**6. Mitochondrial dysfunction play an important role in septic cardiomyopathy**

Mitochondrial dysfunction -present in sepsis but also in many other pathological states such as chronic heart failure, diabetes mellitus, metabolic syndrome and several neurodegenerative diseases- is a syndrome characterized by a picture of reduced oxygen consumption, ineffective ATP production, increased accumulation of ROS and calcium in the organelle leading to the opening in the internal mitochondrial membrane of a high conduction pore allowing the entry of solutes until a molecular weight of 1.500, with mitochondrial swelling, collapse of the electron gradient, release of proapoptotic proteins and death of the organelle (mitoptosis) followed by mitophagy: this phenomenon called mitochondrial permeability transition pore (MPTP) has an important role in sepsis and its inhibition may experimentally improve septic myocardial dysfunction [28,121-123].

Mitochondrial functionality is regulated by the equilibrium between de novo mitochondrial formation (mitochondrial biogenesis) and mitochondrial autophagy (mitophagy), frequently preceding whole cell autophagy. This latter process, taking place only in deenergized organelles, is very important for cell survival because it allows the elimination of dysfunctional mitochondria, producing too much radical oxygen species(ROS) and assures the high quality of the mitochondria provided with a high membrane potential and resistant to MPTP [124]. Moreover autophagy has been shown to be protective in ischemiareperfusion injury (IRI) and inflammation by inhibiting the exposition of mtDNA and the consequent activation of the inflammasome NALP3 [125]. Autophagy involves a complex molecular machinery including more than 30 Atg (AutophaGy-related) proteins and 50 lysosomal hydrolases [126].

The Role of Altered Lipid Metabolism in Septic Myocardial Dysfunction 339

1a mediated biogenesis. For biogenesis to occur it is necessary mitochondria undergo the processes of fusion and fission regulated respectively by mitofusins (Mfn1-2) and by

The first description of mitochondrial dysfunction in sepsis dates back to 1970' [135,136], but the initial studies were not reliable due to inadequate techniques to test mitochondrial function. More recent studies however linked the altered oxygen utilization and ATP production in sepsis to mitochondrial dysfunction (137,138), even though there are substantial differences among the sepsis model induced by LPS administration in which is present a widespread mitochondrial damage and the CLP model characterized by a less evident mitochondrial pathology because damaged organelles undergo autolysis and autophagy [139]. Several alterations have been described in mitochondria in experimental and human sepsis both morphological (swelling, destruction of cristae, decreased electron density, breaks in membrane and altered outer membrane integrity) and biochemical (release of cytochrome c, decreased mtDNA copy number, reduced activity of Complexes I and III) [28,140,141]; their extent has been related to outcome [142]. Several studies showed an initial increase of mitochondrial activity in sepsis, followed by a reduction. Singer [132] hypothesized that, after a metabolic stimulation of mitochondrial function. caused by the excessive inflammatory response, in the successive phase there is a metabolic downregulation determined by the mitochondrial dysfunction. In this latter phase a stimulation of oxygen consumption (VO2) would be harmful because it forces mitochondria, adapted to a reduction of metabolism, to overwork without being prepared for it. The restorative phase of sepsis is linked to the activation of mitochondrial biogenesis; also LPS may have a role in inducing this process [140,143]. The absence of relevant cell death by apoptosis or necrosis in parenchymatous organs and the ability also of the tissues with less regenerative capacity to recover led to think that multiple organ dysfunction syndrome (MODS) were related mainly to a functional compromise mediated by endocrine and metabolic alterations and provided with a protective role [144-146]. In sepsis a paradoxical increase has been observed, in some tissue, in blood oxygen saturation [147], related to outcome and linked to an altered mitochondrial oxygen utilization (cytopathic hypoxia) [148], which has been attributed to nitric oxide (NO) overproduction by mitochondrial nitric oxide synthase (mtNOS) [149,150] or other NOS isoforms: NO competitively inhibits oxygen binding to cytochrome c oxydase (COX) determining a tonic inhibition of mitochondrial oxygen consumption, reversible on inhibiting its excessive production [151]. The exact role of NO inhibition by COX in vivo in physiologic and pathologic conditions is however an unsettled problem [152]. Other gaseous molecules share with NO the ability to interact with COX, to inhibit it competitively and to regulate at this level the body oxygen consumption [153,154]; they are called gasotransmitters and include carbon monoxide (CO) and Hydrogen Sulphide (H2S), this latter molecule was first shown able to induce experimentally in mice, but not in pigs and superior mammals, a particular state of depression of the metabolism and of the resistance to hypoxia, called "suspended animation" (155]. A similar observation has been made with NO and CO in Drosophila [156,157]. Exploiting this property of gasotransmitters in the treatment of myocardial and cerebral hypoxia\ischemia is tempting, considered also the limitations of the current treatment

dynamin-related protein(DRP1 and OPA1) [126].

The mitochondrial biogenesis is orchestrated by PPARand \by the ERR and above all by the network PGC-1PGCand PGC-1 related coactivator (PRC), factors enriched in tissues with high oxidative capacity (heart, muscle, brown adipose tissue and kidneys) and highly inducible by starvation, exercise and cold acclimation (126-128); PGC-is also induced by Ils [129]Mice with single deficiency in PGC-1 or PGC- display only minimal alteration under nonstressed conditions and have a normal mitochondrial mass [130,131] whereas mice with double deletion of both PGC-1 forms died shortly after birth with small hearts, bradycardia, intermittent heart block, and a markedly reduced cardiac output [132], allowing to think that one isoform may compensate for the other. Mitochondrial biogenesis to take place needs the transcription of the genes controlled by mitochondrial DNA, a process activated by the mitochondrial transcription factor A (Tfam), a nuclearencoded transcription factor essential for replication, maintenance and transcription of mitochondrial DNA [133]. The majority of the genes codifying mitochondrial proteins are however present in nuclear DNA and are activated, such as Tfam, by the nuclear respiratory factors 1 and 2 (NRF1 and NRF2). PGC-1 coordinates the function of NRF1, NRF2 and Tfam, regulating mitochondrial biogenesis and, through the complex PPARs-ERRs, controls FAO such that a prompt adaptive response to physiological events as cold acclimation, exercise, starvation etc may be possible with selection of the most suitable level of FFA oxidation needed in the particular situation.

