University of Groningen
Isoflurane induced eNOS signaling and cardioprotection
Baotic, Ines
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2018
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Baotic, I. (2018). Isoflurane induced eNOS signaling and cardioprotection: Preconditioning mechanisms under normal and hyperglycemic conditions. University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter 5
Mitochondrial Bioenergetics in Diabetic
Myocardium – Implications for Protective
Conditioning Strategies
Ines Baotic,1,2 Filip Sedlic,3 Martin Bienengraeber, Ana Sepac,3
Judy R . Kersten,4 Anne H . Epema,5 Robert H . Henning1
1 University of Groningen Department of Clinical Pharmacy and Pharmacology, Groningen, Netherlands
2 Croatian Agency for Medicines and Medical Devices, Zagreb, Croatia 3 University of Zagreb School of Medicine, Zagreb, Croatia
4 Emeritus Professor Department of Anesthesiology Medical College of Wisconsin, Milwaukee, Wisconsin
5 University of Groningen Department of Anesthesiology, Groningen, Netherlands
Table of Content
1 . Normal mitochondrial function
1 .1 . Bioenergetics and Morphology
1 .2 . Generation of reactive oxygen species 1 .3 . Antioxidants and signaling
1.4. Effects of calcium on OXPHOS and mitochondrial ROS generation
1 .5 . Mitochondria and cell death pathways 2 . Nitric oxide in mitochondria
2 .1 . Sources of NO
2 .2 . Potential actions of NO in the mitochondria 3 . Mitochondrial dysfunction and bioenergetics in diabetes 3 .1 . Diabetes alters mitochondrial function
3 .2 . Altered substrate use in diabetic mitochondria – a generator of ROS and other toxic effects
3 .3 . Mitochondrial morphology changes (fusion-fission balance)
in diabetes
3 .4 . Mitochondrial permeability transition pore opening in
diabetic hearts
4 . Implications of hyperglycemia on protective conditioning strate gies: the role of mitochondria
4 .1 . Cardiac & Remote Ischemic Conditioning 4 .2 . Anesthetic Preconditioning
In healthy cardiac tissue, ischemic/anesthetic (pre)conditioning strategies have been successfully used to limit ischemia/reperfusion damage in the experimental setting . However, such strategies failed to provide cardio-protection in diabetic animals or under acute hyperglaecemia and eviden-ce suggests that changes in NO and mitochondrial bioenergetics may fun-damentally underlie susceptibility to myocardial injury during diabetes and acute hyperglycemia (AHG) . In this review, we briefly discuss mitochon-drial function, actions of NO in mitochondria and the various conditioning strategies, and how these are perturbed by diabetes and hyperglycemia .
1. Normal mitochondrial function
1.1.
Bioenergetics and Morphology
The importance of mitochondria for support ing heart function is eviden-ced by their relative abundance in ventricular cardiomyocytes as exem-plified by these organelles occupying roughly 33% of the cellular volu-me (Bers 2001) . The massive energy demand of myocardium is volu-met by mitochondrial ATP synthesis by oxidative phosphorylation (OXPHOS). OXPHOS occurs through interactions among clustered proteins residing within the inner mitochondrial membrane (IMM) that forms the electron transport chain (ETC) . The ETC enables a series of oxidation–reduction reactions that transfer electrons from electron donors, such as nicotina-mide-adenine-dinucleotide (NADH) or flavin-containing enzymes to mole-cular oxygen . NADH-ubiquinone oxidoreductase (complex I) removes two energy-rich electrons (e−) from NADH and transfers them to ubiquinone
(Q), which is coupled with translocation of four protons (H+) across the
IMM (Williams, Boyman et al . 2015) . Complex II then transfers additional e− from FADH
2 to Q without H+ pumping (Williams, Boyman et al . 2015) .
Hydrophobic Q carries electrons to the Q-cytochrome C (Cyt C) oxidore-ductase (complex III), which pumps another four H+ across the IMM . Cyt
C oxidase (complex IV) is the final H+ pump in the ETC and transfers four
H+ across the IMM and four e− from Cyt C to molecular oxygen, generating
two molecules of water (Williams, Boyman et al . 2015) . Translocation of H+
from the matrix to the intermembrane space by the ETC generates a mito-chondrial electrochemical potential (∆µH+) that is described as the
mitoc-hondrial membrane potential (ΔΨm) . This proton gradient is utilized by ATP synthase (complex V) for the phosphorylation of adenosine diphosphate (ADP) to appropriate triphosphate. (Williams, Boyman et al. 2015). ΔΨm is a major force that drives the uptake of Ca2+ into mitochondria . In actively
respiring mitochondria ΔΨm is close to -180 mV, which represents a strong thermodynamic force driving accumulation of cations in the intermembra-ne space (Rizzuto, Bernardi et al. 2000). Fatty acids (FA) are the dominant source of energy in healthy human heart under resting conditions with in-creased contribution from glucose oxidation under workload (Lopaschuk, Ussher et al . 2010) .
Mitochondrial bioenergetics is closely related to their morphology and par-ticularly dependent on whether they are integrated into a network or not
(Rolland, Motori et al . 2013) . Cardiac mitochondria are dynamically shif-ting their morphology between interconnected elongated mitochondrial network and fragmented disconnected mitochondria by a well-balanced process of mitochondrial fusion and fission, which is disrupted under stre-ss and disease (Ong and Hausenloy 2017) . During the past two decades, the important contributors to mitochondrial fusion and fission events have been identified (Pagliuso, Cossart et al . 2018) . To date, two distinct mito-chondrial subpopulations have been described in heart; subsarcolemmal (SSM) mitochondria underneath the sarcolemmal membrane, and interfi-brillar (IFM) mitochondria in between myofibrils. These two subpopulati-ons differ in their protein and lipid composition, as well as in their capacity for respiration and behavior during stress (Riva, Tandler et al . 2005; Hata-no, Okada et al . 2015) . There are however concerns about artefacts being responsible for some of the differences found (Hendgen-Cotta, Esfeld et al . 2018; Koncsos, Varga et al . 2018) .
1.2.
Generation of reactive oxygen species
It has been estimated that under normal conditions up to 1-3% of total mitochondrial oxygen turnover ends up as superoxide anion (O2•), which
gives rise to other reactive oxygen species (ROS) . Particularly complexes I and III may leak electrons to molecular oxygen, reducing it to O2• .
Modera-te generation of ROS under physiological conditions contribuModera-tes to normal cellular signaling, and may underlie adaptive response to various stressors (Sedlic and Kovac 2017) . Excessive generation of ROS occurring in re-perfusion injury is one of major noxious stimuli in myocardial infarction . During reperfusion, superoxide produced at complex I forms exclusevly in the matrix, while complex III releases superoxide mostly into the inner mitochondrial space (IMS) (Bleier and Drose 2013) . In addition to the ETC, donation of single electrons to oxygen by other flavin or heme groups, or iron-sulfur complexes containing enzymes or NADH can generate supe-roxide under physiological and/or pathological conditions (Turrens 2003) . These enzymes include cytochrome P450 monooxygenase (particularly the P450 2C family) located in the IMM, monoamine oxidase B (MAO B) positioned on the outer mitochondrial membrane (OMM) and different ma-trix enzymes and complexes (Finkel 2011). In the mama-trix, NADH is a essen-tial electron donor, however, during inhibition of downstream electron flux, NADH will promote superoxide generation at the flavin site of complex I . Although, predominant are one-electron reactions, two-electron reactions
for example when Cyt C and p66shc interact within the IMS can occur, leading to direct hydrogen peroxide production can occur, (Finkel 2011). In the presence of transition-metal ions, such as Fe2+ or Cu2+, relatively stable
H2O2 can be converted into a highly reactive hydroxyl radical . 1.3.
Antioxidants and signaling
The preservation of intracellular redox homeostasis is reliant on a diverse group of antioxidant molecules . These antioxidants include both non-en-zymatic, low molecular weight molecules such as bilirubin (Ziberna, Marte-lanc et al . 2016), glutathione and α-tocopherol, and an array of enzymatic antioxidants that each have specific subcellular localization and chemical properties (Finkel 2011). Generated superoxide may be converted either spontaneously or enymatically by mitotochondrial manganese superoxide dismutase (SOD) to hydrogen peroxide which in turn can be processed by enzymes including glutathione peroxidase (GPx), catalase (CAT) or pe-roxiredoxin 3 (Prx3) (Finkel 2011) into non-toxic products, i.e. water and oxygen . Three human isoforms of SOD are found in cytoplasm (SOD1), mitochondria (SOD2) and extracellularly (SOD3) . In addition, thioredoxins and glutaredoxins are of significance to maintain redox homeostasis by reducing disulfide bridges in various target proteins (Finkel 2011). It has been suggested that these multiple protein antioxidants may form a loose hierarchical network or redox circuit, that maintains stability of the overall cysteine proteome (Jones and Go 2011) . Consequently, antioxidant prote-ins are increasingly viewed as active participants in redox signaling, rather than passive disposers of intracellular oxidants (Finkel 2011).
