In vivo dynamic nuclear polarization magnetic resonance imaging reveals cardiac mitochondrial redox imbalance as an early indicator of heart failure

TL;DR

A new MRI technique detects early heart failure by tracking mitochondrial redox imbalances in living mice.

Contribution

In vivo DNP-MRI is introduced as a noninvasive method to detect early mitochondrial dysfunction in heart failure.

Findings

DNP-MRI detected accelerated CmP reduction in DOX-treated mice within 30 minutes of drug administration.

No significant changes were observed in epirubicin-treated mice compared to controls.

DNP-MRI can visualize mitochondrial redox imbalance before conventional functional changes appear.

Abstract

Heart failure is a major cause of mortality worldwide. Accumulating evidence indicates that mitochondrial dysfunction, particularly excessive generation of reactive oxygen species (ROS) from the mitochondrial electron transport chain (ETC), plays a vital role in the onset and progression of heart failure. Importantly, mitochondrial dysfunction is believed to emerge at an early stage of heart failure development. However, due to the lack of noninvasive techniques to directly evaluate cardiac mitochondrial function in vivo, the timing and dynamics of mitochondrial functional alterations during the early phase of heart failure development remain unclear. Carbamoyl-PROXYL (CmP) is a membrane-permeable nitroxyl probe that mediates redox reactions within the mitochondrial ETC in the presence of reduced nicotinamide adenine dinucleotide, thereby sensitively indicating mitochondrial electron transfer dynamics. We applied in vivo dynamic nuclear polarization magnetic resonance imaging (DNP-MRI) to a mouse model of doxorubicin (DOX)-induced heart failure to validate its utility. In DOX-treated mice, the CmP reduction rate was significantly accelerated as early as 30 min after drug administration. However, no significant change was detected in epirubicin-treated mice compared with control animals. Considering that DOX induces ROS production through redox cycling at mitochondrial ETC complex I, these results demonstrate that in vivo DNP-MRI enables noninvasive visualization of ETC-associated mitochondrial redox imbalance in the living heart immediately after the onset of cardiotoxic stress, even when ROS generation has just begun and conventional functional changes are unapparent. Therefore, in vivo DNP-MRI represents a powerful noninvasive modality for the early diagnosis of heart failure.

Image 1

Figures

Fig. 5Spatial and temporal changes in cardiac pyruvate metabolism detected using ^13^C-MRS at 24 h after DOX administration. (A) T_1_-weighted magnetic resonance transversal images of the heart are shown. (B)–(D) Chemical shift images (*upper*) were derived from ^13^C-MRS and merged with T_1_-weighted images to confirm the anatomical structure.Fig. 5

Metabolism of CmP by ROS.(A) DOX is converted into a semiquinone radical through the mitochondrial ETC using reduced nicotinamide adenine dinucleotide phosphate (NAD(P)H). The semiquinone radical reduces molecular oxygen to superoxide (O_2_^−^). Superoxide is subsequently converted into hydrogen peroxide (H_2_O_2_) by superoxide dismutase (SOD), and hydrogen peroxide is further converted into hydroxyl radicals (•OH) through the Fenton reaction. (B) The radical form of CmP is normally reduced to hydroxylamine by NADH in mitochondria. However, it is also converted to an oxoammonium cation in the

Full text

Non-standard abbreviations and acronyms

ALTAlanine transaminaseCmPCarbamoyl-PROXYLCSIChemical shift imagingCTComputed tomographyDNPDynamic nuclear polarizationEPRElectron paramagnetic resonanceEPRIElectron paramagnetic resonance imagingFOVField of viewHEHematoxylin and eosinMRIMagnetic resonance imagingMRSMagnetic resonance spectroscopyROIRegion of interestROSReactive oxygen speciesTEMTransmission electron microscopy

Introduction

1

Heart failure is a condition in which the heart cannot pump sufficient blood to meet the demands of peripheral organs. Unfortunately, the prevalence of heart failure is increasing with the aging population, with estimates showing that over 64 million patients worldwide suffered from heart failure in 2017 [1]. Early detection of heart failure can have a significant impact on healthcare economics and patient prognosis [2]. Echocardiography has been frequently used for the imaging evaluation of cardiac function. This Doppler-based noninvasive modality allows assessment of cardiac morphology, motion, and valvular disease. However, examiner proficiency is essential, as interpretation can be complex and variable. Importantly, abnormalities in cardiac function are often difficult to detect until actual changes in cardiac motion or symptoms become apparent. Coronary computed tomography (CT) can be useful for detecting coronary artery disease without catheter-based angiography [3,4]. However, despite being minimally invasive, it involves radiation exposure, carries the risk of contrast-induced allergic reactions, and has limited ability to evaluate cardiac function. Recently, four-dimensional (4D)-flow magnetic resonance imaging (MRI) has gained attention as a noninvasive technique for assessing cardiac hemodynamics [5,6]. However, the number of facilities equipped with this modality remains limited. Furthermore, although 4D-flow MRI can assess the direction and velocity of blood flow, it does not provide direct information about myocardial metabolism or function. Myocardial biopsy is another option for microscopic assessment of myocardial tissue [7]. Nevertheless, this invasive procedure reveals pathological changes only after morphological alterations, such as myocyte degeneration or fibrosis, have already occurred. Although conventional imaging modalities are specialized for capturing cardiac dysfunction and morphological changes, they cannot directly detect causative molecular events such as reactive oxygen species (ROS) production. Electron paramagnetic resonance imaging (EPRI) is highly sensitive and uniquely suited for the direct detection of paramagnetic species and free radicals in vivo [8,9]. However, despite its excellent capability for radical detection, EPRI faces fundamental physical limitations in spatial resolution due to broad electron spin linewidths and rapid relaxation properties. These limitations restrict its ability to achieve high-resolution anatomical imaging and limit its clinical translation [10,11]. Consequently, recognizing early signs of heart failure before symptom onset or a decline in cardiac functions remains difficult. Therefore, we focused on metabolic changes associated with heart failure and aimed to detect them prior to overt functional or morphological deterioration.