Cardiac contractile activity is linked to the mitochondrial biogenesis because an increase in intracellular calcium increases the expression of PGC-1 via calcineurin (CaN) -activating also PPARpromoter and FAO- and Ca-calmodulin dependent kinase (CaMK), a process requiring cAMP response element-binding protein (CREB)(126,133): in this way the contractile activity is strictly linked to the metabolism and to the mitochondrial oxidative function.

PGC-1 expression is also controlled hormonally by tyroid hormones (TH), metabolically via adrenergic system and cAMP, by AMPK and p38 MAPK through phosphorylation [126,133] and by eNOS-produced NO. The cellular mitochondrial content is proportional to the energy requirement [75], mitochondrial biogenesis is therefore controlled by the two main energy sensors in the cell: Sirt1(silent mating type information regulator 2 homolog 1) and AMPK act coordinately to control PGC-1 activity respectively through deacetylation and phosphorylation. Sirtuins are class III histone deacetylating enzymes supplied by AMPK with the adequate level of the NAD+, essential activator of Sirt1 that in turn deacetylates PGC-1 and activates its gene expression [75]. Sirt1 is also an important in vivo regulator of mitophagy by deacetylation of the autophagy genes Atg5, Atg7 and Atg8 [134], linking this protective mechanism, devoted to the clearance of depotentiated mitochondria, with PGC-

1a mediated biogenesis. For biogenesis to occur it is necessary mitochondria undergo the processes of fusion and fission regulated respectively by mitofusins (Mfn1-2) and by dynamin-related protein(DRP1 and OPA1) [126].

338 Lipid Metabolism

lysosomal hydrolases [126].

needed in the particular situation.

the high quality of the mitochondria provided with a high membrane potential and resistant to MPTP [124]. Moreover autophagy has been shown to be protective in ischemiareperfusion injury (IRI) and inflammation by inhibiting the exposition of mtDNA and the consequent activation of the inflammasome NALP3 [125]. Autophagy involves a complex molecular machinery including more than 30 Atg (AutophaGy-related) proteins and 50

The mitochondrial biogenesis is orchestrated by PPARand \by the ERR and above all by the network PGC-1PGCand PGC-1 related coactivator (PRC), factors enriched in tissues with high oxidative capacity (heart, muscle, brown adipose tissue and kidneys) and highly inducible by starvation, exercise and cold acclimation (126-128); PGC-is also induced by Ils [129]Mice with single deficiency in PGC-1 or PGC- display only minimal alteration under nonstressed conditions and have a normal mitochondrial mass [130,131] whereas mice with double deletion of both PGC-1 forms died shortly after birth with small hearts, bradycardia, intermittent heart block, and a markedly reduced cardiac output [132], allowing to think that one isoform may compensate for the other. Mitochondrial biogenesis to take place needs the transcription of the genes controlled by mitochondrial DNA, a process activated by the mitochondrial transcription factor A (Tfam), a nuclearencoded transcription factor essential for replication, maintenance and transcription of mitochondrial DNA [133]. The majority of the genes codifying mitochondrial proteins are however present in nuclear DNA and are activated, such as Tfam, by the nuclear respiratory factors 1 and 2 (NRF1 and NRF2). PGC-1 coordinates the function of NRF1, NRF2 and Tfam, regulating mitochondrial biogenesis and, through the complex PPARs-ERRs, controls FAO such that a prompt adaptive response to physiological events as cold acclimation, exercise, starvation etc may be possible with selection of the most suitable level of FFA oxidation

Cardiac contractile activity is linked to the mitochondrial biogenesis because an increase in intracellular calcium increases the expression of PGC-1 via calcineurin (CaN) -activating also PPARpromoter and FAO- and Ca-calmodulin dependent kinase (CaMK), a process requiring cAMP response element-binding protein (CREB)(126,133): in this way the contractile

PGC-1 expression is also controlled hormonally by tyroid hormones (TH), metabolically via adrenergic system and cAMP, by AMPK and p38 MAPK through phosphorylation [126,133] and by eNOS-produced NO. The cellular mitochondrial content is proportional to the energy requirement [75], mitochondrial biogenesis is therefore controlled by the two main energy sensors in the cell: Sirt1(silent mating type information regulator 2 homolog 1) and AMPK act coordinately to control PGC-1 activity respectively through deacetylation and phosphorylation. Sirtuins are class III histone deacetylating enzymes supplied by AMPK with the adequate level of the NAD+, essential activator of Sirt1 that in turn deacetylates PGC-1 and activates its gene expression [75]. Sirt1 is also an important in vivo regulator of mitophagy by deacetylation of the autophagy genes Atg5, Atg7 and Atg8 [134], linking this protective mechanism, devoted to the clearance of depotentiated mitochondria, with PGC-

activity is strictly linked to the metabolism and to the mitochondrial oxidative function.

