1088

The antibiotic doxycycline compromises cardiac mitochondrial and contractile function in diabetic mice
Rob CI Wust1, Bram F Coolen1, Vida Alizadeh Tazehkandi2, Mariah Daal1, Riekelt H Houtkooper 2, and Gustav Strijkers1

1Preclinical & Translational MRI, Biomedical Engineering and Physics, Amsterdam UMC, Amsterdam, Netherlands, 2Lab Genetic Metabolic Diseases, Amsterdam UMC, Amsterdam, Netherlands

Synopsis

Given the evolutionary similarities between the bacterial and mitochondrial protein synthesis machinery, we tested the hypothesis that the widely used antibiotic doxycycline reduces mitochondrial function, and results in cardiac contractile dysfunction. Indeed, doxycycline exposure resulted in a dose-dependent reduction in maximal uncoupled mitochondrial respiration in cultured H9C2 cells. Maximal mitochondrial respiration was also reduced in mice treated with doxycycline. MRI of treated mice revealed contractile dysfunction as evidenced by marked diastolic and a mild systolic dysfunction. Moreover, doxycycline exacerbated mitochondrial and contractile dysfunction in animals with type 2 diabetes mellitus.

Introduction

In the heart, a high constant energy utilization rate is coupled to a constant energy production by OXPHOS in mitochondria [1,2], and as such, mitochondria serve a crucial role for optimal cardiac function. Because of the evolutionary similarities between the bacterial and mitochondrial protein synthesis machinery, certain classes of antibiotics, such as the widely used tetracyclines, also inhibit mitochondrial function [3], which in turn can affect cardiac contractile function. The diabetic heart is more vulnerable for contractile dysfunction because of mitochondrial dysfunction [4]. This might become even more apparent in combination with tetracycline antibiotics. In this study, we therefore aimed to investigate whether doxycycline impairs mitochondrial and contractile dysfunction in diabetes.

Methods

To establish the metabolic consequences of doxycycline, a tetracycline antibiotic, we tested the mitochondrial function in cultured H9C2 cells with (10 and 30 µg/ml) and without doxycycline treatment, by using an XFe96 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA). To assess mitochondrial and contractile function in mice, drinking water of male non-diabetic (db/+) and diabetic (db/db) C57BL/Ks mice (10 weeks of age) was supplemented with doxycycline, or amoxicillin as a control antibiotic, for 2.5 weeks. Mitochondrial respiration was assessed by Oroboros O2k high-resolution respirometry [2]. Cardiac function in isoflurane-anaesthetized mice was assessed by cardiac MRI, using a 7.0 Tesla MR Solutions small animal scanner (MR Solutions, Guildford, UK) equipped with a 38-mm-diameter mouse volume coil. For left ventricle (LV) systolic function measurement, we used a cardio-respiratory gated multi-slice short-axis cine MRI acquisition using the following parameters: TR/TE = 7/2.8 ms, flip angle = 20°, FOV = 30×30 mm2, matrix size = 192×192, slice thickness = 1 mm, number of slices = 7, number of averages = 5, total acquisition time = 20 min. To assess diastolic function, a single mid-ventricular short-axis slice was acquired using a high frame rate retrospectively gated cardiac sequence [5] with the following sequence parameters: TR/TE = 7/2.35 ms, flip angle = 15°, FOV = 30×30 mm2, matrix size = 192×192, slice thickness: 1 mm, number of k-space repetitions = 400, acquisition time = 13 min. Off-line reconstruction of the diastolic function measurements was performed in MATLAB 8.1 (The Mathworks, Natick, MA, USA) using custom-built routines. In short, imaging data was binned into 60 frames and reconstructed by compressed sensing algorithms using Berkeley Advanced Reconstruction Toolbox (BART) [5, 6] (see Figure 3 for an example). All images were analyzed using MEDIS software (Leiden, The Netherlands). End-diastolic and end-systolic volumes of the left ventricle were used to calculate ejection fraction (EF) as a measure for systolic function. Diastolic function was measured as the ratio of the early or elastic (E’) and atrial (A’) filling rates.

Results and Discussion

Doxycycline exposure resulted in a dose-dependent reduction in mitochondrially-encoded OXPHOS subunits with unaltered expression of nuclear-encoded subunits (Figure 1). Furthermore, doxycycline treatment showed lower basal mitochondrial and maximal uncoupled respiration in H9C2 cardiomyoblasts, which was also accompanied by a shift to glycolysis (data not shown). These data suggest that mitochondrial respiration was severely affected in H9C2 cardiomyoblasts treated with doxycycline, likely due to a disturbed balance between mitochondrial-encoded and nuclear-encoded mitochondrial subcomplexes. Maximal NADH (complex I)-linked mitochondrial respiration in the mouse heart was lower in doxycycline-treated control mice (p<0.001, Figure 2), and was more severely reduced in diabetic (db/db) mice treated with doxycycline (interaction effect: p=0.049). Maximal mitochondrial respiration was 26±4% lower in doxycycline-treated animals (p<0.001). A typical example of the self-gated retrospectively triggered mouse CINE MRI is given in Figure 3. Cardiac MRI confirmed that the untreated db/db animals suffered from heart failure with preserved ejection fraction (HFpEF), as the E’/A’ ratio was lower and EF similar compared to healthy controls. Cardiac MRI revealed a small, but significantly lower EF in both groups of doxycycline-treated animals. This reduction tended to be larger in db/db mice treated with doxycycline (Figure 4). Diastolic dysfunction in both groups treated with doxycycline was evident from the lower E’/A’ ratio compared to control (p=0.001). This reduction was solely due to lower filling rate during the E-phase (p<0.001).