1.4.
Effects of calcium on OXPHOS
and mitochondrial ROS generation
A growing body of evidence suggests that Ca2+ plays a pivotal role in
re-gulating mitochondrial metabolism . Ca2+ is imported across the OMM into
the intermembrane space through the voltage-dependent anion channels (VDAC) (Simamura, Shimada et al . 2008) . Mitochondrial Ca2+ flux across
the IMM involves essentially two pathways: a) ΔΨm–driven uptake via the mitochondrial calcium uniporter (MCU) (Kwong 2017) and b) efflux via ion transporters such as the Na+/ Ca2+ and H+/Ca2+ exchangers, in excitable
cells and in most other tissues, respectively (Giorgi, Agnoletto et al . 2012; Tewari, Camara et al . 2014) . These electroneutral antiporters pump the
Ca2+ from matrix back into the intermembrane space and prevent the
atta-inment of an electrochemical equilibrium (Rizzuto, Bernardi et al . 2000; Giorgi, Agnoletto et al . 2012) . Increased matrix Ca2+ is thought to promote
Krebs cycle activity and ATP synthesis (Jouaville, Pinton et al . 1999; Tara-sov, Griffiths et al . 2012) . Moreover, Ca2+ can regulate mitochondrial
po-tassium channels thereby affecting intramitochondrial ionic homeostasis and ΔΨm . Thus, Ca2+ increase both initiates energy consuming processes
in the cytosol (e .g . contraction) and stimulates mitochondrial dehydroge-nases, balancing aerobic metabolism to the increased energy needs of an active cell (Rimessi, Giorgi et al . 2008; Giorgi, Agnoletto et al . 2012) . An ongoing debate is if mitochondria take up a small fraction of Ca2+ released
at each cytosolic Ca2+ spike during cardiac systole or if they respond to
resting Ca2+ levels as dictated mainly by heart rate (Saotome, Hajnóczky
et al . 2014) . In spite of their high capacity to accumulate Ca2+, the
mitoc-hondrial contribution to overall cytosolic Ca2+ fluxes during cardiac
excita-tion-contraction coupling is considered to be small, because matrix Ca2+
in mitochondria stays relatively low under physiological conditions (Huser, Blatter et al . 2000; Kwong 2017);Saotome, Hajnóczky et al . 2014) .
Consistently with the above, fibroblasts from patients with mitochondrial diseases (MIDs) with defects in complex I show reduced ΔΨm and ATP synthesis, respectively (Visch, Koopman et al . 2006; Giorgi, Agnoletto et al. 2012). Futhermore, the defect contributes to a decreased mitochondrial Ca2+ uptake under physiological stimulation (Visch, Koopman et al . 2006) .
The thus created negative feedback loop further decreases intracellular Ca2+ levels and ATP content in fibroblasts with isolated complex I
deficien-cy (Visch, Koopman et al . 2006) .
Ischemia/reperfusion (I/R) injury and oxidative stress increase cytosolic and subsequently mitochondrial Ca2+, mainly by promoting Ca2+ influx and
attenuating Ca2+ efflux (Sedlic, Sepac et al . 2010), rooted in a decrease in
cellular ATP, ionic disequilibrium, and plasma membrane depolarization . The accumulation of cytosolic Ca2+ is believed to increase mitochondrial
respiration and electron flux along the ETC, thus promoting ROS genera-tion . ROS can further stimulate Ca2+ release from the endoplasmic
reticu-lum via ryanodine receptors (RyR), which are juxtaposed to MCU, allowing the rapid transfer of Ca2+ into mitochondria . Calcium can stimulate its own
release via RyR, which is called calcium-induced calcium release (CICR) . This closes the positive feedback circuit of deleterious accumulation of mitochondrial Ca2+ and mitochondrial ROS production, both of which are
pore (mPTP), a protein formed at the IMM ultimately conferring cell death (Sepac, Sedlic et al . 2010; Abdallah, Kasseckert et al . 2011; Panel, Ghaleh et al . 2017) . Under pathological conditions, mPTP opening can occur as a self-perpetuating series of individual mPTP opening events that have been described as ROS-induced ROS release (RIRR) (Zorov, Filburn et al. 2000) and mitochondrial Ca2+-induced Ca2+ release (mCICR) (Ichas,
Joua-ville et al . 1997) . In both, mPTP opening causes abrupt release of Ca2+ that
is taken up by surrounding mitochondria . Moreover, mPTP opening also causes excessive ROS production that also spreads to neighboring mito-chondria . Both processes favor mPTP opening in adjacent mitomito-chondria, resulting in a continuous mPTP opening event (Figure 1). Cardiac mitoc-hondria are particularly sensitive to continuous mPTP opening since car-diomyocytes are packed with mitochondria in highly ordered crystal-like patterns (Figure 1).
Figure 1. mPTP opening in a cardiomyocyte occurring as a chain reaction.
Oxidative stress-induced mPTP opening was induced in rat cardiomyocyte by laser-induced photoexcitation causing a sudden loss of TMRE fluorescence from individual mitochondria, indicating abrupt dissipation of ∆Ψm. In some instances, as in this example, mPTP opening occurs as a series of events that seem to be spreading from one mitochondrion to another. Unpublished data (F. Sedlic). N. = nucleus.
1.4.
Mitochondria and cell death pathways
Opening of the mPTP is the major trigger for mitochondria-initiated cell death, with the fate of the cardiomyocyte being dependent on the extent of mPTP opening events (Halestrap, Clarke et al . 2004; Lemasters, The-ruvath et al . 2009) . In the case of transient ischemia (e .g . attacks of angina pectoris), when mPTP opening occurs in only few mitochondria, with the apoptosis and necrosis signals remaining below threshold, depolarized (damaged) mitochondria will be digested by mitophagy and the cell may survive . More extensive mPTP opening, allowing passage of molecules up to 1 .5 kDa, induces apoptosis, which is mediated by release of numerous pro-apoptotic factors as reviewed by Crow et al . (Crow, Mani et al . 2004), which in turn initiate caspase-dependent or -independent apoptosis . In contrast to necrosis,apoptosis is reversible up to the point of activation of caspases . Apoptosis is mainly found in the periphery of infarction where ischemic stress is mitigated by collateral circulation allowing for sufficient delivery of oxygen and substrates . Apoptotic cell death is preferable to necrosis since it does not activate inflammation and the ensuing additio-nal cell injury . When the noxious stimulus is large enough, the critical thre-shold for self-perpetuating mPTP opening is reached causing a large num-ber of mPTP opening events, resulting in necrotic cell death due to ROS overproduction, cytosolic Ca2+ overload and hypercontraction . Complete,
abrupt mitochondrial depolarization triggered by continuous mPTP ope-ning causes a reversal of ATP synthesis to ATP hydrolysis in an attempt to maintain ΔΨm . As a result, protons are pumped out of the matrix using ATP as an energy source . While mitophagy and apopotosis might also be initiated in such a scenario (e .g . massive heart attack), the affected car-diomyocytes do not have sufficient stores of ATP to execute both ATP-de-pendent processes .
Generaly within 3 hours after the initiation of coronary occlusion, myocar-dial necrosis is completed (Basso, Rizzo et al . 2010; Thiene and Basso 2010) . In order to salvage myocardium, early reperfusion by coronary in-tervention is required . Paradoxically reperfusion potentiates progression of myocardial injury due to a burst of ROS (Saotome, Hajnóczky et al . 2014). Reoxygenation during reperfusion enables restoration of ΔΨm and ATP supply by recovery of the ETC function (Saotome, Hajnóczky et al . 2014). However, in parallel, regeneration of ΔΨm induces mitochondrial Ca2+ loading due to elevation of cytosolic Ca2+ concentration and a burst
re-perfusion-induced damage is an important determinant in the successful treatment of a coronary occlusion event as evidenced by successful po-stconditioning strategies (Sun, Wang et al . 2005; Riess, Matsuura et al . 2014) .
