). The RNA was isolated using the TRIzol reagent (Invitrogen, 15596026) and reverse transcripted to cDNA using the high-capacity cDNA reverse transcription kits (Thermo Fisher 4368814) in a total mass of 2000 ng. Then qPCR was performed using specific primers (Supplement table 3). The relative gene expression was calculated by the 2^−ΔΔCT method [25].

2.11. Mitochondrial staining

Tetramethylrhodamine methyl ester (TMRM, Invitrogen, I34361), MitoTracker (Invitrogen, M7512), MitoSOX (Invitrogen, M36008), and Dichlorofluorescein (H2DCFDA, Invitrogen, C6827) dyes were used to detect mitochondrial membrane potential, mitochondrial abundancy, and mitochondrial and cytosolic ROS generation respectively. Briefly, after ischemia for 2 h and reperfusion for 4 h as described above, the NRVMs were incubated in culture mediums or PBS containing the probes at optimized concentrations (TMRM 100 nM, MitoTracker 1 μM, MitoSOX 5 μM, H2DCFDA 5 μM) for 15 min at 37 °C. Then images were taken with a fluorescent microscope (Nikon Eclipse Ti) under fixed exposure time. Five pictures of the same group were taken and quantified. The fluorescent intensity was quantified with Image J software (NIH, Bethesda, USA).

2.12. Long and short PCR of mtDNA

The intact and fragmental mtDNA were amplified using specific primers [8]. The same amount of PCR products were loaded to the agarose gel for electrophoresis, and the band intensities were quantified using Image Quant software. The ratio of intact mtDNA (16.2 kb) level to fragmental (0.22 kb) was used to evaluate mtDNA damage [8].

2.13. Mitochondrial OCR measurements

Analyses of cellular bioenergetics were performed using the Seahorse XFe96 Extracellular Flux Analyzer (Agilent, Santa Clara, CA). The NRVMs were plated into an Agilent seahorse XF96 cell culture micro-plate (Product No.101085–004) at a density of 28,000 cells/well after isolation. On the day of the experiment, the cells were exposed to 2 h of ischemia followed by a 4 h reperfusion. β-OHB was added at the time of reperfusion. Then the cells were changed into the XF media (non-buffered DMEM supplemented with 5.5 mM glucose, 1 mM pyruvate, and 4 mM glutamine, pH 7.36) at 37 °C. Oxygen consumption rate (OCR) was measured for basal OCR (OCR before oligomycin minus OCR after antimycin) followed by sequential injections of 1 μg/mL oligomycin, 1 μM FCCP, and 10 μM antimycin A. Mitochondrial parameters were calculated for ATP-linked (OCR before oligomycin minus OCR after oligomycin), proton leak (OCR after oligomycin minus OCR after antimycin), maximal (OCR after FCCP minus OCR after antimycin), reserve capacity (OCR after FCCP minus OCR before oligomycin), and non-mitochondrial (OCR after antimycin). The results of seahorse tests were averaged from 8 wells and normalized to the value per 10,000 cells [26,27].

2.14. Evaluation of autophagic flux in mice

CAG-RFP-GFP-LC3 transgenic mice were used to detect autophagic flux [23]. After 3 h of PBS or 10 mmol/kg β-OHB IP injection, the mouse hearts were harvested and embedded with OCT compound by snap frozen in liquid nitrogen. GFP and RFP signals were detected in frozen sections by confocal microscope and quantified by Image J software. The RFP puncta indicate autolysosome, and the merged GFP and RFP puncta indicate autophagosome. The numbers of autolysosomes and autophagosomes were counted in 4 × 10⁴ μm² images and used to show the autophagic flux in mice.

2.15. siRNA transfection

On day 2 after isolation, the siRNA negative control (MISSION siRNA universal negative control #1, sigma) or ATG7 specific siRNA (SASI_Rn01_00050326, sigma) were transfected to NRVMs following the manufacturer’s protocol. Briefly, 4 μL of 40 μM siRNA stock solution and the same volume of lipofectamine RNAiMAX transfection reagent (Invitrogen, 13778100) were mixed in 1 mL Opti-MEM medium. The cells were cultured in the mixture for 6 h followed by changing to normal culture medium. 36 h after the transfection, the cells were treated and harvested for further analysis, as described above. The knockdown efficiency was evaluated by qRT-PCR and Western blot.