Alterations in mitochondrial structure and function play a key role in the pathogenesis of heart failure in humans, and restoration of mitochondrial function has been proposed as a therapeutic strategy [12,13]. Among available animal models, the mouse model of doxorubicin (DOX)-induced heart failure has been widely used [14,15]. Studies have shown that DOX suppresses complex Ⅰ of the electron transport chain (ETC), is converted to a semiquinone radical form, and induces various ROS, including superoxide (O_2_^−^), hydrogen peroxide (H_2_O_2_), and hydroxyl radicals (·OH) [16,17]. Moreover, DOX-induced cardiotoxicity is attenuated in transgenic mice with enhanced gene expression of antioxidants such as catalase [18], superoxide dismutase [19], and thioredoxin-1 [20]. Although multiple mechanisms have been implicated in DOX-induced cardiotoxicity, oxidative stress and mitochondrial dysfunction are considered critical in the early stages of toxicity [21]. Importantly, ROS production and mitochondrial dysfunction are not only involved in DOX-induced cardiotoxicity but also generally contribute to the pathophysiology of heart failure [22,23]. However, due to the lack of noninvasive techniques to directly assess cardiac mitochondrial function in vivo, the timing and dynamics of mitochondrial functional alterations during early heart failure development remain unclear. As physiological conditions in living organisms, including oxygen partial pressure and enzyme activity, differ substantially from those in vitro, in vivo monitoring of redox alterations in cardiac mitochondria may facilitate early detection of heart failure with potential clinical applications.

Dynamic nuclear polarization (DNP) is a phenomenon in which electron spin polarization is transferred to nuclear polarization by electromagnetic irradiation resonant with electron spins. This process enhances nuclear spin signals, such as those of protons and ^13^C nuclei, during MRI [24,25]. In vivo DNP is a method for inducing the DNP phenomenon at room temperature (Fig. 1A). In this method, molecules containing stable free radicals serve as contrast agents, with nitroxyl probes such as carbamoyl-PROXYL (CmP) commonly used because of their low toxicity [26] and membrane permeability. After administration to mice, CmP is rapidly reduced to its hydroxylamine form via the mitochondrial ETC in the presence of reduced nicotinamide adenine dinucleotide (NADH). The hydroxylamine form lacks free radicals and therefore no longer provides contrast. Accordingly, ETC-associated mitochondrial redox imbalance can be monitored dynamically using in vivo DNP-MRI [[27], [28], [29]]. Compared with EPRI, in vivo DNP-MRI enables imaging with the spatial and temporal resolution of MRI [30,31]. Another approach is dissolution DNP, which induces polarization in vitro at cryogenic temperatures (Fig. 1B). Hyperpolarized ^13^C-pyruvate prepared using the dissolution DNP system can be injected into mice and enzymatically converted to ^13^C-lactate (via lactate dehydrogenase [LDH]), ^13^C-alanine (via alanine transaminase [ALT]), and ^13^C-bicarbonate, which can be evaluated using ^13^C-MRS and chemical shift imaging (CSI) [[32], [33], [34], [35], [36]].Fig. 1Experimental modalities for monitoring redox and enzymatic metabolism. (A) Redox imbalance was assessed using CmP as a redox probe. Radical form of CmP is reduced to its hydroxylamine form with NADH through the mitochondrial electron transport chain. CmP was used as a redox contrast agent for in vivo DNP-MRI (left). A phantom image of 2 mM CmP is shown as a typical example; EPR irradiation, required for the DNP phenomenon produced a contrast effect. EPR spectroscopy can quantify the concentration of CmP (right). A EPR spectrum of CmP is shown. (B) Enzymatic metabolism was evaluated using ^13^C-pyruvate. ^13^C-pyruvate is metabolized to ^13^C-lactate by LDH, to alanine by ALT, and bicarbonate by PDH. The magnetic resonance signal from the nuclear spin of ^13^C can be enhanced by hyperpolarization using in vitro DNP polarizer. Hyperpolarized ^13^C pyruvate was administrated to mice to examine the metabolic alteration in vivo. The spectrum shows ^13^C-MRS of ^13^C-pyruvate and ^13^C-lactate in the mouse heart (upper).Fig. 1

Previous studies using in vivo DNP-MRI have shown that DOX alters redox metabolism, with attenuation observed before echocardiographic or clinical symptoms appear. Daily changes in redox metabolism after DOX administration have already been studied, with evidence suggesting that redox metabolism is attenuated before any changes in echocardiographic findings or symptoms [37]. However, early changes in redox status immediately after DOX administration have not been examined. Therefore, the current study aimed to investigate early-phase alterations in redox imbalance using a mouse model of DOX-induced heart failure. The utility of in vivo DNP-MRI was validated by investigating a commonly used mouse model of DOX-induced heart failure. In parallel, histological changes in the heart were assessed using optical and electron microscopy to correlate metabolic alterations with structural pathology. Epirubicin, another anthracycline with weaker cardiotoxicity, was also investigated for comparison [38]. Furthermore, given that DOX affects mitochondrial function, we examined whether ETC activity was altered in early phase after administration by monitoring oxygen consumption rates in vitro.