2. Nitric oxide in mitochondria
2.1.
Sources of NO
Evidence indicates that nitric oxide (NO) modulates mitochondrial respi-ration through high-affinity binding to Cyt C oxidase (Boveris, Costa et al . 1999) . It is, however, still under intense debate whether NO is actually synthesized in mitochondria . Intramitochondrial production of NO would however require translocation of one of the isoforms of nitric oxide syntha-se (nNOS, eNOS or iNOS) since mtDNA syntha-seems not to encode a conven-tional NO synthase (Lacza, Snipes et al . 2003) . However, some evidence indicates the existence of a mitochondrial NOS variant, referred to as mt-NOS, as reported in liver (Carreras, Converso et al . 2004), brain (Parihar, Nazarewicz et al . 2008) and heart (Nazarewicz, Zenebe et al . 2007; Dedko-va and Blatter 2009) . Mitochondrial location of mtNOS is still elusive; both its binding to Cyt C oxidase (Ghafourifar and Richter 1997; Zaobornyj, Val-dez et al . 2009) and interaction with complex I have been reported (Pari-har, Nazarewicz et al . 2008) . Association of mtNOS with proteins of the ECT is supported by data indicating that mitochondrial NO production is increased by reverse electron transfer (RET) at the expense of ATP produ-ction (Bombicino, Iglesias et al . 2016)
As an alternative to mtNOS, conventional NOS isoforms associated with the OMM (Gao, Chen et al . 2004) or at distance site (Lacza, Pankotai et al . 2009) may provide an external source of NO to mitochondria . Moreover, mitochondrial NO may also be synthetized from the reduction of nitrite (to nitrate) by other enzymes than NOS, e .g . through nitrite reductase activity of heme proteins such as hemoglobin, as observed during extreme hypo-xia in case of ischemia or I/R injury (Shiva 2013)
2.2.
Potential actions of NO in the mitochondria
NO or its derived reactive nitrogen species (RNS) can have profound effe-cts on mitochondrial function under physiological and pathological
con-ditions affecting both OXPHOS complexes and mitochondrial channels (summarized in Figure 2). NO reacts strongly with all heme moieties and thus reversibly reduces the activity of iron-containing enzymes. For exam-ple, NO may compete with oxygen for its binding site on complex IV, an action that might be especially pronounced under hypoxic conditions (Ghafourifar, Bringold et al . 2001; Galkin, Higgs et al . 2007) . Complex II, although having heme moieties, is relative insensitive to NO or RNS but may be inhibited by high levels of RNS, such as nytroxil (HNO) (Riobo, Clementi et al . 2001) While complex I is devoid of heme groups, NO is well known to inhibit complex I and the ensuing ATP production (Brown and Borutaite 2004; Galkin and Moncada 2017). In addition to action on OXP-HOS complexes by direct protein interaction, NO can induce opening of the mitochondrial KATP channel leading to cardioprotection (Queliconi, Woj-tovich et al . 2011) . Mitochondrial KATP channel is being activated by NO•
directly via thiols on channel itself or through complex II (not being inhibi-ted) by secondary RNS like HNO, reactive SNO (RSNO) and nitrolinoleate (LA-NO2; example of nitro-lipid) (Queliconi, Wojtovich et al. 2011). Further-more, RNS generated during I/R injury activate the mitochondrial matrix poly(ADPribose) polymerase (PARP) cascade, thus negatively influencing elctron transport and ATP formation, causatively leading to the impairment or death of the cardiomyocytes (Pacher and Szabo 2007) .
Our previous work has shown that the effects of NO on cardiac mitochon-dria are dose-dependent (Cheng, Sedlic et al . 2011) . A moderate increase in NO bioavailability delayed mPTP opening resulting in cardioprotection in isolated rat cardiomyocytes . This action coincided with an inhibition of VDAC as measured in purified channels reconstituted in a lipid bilayer . Although VDAC does not appear to be key constituent of the mPTP, its inhibition may delay mPTP opening (Baines, Kaiser et al . 2007) . Interestin-gly, we observed that at higher concentrations NO failed to delay mPTP opening and inhibit VDAC, indicating a biphasic effect (Cheng, Sedlic et al . 2011) . The non-linear nature of NO actions is important to keep in mind when designing studies and interpreting “controversial” data. Furthermo-re, tyrosine nitration-induced changes in VDAC conductance during I/R injury have deleterious consequences (Yang, Camara et al . 2012)
S-nitrosylation (SNO) of proteins is another mechanism by which NO may alter function or activity of mitochondria (Ozawa, Komatsubara et al . 2013), as suggested by the extensive SNO of proteins participating in glycolysis, gluconeogenesis, tricarboxylic acid (TCA) cycle, and oxidative phosphorylation (Doulias, Tenopoulou et al. 2013). For instance, SNO can
shield cysteines from irreversible oxidation induced by burst of ROS du-ring the beginning of reperfusion (Murphy, Kohr et al . 2012) . As a specific example, reversible and selective SNO of complex I has multiple benefitil effects during first minutes of reperfusion, as it slows mitochondrial rea-ctivation and thus decreases ROS production and tissue necrosis (Cho-uchani, Methner et al . 2013; Keszler, Brandal et al . 2014) . SNO may also affect mitochondrial function indirectly, e .g . by promoting the availability of substrates through a nitrosylation dependent increase of catalytic activity of very long chain acyl-CoA dehydrogenase (VLCAD), in turn increasing β-oxidation of FAs, a feature absent in eNOS-/- mice (Doulias, Tenopoulou
et al . 2013) . S-nitrosylation of the mitochondrial proteome is mostly de-pendent on eNOS-activity, with the exception of the heart where just one third of the SNO proteome originates from eNOS derived NO, implying the dependence on other functional NOS isoform(s) in cardiomyocytes (Do-ulias, Tenopoulou et al. 2013). Further, studies report that SNO both inhi-bits and activates the L-type Ca2+ channel, again displaying a non-linear
nature of interaction . Inhibition of the L-type Ca2+ channel may be
advan-tageous during I/R, as this would confer cardioprotection by reducing Ca2+
entry and mitigating Ca2+ overload (Sun, Picht et al . 2006) .
A fourth mechanism of NO action on mitochondria is through the forma-tion of peroxynitrite (ONOO-) by its interaction with superoxide anion .
Pe-roxynitrite, like NO, also inhibits mitochondrial respiratory OXPHOS com-plexes, including complexes I, II, IV and V (Brown 1999) . However, while NO inhibition of complexes is reversible, peroxynitrite has the unfavorable property to irreversibly inhibit them . Moreover, generation of peroxynitrite leads to tyrosine-nitration (Castro, Demicheli et al . 2011) and a decrease in cellular NO/SNO which can contribute to redox-nitroso imbalance (Murp-hy, Kohr et al . 2012) .
Finally, the importance of NO for mitochondrial function is underscored by studies showing that deficiency in NO significantly changes profile of mi-tochondrial proteins influencing dynamics, increasing thus mimi-tochondrial fission, decreasing fussion,mitochondrial dynamic protein profiles with de-creased mitochondrial fusion, inde-creased fission, and minimally perturbing autophagy, albeit the exact underlying mechanisms are not yet understo-od (Miller, Knaub et al . 2013) .
Figure 2. Actions of NO in mitochondria. Modified from (Di Giacomo, Rizza et al.
2012) .
Actions of NO in mitochondria need to be distinguished: A) on OXPHOS complexes: Fe-ni-trosylation of CytC (C IV), S-niFe-ni-trosylation of almost all complexes, tyrosine-nitration by per-oxynitrite of all complexes; B) on mitochondrial channels: tyrosine-nitration of VDAC, mPTP opening delay, activation of mK
ATP channel - direct /tiols and indirect/secondary RNS (HNO, RSNO, LA-NO2). C, complex; HNO, nytroxil; LA-NO2, nitrolinoleate; mKATP, mitochondrial K
ATP channel; mPTP, mitochondrial permeability transition pore; RNS, reactive nitrogen spe-cies; RSNO, reactive SNO; SNO, S-nitrosylation; VDAC, voltage-dependent anion channel
3.