2.16. Human plasma β-OHB concentration measurement

The human peripheral blood was collected from 59 AMI patients at University of Alabama at Birmingham (UAB) hospital after Institutional Review Board (IRB) approval. Blood was collected right before PCI when arterial accesses were established and 24 h after PCI with the patients’ consent. Plasmas were isolated after blood collection by centrifugation at 16,000 g for 10 min, and snap frozen in liquid nitrogen until analysis. The concentration of β-OHB in human plasma was measured using β-hydroxybutyrate (ketone body) colorimetric assay kit (Cayman Chemical, 700190). The plasma sample was diluted with β-OHB assay buffer and mixed with freshly made developer solution containing β-OHB enzyme solution. After incubating at room temperature for 30 min, the absorbance was read at 450 nm. Then the concentration was determined according to the standard curve.

2.17. Generating cardiac-specific ATG7 knockout mice

To induce MerCreMer activity, the cardiac-specific ATG7^{flox/flox/αMHC-MerCreMer} mice were injected intraperitoneally with 20 mg/kg tamoxifen daily for consecutive 5 days [28]. MerCreMer negative littermates were served as negative controls. Two days after the last injection, mice were subjected to I/R surgery, and heart tissues were collected for further analysis as described above.

2.18. Statistical analysis

All results are reported as mean ± standard error (mean ± SEM). Statistical analyses were performed with GraphPad Prism software (version 9, GraphPad Software, San Diego, CA, USA). Differences among groups were analyzed by one-way ANOVA or two-way ANOVA followed by a Tukey post hoc test, and paired or unpaired student 2-tailed t-test was used for two-group comparisons. A p-value <0.05 was defined as statistical significance.

3. Results

3.1. The concentration of β-OHB is decreased in the plasma of AMI patients after PCI

β-OHB is a metabolite that responds to stresses [29,30]. We have measured the levels of β-OHB in the plasmas of STEMI patients at 2 time points, right before PCI at the time of arterial access and 24 h after PCI. The levels of β-OHB in the plasmas were significantly induced during ischemia at 0.468 mmol/L, 1.5 times higher compared with 24 h after PCI at 0.306 mmol/L, which is comparable to the average level of 0.28 mmol/L in normal unstressed populations [31]. This suggests that β-OHB may be induced as a stress response during ischemia. Additionally, the level of β-OHB decreased quickly after PCI, indicating β-OHB is a metabolite that subsides quickly after the relief of ischemia (Fig. 1A). We further analyzed the data; among the 59 patients, 17 patients had increased or unchanged β-OHB levels (Fig. 1B), while 42 patients had decreased β-OHB levels (Fig. 1C) (individual levels in Supplemental Table 1). Interestingly, the 17 patients with increasing β-OHB levels had a relatively lower concentration prior to PCI (0.336 ± 0.029 mM). While the 42 patients with decreasing β-OHB levels had an average 0.522 ± 0.079 mM starting concentration. The level of β-OHB is influenced by multiple factors, including age, food intake, diabetic status, symptoms onset time, and so on [32]. We included patient’s baseline characteristics in Supplemental Table 2. Given that blood concentrations of β-OHB up to 1 mM seem not enough to protect the heart in patients after PCI, it is important to investigate whether β-OHB levels higher than 1 mM may protect against cardiac I/R injury and examine the direct effects of β-OHB on cardiomyocytes.

Fig. 1.

The concentration of β-OHB is decreased in the plasma of AMI patients after PCI. Plasma samples were collected at the time of arterial access (baseline) and 24 h after PCI from 59 AMI patients. A, β-OHB concentration at baseline and after PCI. B, Individual β-OHB concentration of 17 patients with increased or unchanged levels at baseline and after PCI. C, Individual β-OHB concentration of 42 patients with decreased levels at baseline and after PCI. PCI, Percutaneous coronary intervention. P values are from paired t-test (A, B, and C).