Material and methods

2

Chemicals

2.1

3-Carbamoyl-2,2,5,5-tetramethyl-1-pyrrolidine-1-oxyl (carbamoyl-PROXYL:CmP) and doxorubicin (DOX) were obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). Epirubicin was obtained from the Pharmaceutical and Medical Device Regulatory Science Society of Japan (Tokyo, Japan). ^13^C-pyruvate was supplied by Cambridge Isotope Laboratories, Inc. (MA, USA). All other chemicals were of reagent grade and obtained from commercial sources.

Animals

2.2

Female C56BL/6 N mice aged 4–5 weeks and weighing 15–20 g were purchased from Charles River Laboratories Japan, Inc. (Yokohama, Japan). Mice were maintained under a 12-h light/dark cycle with free access to water and food (MF diet, Oriental Yeast Co., Tokyo, Japan). All animal care and experimental procedures were approved by the Gifu University Animal Experiment Ethics Committee and were conducted in accordance with the Gifu University Animal Experiment Guidelines (approval number: AG-P-N-20230086). We also clarified compliance with international guidelines for animal care and use.

Mouse model of DOX-induced heart failure and control experiment

2.3

After dissolving 1.5 mg of DOX or epirubicin in 1.0 mL saline, it was intraperitoneally injected once into mice at a volume of 10 mL/kg body weight. The control group was intraperitoneally injected with the same amount of saline.

Monitoring redox imbalance in the heart using in vivo DNP-MRI

2.4

In vivo redox status monitoring was performed using an in vivo DNP-MRI system (Keller, Japan REDOX Co., Ltd., Japan). The external magnetic field (B_0_) for electron paramagnetic resonance (EPR) irradiation and MRI was fixed at 15 mT. The EPR irradiation and MRI frequencies were 458 MHz and 689 kHz, respectively. For cardiac imaging, a single-turn bent surface coil (length, 20 mm; width, 32 mm) was used. Imaging was performed while mice were made to anesthetize with 1.5%–2.0% isoflurane. In vivo DNP-MRI scanning started immediately after intravenously injecting 300 mM CmP at 5 mL/kg (body weight) and images were acquired every 30 s from 0.5 to 2.5 min. Images were obtained from the mice. One group received DOX before 30 min (n = 5) or 24 h (n = 5), the other group received epirubicin before 30 min (n = 5), and the control group received saline (n = 5). We also confirmed the location of the heart using a 1.5-T animal MRI scanner with the same field of view (FOV) as an in vivo DNP-MRI. Thereafter, the average intensity of the region of interest (ROI) was analyzed using ImageJ software, and decay rates were determined by plotting it in logarithmic form. Moreover, based on the obtained images, a redox map visualizing the decay rate of the ROI at each pixel was generated using a custom Excel macro program. The in vivo DNP-MRI scanning conditions were as follows: power of EPR irradiation = −5 dB, flip angle = 90°, repetition time × echo time × EPR irradiation time (TEPR) = 500 × 20 × 300 ms, slice thickness = 100 mm, including the whole thickness of the mouse; phase-ending steps = 32, and FOV = 32 × 32 mm.

1.5-T MRI of the mouse heart

2.5

To confirm the anatomical location of the heart, T_1_-weighted images were obtained using a 1.5-T MRI scanner with the same FOV as the in vivo DNP-MRI. These coronal images were obtained using a fast spin echo sequence with the following parameters: echo time = 12 ms, repetition time = 500 ms, slices = 6, E factor = 4, and resolution = 0.23 × 0.23 mm^2^.

Fluorescence microscopy of cardiac tissues to confirm DOX and epirubicin accumulation

2.6

After intraperitoneal injection of DOX into mice, the hearts were excised after 30 min (n = 5) and 24 h (n = 5). Similarly, after intraperitoneally injection of epirubicin (n = 5) or saline (n = 5) into mice, the hearts were excised after 30 min. The obtained hearts were then placed in a container, which was then filled with optimal cutting temperature compound (Surgipath FSC22, Leica Biosystems, Germany) to make frozen blocks. To confirm the presence of DOX or epirubicin in the tissue, frozen tissue sections without any staining were observed under a fluorescence microscope (BZ-X 800, KEYENCE, Osaka, Japan) using Texas Red filter given that DOX and epirubicin emit red fluorescence [39].