The first description of mitochondrial dysfunction in sepsis dates back to 1970' [135,136], but the initial studies were not reliable due to inadequate techniques to test mitochondrial function. More recent studies however linked the altered oxygen utilization and ATP production in sepsis to mitochondrial dysfunction (137,138), even though there are substantial differences among the sepsis model induced by LPS administration in which is present a widespread mitochondrial damage and the CLP model characterized by a less evident mitochondrial pathology because damaged organelles undergo autolysis and autophagy [139]. Several alterations have been described in mitochondria in experimental and human sepsis both morphological (swelling, destruction of cristae, decreased electron density, breaks in membrane and altered outer membrane integrity) and biochemical (release of cytochrome c, decreased mtDNA copy number, reduced activity of Complexes I and III) [28,140,141]; their extent has been related to outcome [142]. Several studies showed an initial increase of mitochondrial activity in sepsis, followed by a reduction. Singer [132] hypothesized that, after a metabolic stimulation of mitochondrial function. caused by the excessive inflammatory response, in the successive phase there is a metabolic downregulation determined by the mitochondrial dysfunction. In this latter phase a stimulation of oxygen consumption (VO2) would be harmful because it forces mitochondria, adapted to a reduction of metabolism, to overwork without being prepared for it. The restorative phase of sepsis is linked to the activation of mitochondrial biogenesis; also LPS may have a role in inducing this process [140,143]. The absence of relevant cell death by apoptosis or necrosis in parenchymatous organs and the ability also of the tissues with less regenerative capacity to recover led to think that multiple organ dysfunction syndrome (MODS) were related mainly to a functional compromise mediated by endocrine and metabolic alterations and provided with a protective role [144-146]. In sepsis a paradoxical increase has been observed, in some tissue, in blood oxygen saturation [147], related to outcome and linked to an altered mitochondrial oxygen utilization (cytopathic hypoxia) [148], which has been attributed to nitric oxide (NO) overproduction by mitochondrial nitric oxide synthase (mtNOS) [149,150] or other NOS isoforms: NO competitively inhibits oxygen binding to cytochrome c oxydase (COX) determining a tonic inhibition of mitochondrial oxygen consumption, reversible on inhibiting its excessive production [151]. The exact role of NO inhibition by COX in vivo in physiologic and pathologic conditions is however an unsettled problem [152]. Other gaseous molecules share with NO the ability to interact with COX, to inhibit it competitively and to regulate at this level the body oxygen consumption [153,154]; they are called gasotransmitters and include carbon monoxide (CO) and Hydrogen Sulphide (H2S), this latter molecule was first shown able to induce experimentally in mice, but not in pigs and superior mammals, a particular state of depression of the metabolism and of the resistance to hypoxia, called "suspended animation" (155]. A similar observation has been made with NO and CO in Drosophila [156,157]. Exploiting this property of gasotransmitters in the treatment of myocardial and cerebral hypoxia\ischemia is tempting, considered also the limitations of the current treatment with hypothermia of the hypoxic organ injury, even though they also have a high toxicity [158]. To the picture of metabolic hibernation may contribute, besides NO, also CO and H2S increased during sepsis [159-161].

The Role of Altered Lipid Metabolism in Septic Myocardial Dysfunction 341

role of PARP has also been showed by Soriano in the setting of human septic myocardiopathy [166], where the degree of myocardial compromise related well with the staining for the product of activated PARP [poly(ADP-ribose) (PAR)], and in many experimental model of cardiovascular diseases[167] in which PARP inhibitors provide significant benefits. Interestingly the metabolic hibernation in sepsis according to our hypothesis has a different starting point from the true mammals hibernation in which the scarcity of nutrients leads to increased levels of NAD+ and AMP with AMPK and SIRT1 activation and deacetylation of clock protein BMAL1 and the opposite switch from glucose to lipid metabolism [168]. The connection between bioenergetics impairment and inflammation in sepsis has also been studied by Liu and colleagues [169-170] that underline the role of Sirt1 in reprogramming immune response through the silencing of the proinflammatory genes and the coordination of the metabolic response. In the experimental setting many attempts have been made to treat mitochondrial dysfunction by supplying substrates, cofactors, mitochondrial-targeted antioxidants, scavengers of mitochondrial ROS and membrane stabilizers displaying improvements at many levels: ATP production, reduced production of ROS in mitochondrial ETC, reduction of apoptosis and prevention of MPTP opening [171-174]. Apoptosis may also have a role in septic mitochondrial dysfunction as it has in CHF. Lancel and colleagues demonstrated in cardiomyocytes isolated 4 hours after endotoxin injection in rats, an increase in the release of caspase-3, -8, and -9-like activities, associated with destruction of the sarcomeric structure and cleavage of components of the cardiac myofilament [175]. Simvastatin, whose use has been proposed in sepsis, may protect against apoptosis induced by

Staphylococcus aureus alpha-toxin [176].

**NF-kB PGC-1**

**AMPK**

**Figure 5.** Mitochondrial dysfunction in sepsis. See text for explanation

**SIRT1**

**Prolonged depression of mitochondrial biogenesis and function**

**NAD+**

**HIF-1 activation**

**Sepsis, LPS Inflammation**

> **PARP activation**

**Increased oxidative and nitrosative stress**

> **PPAR, ERR LXR, FXR**

In addition to the antagonism against COX the higher levels of NO, in presence of an increased oxidative stress, by reacting with superoxide, may further injure mitochondria determining the formation of the toxic peroxynitrite able to damage the Complex I and III of the ETC, the membrane lipid and mitochondrial and nuclear DNA. The available data however cannot be explained only by this mechanism; the metabolic ebb phase, coincident with CARS and silencing of proinflammatory genes, is characterized by a profound state of mitochondrial depression, reduced oxidative metabolism and oxygen consumption leading to a state of metabolic hibernation responsible for the depressed cell function during MODS in sepsis; mitochondrial depression is however completely reversible on recovery from sepsis. Recently Schilling and colleagues showed that LPS administration through TLR4 stimulation may induce a fuel switch from preferential myocardial use of lipid to glycolysis, similar to the one observed in the failing heart, with decreased mitochondrial oxidation, lipid accumulation within them and a suppression of the genes encoding PGC-1 and : the genetic overexpression of PGC-1 may restore mitochondrial function independently of the inflammatory response [162]. Smeding and colleagues [163], on the other hand, observed that the sirtuin activator resveratrol protect mitochondria in a mice model of CLP improving myocardial dysfunction, preserving energy production but without improving survival. At this point some question should be raised: could we think that the mitochondrial dysfunction is a reversible event even in the presence of a continuing sepsis? Which may be the correct link and the ordinated succession among all the alterations described?