3.2. β-OHB reduces infarct size and preserves cardiac function when given at the time of reperfusion in mice

To determine whether β-OHB can alleviate myocardial I/R injury when given at reperfusion, β-OHB was given to C57BL/6 J mice subjected to cardiac I/R surgery. The protocol of the animal study is shown in Fig. 2A. Circulating β-OHB levels were measured at baseline and different time points after injection at reperfusion in mice (Fig. 2B and C). After 45 min of ischemia, circulating β-OHB level increased 2 folds of its baseline level, indicating it is a fast-response metabolite upon heart stress. With one dose of 10 mmol/kg injection, β-OHB reached its peak level at 5.48 ± 0.372 mM 30 min after injection. It returned to baseline level 8 h after injection (Fig. 2B). In 5 mice, the 8 h area under curve (AUC), which means the total amount that comes into circulation after administration, was 11.06 ± 0.59 mmol/L•h, and the 24 h AUC was 19.7 ± 1.72 mmol/L•h (Fig. 2C). To explore whether administration of β-OHB is cardioprotective, TTC staining was used to determine infarct size. Representative images in the two groups are shown in Fig. 2D, in which the part stained in blue color is defined as non-ischemic zone, the part without blue dye is defined as area at risk, and the part in pale or white is defined as infarcts. The area at risk normalized to the total left ventricle area was comparable between the normal saline group and treatment group, indicating similar levels of damage caused by I/R surgery (Fig. 2E). β-OHB significantly reduces infarct normalized by area at risk or left ventricle area by around 50% (Fig. 2F and G). Furthermore, echocardiography showed that β-OHB administration preserves systolic function, as indicated by preserved left ventricle ejection fraction (Fig. 2H and J). Heart rate was not affected in the perioperative period (Fig. 2I). These results show that β-OHB protects the heart from I/R injury when administered at reperfusion at the level of around 4–6 mM within one hour of one dose of IP injection.

Fig. 2.

In mice, β-OHB reduces myocardial infarct size and preserves cardiac function after I/R when given at reperfusion. C57BL6/J mice were subjected to I/R surgery (ischemia 45 min and reperfusion 24 h). Different treatment (normal saline as control or β-OHB) was given via intraperitoneal injection at the time of reperfusion. TTC staining was performed to measure infarct size and echocardiography was used to evaluate cardiac functions. A, Study protocol. B and C, Circulating β-OHB levels in the perioperative period in mice. D, Representative TTC staining images of normal saline and β-OHB group, the black dotted line indicated the infarct size (Bar = 1 mm). E-G, Quantification of area at risk/left ventricle size, infarct size/area at risk, and infarct size/left ventricle size. Data are shown as percentages. (N = 12 in the normal saline group, N = 11 in the β-OHB group.) H, Representative strain images pre- and post-I/R surgery. I, Heart rate pre- and post-I/R surgery. J, Quantification of left ventricular ejection fraction. (N = 7 in each group.). (NS, normal saline. β-OHB, β-hydroxybutyrate. ECHO, echocardiography. I/R, ischemia reperfusion.) All data are presented as mean ± SEM. P values are from unpaired t-test (B, C, E, F, and G), or two-way ANOVA (I and J). (** p < 0.01, **** p < 0.0001).

3.3. β-OHB prevents I/R-induced cell death, enhances autophagy, and increases mitochondrial DNA in NRVMs