EPR spectroscopy for the detection of CmP radicals in ex vivo and in vitro studies

2.7

The concentration of free radicals such as CmP can be measured using EPR spectroscopy (Fig. 1A). For ex vivo study, heart homogenates were prepared 30 min after intraperitoneal administration of DOX (n = 7) or saline (n = 7). The homogenates were diluted fourfold with MAS buffer to the weight of the hearts. Thereafter, 48-μL of each solution was mixed with 1.0 μL of distilled water and 1 μL of 1 mM CmP. In another group, the same amount of potassium cyanide (KCN, 500 mM) was added instead of distilled water to inhibit mitochondrial function. To examine the redox reaction of CmP, EPR measurements were conducted every 5 min up to 30 min, and decay rates were calculated. For in vitro studies, hearts were obtained and similarly diluted fourfold with MAS buffer to create the heart solution. For the DOX group (n = 4), the 48 μL solution was mixed with 2.5 μL of DOX, 1 μL of distilled water, and 1.0 μL of 1 mM CmP. For the control group (n = 4), the solution was mixed with 3.5 μL of distilled water and 1.0 μL of 1 mM CmP. For the group with reduced mitochondrial function (each, n = 4), the solution was mixed with 2.5 μL of DOX or distilled water, 1.0 μL of KCN, and 1.0 μL of 1 mM CmP. All reagents were used at the same concentrations as in the ex vivo study. EPR measurements were performed every 2 min for 14 min, and the decay rates were calculated. An X-band EPR spectrometer (JEOL Ltd., Tokyo, Japan) was used for EPR spectroscopy with the following conditions: microwave frequency = 9.4 GHz (336 mT), microwave power = 1 mW, modulation width = 0.06 mT, sweep time = 1 min, sweep width = ±5 mT, and time constant = 0.03 s.

Hyperpolarized 13C-MRS in vivo

2.8

A probe solution containing 22 μL of 11 M^13^C-pyruvate, 15 mM OX063, and 2.5 mM ProHance (Eisai, Tokyo, Japan) was prepared. The solution was hyperpolarized using a dissolution DNP polarizer (SpinAligner, Polarize, Denmark) with microwave irradiation at 93.42 GHz and 100 mW at 1.4 K for 45 min. The hyperpolarized solution was neutralized by mixing it with the dissolution buffer composed of 0.3 mM EDTA and 60 mM NaOH and warmed for immediate intravenous injection. After injection of the hyperpolarized ^13^C-pyruvate (15 μL/g) into the tail vein, ^13^C-MRS (magnetic resonance spectroscopy) was performed using a 1.5-T MRI scanner under 1.5%–2.0% isoflurane anesthesia. During the scanning, the chemical shifts, such as ^13^C-lactate to ^13^C-pyruvate ratio, ^13^C-alanine to ^13^C-pyruvate ratio, and ^13^C-bicarbonate to ^13^C-pyruvate, were examined over time. The mice at 30 min (n = 3) and 24 h (n = 4) after DOX administration and control group (n = 4) were then evaluated. Thereafter, chemical shift images for both the control and 24-h post-DOX administration groups were generated using ^13^C-MRS and merged to confirm the location of the changes.

Histopathology

2.9

After administration of DOX, epirubicin, or saline, mouse hearts were excised at the indicated time points. Heart samples were collected 30 min (n = 4) and 24 h (n = 4) after DOX administration, 30 min after epirubicin administration (n = 4), and 30 min after saline administration (n = 4). The hearts were fixed using 10% formalin natural buffer solution and embedded in paraffin blocks. The blocks were then cut into 3-μm sections, which were then stained using hematoxylin and eosin (HE) and Masson's trichrome. The stained tissues were observed using the microscope (BZ-X 800, KEYENCE, Osaka, Japan).

Sample preparation for transmission electron microscopy (TEM)

2.10

After making an incision in the right heart appendages, 2.5% glutaraldehyde was perfused systematically from the left ventricle to fix tissues. The hearts were then removed and stored in 2.5% glutaraldehyde under light shielding at 4 °C. The control (n = 2), 30-min (n = 2), post-DOX administration, and 24 h (n = 2) post-DOX administration groups were prepared.

Measuring oxygen consumption rates in the cardiac mitochondria in vitro

2.11

Oxygen consumption rates in the cardiac mitochondria were measured using an oxygen measuring system (Oxytherm+, Hansatech, UK). After cardiac homogenates were prepared, protein concentrations were measured using absorbance meter (SH-1000Lab, CORONA ELECTRIC Co., Japan). Thereafter, the concentration of protein was adjusted to 3 mg/mL through dilution with MAS buffer. Before measuring the oxygen concentration, 850 μL of MAS buffer, 20 μL of BSA (10%, diluted with MAS buffer), 10 μL of rotenone (200 μM, diluted with dimethyl sulfoxide [DMSO]), and 50 μL of the samples were mixed in order that the protein concentration was adjusted to 3 mg/mL. During monitoring of oxygen concentration, 20 μL of succinate (500 mM, diluted with MAS buffer), 20 μL of ADP (60 mM, diluted with MAS), 10 μL of oligomycin (1 mM, diluted with DMSO), and 20 μL of carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) (400 μM, diluted with DMSO) were added in the same order. Thereafter, the oxygen consumption rate per minute was measured in each phase. The following phases were analyzed: phase I (rotenone and succinate) shows the function of complex II, phase II (rotenone, succinate, and ADP) shows the function of complex V, phase III (rotenone, succinate, ADP, and oligomycin) shows the suppressed state of complex V, and phase IV (rotenone, succinate, ADP, oligomycin, and FCCP) shows the maximum respiratory capacity.