We try to reconstruct the successive steps leading to mitochondrial dysfunction and to metabolic hibernation in sepsis (fig. 5). The first step is the hypercatabolic phase with the transient activation of mitochondrial oxidative phosphorylation (OXPHOS) activated by the acute hormonal stress response in the very early phase, with increased oxidative and nitrosative stress damaging ETC and inducing a downregulation of many of its proteins. Immediately after, in the inflammatory setting of sepsis, occurs the shift to glycolysis and the activation of mito autophagy. SIRT1 is rapidly deactivated by the reduced function of AMPK, the decreased NAD+ levels due to the activation of PARP and by the increased expression of miR199a [66]. The stimulation of this enzyme in sepsis was first showed by Fink [148]; recently Zhang and colleagues, after LPS administration in mice, observed that the plasma levels of ATP, SIRT1 and AMPK decline precipitously after 10' and remain below the detection level for up to 12 h while HIF-1 increases contemporarily along with an increase in mitochondrial autophagy. Importantly all these effects are absent in wild type mice treated with resveratrol but not in Sirt1 liver knockout mice [164]. This study further confirm the critical role of SIRT1 that would be downregulated in sepsis for many causes: increased HIF-1 expression, increased miR199a expression**,** drastic reduction of NAD+ by activation of PARP consuming it and reduced activity of AMPK supplying it. The main consequence of these modifications are the reduced deacetylation by SIRT1 and phosphorylation by AMPK of PGC-1, also PGC seems to be reduced in muscle during sepsis [162,165] making impossible a compensation. The role of PARP has also been showed by Soriano in the setting of human septic myocardiopathy [166], where the degree of myocardial compromise related well with the staining for the product of activated PARP [poly(ADP-ribose) (PAR)], and in many experimental model of cardiovascular diseases[167] in which PARP inhibitors provide significant benefits. Interestingly the metabolic hibernation in sepsis according to our hypothesis has a different starting point from the true mammals hibernation in which the scarcity of nutrients leads to increased levels of NAD+ and AMP with AMPK and SIRT1 activation and deacetylation of clock protein BMAL1 and the opposite switch from glucose to lipid metabolism [168]. The connection between bioenergetics impairment and inflammation in sepsis has also been studied by Liu and colleagues [169-170] that underline the role of Sirt1 in reprogramming immune response through the silencing of the proinflammatory genes and the coordination of the metabolic response. In the experimental setting many attempts have been made to treat mitochondrial dysfunction by supplying substrates, cofactors, mitochondrial-targeted antioxidants, scavengers of mitochondrial ROS and membrane stabilizers displaying improvements at many levels: ATP production, reduced production of ROS in mitochondrial ETC, reduction of apoptosis and prevention of MPTP opening [171-174]. Apoptosis may also have a role in septic mitochondrial dysfunction as it has in CHF. Lancel and colleagues demonstrated in cardiomyocytes isolated 4 hours after endotoxin injection in rats, an increase in the release of caspase-3, -8, and -9-like activities, associated with destruction of the sarcomeric structure and cleavage of components of the cardiac myofilament [175]. Simvastatin, whose use has been proposed in sepsis, may protect against apoptosis induced by Staphylococcus aureus alpha-toxin [176].

340 Lipid Metabolism

increased during sepsis [159-161].

with hypothermia of the hypoxic organ injury, even though they also have a high toxicity [158]. To the picture of metabolic hibernation may contribute, besides NO, also CO and H2S

In addition to the antagonism against COX the higher levels of NO, in presence of an increased oxidative stress, by reacting with superoxide, may further injure mitochondria determining the formation of the toxic peroxynitrite able to damage the Complex I and III of the ETC, the membrane lipid and mitochondrial and nuclear DNA. The available data however cannot be explained only by this mechanism; the metabolic ebb phase, coincident with CARS and silencing of proinflammatory genes, is characterized by a profound state of mitochondrial depression, reduced oxidative metabolism and oxygen consumption leading to a state of metabolic hibernation responsible for the depressed cell function during MODS in sepsis; mitochondrial depression is however completely reversible on recovery from sepsis. Recently Schilling and colleagues showed that LPS administration through TLR4 stimulation may induce a fuel switch from preferential myocardial use of lipid to glycolysis, similar to the one observed in the failing heart, with decreased mitochondrial oxidation, lipid accumulation within them and a suppression of the genes encoding PGC-1 and : the genetic overexpression of PGC-1 may restore mitochondrial function independently of the inflammatory response [162]. Smeding and colleagues [163], on the other hand, observed that the sirtuin activator resveratrol protect mitochondria in a mice model of CLP improving myocardial dysfunction, preserving energy production but without improving survival. At this point some question should be raised: could we think that the mitochondrial dysfunction is a reversible event even in the presence of a continuing sepsis? Which may be

the correct link and the ordinated succession among all the alterations described?

We try to reconstruct the successive steps leading to mitochondrial dysfunction and to metabolic hibernation in sepsis (fig. 5). The first step is the hypercatabolic phase with the transient activation of mitochondrial oxidative phosphorylation (OXPHOS) activated by the acute hormonal stress response in the very early phase, with increased oxidative and nitrosative stress damaging ETC and inducing a downregulation of many of its proteins. Immediately after, in the inflammatory setting of sepsis, occurs the shift to glycolysis and the activation of mito autophagy. SIRT1 is rapidly deactivated by the reduced function of AMPK, the decreased NAD+ levels due to the activation of PARP and by the increased expression of miR199a [66]. The stimulation of this enzyme in sepsis was first showed by Fink [148]; recently Zhang and colleagues, after LPS administration in mice, observed that the plasma levels of ATP, SIRT1 and AMPK decline precipitously after 10' and remain below the detection level for up to 12 h while HIF-1 increases contemporarily along with an increase in mitochondrial autophagy. Importantly all these effects are absent in wild type mice treated with resveratrol but not in Sirt1 liver knockout mice [164]. This study further confirm the critical role of SIRT1 that would be downregulated in sepsis for many causes: increased HIF-1 expression, increased miR199a expression**,** drastic reduction of NAD+ by activation of PARP consuming it and reduced activity of AMPK supplying it. The main consequence of these modifications are the reduced deacetylation by SIRT1 and phosphorylation by AMPK of PGC-1, also PGC seems to be reduced in muscle during sepsis [162,165] making impossible a compensation. The