Mitochondrial dysfunction
and bioenergetics in diabetes
Mitochondrial dysfunction plays a central role in the pathogenesis of both type I and II diabetes, ranging from impaired insulin secretion to the de-velopment of diabetic complications, including myocardial injury . In type II diabetes, glucotoxicity is associated with excessive ROS production by mitochondrial and non-mitochondrial sources, which inflicts damage to insulin-secreting beta cells (Robertson 2004), ultimately resulting in a decrease in insulin secretion in later stages of type II diabetes . Moreover, mutations in mitochondrial DNA have been linked to diabetes type II (Jia-ng, Li et al . 2017), and are causative of maternally inherited diabetes and
deafness (MIDD) (Naing, Kenchaiah et al . 2014), suggesting that mitoc-hondrial dysfunction can also cause diabetes. Further, the alterations in mitochondrial function in various organs inflict damage on these organs and contribute to diabetic complications . Consequently, the mitochondrial impairment plays a crucial role in the high incidence of acute myocardial infarction and cardiomyopathy in diabetics . In the following sections we will elaborate on several mechanisms by which diabetes alters mitochon-drial function, thus driving damage to the heart .
3.1
Diabetes alters mitochondrial function
Several mechanisms contribute to mitochondrial dysfunction in the cour-se of diabetes, including insufficient ATP generation, excessive generati-on of ROS and oxidative stress, release of multiple pro-apoptotic stimuli, cytosolic Ca2+ overload and impairment of the synthesis of biomolecules
(Ahn and Metallo 2015) . Increased cellular and mitochondrial production of ROS or reduction in antioxidant defense can induce DNA damage and subsequent mutations in nuclear and mitochondrial DNA (mtDNA) (Hicks, Labinskyy et al . 2013), giving rise to dysfunctional mitochondrial proteins causing mitochondrial dysfunction (Zaragoza, Brandon et al. 2011). Furt-her, mtDNA gene expression in diabetes may also be altered by epigenetic mechanisms (van der Wijst, van Tilburg et al . 2017), as exemplified by the impairment of mtDNA transcription due to hypermethylation of its D-loop induced by upregulation of DNA (cytosine-5)-methyltransferase 1 (Dnmt1) (Mishra and Kowluru 2015) . In addition to DNA damage, oxidative stress in diabetes can also cause lipid peroxidation of cellular membranes (Davi, Falco et al. 2005), which may result in mitochondrial dysfunction and
im-pairment of ATP-synthase activity and ATP production (Soto-Urquieta, Lo-pez-Briones et al . 2014) . Moreover, diabetes-associated oxidative stress also affects the function of proteins, including mitochondrial proteins, by posttranslational changes arising from tyrosine nitration, thiol group oxida-tion (Ates, Kaplan et al . 2016), protein carbonylaoxida-tion (Hecker and Wagner 2018), O-GlcNAcylation, and others (Turko, Li et al . 2003; Hu, Suarez et al . 2009; Zhao, Feng et al. 2016; Korovila, Hugo et al. 2017). O-GlcNAcylation following exposure to high glucose alters the function of various cardiac proteins and contributes to altered homeostasis (Chatham and Marchase 2010), which concerns several mitochondrial proteins, including complex I subunit NDUFA9, complex III subunits core 1 and core 2, and subunit I of complex IV (Hu, Suarez et al . 2009) .
Diabetes may also affect the maintenance of proper mitochondrial functi-on and number by affecting mitophagy, i .e . the autophagosomal removal of damaged or dysfunctional mitochondria (Kumari, Anderson et al . 2012; Xu, Kobayashi et al. 2013; Higgins and Coughlan 2014; Sarparanta, Gar-cia-Macia et al . 2017; Jin, Li et al . 2018) . However, there is some con-troversy pertaining to the exact effect of diabetes on mitophagy . Some reports indicate increased mitophagy in diabetic heart (Xu, Kobayashi et al . 2013) and brain (Kumari, Anderson et al . 2012), possibly representing a compensatory mechanism aimed at clearing dysfunctional mitochondria . However, overly active mitophagy may also be responsible for reduced mi-tochondrial content in the diabetic heart, diminishing ATP production ca-pacity (Jin, Li et al . 2018) . Conversely, other studies report impairment of mitophagy in diabetes that would be responsible for accumulation of dys-functional mitochondria (Higgins and Coughlan 2014; Durga Devi, Babu et al . 2017) . The latter is supported by data showing that the higher inciden-ce of postinfarction heart failure in type II diabetic miinciden-ce is associated with impaired cardiac mitophagy (Durga Devi, Babu et al . 2017) . In addition to excessive mitophagy, the diminished mitochondrial content in the diabetic heart may so be caused by impaired mitochondrial biogenesis . In support of this, epigenetic downregulation of the main mitochondrial biogenesis regulator, PGC-1α, in skeletal muscles of diabetic subjects correlates with a reduced expression of mtDNA, indicating decrease in mitochondrial density (Barres, Osler et al . 2009) . This is in agreement with other stu-dies showing reduced expression of mtDNA copy number and respiratory chain activity (Boushel, Gnaiger et al . 2007), reduced expression of mitoc-hondrial oxidative phosphorylation genes controlled by PGC-1α in skeletal muscle (Mootha, Lindgren et al . 2003) and hearts (Yu, Gong et al . 2017),
and reduced synthesis of mitochondrial proteins driven by mitochondrial
transcription factor A (Kanazawa, Nishio et al . 2002) . In contrast, Vasquez
et al . showed unaltered density and mitochondrial enzyme activity in heart of rats with STZ-induced diabetes (Vazquez, Berthiaume et al . 2015), whi-ch seems at odds with the reduction in total activity of respiratory whi-chain in skeletal muscles of type II diabetic patients (Boushel, Gnaiger et al . 2007) . It remains to be elucidated whether this discrepancy reflects differences in type of diabetes, type of tissue or species .
3.2
Altered substrate use in diabetic mitochondria – a
generator of ROS and other toxic effects
Oxidative stress is widely recognized as a key disorder in the diabetic heart and stems from elevated generation of ROS from various sources (Kayama, Raaz et al . 2015) and attenuated antioxidant defenses, such as reduced thioredoxin expression in cardiomyocytes induced by hypergly-cemia (Luan, Liu et al . 2009) . Oxidative stress can induce a plethora of detrimental effects as discussed above, including oxidation of lipids, re-sulting in dysfunction of cellular membranes (Davi, Falco et al. 2005), oxi-dative damage to nDNA and mtDNA, leading to mutations causing mitoc-hondrial dysfunction (Shokolenko, Venediktova et al . 2009), and oxidation of proteins, generating proteotoxic stress and cell dysfunction . Moreover, ROS may directly affect mitochondrial function by accelerating mPTP opening (see below) (Korovila, Hugo et al . 2017) . Mitochondria represent one the most prominent sources of ROS in the diabetic heart, due to al-tered substrate metabolism characterized by elevated utilization of FAs or glucose . Additional mechanisms of elevated oxidative stress in diabetes are discussed in section 3 .4 .
Altered glucose/pyruvate utilization. High glucose enhances ROS pro-duction by mitochondria and other sources like NAD(P)H oxidase (Yu, Ro-botham et al . 2006; Shen 2010; Balteau, Tajeddine et al . 2011) . Glucose may enter cardiomyocytes independent of insulin by GLUT12 (Waller, Ge-orge et al . 2013), or by GLUT4, which is translocated to the sarcolemma during stress, such as ischemia (Sun, Nguyen et al . 1994), and possibly by the highly expressed Na+/glucose cotransporter 1 (Zhou, Cryan et al .
2003) . The ensuing acute increase in glucose enhances glucose oxidation and lactate production in humans during hyperglycemia (Wisneski, Stan-ley et al . 1990) . In keeping with this interpretation, we recently showed that high glucose is promptly metabolized in isolated rat cardiomyocytes
by enhancing mitochondrial respiration (Sedlic, Muravyeva et al . 2017), with an acute increase in oxygen consumption, NADH generation, ΔΨm and ROS production, ultimately resulting in acceleration of mPTP opening and reduced resistance of cells to oxidative stress . The reversal of these actions by the mitochondrial uncoupler dinitrophenol (DNP) underlines a critical role of mitochondrial hyperpolarization in response to high glucose . The combination of high glucose and DNP conferred cellular protection similar to the hyperosmotic mannitol control, which was attributed to a preconditioning-like phenomenon triggered by small amount of ROS pro-duction (see section 4) (Sedlic, Muravyeva et al . 2017) . While these data demonstrate ROS-mediated acute toxicity of high glucose, other studies suggested that heart may protect itself by limiting glucose metabolism and favoring oxidation of FAs and other substrates (Vazquez, Berthiaume et al. 2015). Unfortunately, such substitution of glucose by FAs also yields increased amounts of ROS, as discussed below .