Statistical analysis

2.12

All data are presented as mean ± standard deviation (SD). The number of biological replicates (n) is indicated in the corresponding figure legends. Differences between two groups were analyzed using a two-tailed Student's t-test. A p value of <0.05 or <0.01 was considered statistically significant. Statistical analyses were conducted using GraphPad Prism 9 (Dotmatics, Boston, USA). All experiments were independently repeated at least three times.

Data availability

2.13

All materials, data, and protocols used in this study are available from the corresponding author upon reasonable request.

Results

3

Pharmacokinetic DNP-MR images for control or DOX-treated mice are presented in Fig. 2A. DNP enhancement was maximal at 0.5 min, and the enhanced DNP signal decreased depending on CmP reduction in the heart region (Fig. 2A). The ROI for the heart was selected based on 1.5-T MRI images (Fig. 2B). At 30 min after DOX administration, the heart metabolized CmP the fastest (Fig. 2C). In contrast, the decay rate at 30 min after epirubicin administration did not differ from that in the control group (Supplemental Fig. S1A). The redox map, which visualizes the decay rate in each pixel, clearly showed a spatiotemporal increase in the decay rate at 30 min (Fig. 2D). Moreover, to confirm the localization of the administered DOX in the heart, fluorescence microscopy of frozen sections showed that drug accumulation in the heart was greatest at 30 min and were washed out at 24 h (Fig. 2D). Interestingly, epirubicin also showed similar accumulation in the heart region. Therefore, the redox imbalance is most evident when DOX accumulated in the heart, although epirubicin did not show redox alteration. Moreover, to determine mitochondrial ETC function, oxygen consumption rates were examined in the mice group 30 min after DOX or saline administration (Supplemental Fig. S1B). However, mitochondrial function did not significantly differ between the groups. These data suggest that mitochondrial function was normal at this time point.Fig. 2In vivo DNP-MRI showing enhanced CmP reduction at 30 min after DOX administration. (A) In vivo DNP-MRI was performed in control and DOX-treated groups, and DOX-treated mice were examined at 30 min and 24 h after administration. Images were taken every 0.5 min up to 2.5 min after DOX administration. Representative images with EPR off and a color bar are shown. (B) The 1.5T conventional MRI image and in vivo DNP-MRI image of the same mouse are shown. The region of interest (ROI) for in vivo DNP-MRI at the heart was placed on the heart for quantitative analysis. (C) The average ROI signal intensity was plotted logarithmically, and the decay rate was calculated. Data are presented as means ± SD (n = 5, each group), ∗p < 0.05. (D) Redox maps visualizing the decay rate on a pixel-by-pixel basis (upper). Accumulation of the drugs was also examined in frozen sections through observation with a red fluorescence filter (lower). Images are shown for the control, 30-min post-DOX administration, 24-h post-DOX administration, and 30-min post-epirubicin injection groups.Fig. 2

Quantitative redox evaluation in heart homogenates was examined using X-band EPR spectroscopy. In both ex vivo and in vitro studies, the spectra decreased with time as CmP was decreased. The EPR spectra in the ex vivo study were recorded at 40 s and 30 min after the samples were mixed with 1 mM CmP, with or without KCN. The spectrum tended to decrease more quickly in the DOX-treated group than in the control group, although the difference was not significant (p = 0.069; Fig. 3A). KCN treatment significantly suppressed redox imbalance in both the control (p = 0.047) and DOX-treated (p = 0.029) groups. The EPR spectra in the in vitro study were observed at 40 s and 14 min after mixing with CmP. The CmP was significantly more rapidly metabolized in the DOX-treated heart homogenates than in controls (p < 0.01; Fig. 3B). Moreover, heart homogenates mixed with KCN significantly suppressed reduction rate of CmP compared to both the control (p = 0.039) and DOX groups (p < 0.01). Regarding overall tendency, the results were similar in ex vivo and in vitro studies.Fig. 3EPR spectroscopy revealing quantitatively enhanced CmP metabolism 30 min after DOX administration on both ex vivo and in vitro studies. (A) For the ex vivo experiments, heart homogenates obtained 30 min after saline or DOX administration were used as samples. Each sample was examined with and without the addition of KCN. The EPR spectrum measured at 40 s and 30 min after mixing with CmP are shown (left). The EPR signal decay rates were quantitatively calculated from these spectra (right). Data are presented as means ± standard deviation (SD) (n = 7, each group), ∗p < 0.05. (B) For the in vitro experiments, heart homogenates were mixed with saline, DOX, saline and KCN, or DOX and KCN. The EPR spectra obtained at 40 s and 14 min (left) are shown, and the decay rates were calculated in the same manner as that in the ex vivo study. Data present as means ± SD (n = 4, each group), ∗∗p < 0.01.Fig. 3