**Figure 5.** Mitochondrial dysfunction in sepsis. See text for explanation

#### **6.1. Cardiac function in sepsis**

The pathogenesis of the myocardial dysfunction in sepsis is not definitively explained and is probably caused by a myriad of alterations. The morphological picture recognized in sepsis was formerly considered unspecific but more recently many interesting alterations have been described even though certainly some of them are not the cause but the consequence of the acute cardiac damage. Beside not specific alterations such as cell edema, the already described mitochondrial lesions, damage of cytoplasmic membrane and myofibrils, inflammatory cells infiltration [139,177], more definite biochemical and microscopic changes have been observed with lipid infiltration [178], increased deposits of glycogen and increased levels of glucose uptake, GLUT1 and GLUT4 [141], increased sarcolemmal permeability [179], disruption of sarcolemmal dystrophin and beta-dystroglycan [180].

The Role of Altered Lipid Metabolism in Septic Myocardial Dysfunction 343

independent indexes of LV contractility leading the authors to think that the depressed

The infusion of small doses of LPS in healthy volunteers may reproduce the features of the septic circulatory dysfunction with low systemic resistances and a depressed myocardial function [187]. The mechanism of the altered myocardial function, at first attributed to a myocardial depressant factor (MDF) [192], began to be understood when in 1985 Parrillo [193] showed that the serum of septic humans during the acute phase may depress the contractile function of isolated myocardiocytes; this effect may be prevented, as subsequently demonstrated by Kumar, by immunoadsorption of IL-1 andF 194 Ils may cause myocardial depression both directly or through the NO produced by the induction of iNOS [195]. In the recent years the focus of the research has been turned on the upstream mediators of ILs, the TLRs, that have been shown to be main mediators of myocardial damage in sepsis models of both LPS infusion and CLP and in human sepsis [196]. Their effect as already seen is due to the release, consequent to the inflammatory reaction and to the tissue damage, of DAMP further increasing tissue injury. The main TLRs types expressed in the heart are TLR2 (recognizing lipoteichoic acid) and TLR4 (LPS receptor) but also TLR 3 and 9 are present [196]. Both TLR2 and TLR4 have been demonstrated as essential mediators of the septic myocardial damage: in mice with their genetic deficiency there was a better preservation of cardiac function and a decreased mortality[14,197,198]. Also TLR5, receptor for bacterial flagellin, TLR9 (endocellular recognizing bacterial DNA) and very recently TLR3, endocellular receptor recognizing viral double stranded RNA, have been showed to play a deleterious role in mediating cardiac dysfunction in sepsis [199,200]. This depression of myocardial function is linked to the response of myocardiocytes to the release of alarmins with increased production of pro- and anti-inflammatory ILs, of chemokines and cell adhesion molecules [196]. The connection with the reduced contractile function is represented by the increased production of the 2 small calcium regulated molecules (S100A8 and S100A9) that have been showed with coimmunoprecipitation, able to suppress calcium flux interacting with the RAGE receptor [201]. TLRs lead to iNOS stimulation, NF-kB upregulation and to increased expression of TNF, IL1, IL6, IFN chemokines and cell adhesion molecules responsible for the cellular damage but also for the initiation of the reparative process [202]. The inflammatory reaction, as already seen, is modulated by the appearance of LPS tolerance and proinflammatory genes silencing. As already said, a long term stimulation with LPS or a TLR4 ligand (such as HMBG1) may induce preconditioning [9] and in decreased brain infarct size in mice that

contractility after LPS infusion was due only to loading conditions [191].

were subjected to focal cerebral ischaemia/reperfusion injury [10].

In sepsis there is an increased production of NO by inflammatory cells infiltrating myocardium but also by endothelial and smooth muscle vascular cells due to stimulation of iNOS. The increased NO production causes an impressive vasodilation, a reduced sensitivity to the vasoconstrictors and an increase in capillary permeability. NO has also a depressive action upon the myocardium mediated by the production of cGMP. In fact NO

**6.3. Pathophysiological mechanism** 

#### **6.2. Hemodynamic pattern**

Parker and colleagues in 20 p. in septic shock [181], adequately resuscitated, with the combined use of radionuclide-gated cardiac cineangiography and thermodilution observed important differences in the hemodynamic pattern of survivors and non survivors. Both groups had a hyperdynamic pattern with high cardiac index (CI) and low systemic vascular resistance (SVR) but survivors had an important LV dilatation and a reduced left ventricle ejection fraction (LVEF) normalizing in 7-10 days while non survivors paradoxically showed higher LVEF but did not show ventricular dilatation. On the basis of these data Parker thought that LV dilatation were an adaptive mechanism apt to increase systolic stroke through a better utilization of Starling mechanism (early preload adaptation). Successive studies by echocardiography have not confirmed however a significant LV dilatation while observing an altered biventricular relaxation and a systolic dysfunction of the right ventricle in 41% of the p, isolated or in conjunction with a LV dysfunction [182-185].The apparently better cardiac function in non survivors is currently perceived as a consequence of the persistent septic hemodynamic pattern with decreased SVR causing higher CI and LVEF, both being heavily load-dependent indexes [186]. A LV depressed contractile performance is constant in human and experimental sepsis [187] if related to pre and after-load, it is explained by the global ventricular hypokinesia observed in 60% of the p. in septic shock [188]: the impairment of LV contractility may be unravelled also by the reduction of CI in response to norepinephrine infusion [181]. A feature of septic myocardiopathy is the absence of the increased LV filling pressure, characteristic of cardiogenic shock, due to the dysfunction of the right ventricle [189] and to the slightly increased LV compliance [190]

Currently the hyperkinetic syndrome (small left ventricle, supranormal ejection fraction, tachycardia, high cardiac index) had a 100% mortality rate [186]. In contrast with the majority of other reports Jianhui and colleagues ( describe in mice infused with intraperitoneal LPS and studied with left ventricular (LV) pressure-volume catheters significant drop in LV stroke volume with a significant decrease in LVEF with no apparent change in LV afterload. load-dependent indexes of LV function were markedly reduced at 6 h, including EF, stroke work, and dP/dtmax. In contrast, there was no reduction of loadindependent indexes of LV contractility leading the authors to think that the depressed contractility after LPS infusion was due only to loading conditions [191].

