Unique morphological characteristics of mitochondrial subtypes in the heart: the effect of ischemia and ischemic preconditioning

Rationale— Three subsets of mitochondria have been described in adult cardiomyocytes intermyofibrillar (IMF), subsarcolemmal (SSM), and perinuclear (PN). They have been shown to differ in physiology, but whether they also vary in morphological characteristics is unknown. Ischemic preconditioning (IPC) is known to prevent mitochondrial dysfunction induced by acute myocardial ischemia/reperfusion injury (IRI), but whether IPC can also modulate mitochondrial morphology is not known. Aims— Morphological characteristics of three different subsets of adult cardiac mitochondria along with the effect of ischemia and IPC on mitochondrial morphology will be investigated. Methods— Mouse hearts were subjected to the following treatments (N=6 for each group): stabilization only, IPC (3x5 min cycles of global ischemia and reperfusion), ischemia only (20 min global ischemia); and IPC and ischemia. Hearts were then processed for electron microscopy and mitochondrial morphology was assessed subsequently. Results— In adult cardiomyocytes, IMF mitochondria were found to be more elongated and less spherical than PN and SSM mitochondria. PN mitochondria were smaller in size when compared to the other two subsets. SSM mitochondria had similar area to IMF mitochondria but their sphericity measures were similar to PN mitochondria. Ischemia was shown to increase the sphericity parameters of all 3 subsets of mitochondria; reduce the length of IMF mitochondria, and increase the size of PN mitochondria. IPC had no effect on mitochondrial morphology either at baseline or after ischemia. Conclusion— The three subsets of mitochondria in the adult heart are morphologically different. IPC does not appear to modulate mitochondrial morphology in adult cardiomyocytes.


INTRODUCTION
There are three distinct sub-populations of mitochondria present in adult cardiomyocytes: Intermyofibrillar (IMF) mitochondria, which reside between the myofibers and mainly provide energy for myocardial contraction; subsarcolemmal (SSM) mitochondria that are localized beneath the plasma membrane and provide energy for ion channel function; and perinuclear (PN) mitochondria that are juxtaposed to the cell nucleus and mainly provide energy for gene transcription 1,2 . These 3 mitochondrial subtypes in the heart have been shown to differ in function in terms of oxidative capacity, calcium uptake and sensitivity to mitochondrial permeability transition (MPTP) opening [3][4][5][6][7] . In spite of the presence of large volume of literature concerning the physiology of cardiac mitochondria, the morphological differences between the 3 mitochondrial subtypes have not been comprehensively investigated in the adult heart. Changes in mitochondrial morphology and function have been shown to be critical to the susceptibility of the heart to acute ischemia/reperfusion injury (IRI), and the cardioprotective phenomenon of ischemic preconditioning (IPC), an intriguing phenomenon in which one or more cycles of brief ischemia and reperfusion render the myocardium tolerant to lethal IRI [8][9][10] . The mechanisms underlying IPC have been extensively investigated and comprise a complex array of signal transduction pathways many of which converge on and preserve the function of mitochondria following acute IRI. A number of prosurvival kinases implicated in IPC signaling such as AKT and PKA have been shown to modulate the mitochondrial function 11,12 . Hence, it is plausible to hypothesize that IPC induces its cardioprotective effect by modulating the morphology of cardiac mitochondria.
In this study, we investigate the morphological differences between the three subtypes of mitochondria in adult murine cardiomyocytes and investigate the changes induced by ischemia. We then investigate the effect of IPC on mitochondrial morphology at baseline and following ischemia.

MATERIALS AND METHODS
Experimental procedures were performed in accordance to the Animals (Scientific Procedures) Act 1986 published by the UK Home Office. Hearts obtained from male 8-10 weeks old C57/BL6 mice were used throughout.

TEM specimen processing and imaging
Hearts were fixed using EM grade 1% paraformaldehyde, 2% glutaraldehyde in 0.1M sodium cacodylate buffer. Sections from left ventricle were obtained from the hearts and were post-fixed using 2% osmium tetraoxide for 2 hours at 4°C. Each segment then underwent sequential dehydration using different concentration of ethanol (25%, 50%, 70%, 90%, 100%) and washed with 1,2-epoxypropane, before being embedded in Epoxy resin (Poly/Bed-21844-1). Ultrathin sections of 70nm were cut using an ultra-microtome (Reichert) and were further stained using lead citrate. Images were acquired using a Jeol 1010 transmission electron microscopes equipped with Gatan Orius and had a pixel dimensions of 5 x 5 nm in both X and Y. For each subtype of mitochondria in each heart, 6 representative cells in their longitudinal orientation were captured in a blinded fashion. In total, 7,301 IMF mitochondria, 5,346 PN mitochondria and 3739 mitochondria were segmented and analyzed. Cells with intact fibers were selected for the analysis and hypercontracted cells as well as cells with myofibrillar disarray were excluded from the analysis.