We hypothesized that energy metabolism would change due to mitochondrial dysfunction caused by DOX. In vivo study using dissolution DNP-MRS showed hyperpolarized MR signals from ^13^C-pyruvate to ^13^C-lactate, ^13^C-hydrate pyruvate, ^13^C-alanine, and ^13^C-bicarbonate. The representative hyperpolarized MR spectra and time courses are shown in Fig. 4A. The chemical shift in hyperpolarized pyruvate, lactate, alanine, hydrate pyruvate, and bicarbonate resonated at 170, 183, 177, 179, and 161 ppm, respectively [40]. The hydrate pyruvate is metabolically inactive [41]. Thereafter, the metabolite ratio of lactate to pyruvate, alanine to pyruvate, and bicarbonate to pyruvate were calculated using dynamic hyperpolarized ^13^C spectra. Given that the pyruvate is metabolized to lactate via LDH, to alanine via ALT, and to bicarbonate via pyruvate dehydrogenase (PDH), the ratio shows enzyme activity, respectively. DOX treatment significantly decreased all lactate/pyruvate, alanine/pyruvate, and bicarbonate/pyruvate ratios (Fig. 4B–D). CSI showed the spatial distribution of each ^13^C molecule within the heart region (Fig. 5A–D). Maps of ^13^C-pyruvate, ^13^C-lactate, and ^13^C-bicarbonate clearly demonstrated that DOX treatment significantly decreased ^13^C metabolites within the heart.Fig. 4In vivo^13^C-MRS showing changes in the enzymatic metabolism at 30 min and 24 h after DOX administration. For in vivo studies, hyperpolarized ^13^C-pyruvate was intravenously injected into mice at 30 min (n = 3) and 24 h (n = 4) after intraperitoneal administration of saline or DOX. The time course of MR signals according to each metabolite is shown (left). The peaks of ^13^C-pyryvate, ^13^C-lactate, ^13^C-alanine, and ^13^C-bicarbonate were observed at 170, 183, 177, and 179 ppm, respectively, after which temporal changes were analyzed. (B)–(D) Based on ^13^C-MRS data, the ratios of ^13^C-pyruvate/^13^C-lactate, ^13^C-pyruvate/^13^C-alanine, and ^13^C-pyruvate/^13^C-bicarbonate were calculated. Data are presented as means ± SD, ∗p < 0.05, and ∗∗p < 0.01.Fig. 4. Fig. 5Spatial and temporal changes in cardiac pyruvate metabolism detected using ^13^C-MRS at 24 h after DOX administration. (A) T_1_-weighted magnetic resonance transversal images of the heart are shown. (B)–(D) Chemical shift images (upper) were derived from ^13^C-MRS and merged with T_1_-weighted images to confirm the anatomical structure.Fig. 5

HE staining showed no significant changes in cardiomyocytes at 30 min after DOX or epirubicin administration compared with the control group. However, a portion of the cardiomyocytes degenerated within 1 day, with the degenerated cardiomyocytes showing abnormal alignment and vacuolation at high magnification (Fig. 6A). At low magnification, no massive fibrosis or inflammatory cell infiltration was noticeable in all samples. The inflammatory cell infiltration was not evident in all groups. Masson's trichrome staining showed no significant fibrosis in the heart 30 min after DOX, epirubicin, or saline administration. In contrast, thin collagen fibers proliferated between some degenerated cardiomyocytes within 1 day after DOX administration (Fig. 6A).Fig. 6Optical histopathology of DOX-treated hearts showing degenerated cardiomyocytes, and transmission electron microscopy (TEM) histopathology showing the mitochondrial morphological changes in DOX-treated group at 24 h after administration. (A) Cardiac histology of the control, DOX-treated, and epirubicin-treated groups. The DOX-treated group was examined at 30 min and 24 h after administration (n = 4, each group). Representative HE stains are shown at 40 × and 400 × magnification. The area outlined by the yellow dotted line indicates degenerated cardiomyocytes. Representative Masson's trichrome stains are shown at 400 × magnification. Arrowheads indicate proliferated thin collagen fibers. (B) Cardiac TEM images for control and DOX-treated groups. The DOX-treated group was examined at 30 min and 24 h after administration (n = 2, each group). Representative TEM images are shown at 1500 × (upper) and 5000 × (lower) magnification. The indicators highlight the following structures: white arrowheads: mitochondria; blue arrowheads: lipid droplets; red dotted line: RBC; orange dotted line: nucleus; and purple bi-directional arrow: sarcomere.Fig. 6

Cardiac TEM histology of the control and DOX-treated groups are shown in Fig. 6B. The DOX-treated groups were assessed at 30 min and 1 day after administration. The lower limit of magnification was set at 1500 × , whereas the upper limit was set at 5000 × . At 30 min after DOX administration, cardiac mitochondria were irregularly shaped (white arrowhead), and lipid droplets appeared (blue arrowhead). At 1 day, the mitochondria were swollen, and the alignment became obscured compared with that in the control group, which showed regularly arranged mitochondria between cardiac muscle fibers. Moreover, the internal structures of the mitochondria were unclear compared to that in the control group.