Altered FA utilization. A shift of metabolism from glucose to FAs was demonstrated in models with chronically elevated glucose levels, docu-menting decreased oxidation of glutamate, pyruvate and succinate, incre-ased oxidation of long-chain FA substrate palmitoyl-carnitine specifically in SSM, which was accompanied by increased activity of medium- and long-chain acylCoA dehydrogenases (Vazquez, Berthiaume et al . 2015), and increased oxidation of FAs (Mather, Hutchins et al. 2016). Decreased glucose oxidation may be rooted in a reduced activity of complexes I and II of the ETC (Vazquez, Berthiaume et al. 2015). Enhanced FA oxidation can result from elevated peripheral lipolysis and upregulation of plasma membrane FA transport (FAT/CD36, FABPpm) as demonstrated in db/db hearts (Carley and Severson 2008) .
The shift toward FA oxidation, however, also elevates ROS, as documen-ted in models with type I DM (Vazquez, Berthiaume et al . 2015; Cortassa, Sollott et al. 2017), as supraphysiological levels of FAs elevate ROS by uncoupling mitochondria, impairing the capacity of major mitochondrial antioxidant systems, glutathione and thioredoxin (Cortassa, Sollott et al . 2017). High fat diet that elevates FA oxidation elevates cardiac uncoupling protein 3 expression, decreases efficiency of energy production in the he-art and reduces cardiac function, corroborating the detrimental effects of high FA utilization (Cole, Murray et al. 2011). It follows that, inhibition of FA oxidation by various approaches, such as lowering of mitochondrial FA uptake by inhibition of carnitine palmitoyl-transferases, improves cardiac
function in diabetic rats (Wall and Lopaschuk 1989; Schmitz, Rosen et al . 1995; Jaswal, Keung et al . 2011) .
In addition to increased ROS production, enhanced FA oxidation induces its deleterious effects by affecting other pathways, including the impair-ment of adenine nucleotide translocase (ANT) activity that mediates ATP efflux from mitochondria (Roussel, Thireau et al . 2015), augmentation of Ca2+ release from the sarcoplasmic reticulum by oxidation and
S-nitrosyla-tion of RyR (Roussel, Thireau et al . 2015), and possibly by excessive mi-tochondrial uncoupling via uncoupling proteins and activation of peroxiso-me proliferator–activated receptor (PPAR)α (Murray, Panagia et al . 2005) . While we showed acutely elevated glucose utilization by cardiomyocytes and subsequent ROS overproduction (Sedlic, Muravyeva et al . 2017), data from the chronic hyperglycemia setting suggest a compensatory attenu-ation of glucose oxidattenu-ation at the expense of deleterious increase in FA utilization (Vazquez, Berthiaume et al . 2015) . Either way, cardiomyocytes appear at risk of injury during overt diabetes, as both excessive glucose oxidation (forced by hyperglycemia) and excessive FA oxidation increa-se ROS generation . However, it remains to be determined to what extent glucose and FA oxidation contribute to the excessive ROS generation in cardiac mitochondria in diabetes during acute or chronic conditions . Ketone body and lactate utilization. High rates of oxidation of glucose and FAs produce large amounts of ROS. However, it appears that the dia-betic heart finds some salvage in consuming ketone bodies and lactate . In contrast to ex vivo experiments where hearts are perfused with glucose and palmitate fixed concentrations, but without insulin and other substra-tes essential 10for oxidative metabolism (e .g ., ketone bodies and lacta-te), in vivo human data suggest an absence of increased FA utilization in diabetic hearts (Mizuno, Harada et al . 2017) . Most likely, absence of in-creased cardiac FA utilization in vivo is due to substrate competition by lactate and ketone bodies, as was found by NMR spectroscopy evaluating substrate contributions to acetyl CoA production in mouse heart (Stowe, Burgess et al . 2006) . This study documented that the contribution from lactate and ketone bodies (24% and 34%) to total acetyl CoA producti-on grossly outweighed that of FA utilizatiproducti-on (18%). Further, plasma lactate and ketone levels are increased in db/db mice, suggesting that they provi-de alternative substrates to glucose and FAs for the diabetic heart (Avoga-ro, Nosadini et al . 1990; Oakes, Thalen et al . 2006; Janardhan, Chen et al . 2011) . Beneficial effects of ketone body utilization are reflected by a higher
production of ATP per mole of oxygen from ketone bodies compared to glucose and FA (Veech 2004). Importantly, ketone bodies may reduce ROS generation by scavenging free radicals, and unlike FAs, oxidation of keto-ne bodies does not produce uncoupling (Veech 2004) .
Lactate represents another alternative to glucose or FA combustion as it is quickly metabolized in mitochondria, following intracellular or intramitoc-hondrial conversion to pyruvate (Brooks, Dubouchaud et al. 1999) Further-more, the fibroblast–cardiomyocyte lactate shuttle hypothesis, reviewed by Brookes (Brooks 2009), emerged as important element influencing la-ctate metabolism in cardiomyocytes (Rakus, Gizak et al . 2016) . However, data confirming a role for the fibroblast–cardiomyocyte lactate shuttle in diabetes are lacking . In contrast, data from STZ rat heart indicate that dia-betes affects lactate oxidation even to a larger extent than glucose oxi-dation (Chatham, Gao et al . 1999), although this may relate to the experi-mental setting not mimicking in vivo conditions . Evidence that lactate is an alternative fuel to glucose in diabetes is very weak or even absent
Taken together, evidence indicates that changes in glucose and FA oxida-tion in diabetic heart elevates ROS producoxida-tion and causes other negative effects which may be attenuated by ketone body oxidation while the posi-tive contribution on lactate oxidation still needs to be elucidated .
Metabolic inflexibility and remodelling. At rest, cardiac energy con-sumption relies mostly on FA oxidation (60-70%), while glucose (30-40%) and other substrates are less important (Stanley, Lopaschuk et al . 1997; Carley and Severson 2005) . However, when exercising, healthy human he-art significantly increases glucose oxidation without changing FA oxidati-on (Neglia, De Caterina et al . 2007) . This shift in substrate metabolism in response to exercise or substrate availability is called metabolic flexibility, which improves energy efficiency (Goodpaster and Sparks 2017) . In dia-betes, owing to insulin resistance, many tissues are unable to increase glucose oxidation (Goodpaster and Sparks 2017), which may result in ina-dequate ATP production during increased cardiac work and myocardial injury .
However, the exact molecular mechanism underlying metabolic inflexibility in the diabetic heart is not defined. Fundamental mechanisms involved are thought to include a decrease in insulin-dependent glucose uptake and an increase in FA supply and oxidation due to increased peripheral lipo-lysis and FA uptake (Carley and Severson 2008). Decreased insulin de-pendent glucose uptake seems rooted in a downregulation of the
domi-nant insulin-dependent glucose transporter, GLUT4, since translocation of GLUT4 to plasma membrane immediately increases (10- to 20-fold) gluco-se uptake (Shao and Tian 2015) . Insulin facilitates glucogluco-se transport in part through NO, whereby AMP-activated protein kinase (AMPK) phosphoryla-tes eNOS on Ser1177, leading to its activation . Consequently, eNOS
knoc-kout mice or isolated hearts treated with an inhibitor of NOS, L-NAME, exhibit diminished insulin-stimulated glucose uptake (Li, Hu et al . 2004) . This further ties deficient NO bioavailability to impaired metabolism and limited glucose uptake in diabetic cardiomyocytes .
Impaired mitochondrial Ca2+ handling due to downregulation of the
mitoc-hondrial Ca2+ uniporter likely represents a second mechanism contributing
to metabolic inflexibility . Diaz-Juarez et al . recently showed that restoring the mitochondrial Ca2+ uniporter levels by adenoviral expression, which
fa-cilitates mitochondrial Ca2+ uptake, improved the metabolic profile of high
glucose treated cardiomyocytes by increasing glucose oxidation and re-ducing FA oxidation to control levels (Diaz-Juarez, Suarez et al. 2016). In addition, restoring mitochondrial Ca2+ uniporter levels augmented
pyruva-te dehydrogenase activity and mitochondrial membrane popyruva-tential, and re-duced oxidative stress and apoptosis .
Taken together, altered substrate metabolism in a form of metabolic in-flexibility/remodeling in cardiac mitochondria of diabetics may lead to di-fferent toxic effects, including excessive ROS generation, diminished ATP metabolism, excessive uncoupling and increased Ca2+ release from
mito-chondria .