#### **6.3. Pathophysiological mechanism**

342 Lipid Metabolism

**6.1. Cardiac function in sepsis** 

**6.2. Hemodynamic pattern** 

The pathogenesis of the myocardial dysfunction in sepsis is not definitively explained and is probably caused by a myriad of alterations. The morphological picture recognized in sepsis was formerly considered unspecific but more recently many interesting alterations have been described even though certainly some of them are not the cause but the consequence of the acute cardiac damage. Beside not specific alterations such as cell edema, the already described mitochondrial lesions, damage of cytoplasmic membrane and myofibrils, inflammatory cells infiltration [139,177], more definite biochemical and microscopic changes have been observed with lipid infiltration [178], increased deposits of glycogen and increased levels of glucose uptake, GLUT1 and GLUT4 [141], increased sarcolemmal permeability [179], disruption of sarcolemmal dystrophin and beta-dystroglycan [180].

Parker and colleagues in 20 p. in septic shock [181], adequately resuscitated, with the combined use of radionuclide-gated cardiac cineangiography and thermodilution observed important differences in the hemodynamic pattern of survivors and non survivors. Both groups had a hyperdynamic pattern with high cardiac index (CI) and low systemic vascular resistance (SVR) but survivors had an important LV dilatation and a reduced left ventricle ejection fraction (LVEF) normalizing in 7-10 days while non survivors paradoxically showed higher LVEF but did not show ventricular dilatation. On the basis of these data Parker thought that LV dilatation were an adaptive mechanism apt to increase systolic stroke through a better utilization of Starling mechanism (early preload adaptation). Successive studies by echocardiography have not confirmed however a significant LV dilatation while observing an altered biventricular relaxation and a systolic dysfunction of the right ventricle in 41% of the p, isolated or in conjunction with a LV dysfunction [182-185].The apparently better cardiac function in non survivors is currently perceived as a consequence of the persistent septic hemodynamic pattern with decreased SVR causing higher CI and LVEF, both being heavily load-dependent indexes [186]. A LV depressed contractile performance is constant in human and experimental sepsis [187] if related to pre and after-load, it is explained by the global ventricular hypokinesia observed in 60% of the p. in septic shock [188]: the impairment of LV contractility may be unravelled also by the reduction of CI in response to norepinephrine infusion [181]. A feature of septic myocardiopathy is the absence of the increased LV filling pressure, characteristic of cardiogenic shock, due to the dysfunction of the right ventricle [189] and to the slightly increased LV compliance [190]

Currently the hyperkinetic syndrome (small left ventricle, supranormal ejection fraction, tachycardia, high cardiac index) had a 100% mortality rate [186]. In contrast with the majority of other reports Jianhui and colleagues ( describe in mice infused with intraperitoneal LPS and studied with left ventricular (LV) pressure-volume catheters significant drop in LV stroke volume with a significant decrease in LVEF with no apparent change in LV afterload. load-dependent indexes of LV function were markedly reduced at 6 h, including EF, stroke work, and dP/dtmax. In contrast, there was no reduction of loadThe infusion of small doses of LPS in healthy volunteers may reproduce the features of the septic circulatory dysfunction with low systemic resistances and a depressed myocardial function [187]. The mechanism of the altered myocardial function, at first attributed to a myocardial depressant factor (MDF) [192], began to be understood when in 1985 Parrillo [193] showed that the serum of septic humans during the acute phase may depress the contractile function of isolated myocardiocytes; this effect may be prevented, as subsequently demonstrated by Kumar, by immunoadsorption of IL-1 andF 194 Ils may cause myocardial depression both directly or through the NO produced by the induction of iNOS [195]. In the recent years the focus of the research has been turned on the upstream mediators of ILs, the TLRs, that have been shown to be main mediators of myocardial damage in sepsis models of both LPS infusion and CLP and in human sepsis [196]. Their effect as already seen is due to the release, consequent to the inflammatory reaction and to the tissue damage, of DAMP further increasing tissue injury. The main TLRs types expressed in the heart are TLR2 (recognizing lipoteichoic acid) and TLR4 (LPS receptor) but also TLR 3 and 9 are present [196]. Both TLR2 and TLR4 have been demonstrated as essential mediators of the septic myocardial damage: in mice with their genetic deficiency there was a better preservation of cardiac function and a decreased mortality[14,197,198]. Also TLR5, receptor for bacterial flagellin, TLR9 (endocellular recognizing bacterial DNA) and very recently TLR3, endocellular receptor recognizing viral double stranded RNA, have been showed to play a deleterious role in mediating cardiac dysfunction in sepsis [199,200]. This depression of myocardial function is linked to the response of myocardiocytes to the release of alarmins with increased production of pro- and anti-inflammatory ILs, of chemokines and cell adhesion molecules [196]. The connection with the reduced contractile function is represented by the increased production of the 2 small calcium regulated molecules (S100A8 and S100A9) that have been showed with coimmunoprecipitation, able to suppress calcium flux interacting with the RAGE receptor [201]. TLRs lead to iNOS stimulation, NF-kB upregulation and to increased expression of TNF, IL1, IL6, IFN chemokines and cell adhesion molecules responsible for the cellular damage but also for the initiation of the reparative process [202]. The inflammatory reaction, as already seen, is modulated by the appearance of LPS tolerance and proinflammatory genes silencing. As already said, a long term stimulation with LPS or a TLR4 ligand (such as HMBG1) may induce preconditioning [9] and in decreased brain infarct size in mice that were subjected to focal cerebral ischaemia/reperfusion injury [10].