Conclusion

Doxycycline antibiotics reduce mitochondrial function in cardiac cells and mouse hearts in vivo. As a result, mice treated with doxycycline exhibited marked diastolic and mild systolic dysfunction. This negative effect of doxycycline on cardiac mitochondrial function was exacerbated in diabetic animals, resulting in a more severely affected cardiac contractile function.

Acknowledgements

This work was in part supported by a grant from Amsterdam Cardiovascular Sciences and NWO Open Technologieprogramma MUSICIAN.

References

1. Wüst RCI, Helmes M, Stienen GJM. Rapid changes in NADH and flavin autofluorescence in rat cardiac trabeculae reveal large mitochondrial complex II reserve capacity. J Physiol. 2015;593(8):1829-1840.

2.Wüst RCI, de Vries HJ, Wintjes LT, Rodenburg RJ, Niessen HW, Stienen GJ. Mitochondrial complex I dysfunction and altered NAD(P)H kinetics in rat myocardium in cardiac right ventricular hypertrophy and failure. Cardiovasc Res. 2016;111(4):362-372.

3. Moullan N, Mouchiroud L, Wang X, et al. Tetracyclines Disturb Mitochondrial Function across Eukaryotic Models: A Call for Caution in Biomedical Research. Cell Rep. 2015;10(10):1681-1691.

4. Smith RL, Soeters MR, Wüst RCI, Houtkooper RH. Metabolic Flexibility as an Adaptation to Energy Resources and Requirements in Health and Disease. Endocr Rev. 2018;39(4):489-517.

5. Motaal AG, Coolen BF, Abdurrachim D, et al. Accelerated high-frame-rate mouse heart cine-MRI using compressed sensing reconstruction. NMR Biomed. 2013;26(4):451-457.

6. Uecker M, Ong F, Tamir J, et al. Berkeley Advanced Reconstruction Toolbox. Proc Intl Soc Mag Reson Med. 2015;23:2486.

Figures

Figure 1. Mitonuclear protein imbalance and reduced respiration in H9C2 cardiomyoblasts exposed to doxycycline. H9C2 cardiomyoblasts were exposed to doxycycline at various concentration: 0, 10 or 30 μg/ml for 72-96 hrs. A-C, Western blot showing protein expression of mitochondrially-encoded cytochrome c oxidase I (MT-CO1) and nuclear-encoded succinate dehydrogenase complex, subunit A (SDHA). D, TOM20 protein content was not significantly different between conditions. E, Typical Seahorse run. F-G, Routine and maximal respiration was reduced in doxycycline-treated cells. H, Oxygen consumption used for ATP production (Basal – Leak respiration) was lower in doxycycline-treated cells. Bar graphs are expressed as mean+SEM. *: p<0.05, one-way ANOVA.

Figure 2. Impaired mitochondrial respiration in hearts of healthy and diabetic animals treated with doxycycline. A, leak respiration did not differ. B, NADH (complex I)-linked oxidative phosphorylation (OXPHOS) was lower after doxycycline in normal animals, but to a larger extent in db/db animals. C, Lower maximal OXPHOS with complex I (NADH-linked) and complex II (succinate) substrates in doxycycline-treated animals. D, Maximal succinate-driven (complex II) respiration tended to be lower after doxycycline treatment. E-F, Mild complex I dysfunction was observed by normalizing respiration to maximal respiration (flux control ratio; FCR). *: p<0.05 vs. amoxicillin; †: p=0.06 vs. amoxicillin; +: interaction (diabetes x antibiotic)-effect p<0.05, two-way ANOVA.

Figure 3. Self-gated retrospectively triggered mouse CINE MRI. A GIF file of one mid-ventricular 1 mm thick SA slice was reconstructed using a high frame rate retrospectively gated cardiac sequence (see text for details). Data was binned into 60 frames and reconstructed using custom-made scripts.

Figure 4. Cardiac contractile dysfunction in animals treated with doxycycline, which worsened in animals with pre-existing diabetes (db/db). A, two representative examples. EDV: end-diastolic volume; ESV: end-systolic volume; E’: filling rate during elastic phase during diastole; A’: filling rate during atrial kick phase during diastole. B, Ejection fraction (EF) was significantly lower after doxycycline treatment, but only marginally in control animals, with low EF seen in some diabetic animals. C-E, Diastolic dysfunction (lower E’/A’) was observed in control and diabetic animals after doxycycline treatment due to lower E’, suggestive of a stiffer ventricle after doxycycline treatment. *: p<0.05 vs. amoxicillin; #: p<0.05 vs. control, two-way ANOVA.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1088