To investigate whether β-OHB protects the cardiomyocytes directly and to delineate the mechanisms, NRVMs were treated with β-OHB in vitro at the time of reperfusion; measurements of cell death (LDH assay) and viability (MTT assay) were performed after reperfusion. We first did a dose response and found that 4 mM of β-OHB reduced the I/R-induced cell death the most and 8 mM of β-OHB start causing more cell death (Supplemental Fig. 1). Most importantly, 4 mM of β-OHB level can be achieved in physiological conditions [33]. Hence, we chose 4 mM of β-OHB to perform the following experiments. Results showed that β-OHB treatment significantly decreased cell death and preserved cell viability both under normoxia and I/R conditions (Fig. 3A and B). To determine whether these effects were linked to increased autophagy, microtubule-associated protein 1A/1B-light chain 3 (LC3) and p62 protein levels were measured using western blot; LC3 is required in the process of autophagosome maturation, and p62 is an autophagic cargo adaptor, which are used as autophagy markers to show autophagic flux. β-OHB treatment increased the LC3 II levels while decreasing p62 level, indicative of enhanced autophagic flux (Fig. 3C and D). The increased autophagic flux has been verified by using Bafilomycin A1, which prevents the process of autophagosome and lysosome fusion and increased LC3 II further after β-OHB treatment (Fig. 3C and D). Inhibition of mTOR complex I, which is a master regulator of autophagy, is necessary to initiate autophagy [34]. Since HDAC inhibitors block mTOR activity and may induce autophagy through this inhibition and β-OHB is an endogenous HDAC inhibitor, we assayed the mTOR activity with β-OHB treatment [28,35]. We found that β-OHB significantly reduced the phosphorylation of mTOR both under normoxia and I/R conditions (Fig. 3E and F). Ribosomal protein S6 is a downstream target of mTOR, and the phosphorylation level of S6 is an indicator of mTOR activity, which regulates autophagy [36]. β-OHB treatment decreased the phosphorylation level of S6 (pS6) (Fig. 3E and F). Eukaryotic translation initiation factor 4E-binding protein 1 (4EBP1) is another characterized downstream target of mTOR, which can regulate mRNA translation [37]. β-OHB treatment also reduced mTOR-4EBP1 signaling, indicating inhibition of mTOR activity (Fig. 3E and F). AMP-activated protein kinase (AMPK) also participates in regulating autophagy in multiple pathways, including inhibiting mTOR and activating ULK1 complex activity in several sites [38,39]. β-OHB given at reperfusion increased the phosphorylation of AMPK and ULK1 (Fig. 3E and F). These results demonstrate that β-OHB activates autophagy via inhibiting mTOR activity.

Fig. 3.

In NRVMs, β-OHB prevents I/R-induced cell death, enhances autophagy, and increases mitochondrial DNA. NRVMs were subjected to simulated I/R and treated with PBS or β-OHB. MtDNA was isolated for qPCR, and protein was isolated for western blots. LDH and MTT assays were performed to detect cell death and cell viability. A, Cell death assay using LDH assay. (N = 4). B, Cell viability assay using MTT assay. (N = 6). C and D, Representative images and quantifications of western blot analysis of LC3 and p62 levels. E and F, Representative images and quantifications of western blot analysis of mTOR, pmTOR, S6, pS6, AMPK, pAMPK, ULK1, pULK1, 4EBP1, and p4EBP1. G, mtDNA copy numbers were analyzed by qRT-PCR using specific ATP6 and COX-II primers. All data are presented as mean ± SEM. P values are from RM one-way ANOVA with the Geisser-Greenhouse correction (A, B, D, F, and G).

I/R injury induces damage to cellular structures and causes irreversible cell death, in which mitochondrion is an important mediator [40]. We measured the mitochondrial DNA content in β-OHB-treated NRVMs. ATP6 and COX-II are two important functional genes in the mitochondrial genome [41]. Again, we did a dose-response, and showed that 4 mM β-OHB treatment in reperfusion increased the mitochondrial DNA levels the most (Supplemental Fig. 2). We used 4 mM of β-OHB to perform the subsequent experiments. β-OHB at 4 mM significantly increases mtDNA copy numbers (Fig. 3G) given at reperfusion, indicating that β-OHB possibly increases the mitochondrial number during I/R injury. These data indicated that β-OHB rescues the cardiomyocytes from I/R injury and maintains mitochondrial mass through inducing autophagic flux by inhibiting the mTOR signaling pathway.