Image analysis
Three subsets of cardiac mitochondria were segmented and traced using image J (V. 1.47). Shape descriptors of individual mitochondria were measured in 2D including:
"perimeter (the distance surrounding the actual shape of mitochondria in 2D, expressed in μm)";

d.
"Feret's diameter (defined as the longest distance between any two points within the selected mitochondrion, expressed in μm)"; e. "roundness (calculated by the formula: 4 x Area/π x Major axis 2 )"; and f. "circularity (calculated by the formula: 4π x Area/Perimeter 2 ). Mitochondria exhibiting a perfect circular shape have a circularity value close to 1.0 whereas more elongated mitochondria have a circularity value that is closer to 0.0)".

Statistics
All statistical tests were performed using Stata (v.13). Mitochondrial shape descriptors were analyzed using the mixed effects model with random effects accounting for variability among cells and hearts of the nested structure, and the fixed effects of mitochondrial types or different treatments. Mitochondrial Area and AR were logarithmically transformed in order to perform the analysis on normally distributed data. Graphs illustrating modelpredicted means and 95% confidence intervals were made using Microsoft Excel 2010. Box and whiskers plot, illustrating the median and 5% to 95% percentile, were used to show frequency distribution of individual shape descriptors. p value of ≤0.05 was considered significant.

Morphological Differences of Three Subtypes of Cardiac Mitochondria
The precise location of cardiac mitochondria determines their morphological constitution 13 . In mouse cardiac myocytes IMF mitochondria are often rectangular and are extended between the myofibers allowing them to make contact with their two juxtaposed mitochondria (indicated by white arrows in Figure 2) whereas SSM mitochondria reside by the subsarcolemmal region and form clusters that are less organized than the IMF mitochondria (indicated by black arrows in Figure 2).
Similarly, PN mitochondria also form clusters around the nucleus of cardiac myocytes and extend between the myofibers but lack the typical elongated morphology of IMF mitochondria (indicated by yellow arrow and star in Figure 2).
Morphological comparisons of different subsets at ST indicated that the IMF mitochondria are not larger than SSM mitochondria (Figure 3a), however, they have significantly larger perimeter (Figure 3b), which corresponds to their rectangular shape and greater Feret's diameter ( Figure 3c). These facts are also evident from the distribution of shape descriptors of these two individual subtypes (see Supplementary Figure 1).
In addition, the rounder ( Figure 3d) and more circular shape ( Figure 3). Unlike IMF mitochondria, ischemia alone does not change the Feret's diameter of SSM mitochondria whereas hearts receiving IPC before ischemia exhibit a significant reduction of Feret's diameter (Figure 7c). In addition, no significant differences are observed in terms of Feret's diameter between groups receiving ischemia with or without IPC (Figure 7c).
In line with sphericity data of IMF mitochondria, ischemia alone induces a 13% rise in roundness of SSM mitochondria (Figure 7d). This in turn is accompanied by the reduction of AR ( Figure 7f) and an increase in circularity (Figure 7e) thereby indicating a reduction of elongated SSM mitochondria after ischemia. In parallel to the data from IMF mitochondria, IPC given prior to ischemia does not induce any meaningful significant changes in respect to the morphology of SSM mitochondria after ischemia (Figure 7 & Supplementary Figure 3).

PN mitochondria respond differently to ischemia and IPC
Perinuclear mitochondria are the least studied mitochondria in adult cardiac myocytes. EM micrographs of PN mitochondria undergoing IPC-only treatment illustrate an absence of any substantial positional and morphological changes (Figure 8a, 8b). In comparison, PN mitochondria from the ischemia-only group or ischemia group receiving IPC exhibit larger as well as more circular PN mitochondria (Figure 8c and 8d -yellow arrows). Besides, a distinct loss of endoplasmic reticulum structure around the nuclear region is evident in these ischemic groups (Figure 8c, 8d). PN mitochondria do not exhibit morphological alterations after IPC only treatment ( Figure 9 and Supplementary 4). In contrast to the morphometric data from SSM and IMF, an increase in PN mitochondrial area (Figure 9a) can indicate potential swelling of PN mitochondria following ischemia. Besides, IPC treatment given prior to ischemia (Figure 9a) does not prevent the significant rise in the area and perimeter of this subset (Figure 9a and 9b). Together, with the lack of any significant difference in respect to perimeter and Feret's diameter of the two ischemic groups (Figure 9b