In sepsis there is an increased production of NO by inflammatory cells infiltrating myocardium but also by endothelial and smooth muscle vascular cells due to stimulation of iNOS. The increased NO production causes an impressive vasodilation, a reduced sensitivity to the vasoconstrictors and an increase in capillary permeability. NO has also a depressive action upon the myocardium mediated by the production of cGMP. In fact NO behaves as a double-edged sword because if on one hand depresses myocardial contractility increasing cGMP production, reduces oxygen consumption by antagonizing COX and may induce apoptosis on the other hand it supports relaxation normalizing the stiffness of the septic myocardium, antagonizes the action of the increased endothelin and has a cardioprotective effect through the induction of preconditioning [203]: unselective antagonism of NOS with NNMA is harmful in sepsis because even though it raises blood pressure it decreases cardiac, hepatic and renal blood flow [204]. Many other alterations have been described in septic myocardium: an altered phosphorylation of Troponin I (TnI) in serin 23-24 by Protein Kinase A (PKA), activated by receptors or an upregulation of Protein Phosphatase 2A (PP2A). dephosphorylating TnI [205,206]. The cause of the increased phosphorylation of Tn seems to be the increased -adrenergic activity during the early phases of sepsis followed by a reduction during the late phase characterized by the hypodynamic cardiac pattern. A calcium leak from sarcoplasmic reticulum has been also described in sepsis similar to that observed in CHF and determined also by PKA activation and increased phosphorylation of the calstabin, a protein maintaining in its stable form the sarcoplasmic ATPase assigned to the re-entry of Calcium in sarcoplasmic reticulum after systole [207,208]. Such an alteration could justify both diastolic dysfunction due to the diastolic calcium leak and the systolic dysfunction showing itself in the late phase as expression of depletion of Calcium reserves. An altered autonomic tone with increased in sympathetic activity in the early phases of sepsis and decreased heart rate variability has been described [209].

The Role of Altered Lipid Metabolism in Septic Myocardial Dysfunction 345

for cardiac function in sepsis relating well with left ventricular dysfunction and a poor prognosis, BNP is related to the outcome but only weakly with cardiac filling pressures [212]. A high level of BNP >144 pg/ml predicts cardiac dysfunction with high sensitivity (92%) and high specificity (86%)[213]. BNP upregulation takes part in the reactivation of the fetal gene program and its production has been shown to be increased during inflammatory processes related to the ability of inflammatory ILs and of p38MAPK to increase its secretion increasing its promoter activity [214]. Differently from the regulation of the similar molecule ANP, BNP secretion is regulated more by inflammation than by

The chronically failing heart is certainly energy-deprived [215], as has been ascertained by studies with 31P magnetic resonance (MR) spectroscopy [216] showing the reduced phosphocreatine: ATP ratio, important index of the energetic state of the heart correlated with the indexes of systolic and diastolic function [217]. The cardiac contractile function is inextricably linked to the metabolic function so that the alterations of cardiac metabolism observed in chronic heart failure must affect also the contractile function. As heart failure progresses mithocondrial respiration is gradually compromised preventing that a normal level of FAO may take place. In samples taken from explanted hearts, during transplantation in patients with advanced heart failure, the level of the enzymes of FAO was low compared to the non-failing hearts [218]. The reduction of mitochondrial biogenesis and of FFA oxidation in the advanced phase of CHF is, as already seen, caused by the reactivation of the fetal gene program through the upregulation of HIF-1 and the downregulation of PPAR, \ERR, of SIRT1 and of the PGC-1network demonstrated in myocardial samples of patients with end-stage heart failure compared with explanted non failing hearts [219]. The expression of these transcription factors and coactivators is reduced in pathological hypertrophy and heart failure whereas it is upregulated in physiological forms of hypertrophy related to postnatal growth or exercise training [220] so furnishing a precise boundary between physiologic and pathologic cardiac growth. Interestingly also a constitutive cardiospecific overexpression of PGC-1 in mice (MHC–PGC-1 mice) may damage the heart leading to myocardiopathy with uncontrolled mitochondrial proliferation but only in the neonatal period, whereas the overexpression in adult mice provoked only modest mitochondrial proliferation [221,222]. PGC-1 -\- mice show a reduced capacity both of exercising and of acclimating to cold

The fetal gene program is mediated by an incompletely known but modified expression of various microRNA [224-226]: Sucharov described that changes in the expression level of miR 100 and 133b contribute to the regulation of the fetal gene program [225]. The reduced expression of SIRT1 underlies the depressed mitochondrial biogenesis and FAO in the chronically failing heart. SIRT1 seems to be expressed at higher levels in the first phases of hypertrophy, the first physiologic response to the overload of the heart, as a

hemodynamic compromise.

temperature [223].

**7. Metabolic dysfunction in chronic heart failure** 

Levy and colleagues [141], as already seen, showed in the septic myocardium important alterations (reduced cardiac performances, increased glucose uptake, increased myocardial glucose transmembrane transporters GLUT1 and GLUT4, increased deposits of glycogen). Because these modifications have been described as typical of the myocardial hibernation, Levy proposed that this latter may take part in the septic cardiac dysfunction. It is however debatable that the "true" ischemic myocardial hibernation is responsible for cardiac dysfunction in sepsis for several reasons: 1)myocardial hibernation is no longer considered as a down-regulation of the contractile and metabolic activity aiming at reducing the oxygen consumption and at adapting it to the reduced blood supply (smart heart hypothesis). It is instead attributed to repetitive episodes of myocardial stunning [210] 2) reduced myocardial perfusion has never been demonstrated in sepsis [211], the biochemical modifications showed in septic myocardium are not features of hibernating myocardium, they are instead present in the fetal heart and also in the stressed myocardium as expressions of the reactivation of the fetal gene program [47,52]. The alterations described in septic myocardium would therefore be unspecific and their presence could be explained in sepsis by the inflammatory state that characterizes it. The picture of metabolic hibernation present in sepsis should therefore be separated from that of the true myocardial ischemic hibernation whose existence, in our opinion, has not been convincingly demonstrated in sepsis.