3.4. β-OHB maintains mitochondrial homeostasis, preserves mitochondrial function, and reduces ROS production in NRVMs after simulated I/R

Since mitochondria are the most vulnerable components and sources of ROS generation in I/R injury, increasing the removal of damaged mitochondria by activating autophagy helps to reduce the injury [13,28,42]. To test whether β-OHB treatment can preserve mitochondrial homeostasis in I/R, we measured the mitochondrial homeostasis and function in NRVMs using TMRM (detection of mitochondrial membrane potential), MitoTracker (mitochondria mass), H2DCFDA (total ROS produced in the cells) and Mito-SOX (ROS produced specifically in mitochondria). β-OHB treatment at 4 mM markedly preserved mitochondrial membrane potential and mitochondrial abundance after I/R (Fig. 4A and 4B), consistent with our previous results showing increased mtDNA copy number with β-OHB treatment (Fig. 3G). Notably, β-OHB significantly reduced ROS production both in the cells and mitochondria (Fig. 4A and 4B). During I/R injury, mtDNA is easily damaged and forms fragments, resulting in impaired mitochondrial function [43]. The ratio of intact mtDNA (16.2 kb) and fragmented mtDNA (0.22 kb) indicates mtDNA damage level. When there is less mitochondrial damage, the intact mtDNA level is higher. β-OHB treatment significantly increased the ratio (Fig. 4C and 4D), showing reduced mtDNA damage. Seahorse assay is a sensitive tool to evaluate mitochondrial functions and we measured the mitochondrial function in NRVM subjected to I/R injury. β-OHB significantly increased basal OCR under normoxia and after I/R, ATP-linked OCR under normoxia condition (there is a trend to increase ATP-linked OCR after I/R), proton leak and non-mitochondrial respiration (Fig. 4E), while reduced the reserve capacity under normoxia. There was no significant difference in the maximal respiration level (Fig. 4E). These data indicate that β-OHB increased the basal OCR and ATP production. Through these effects of increased ATP production and autophagy, β-OHB preserves mitochondrial homeostasis in cardiomyocytes effectively after I/R, leading to significantly reduced cell death.

Fig. 4.

β-OHB maintains mitochondrial homeostasis and preserves mitochondrial function in NRVMs after I/R. NRVMs were subjected to simulated I/R, and stained with TMRM, mitoTracker, H2DCFDA, and MitoSOX. Images were taken with a fluorescent microscope under a fixed exposure time. Primers for intact and short mtDNA were used for semi-quant PCR. After I/R, NRVMs OCR was measured by sequentially adding oligomycin, FCCP, and antimycin A. A and B, Representative images of mitochondrial staining and quantification. C and D, Representative images of intact and fragmented mtDNA PCR product gel and quantification. E, Representative images of seahorse assay and OCR calculation. All data are presented as mean ± SEM. P values are from RM one-way ANOVA with the Geisser-Greenhouse correction (B, D, and E).

3.5. β-OHB promotes autophagic flux and increases mtDNA contents in mouse myocardium when given at the time of reperfusion

After showing that β-OHB can induce autophagy and promote mitochondrial homeostasis in cardiomyocytes, we measured the effects of β-OHB given at the time of reperfusion on autophagic flux and mitochondrial homeostasis in mouse hearts subjected to I/R surgery. After ischemia (45 min) and reperfusion (24 h), the heart was harvested and divided into three parts: ischemia zone, border zone, and remote zone. In the border zone, β-OHB induced LC3 II and lowered p62 protein levels, indicating increased autophagic flux (Fig. 5A and 5B). To better measure β-OHB induced autophagic flux in the myocardium, we used the autophagic flux reporter CAG-RFP-GFP-LC3 transgenic mice [23]. Autophagosomes are marked as yellow puncta by the combination of GFP and RFP fluorescence. While autolysosomes are marked as red puncta as the low pH in lysosomes quenches GFP fluorescence. The increase of both autophagosomes and autolysosomes indicates increased autophagic flux. With β-OHB treatment, the number of autophagosomes and autolysosomes increased significantly in the myocardium (Supplemental Fig. 3), indicating activation of autophagic flux. The copy numbers of three mtDNA encoded genes, ATP6, ND1, and ND6, were increased significantly in the border zone by β-OHB treatment, while there is no significant difference in the ischemia and remote zones compared with normal saline control (Fig. 5C). Therefore, when given at reperfusion, β-OHB significantly enhanced autophagic flux and increased mtDNA content in the border zone of the mouse hearts subjected to I/R injury, which may explain β-OHB’s cardioprotective effects.