Discussion

4

Previous studies have reported that DOX acts on the ETC, where it is converted into a semiquinone form, leading to the generation of ROS [17,42,43] (Fig. 7A). Furthermore, CmP is reduced in mitochondria and also exhibits antioxidant properties [44,45]. When CmP reacts with ROS such as O_2−, •OH, and ROO•, it is oxidized to its oxoammonium cation form. This oxoammonium cation is rapidly reduced to the hydroxylamine form by reduced glutathione (GSH) [46,47]. Neither the oxoammonium cation nor the hydroxylamine induces the DNP phenomenon because they lack a free radical. Based on these findings, we hypothesized that ROS produced by DOX in the early phase rapidly react with CmP (Fig. 7B). This hypothesis explains why the DNP effect observed by in vivo DNP-MRI was attenuated more rapidly in the 30-min post-DOX administration group than in the control group. Our experimental results also showed that a greater amount of DOX was detected in the frozen sections at 30 min than at 24 h after administration and that the decay rate at 24 h, when DOX had been washed out from the tissue, did not differ significantly from that observed under control conditions. Considering that cardiomyocytes are constantly contracting and relaxing, adenosine triphosphate (ATP) consumption is much higher than that in other cells [48], with approximately 90% of their ATP being supplied through oxidative metabolism [49]. Moreover, oxygen consumption rates in cardiac mitochondria did not significantly change 30 min after DOX administration. We also directly examined the concentration of CmP in cardiac homogenates using EPR spectroscopy. Based on the results of in vivo DNP-MRI and fluorescence microscopy findings, cardiac homogenates prepared from hearts excised 30 min after intraperitoneal administration likely contained both ROS and DOX. In homogenates mixed with CmP, free radicals tended to be reduced more rapidly, with this tendency being more evident in the in vitro study. We believe that CmP was reduced to its hydroxylamine form as discussed earlier in relation to in vivo DNP-MRI (Fig. 7B). Notably, addition of KCN markedly suppressed CmP reduction irrespective of DOX treatment, suggesting that CmP reduction reflects mitochondrial function. Blocking mitochondrial ETC may inhibit mitochondrial ROS production [50]. These in vitro results of EPR spectroscopy are consistent with the findings of the in vivo DNP-MRI. Another important observation was that the EPR spectrum in the in vitro study was measured immediately after cardiac homogenates were mixed with DOX. This indicates that DOX-induced ROS are generated earlier than 30 min in cardiomyocytes. This consideration is valid because in vitro experimental results were consistent with those of the ex vivo study.Fig. 7. Metabolism of CmP by ROS.(A) DOX is converted into a semiquinone radical through the mitochondrial ETC using reduced nicotinamide adenine dinucleotide phosphate (NAD(P)H). The semiquinone radical reduces molecular oxygen to superoxide (O_2^−^). Superoxide is subsequently converted into hydrogen peroxide (H_2_O_2_) by superoxide dismutase (SOD), and hydrogen peroxide is further converted into hydroxyl radicals (•OH) through the Fenton reaction. (B) The radical form of CmP is normally reduced to hydroxylamine by NADH in mitochondria. However, it is also converted to an oxoammonium cation in the presence of excessive ROS through DOX treatment. Given that the oxoammonium cation lacks an unpaired electron, it does not induce the DNP phenomenon. It is rapidly reduced to the hydroxylamine form by reductants such as NADH and GSH. Therefore, in the early phase after DOX administration, ROS is responsible for the attenuation of the DNP effect.Fig. 7

Regarding enzymatic reactions, our in vivo ^13^C-MRS studies demonstrated alterations in the ^13^C-pyruvate/bicarbonate ratio 24 h after DOX administration, which were localized to the heart as confirmed by CSI. Similar results have been reported previously and have been mentioned to change in pyruvate dehydrogenase flux [40,51]. The enzyme plays a role in the conversion from pyruvate to acetyl-CoA, which is important for cardiac ATP production [52]. In the clinical fields, the ^13^C-pyruvate/bicarbonate ratio has been investigated as a biomarker for monitoring cardiotoxicity in breast cancer patients receiving DOX [53]. Moreover, both the ^13^C-lactate/pyruvate and ^13^C-alanine/pyruvate ratios were decreased, and CSI confirmed that these metabolic alterations also localized to the heart. These reductions likely reflect alterations in enzyme flux, such as LDH and ALT activity in the myocardium. In contrast, the lactate/pyruvate ratio increased at 30 min after DOX administration. We believe that ^13^C-pyruvate tended to be metabolized to lactate rather than acetyl-CoA, which entered mitochondria, as DOX impaired mitochondrial function and induced oxidative stress.

Histologically, some cardiomyocytes at 24 h after DOX administration exhibited vacuolar degeneration, alignment irregularities, and interstitial fibrosis, which are generally prominent several weeks after DOX injection [37,54]. In contrast, no significant optical microscopic changes were observed at 30 min after DOX or epirubicin administration. Scanning electron microscopy revealed that cardiac mitochondria in the control group were aligned between myofibers and showed a clear matrix. The mitochondria at 24 h after DOX administration appeared swollen, randomly aligned, and irregularly shaped, with an unclear matrix. Similar changes in cardiac mitochondria have been reported even when small doses of DOX are administered weekly to mice [55]. In contrast, the 30-min group showed partially irregularly shaped mitochondria, although these changes were not significant compared with those observed in the 24-h group. These findings indicate that cardiac TEM enables detection of ultrastructural morphological changes at earlier stages than optical microscopy, whereas in vivo DNP-MRI can simultaneously detect metabolic alterations at the same time point. Consequently, we believe that ROS were already present in the heart despite the absence of disruption in the mitochondrial membranes. In contrast, the decay rate in epirubicin-treated hearts did not significantly differ from that in the control group, although the drug accumulated in cardiac tissue 30 min after administration. A recent study reported that CmP reduction was suppressed at 2 or 6 days after DOX administration, attributed to mitochondrial oxidative stress and dysfunction, with a concomitant decrease in left ventricular ejection fraction (LVEF) observed at 6 days [37]. In contrast to these late-phase findings, we observed increased CmP reduction at 30 min after DOX administration. Therefore, our results likely reflect acute DOX-induced ROS generation rather than secondary mitochondrial dysfunction. These findings suggested that early-phase cardiac in vivo DNP-MRI enables noninvasive detection of DOX-induced free radical production before the development of overt cardiac dysfunction, including LVEF decline in clinical settings.