3.3
Mitochondrial morphology changes (fusion-fission
balance) in diabetes
Disorders in mitochondrial network morphology (fusion-fission balance) emerge as important component of diabetes-induced cardiomyocyte pat-hophysiology . In general, mitochondrial fission is associated with exces-sive ROS generation, activation of apoptosis and decreased resistance of cardiomyocytes to injury, while mitochondrial fusion is linked to a cardio-protective profile (Willems, Rossignol et al . 2015; Jin, Li et al . 2018) . Insulin resistance, high glucose and potentially increased FA oxidation induce mi-tochondrial fission in diabetes .
Mitochondrial fission/fragmentation. High glucose promotes mitoc-hondrial fission by downregulating pro-fusion proteins and upregulating
pro-fission proteins . Diabetes promotes mitochondrial fission by increa-sing activity of Drp-1 in various cell types (Gawlowski, Suarez et al . 2012; Wang, Wang et al . 2012; Ayanga, Badal et al . 2016; Li, Zhang et al . 2016; Wang, Zhang et al . 2017) . In STZ-induced diabetic mice hearts, DRP1 activity is increased because of O-GlcNAcylation of Drp-1 and decreased phosphorylation of DRP-1 at Ser637 with mitochondria exhibiting enhanced
fragmentation and depolarization (Gawlowski, Suarez et al . 2012) . Hyper-glycemia applied prior to I-R injury increased levels of Drp-1 and Fis1 and decreased levels of Opa1 in mice brains (Kumari, Anderson et al . 2012) . Likewise, following exposure to high glucose, mitochondria undergo rapid Drp-1-mediated fragmentation with increase in ROS formation (Yu, Robot-ham et al . 2006) . As inhibition of network fragmentation attenuated ROS production, mitochondrial fission likely precedes ROS overproduction rat-her than vice-versa (Yu, Robotham et al . 2006) . The association between mitochondrial fission and excessive ROS generation in diabetes is also supported by a recent study using STZ-induction in mice showing inhibi-tion of Drp-1 to attenuate mitochondrial fragmentainhibi-tion and ROS produ-ction in endothelial cells (Wang, Zhang et al . 2017) . In addition, inhibition of Drp-1 attenuates myocardial I/R injury in STZ-induced diabetic mice (Ding, Dong et al . 2017) . However, effective blocking of ROS increase by inhibition of mitochondrial pyruvate uptake, did not prevent mitochondrial fragmentation in high glucose conditions, suggesting a complex interplay among ROS mitochondrial fragmentation and mitochondrial pyruvate me-tabolism (Yu, Robotham et al . 2006) .
Altered FA metabolism in diabetes may also contribute to the imbalan-ce in mitochondrial fusion/fission. Increasing FA oxidation in cardiomyo-cytes (by overexpressing long-chain acyl-CoA synthetase 1 or by palmita-te exposure) enhanced A-kinase anchor propalmita-tein ubiquitination, decreased Drp-1 phosphorylation at Ser637 and altered the proteolytic processing of
Opa1 (Tsushima, Bugger et al . 2018), collectively leading to mitochondrial fission and ROS overproduction . Interestingly, in this study, ROS scaven-ging restored mitochondrial morphology, indicating that ROS precede mitochondrial fission in contrast to the above-mentioned study (Yu, Ro-botham et al . 2006) . Thus, the relationship among ROS and fusion-fission machinery needs further investigation .
Other pro-fission factors such as Fis1, Mff, MiD49, and MiD51 are less
well studied in the diabetic heart (Loson, Song et al . 2013) . In addition to
increased expression of Drp-1, high glucose induces greater abundance of Fis1 in rat cardiomyocytes (Wang, Gao et al. 2017) and cardiac
pro-genitor cells (Choi, Park et al . 2016) . Moreover, increased expression of Drp-1 and FIS1 was also found in endothelial cells isolated from diabetic patients and in cells treated with high glucose (Shenouda, Widlansky et al . 2011) . A study in mouse hearts exposed to I-R injury found increased Mff-mediated mitochondrial fragmentation and mitochondrial apoptosis (Jin, Li et al . 2018), further linking the mitochondrial fusion-fission balance to resistance of cardiomyocytes to I-R injury .
Mitochondrial fusion. The shift toward mitochondrial fission in diabetes is often accompanied by a decrease in pro-fusion factors, such as Opa1 (Ma-kino, Scott et al . 2010) . High abundance of Opa1 and mitochondrial fusi-on are perceived as beneficial for cardiac homeostasis. For example, Opa1 overexpression improves mitochondrial function in mouse models with mutations in complex I and IV associated proteins (Civiletto, Varanita et al . 2015) . Conversely, increased processing and inactivation of Opa1 leads to heart failure in mice (Wai, Garcia-Prieto et al . 2015) . Given that insulin in-duces mitochondrial fusion in cardiomyocytes via pathways including Akt, mTOR, NF-κB and Opa1 (Parra, Verdejo et al . 2014), insulin resistance in cardiomyocytes may contribute to mitochondrial fission in diabetes . In addi-tion, exposure of neonatal cardiomyocytes to high glucose led to O-GlcNA-cylation and dysfunction of Opa1, resulting in mitochondrial fragmentation, depolarization and reduced complex IV activity, which was rescued by Opa1 overexpression (Makino, Suarez et al. 2011). Further, acute hyperglycemia decreased the expression of Opa1 in brain of mice (Kumari, Anderson et al . 2012) . In addition to changes in their expression, diabetes inhibits activity of pro-fusion factors via post-translational modifications . As mentioned above, excessive FA oxidation decreases Opa1 activity because of enhanced pro-teolytic processing (Tsushima, Bugger et al. 2018). Further, reduced activity of Opa1 and concomitant mitochondrial fission was found dependent on Opa1 hyperacetylation in hearts of db/db mice (Samant, Zhang et al . 2014), possibly dependent on downregulation of a major cardiac deacetylase, SIRT3 (Caton, Richardson et al . 2013) . Moreover, two other fusion pro-teins, Mfn1 and Mfn2 were downregulated by high glucose in primary neo-natal rat cardiomyocytes (Wang, Gao et al . 2017), while Mfn1 was downre-gulated in right atrial myocardium obtained from type II diabetic patients (Montaigne, Marechal et al . 2014) .
Collectively, these data indicate diabetes to provoke a marked dysregula-tion of the mitochondrial fusion/fission balance by affecting expression of pro-fission/-fusion factor and their post-translational modifications, which seems related to increased ROS production .
3.4
Mitochondrial permeability transition pore opening
in diabetic hearts
Diabetes promotes mPTP opening (Wang, Gao et al . 2017) (Williamson, Dabkowski et al . 2010; Sloan, Moukdar et al . 2012), which is associated with increased cell death during I-R or oxidative stress in cardiomyocytes (Sedlic, Pravdic et al . 2010; Gedik, Heusch et al . 2013) . Accelerated mPTP opening may occur due to excessive ROS production and/or mitochon-drial Ca2+ overload, or by increased sensitivity of mPTP
components/regu-lators to these triggers .
Diabetes accelerates mPTP opening. We recently showed that acute high glucose accelerates mPTP opening and increases cardiomyocyte death due to excessive ROS production (Sedlic, Muravyeva et al . 2017) . These findings corroborate the accelerated mPTP opening in cardiac mi-tochondria isolated from STZ-induced diabetic rats (Sloan, Moukdar et al. 2012). Further, ROS sensitizes mPTP to Ca2+-induced opening as
de-monstrated in STZ rats . Likewise, increased sensitivity of mPTP to Ca2+
was found in sedentary hypercholesterolemic pig, possibly related to the downregulation of antioxidant enzymes such as manganese SOD, thiore-doxin-2 and peroxiredoxin-3 (McCommis, McGee et al. 2011). Further, high palmitate can also accelerate mPTP opening, possibly by augmen-ting ROS generation (Taddeo, Laker et al . 2014) . Increased sensitivity to mitochondrial Ca2+-induced mPTP opening was also demonstrated in
atrial appendages isolated from diabetic patients (Anderson, Rodriguez et al . 2011) . Conversely, activation of the Ca2+ sensing receptor, which is
downregulated in diabetes, improves Ca2+ homeostasis and delays mPTP
opening induced by high glucose in neonatal rat cardiomyocytes (Wang, Gao et al . 2017), while insulin treatment and exercise attenuate accele-rated mPTP opening in cardiac mitochondria from diabetic rats (da Silva, Natali et al . 2015) . Interestingly, mitochondrial subpopulations seem to differentially react to diabetic conditions . In STZ-induced diabetic mice hearts, interfibrillar mitochondria exhibit enhanced mPTP opening, in-creased cyclophilin D and apoptosis, in contrast to normal function of subsarcolemmally located mitochondria (Williamson, Dabkowski et al . 2010) .