Fig. 5.

β-OHB promotes autophagic flux and increases mtDNA contents in mouse myocardium when given at the time of reperfusion. C57Bl6/J mice were subjected to I/R surgery, with normal saline or β-OHB given at reperfusion. The harvested heart was divided into ischemia, border, and remote zones. CAG-RFP-GFP-LC3 mice were given normal saline or β-OHB. After 3 h, the heart was harvested and processed to detect autophagic flux. A and B, Representative images of western blot of LC3 and p62 levels in the IZ, BZ, and RZ and quantifications of BZ. C, ATP6, ND1, and ND6 mtDNA levels in the IZ, BZ, and RZ measured by qPCR. (IZ, ischemia zone. BZ, border zone. RZ, remote zone.) All data are presented as mean ± SEM. P values are from unpaired t-test (B and C).

3.6. The cardioprotective effect of β-OHB during I/R injury is dependent on autophagy

ATG7 is an essential autophagy-related gene that expands autophagosome and facilitates ATG8 lipidation through its E1-like enzymatic activity [44]. To evaluate whether the cardioprotective effect of β-OHB is dependent on autophagic flux, NRVMs were transfected with specific siRNA to knock down ATG7. The ATG7 protein expression level was significantly reduced compared with the control group by around 60% (Fig. 6A and B). In the cardiomyocytes treated with control siRNA, β-OHB increased LC3 II both at normoxia and after I/R. In contrast, cardiomyocytes treated with ATG7 siRNA plus β-OHB failed to induce LC3 II (Fig. 6A and B), consistent with the concept that β-OHB induced autophagy is dependent on ATG7. Similarly, ATG7 siRNA treatment blocked the β-OHB associated increases in mtDNA levels under both normoxia and I/R conditions (Fig. 6C), as well as abolishing the protective effect of β-OHB on cell death measured both by LDH assay and MTT assay (Fig. 6D and E). Collectively, these results suggest that the protective effects of β-OHB on mitochondrial homeostasis and cell death are dependent on autophagy in cardiomyocytes.

Fig. 6.

The cardioprotective effect of β-OHB during I/R injury is dependent on autophagy. NRVMs were transfected with ATG7 specific siRNA and subjected to I/R. A and B, Representative western blot and quantification of ATG7 (upper panel) and LC3 II (lower panel). C, mtDNA levels were measured by qRT-PCR using specific ATP6 (upper panel) and COX-II (lower panel) primers. D, Cell death assay using LDH assay. E, Cell viability assay using MTT assay. All data are presented as mean ± SEM. P values are from unpaired t-test (B upper), or RM one-way ANOVA with the Geisser-Greenhouse correction (B lower, C, D, and E).

3.7. The cardioprotective effect of β-OHB is dependent on autophagy in mice

Cardiomyocyte-specific ATG7 knockout mice were used to determine whether the cardioprotective effect of β-OHB is via activating autophagy. With 5 days of tamoxifen injection and two days of washout, we successfully generated inducible cardiomyocyte-specific ATG7 knockout mice using αMHC-merCremer mice (Fig. 7A and B). Some inducible cardiomyocyte-specific ATG7 knockout mice have reduced left ventricular ejection fraction (est 20–30%), but they tolerated I/R surgery well as reported before [45]. In the wild-type control mice with tamoxifen injection, β-OHB significantly reduced infarct size when given at the time of reperfusion and partially preserved the systolic function. In contrast, in the ATG7 knockout mice, ATG7 deficiency abolished β-OHB’s effect on reducing infarct size (Fig. 7C-7F) and preserving cardiac function (Fig. 7G and H), illustrating β-OHB’s cardioprotective effect is via autophagy.

Fig. 7.