CONCLUSION
In this study, we showed for the first time that the three subsets of mitochondria in the adult heart differ in their morphological characteristics and respond differently to ischemic stress. We also found that IPC had no effect on mitochondrial morphology at baseline and following ischemia. Subtype heterogeneity of cardiac mitochondria in terms of physiology and morphology has been described to be critical for the effective response to pathological stimuli and is highly preserved in mammals 2, [15][16][17][18][19] . Although the role of mitochondrial morphology in cellular physiology has become clearer in recent years, the exact morphological behavior of cardiac mitochondria under normal and pathological settings have not been comprehensively studied and are For mainly confined to IMF and SSM mitochondria 13,20-23 . Furthermore, the morphological status of mitochondria within cells are directly linked with their function 24,25 instance, H2O2 insult of neonatal rat ventricular myocytes induces a dysregulation of Ca 2+ handling and dissipation of mitochondrial membrane potential which in turn increases the measure of mitochondrial roundness 26 .
Similarly, IRI simulation in H9C2 cells has been shown to induce the formation of circular donut shaped mitochondria which possess lower membrane potential and altered K + cannel activity in comparison to their elongated counterparts which further supports the link between the morphology of mitochondria and their function 27 . Consistent with our findings, a study of myocardial cells isolated from the monkey's heart showed that the perinuclear mitochondria are the least elongated and most spherical subtype of mitochondria present in cardiac myocytes. Interestingly, the same group also showed that the SSM mitochondria are longer in diameter than the other subtypes of mitochondria whereas our study showed a greater diameter in IMF mitochondria in comparison to other subtypes 18 . The shape differences that we observed between SSM and IMF mitochondria of cardiac myocytes are also present in the same subtypes in mouse skeletal muscle cells. Skeletal muscle SSM mitochondria are shorter in respect to perimeter and Feret's diameter as well as being rounder in comparison to IMF mitochondria. However, both SSM and IMF mitochondria from skeletal muscle mitochondria are shorter in respect to Feret's diameter and perimeter compared with the same subtypes of cardiac mitochondria 17 . Collectively, these studies stress that the shape differences observed in different subtypes of cardiac mitochondria are influenced by their regional location. Moreover, the presence of differences in terms of shape parameters of same subtypes of mitochondria further emphasizes on the extent of species and cell-specific nature of mitochondrial morphology.
The unique metabolic and potential structural differences between various subtypes of cardiac mitochondria appear to mediate subtype-specific feedback in response to IRI as well as cardioprotective strategies conducted against this insult including IPC. For instance, an induction of global no-flow ischemia in rabbit hearts can hamper SSM oxidative phosphorylation whereas the IMF mitochondria remains unaffected 28 . Isolated SSM mitochondria from male Sprague Dawley rats also exhibit lower oxidative phosphorylation and reduced NADH oxidation after a 40 minutes hypoxia/reoxygenation insult thereby indicating the unique properties of mitochondrial subtypes 21 . Likewise and in partial agreement with our data, myocardial infarction via left coronary artery ligation of male Sprague Dawley rats has been shown to specifically alter the topography 29 . In parallel, IPC has been shown to induce its cardioprotective effects by increasing the S-nitrosylation, promoting the complex I respiration and preserving the activity of electron transport chain enzymes of SSM mitochondria 30,31 . These results indicate the importance of both morphology and physiology of different mitochondrial subtypes in cardiomyocytes function.
Major impairment of mitochondrial structure begins before the induction of reperfusion and during ischemia and preserving the abrupt fragmentation has been shown to confer cardioprotection [32][33][34][35] . Nevertheless, the mechanism behind morphological alterations of different subtypes of cardiac mitochondria following acute ischemia is poorly understood. In line with previous reports, we observed granulation and increase in mitochondrial roundness in all subtypes of mitochondria; however, only PN mitochondria showed significant swelling after ischemia 36,37 .
Our results are consistent with our previous report where we showed a decrease in the length of IMF mitochondria after treating mouse hearts with 20 minutes of ischemia 34 .
Within the same report we also showed that the mitochondria division inhibitor can preserve the morphology of IMF mitochondria thereby indicating the therapeutic potential of manipulating this subtype of cardiac mitochondria 34 . Moreover, preservation of mitochondrial morphology via IPC has been previously shown to confer cardioprotection in the setting of IRI 11,[38][39][40] . Although IPC has been suggested to preserve the function of SSM mitochondria following IR, we did not observe any IPC-induced preservation of SSM mitochondria morphology following ischemia-only treatment 30,31 . The absence of shapepreservation in all subtypes of cardiac mitochondria in our study may indicate that the possible protective effects of IPC on mitochondrial morphology are not mediated during ischemia, but are actually initiated at the start of reperfusion.
In conclusion, we show for the first time that the three subsets of cardiac mitochondria have unique shape parameters. Ischemia induced morphological alterations in all three subsets of cardiac mitochondria and its effects were not rescued by IPC.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Acknowledgments
The authors would like to thank Mr.           Comparison of predicted mean values ± S.E of different shape descriptors of Intermyofibrillar (IMF), Subsarcolemmal (SSM) and Perinuclear (PN) mitochondria at baseline (* and *** denote p≤0.05 and p<0.0001, respectively).  Table 3 Pairwise comparison of predicted mean values ± S.E of different shape descriptors of SSM mitochondria after individual treatments (* and *** denote p≤0.05 and p<0.0001, respectively).