ROS and mPTP. Diabetes enhances ROS production by different mec-hanisms, including alteration in substrate utilization, as described above, which may lead to accelerated/enhanced mPTP opening (Sedlic, Pravdic et al . 2010) Additional mechanisms include high glucose-induced NADPH
oxidase activation in cardiomyocytes, which produces superoxide ani-on (Privratsky, Wold et al . 2003; Balteau, Tajeddine et al . 2011; Maalouf, Eid et al . 2012) . Lipoxygenases may also generate excessive amounts of ROS in the diabetic (STZ-induced) heart, which is linked to diabetic car-diomyopathy (Suzuki, Kayama et al . 2015) . STZ-induced diabetes lowers SOD expression and increases expression of NADPH oxidase, effects lin-ked to accelerated mPTP opening (da Silva, Natali et al . 2015; Yu, Gong et al . 2017) . Increased oxidative stress in diabetic heart may also arise from downregulation of other antioxidants, including glycation-induced inactivation of thioredoxin-1 (Wang, Lau et al . 2010) . The importance of antioxidative defense is underlined by catalase overexpression reducing diabetes-induced heart injury and dysfunction (Ye, Metreveli et al . 2004; Cong, Ruan et al . 2015; Wang, Tao et al . 2017) . This is further supported by findings that dysfunction of mitochondrial aldehyde dehydrogenase 2 (ALDH2) promotes myocardial damage (reviewed in (Chen, Ferreira et al. 2014) . ALDH2 participates in alcohol metabolism, but also detoxifies ROS products such as 4-hydroxynonenal and ameliorates cardiac injury infli-cted by oxidative stress (Ma, Guo et al . 2011) . STZ-diabetic mice that are knockout for ALDH2 have impaired glucose uptake, altered metabolic pro-file and diastolic dysfunction (Wang, Fan et al. 2016).
Calcium and mPTP. Disorders of mitochondrial Ca2+ homeostasis may
not only trigger mPTP opening in diabetic hearts, but it may also impair normal mitochondrial function and respiration (Griffiths and Rutter 2009) . High glucose can acutely increase cytosolic Ca2+ in heart and brain by
in-ducing its release from the sarcoplasmic reticulum (Erickson, Pereira et al . 2013; Sorrentino, Borghetti et al . 2017) . The release of Ca2+ from
sarcopla-smic reticulum is mediated by O-GlcNAcylation of calcium-calmodulin de-pendent protein kinase II (CaMKII) in turn phosphorylating and activating RyR in rat myocytes (Erickson, Pereira et al . 2013) . In addition to arrhyt-hmogenic effects, such Ca2+ release may also cause mitochondrial Ca2+
overload triggering mPTP opening, because of efficient transport of Ca2+
from sarcoplasmic reticulum to mitochondria governed by the juxtaposi-tion of sarcomplasmic inositol 1, 4, 5-trisphosphate receptors and RyR and the mitochondrial Ca2+ uniporter (mCU) (Ruiz-Meana, Fernandez-Sanz
et al . 2010; Qi, Li et al . 2015) . Elevated CaMKII expression is observed in patients with a failing heart, further supporting its role in cardiac pathop-hysiology (Hoch, Meyer et al . 1999) . Cytosolic Ca2+ overload in diabetics
may also stem from a reduced diastolic clearance, as indicated by prolon-ged Ca2+ transients in high glucose-treated neonatal rat cardiomyocytes
(Clark, McDonough et al . 2003), possibly mediated by the O-GlcNAcylati-on-dependent downregulation of sarcoendoplasmic reticulum Ca2+
-ATPa-se 2a (SERCA 2a), the ion pump responsible for diastolic re-uptake of Ca2+
into the sarcoplasmic reticulum .
However, controversies about the regulation of mitochondrial Ca2+
rema-in . High glucose was found to downregulate mCU (Diaz-Juarez, Suarez et al . 2016), which decreases cardiac performance and increases cardio-myocytes death during I-R (Rasmussen, Wu et al . 2015) . Moreover, mito-chondria with lower mCU numbers have increased cytosolic Ca2+ levels,
possibly due to inadequate ATP production and increased oxygen con-sumption, which may serve to lower mitochondrial Ca2+ uptake (Suarez,
Hu et al . 2008) as a compensatory mechanism to limit mitochondrial Ca2+
overload and mPTP opening (Diaz-Juarez, Suarez et al . 2016) . Likewise, insulin resistance may decrease mitochondrial Ca2+ uptake, as insulin
fa-iled to increase Ca2+ transient amplitude in paced cardiomyocytes of ob/
ob mice (Fauconnier, Lanner et al. 2005). However, rat H9c2 myoblasts,
endothelial and HeLA cells, and STZ-induced diabetic rat heart showed an opposite response, i .e . an increase in mitochondrial Ca2+ in response
to high glucose (Hou, Zhang et al . 2013; da Silva, Natali et al . 2015; Duan, Yin et al . 2015; Zu, Wan et al . 2015) . In addition, high palmitate treatment induced mitochondrial depolarization and increased mitochondrial Ca2+
uptake after release from the sarcoplasmic reticulum in isolated cardio-myocytes (Joseph, Barca et al . 2016) .
Collectively, these data corroborate an impairment of mitochondrial Ca2+
handling in diabetes . Due to inconsistency between effects of high glu-cose and FAs on mitochondrial Ca2+ uptake, with studies shown both an
increase and decrease, it needs to be determined whether accelerated mPTP opening in diabetes occurs (partly) in response to enhanced mitoc-hondrial Ca2+ uptake .
4. Implications of hyperglycemia on protective
conditioning strategies: the role of mitochondria
Hyperglycemia has been demonstrated to be an independent predictor of death after myocardial infarction . Similarly, hyperglycemia has been shown to adversely impact outcome in patients undergoing surgery and anesthe-sia . However, clinical strategies to address the increased cardiovascular risk in patients with diabetes and AHG have been disappointing .
Over the last decades, various conditioning strategies have been discove-red that share the potential to substantially lower the impact of ischemic events on the heart . Such strategies include ischemic pre-and post-condi-tioning (IPC/ IPoC), remote ischemic condipost-condi-tioning (RIC) and pharmacolo-gical conditioning, including anesthetic pre- and post-conditioning (APC/ APoC) . While these strategies offered cardioprotection in the experimental setting in normal animals, they generally failed to protect the heart in dia-betic animals and in animals with induced acute hyperglycemia . Moreover, in patients either with or without diabetes, the results of the majority of the clinical trials exploring the effect of various preconditioning strategies have been inconsistent (Kunst and Klein 2015; Xie, Zhang et al. 2018). There are many potential explanations for the negative findings of these clinical studies, most likely involving altered intracellular cytoprotective si-gnaling mechanisms . In fact, evidence suggests that changes in NO and mitochondrial bioenergetics may fundamentally underlie the susceptibility to myocardial injury in diabetes and AHG and likely play critical roles in mediating the loss of IPC/APC induced cardioprotection .
4.1.
Cardiac & Remote Ischemic Conditioning
Cardiac ischemic preconditioning (IPC) or postconditioning (IPoC), i .e . the induction of one or a series of brief period(s) of coronary artery occlusi-on prior to or after a myocardial infarctiocclusi-on, respectively, cocclusi-onfers substan-tial cardioprotection and limits infarct size (Murry, Jennings et al . 1986; Yellon, Alkhulaifi et al . 1993) . Its cardioprotective cellular signaling encom-passes a complex machinery and several endpoints, one of which is re-duced mPTP opening (Murry, Jennings et al . 1986; Murphy and Steen-bergen 2008) . Conditioning signaling pathways are initiated by several upstream factors, including the intracellular molecules NO and ROS and extracellular autacoid factors (mainly adenosine, bradykinin and opioid peptides) through specific receptors . Briefly, cardioprotection elicited by IPC and IPoC is mediated by activation of different adenosine receptor subtypes (A1 and A3) (Heusch 2010) expressed on cardiomyocytes (Xin, Yang et al . 2012) (Liu, Thornton et al . 1991; Liu, Richards et al . 1994; Mc-Cully, Toyoda et al . 2001) . In contrast to IPC, A2A and A2B subtypes mediate protection by postconditioning, with A1 and A3 subtypes being irrelevant (McIntosh and Lasley 2012) . IPoC is also mediated through bradykinin (B) receptors (Penna, Mancardi et al . 2008) . The cytoprotective signal is furt-her transduced through one of the three major pathways: the reperfusion
injury survival kinases or RISK pathway (PI3K/Akt and MEK/ERK), the sur-vivor activating factor enhancement or SAFE pathway (JAK/STAT3) and (eNOS)/cGMP/PKG pathway . These signaling pathways have been discu-ssed extensively in other reviews (Yellon and Downey 2003; Cohen and Downey 2015). Finally, the signaling trough these receptor-initiated pat-hways may be critically dependent on the concentration and organization of its components by caveolins, i .e . scaffolding proteins within speciali-zed, cholesterol-rich domains in the plasma membrane called caveolae . Interestingly, the disruption of caveolar structure both blocks IPC-induced activation of the Akt/eNOS/NO/SNO signaling pathway (Sun, Kohr et al . 2012) and abolishes the cardioprotective effect of IPoC by eliciting mito-chondrial dysfunction due to inhibition of the interaction of ERK1/2 with mitochondria (Garcia-Nino, Correa et al . 2017) .