The cardioprotective effect of β-OHB is dependent on autophagy in mice. A and B, Representative images and quantifications of western blot analysis of ATG7 expression in mouse hearts. C, Representative TTC staining images of WT and ATG7 cKO mice treated with normal saline or β-OHB, the black dotted line indicated the infarct size (Bar = 1 mm). D–F, Quantification of area at risk/left ventricle size, infarct size/area at risk, and infarct size/left ventricle size. Data are shown as percentages. G, Representative strain images pre- and post-I/R surgery. H, Quantification of left ventricular ejection fraction. All data are presented as mean ± SEM. P values are from unpaired t-test (B), or two-way ANOVA (D, E, F, and H).

In summary, these results revealed β-OHB can reduce infarct size and preserve cardiac functions by improving mitochondrial homeostasis when administered at reperfusion. Additionally, these cardioprotective effects are achieved through activating autophagy via mTOR inhibition.

4. Discussions

The cardioprotective effect of β-OHB given before ischemia has been previously reported; increasing the levels of β-OHB by fasting or IP injection before I/R surgery can significantly limit the infarct size in rodents [15,16]. In addition, given β-OHB continuously by osmotic pumps for 24 h at reperfusion can protect the heart from I/R injury [17]. The studies herein report that one convenient dose β-OHB treatment at the time of reperfusion has similar protective effects, and thus, has translational applications in the clinic with a single dose of medication. Importantly, our study highlights the high efficacy of β-OHB treatment during reperfusion (reduce infarct size by around 50% at a clinically realistic time point), and that its cardioprotective effects are via enhanced autophagy and maintenance of mitochondrial homeostasis. Furthermore, we have shown these protective effects are through direct effect on cardiomyocytes using cardiomyocyte tissue culture and cardiomyocyte-specific ATG7 knockout mice. As shown in Fig. 1, β-OHB levels increase in response to cardiac ischemia in STEMI patients and return to baseline levels 24 h after the revascularization. However, the levels of β-OHB induced by cardiac ischemia do not appear to be sufficient to provide cardiac protection (0.5 mM in STEMI patients versus 4–6 mM of plasma level in mice); herein, we show that a single intraperitoneal dose of β-OHB given after 45 min ischemia, at the time of reperfusion significantly reduced infarct size and preserved systolic function in mice. Similarly, simulated in vitro I/R experiments using NRVMs with 4 mM β-OHB treatment also showed reduced cell death and increased cell viability, consistent with the protective effects of β-OHB in mice.

In this study, we showed that β-OHB activates autophagic flux. HDACs have important functions in regulating the expression of lots of genes by their deacetylation activity [32]. Studies have shown that using small molecule compounds to inhibit the activity of HDACs exert beneficial effects during I/R injury [21]. β-OHB is a signaling molecule affecting multiple pathways and is an endogenous HDAC inhibitor [19]. Thus, it’s possible that β-OHB can protect the heart from I/R injury, similar to HDAC inhibitors. One mechanism by which HDAC inhibitors protect the heart from I/R injury is through increasing autophagy [23]. In this paper, we demonstrated that β-OHB inhibits the activity of mTOR, thus activating autophagy.

We also found in NRVMs, when β-OHB is given at reperfusion, the mitochondrial membrane potential and mitochondrial mass both improved under normoxia and I/R conditions. In addition, the ROS generated in the cells and mitochondria was significantly reduced, indicating less mitochondrial damage [46]. As a part of general autophagy, mitophagy is a selective degradation of damaged mitochondria by autophagy, which is essential in regulating cell energy production and mitochondrial quality control [47]. β-OHB contributes to maintaining mitochondrial homeostasis, at least partially depending on increasing mitophagy to remove damaged mitochondria during I/R injury. It has been shown that increasing autophagy by Tat-Beclin is cardioprotective during I/R injury [28]. However, whether increasing mitophagy has similar effects still needs to be tested.

To assess the mitochondrial respiratory function, we performed Seahorse assays in NRVMs after simulated I/R. β-OHB increases the basal and ATP-linked OCR at normoxia. The increased basal and ATP-linked OCR indicates ATP production was increased in the cells after β-OHB treatment at normoxia [48]. After I/R treatment, β-OHB increased basal OCR, and β-OHB has a trend to increase ATP-linked OCR, indicating less damaged mitochondria. We also found the proton leak is increased with β-OHB treatment, which may be due to uncoupling or more permeable mitochondrial inner membrane [48]. However, we showed that the mitochondrial membrane potential is preserved with β-OHB. Thus, the increase in proton leak is possibly due to the increased uncoupling protein activity. The increased proton leak and non-mitochondrial respiration by β-OHB is very intriguing since it showed that β-OHB may increase ATP production while not stressing damaged mitochondria during I/R. These data support the oxidative role of β-OHB in addition to its role of regulating autophagy, which makes β-OHB a better therapeutic than other drugs that only regulate a specific pathway.