Early ROS production may initially function as a redox signaling mechanism rather than a direct mediator of oxidative damage [56]. At this stage, compensatory antioxidant systems, such as the Nrf2 pathway, may be transiently activated to buffer oxidative stress and maintain redox homeostasis [57]. Consistent with this interpretation, the absence of a decline in oxygen consumption suggests that mitochondrial bioenergetics remain largely preserved during the early phase. Nevertheless, sustained or excessive ROS production may eventually overwhelm these feedback mechanisms, resulting in progressive mitochondrial dysfunction and structural injury [58]. Although such compensatory and signaling processes may occur after ROS generation, the present findings suggest that DOX-induced excessive ROS production disrupts redox homeostasis, causing a redox imbalance that is detectable by in vivo DNP-MRI. The redox alterations detected by in vivo DNP-MRI in this study are consistent with accumulating evidence that mitochondrial dysfunction [59] and impaired mitochondrial quality control (MQC) [60] contribute to the early stages of heart failure. Excessive ROS production, dysregulated mitophagy [61], imbalance of the mitochondrial unfolded protein response (mtUPR) [62], abnormal mitochondrial fission [63] drive inflammation [64], and cardiomyocyte injury [[65], [66], [67], [68]]. In this context, the NDUFS4–SIRT5–DUSP1 axis [69,70], endoplasmic reticulum (ER)-mitochondria stress signaling [71], and stress-responsive cell death programs [72] have emerged as key regulators of mitochondrial homeostasis. Although further validation is required, in vivo DNP-MRI-based redox imaging may represent a noninvasive strategy to evaluate subclinical oxidative and mitochondrial stress during the progression of heart failure.

Mitochondrial redox homeostasis represents a more promising therapeutic target than systemic antioxidant supplementation [73]. The failure of several antioxidant therapies in clinical trials may be attributed to indiscriminate ROS scavenging without subcellular compartment specificity [74]. Conversely, strategies aimed at preserving mitochondrial function and maintaining physiological redox signaling balance—rather than simply eliminating ROS—may provide greater translational potential [75,76]. Our findings suggest that targeting mitochondrial redox homeostasis could provide a clinically relevant approach for early-stage intervention before the onset of overt cardiac dysfunction.

An important limitation of this study is that human low-field in vivo DNP-MRI systems remain under technical development and clinical validation has not yet been established. Moreover, spatial resolution, reproducibility, and correlation with long-term cardiac outcomes require further investigation. Nonetheless, in vivo monitoring of redox imbalance holds substantial clinical significance. Under conventional in vitro conditions, cardiomyocytes are exposed to supraphysiological oxygen concentrations, which may distort oxidative stress responses and partially explain why numerous antioxidant therapies have failed in clinical trials. By enabling real-time, noninvasive evaluation of redox status under physiological oxygen tension, DNP-MRI may provide a more accurate platform for stratifying patients, optimizing mitochondrial-targeted therapies, and monitoring therapeutic response in cardiovascular disease.

Conclusion

5

Our findings indicate that ROS play a critical role in the pharmacological effects of DOX. In vitro EPR spectroscopy demonstrated that DOX generates ROS immediately after interacting with cardiac tissues. This finding suggests that redox metabolic changes occur earlier than enzymatic metabolic alterations. Furthermore, in vivo DNP-MRI enabled noninvasive visualization of ROS in cardiomyocytes prior to the appearance of histopathological changes detectable by optical microscopy or measurable alterations in cardiac mitochondrial function. In addition, ^13^C metabolic imaging revealed that changes in redox imbalance precede enzymatic flux alterations, such as PDH, LDH, and ALT. Taken together, these results suggest that in vivo DNP-MRI can serve as an early diagnostic tool for detecting ROS production in the heart. Given that oxidative stress and mitochondrial dysfunction contribute to the early stages of heart failure beyond DOX-induced cardiotoxicity, this technology may play a powerful role for the early detection of heart failure.

Source of funding

These experiments were supported by JSPS KAKENHI (grant number 22K16135).

CRediT authorship contribution statement

Koki Ichihashi: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft. Fuminori Hyodo: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing. Abdelazim Elsayed Elhelaly: Data curation, Formal analysis, Investigation, Software, Visualization. Hiroyuki Tomita: Data curation, Formal analysis, Methodology, Resources, Supervision. Shoya Shiromizu: Formal analysis, Investigation, Resources. Keita Fujimoto: Data curation, Funding acquisition, Resources. Hirohiko Imai: Formal analysis, Investigation, Methodology. Yoshifumi Noda: Methodology, Resources, Software. Hiroki Kato: Data curation, Investigation, Methodology. Akira Hara: Conceptualization, Methodology, Supervision. Masayuki Matsuo: Conceptualization, Funding acquisition, Resources, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.