Remote ischemic conditioning is a distinct conditioning strategy - without the need for temporary coronary artery occlusion – showing similar car-dioprotective effects as IPC and IPoC in experimental studies . This no-ninvasive procedure constitutes a repeated inflation and deflation of the blood-pressure cuff to induce transitory ischemia and reperfusion in the arm or leg (Cheung, Kharbanda et al . 2006; Ali, Callaghan et al . 2007; Bot-ker, Kharbanda et al . 2010; Hausenloy, Candilio et al . 2015) . The mecha-nistic concept is that RIC induces release of humoral factor(s)/protecti-ve signal(s), which subsequently confactor(s)/protecti-vey(s) protection to the myocardium (Hausenloy and Yellon 2008) . Recent evidence in humans suggests that the humoral factor is nitrite, which is derived from NO released during re-active hyperemia because of sheer stress mediated activation of eNOS (Rassaf, Totzeck et al . 2014) . On the cardiomyocyte level, RIC activates similar downstream signaling pathways as IPC and IPoC (Tamareille, Ma-teus et al . 2011; Shi and Vinten-Johansen 2012), but is not mediated by adenosine receptors, at least in pig (Hausenloy, Iliodromitis et al . 2012) . Rather, RIC seems to be conferred by bradykinin, as cardioprotection by mesenteric RIC is abrogated by B2 receptor blockade in rat (Schoemaker and van Heijningen 2000), while RIC of the forearm down-regulates B1 and B2 receptors (in neutrophils) for up to 24 h in humans undergoing coronary artery bypass surgery (Saxena, Aggarwal et al . 2013) . Bradykinin signaling results in the formation of PI3K, which as described above both activates Akt (RISK), as well as NOS/cGMP/PKG, and results in cytoprote-ction (Saxena, Shaw et al . 2011) . The role of kinins in this setting appears to be two-fold . Depending on the pattern of bradykinin release and inte-ractions with other mediators involved in conditioning, bradykinin elicits
an inflammatory response through the induction of chemotaxis and the activation of neutrophils (Ehrenfeld, Millan et al . 2006), yet also triggers an anti-inflammatory response through downregulation of the kinin receptors in neutrophils, resulting in protection against IR injury . Despite these posi-tive findings, a recent meta-analysis of patients undergoing cardiac opera-tions indicates that RIPC does not reduce acute myocardial infarction and mortality (Xie, Zhang et al. 2018).
Role of mitochondria in condition strategies. Although the downstream protective mechanisms of conditioning strategies are still incompletely un-derstood, mitochondria constitute important players/mediators on which various signal transduction routes converge . Several important pro-sur-vival mediators have been identified so far, including protein SNO, mitoc-hondrial KATP channel activation, mitochondrial hexokinase 2 (mtHK2) bin-ding/association, mitochondrial Ca2+ loading, and ROS/NO production (or
nitroso-redox balance) .
Protection via protein SNO particularly involves SSM mitochondria, in whi-ch IPC induced nitrosylation of mitowhi-chondrial proteins increases associa-tion with caveolin-3/eNOS/NO in mice hearts (Sun, Nguyen et al . 2015) . Furthermore, SSM are more responsive to IPoC (Chen, Paillard et al. 2012) and pharmacological signaling compared to IFM (Holmuhamedov, Ober-lin et al. 2012). Further, a number of mitochondrial targets of SNO in IPC have been identified (eg. F1-ATPase) that reduce intracellular Ca2+ loading,
restore nitroso-redox balance during I/R and inhibit mPTP opening (Sun, Morgan et al . 2007; Kohr, Sun et al . 2011) . The role of SNO mediated car-dioprotective effects is further substantiated in RIC, in which nitrite relea-se from femoral arteries is converted to SNO by myoglobine (Mb) in car-diomyocytes, resulting in nitrosylation of complex I (Rassaf, Totzeck et al . 2014) . Local activation of nitrite by Mb was confirmed by the abrogation of RIC induced cardioprotection in Mb-/- knockout mice .
Mitochondrial connexin 43 (mtCx43) is activated by IPC via SNO, which serves to elevate ROS that participate in the preconditioning signaling cascade (Soetkamp, Nguyen et al . 2014) . A 7-day long treatment with HG significantly reduced mtCx43 expression in retinal endothelial cells, which also showed increased mitochondrial fragmentation and cytochrome c re-lease (Trudeau, Muto et al . 2012) . Whether mtCx43 activity in the heart is also impaired by diabetes/HG remains to be examined .
Apart from SNO, different conditioning strategies may also offer protecti-on by activatiprotecti-on of mitochprotecti-ondrial KATP channels . Whereas its involvement
in preconditioning has been unequivocally demonstrated, it is still unclear whether KATP channel opening initiates the protection or merely mediates protection via the surface receptors initiated NO/cGMP/PKG/PKC signa-ling, leading to its opening, as suggested in more recent literature (Cuong, Kim et al . 2006; Costa, Pierre et al . 2008) reviewed in (Cohen and Downey 2015). Further data suggest that PKC-ε activation links cytosolic PKG to opening of mitochondrial KATP channels (Ohnuma, Miura et al . 2002) . This explains why opening of the mitochondrial KATP is PKG-dependent, des-pite their location on IMM making them inaccessible to cytosolic PKG . KATP channel opening subsequently increases K+ conductance, stabilizing
the resting membrane potential with shortening of the action potential and reduction of Ca2+ influx, resulting in increased intracellular ATP because
of reduced contractility and prevention of Ca2+ overload (Tinker, Aziz et al .
2014) .
In recent years, the glycolytic enzyme mtHK2 emerged as determinant of infarct size as mitochondria with bound mtHK2 are more resistant to mPTP opening (Chiara, Castellaro et al . 2008) . Similarly, a positive corre-lation between dissociation of mitochondrial HK2 and infarct size has been found (Pasdois, Parker et al . 2013) Transient ischemia, such as IPC, induces the translocation and physical binding of HK2 to mitochondria, which plays an essential role in maintenance of mitochondrial polarizati-on (Smeele, Southworth et al . 2011) . mtHK2 cardioprotective mechanisms preserve OMM permeability during I/R injury ensuring stabilization of ΔΨm, prevention of OMM rupture and Cyt C release, and reduction of ROS (Ne-derlof, Eerbeek et al. 2014). Furthermore, mtHK2 may also determine the direction of cardiac metabolic flux and modulate metabolic (in)flexibility, by preventing acidosis precipitated by improved coupling of glycolysis and glucose oxidation and inhibition of FA oxidation (Nederlof, Eerbeek et al. 2014) . Thus, mtHK2 constitutes one of the important end-effectors of IPC (Pasdois, Parker et al . 2012; Nederlof, Gurel-Gurevin et al . 2016) .
Loss of effectiveness of iPC in diabetes. Ample evidence exists that IPC is less effective in diabetes and AHG, as reviewed by Miki, T . et al . (Miki, Itoh et al . 2012) . Kersten et al . (2000) were the first to report that hyperglycemia abolished the IPC protection (Kersten, Schmeling et al . 1998; Kersten, Toller et al. 2000). Failure of IPC in diabetes and hypergly-cemia is hallmarked by a loss of adenosine receptor-Gi protein coupling (Green and Johnson 1991), absence of activation of mitochondrial KATP (Kersten, Montgomery et al. 2001; Fancher, Dick et al. 2013) and decrease in mRNA/protein expression of K channels in diabetic rats (Chen, Cheng