In this study, we also tested the optimal concentration of β-OHB for cardiomyocyte protection during I/R injury. Although β-OHB is a cardioprotective metabolite, there is likely a threshold of its level that, if surpassed, would induce damage instead of protection. In physiologic conditions, the circulating ketone level is under 6–8 mM even after prolonged fasting and vigorous exercise. While in pathological conditions, such as diabetic ketoacidosis, it can accumulate up to 20 mM [32]. In our study, 4 mM concentration of β-OHB used at reperfusion can effectively activate autophagic flux and prevent cell death. The beneficial effect of β-OHB on mitochondrial DNA diminished at the level of 8 mM. Thus, we propose to keep the blood concentration of β-OHB at around 4 mM, and the safe and therapeutic concentration of the β-OHB can be easily achieved by monitoring β-OHB blood level in patients. Most of the studies focus on chemically synthesized small molecules in treating AMI [46]. Compared to them, β-OHB can be easily delivered safely, and it has been tested in human heart failure patients [49]. Therefore, β-OHB is a very feasible therapeutic to treat AMI patients at the time of PCI during revascularization.

There are several limitations of this study. First, inflammation plays an important role in the development of myocardial ischemia [50]. Upon reperfusion, a cascade of inflammatory signals is activated by restoring blood flow. These include the activation of pattern recognition receptors as well as the Nod-, LRR- and pyrin domain-like receptors (NLRs) [51], which play essential roles in many cardiovascular diseases [52]. Pharmacological inhibition of NLPR3 inflammasome activation can ameliorate myocardial remodeling in AMI [53]. It has also been reported that β-OHB can reduce inflammation by attenuating the NLRP3 inflammasome formation [54,55]. Since the block of autophagic flux abolished the protective effects of β-OHB, the inflammation process might be downstream of β-OHB-induced autophagy. However, we did not evaluate the inflammation level in our study. Further studies may assess whether β-OHB can attenuate the inflammatory responses during I/R injury and whether it is dependent on autophagy. Recently, ketogenesis by the increase of HMGCS2 (rate limiting enzyme for ketone generation) results in cell dedifferentiation and proliferation of cardiomyocytes and reduced infarct size [56]. Due to the scope of this paper, we did not check the cardiomyocyte dedifferentiation and proliferation. Another limitation of our study is that we did not assess the acetylation levels after I/R injury. It is reported that acetylated proteins within the mitochondria are increased in response to I/R injury [57], and β-OHB can reduce acetylation levels of autophagy-related proteins by decreasing acetyl Co-A pool to activate autophagy in HFpEF mouse models [55]. It is possible that β-OHB can maintain mitochondrial homeostasis by regulating mitochondrial proteins acetylation. This needs to be further investigated in future studies.

5. Conclusions

β-OHB reduces infarct size administered at reperfusion through increasing autophagic flux and maintaining mitochondrial homeostasis by inhibiting the mTOR pathway. Since β-OHB has been safely tested in heart failure patients, it may be a viable therapeutic to reduce infarct size in STEMI patients.

Abbreviations:

STEMI

ST-elevation myocardial infarction

PCI

Percutaneous coronary intervention

I/R

Ischemia-reperfusion

AMI

Acute myocardial infarction

mtDNA

Mitochondrial DNA

ROS

Reactive oxygen species

β-OHB

β-hydroxybutyrate

HDAC

Histone deacetylase

LVEF

Left ventricular ejection fraction

ATG7

Autophagy-related gene 7

NRVMs

Neonatal rat ventricular myocyte

OCR

Oxygen consumption rate

mTOR

mammalian target of rapamycin