Myocardial injury in sepsis is also shown by the increase of many markers such as troponin (Tn) and natriuretic peptides. Whereas Tn I and TnT levels are good indicators for cardiac function in sepsis relating well with left ventricular dysfunction and a poor prognosis, BNP is related to the outcome but only weakly with cardiac filling pressures [212]. A high level of BNP >144 pg/ml predicts cardiac dysfunction with high sensitivity (92%) and high specificity (86%)[213]. BNP upregulation takes part in the reactivation of the fetal gene program and its production has been shown to be increased during inflammatory processes related to the ability of inflammatory ILs and of p38MAPK to increase its secretion increasing its promoter activity [214]. Differently from the regulation of the similar molecule ANP, BNP secretion is regulated more by inflammation than by hemodynamic compromise.

#### **7. Metabolic dysfunction in chronic heart failure**

344 Lipid Metabolism

been described [209].

sepsis.

behaves as a double-edged sword because if on one hand depresses myocardial contractility increasing cGMP production, reduces oxygen consumption by antagonizing COX and may induce apoptosis on the other hand it supports relaxation normalizing the stiffness of the septic myocardium, antagonizes the action of the increased endothelin and has a cardioprotective effect through the induction of preconditioning [203]: unselective antagonism of NOS with NNMA is harmful in sepsis because even though it raises blood pressure it decreases cardiac, hepatic and renal blood flow [204]. Many other alterations have been described in septic myocardium: an altered phosphorylation of Troponin I (TnI) in serin 23-24 by Protein Kinase A (PKA), activated by receptors or an upregulation of Protein Phosphatase 2A (PP2A). dephosphorylating TnI [205,206]. The cause of the increased phosphorylation of Tn seems to be the increased -adrenergic activity during the early phases of sepsis followed by a reduction during the late phase characterized by the hypodynamic cardiac pattern. A calcium leak from sarcoplasmic reticulum has been also described in sepsis similar to that observed in CHF and determined also by PKA activation and increased phosphorylation of the calstabin, a protein maintaining in its stable form the sarcoplasmic ATPase assigned to the re-entry of Calcium in sarcoplasmic reticulum after systole [207,208]. Such an alteration could justify both diastolic dysfunction due to the diastolic calcium leak and the systolic dysfunction showing itself in the late phase as expression of depletion of Calcium reserves. An altered autonomic tone with increased in sympathetic activity in the early phases of sepsis and decreased heart rate variability has

Levy and colleagues [141], as already seen, showed in the septic myocardium important alterations (reduced cardiac performances, increased glucose uptake, increased myocardial glucose transmembrane transporters GLUT1 and GLUT4, increased deposits of glycogen). Because these modifications have been described as typical of the myocardial hibernation, Levy proposed that this latter may take part in the septic cardiac dysfunction. It is however debatable that the "true" ischemic myocardial hibernation is responsible for cardiac dysfunction in sepsis for several reasons: 1)myocardial hibernation is no longer considered as a down-regulation of the contractile and metabolic activity aiming at reducing the oxygen consumption and at adapting it to the reduced blood supply (smart heart hypothesis). It is instead attributed to repetitive episodes of myocardial stunning [210] 2) reduced myocardial perfusion has never been demonstrated in sepsis [211], the biochemical modifications showed in septic myocardium are not features of hibernating myocardium, they are instead present in the fetal heart and also in the stressed myocardium as expressions of the reactivation of the fetal gene program [47,52]. The alterations described in septic myocardium would therefore be unspecific and their presence could be explained in sepsis by the inflammatory state that characterizes it. The picture of metabolic hibernation present in sepsis should therefore be separated from that of the true myocardial ischemic hibernation whose existence, in our opinion, has not been convincingly demonstrated in

Myocardial injury in sepsis is also shown by the increase of many markers such as troponin (Tn) and natriuretic peptides. Whereas Tn I and TnT levels are good indicators The chronically failing heart is certainly energy-deprived [215], as has been ascertained by studies with 31P magnetic resonance (MR) spectroscopy [216] showing the reduced phosphocreatine: ATP ratio, important index of the energetic state of the heart correlated with the indexes of systolic and diastolic function [217]. The cardiac contractile function is inextricably linked to the metabolic function so that the alterations of cardiac metabolism observed in chronic heart failure must affect also the contractile function. As heart failure progresses mithocondrial respiration is gradually compromised preventing that a normal level of FAO may take place. In samples taken from explanted hearts, during transplantation in patients with advanced heart failure, the level of the enzymes of FAO was low compared to the non-failing hearts [218]. The reduction of mitochondrial biogenesis and of FFA oxidation in the advanced phase of CHF is, as already seen, caused by the reactivation of the fetal gene program through the upregulation of HIF-1 and the downregulation of PPAR, \ERR, of SIRT1 and of the PGC-1network demonstrated in myocardial samples of patients with end-stage heart failure compared with explanted non failing hearts [219]. The expression of these transcription factors and coactivators is reduced in pathological hypertrophy and heart failure whereas it is upregulated in physiological forms of hypertrophy related to postnatal growth or exercise training [220] so furnishing a precise boundary between physiologic and pathologic cardiac growth. Interestingly also a constitutive cardiospecific overexpression of PGC-1 in mice (MHC–PGC-1 mice) may damage the heart leading to myocardiopathy with uncontrolled mitochondrial proliferation but only in the neonatal period, whereas the overexpression in adult mice provoked only modest mitochondrial proliferation [221,222]. PGC-1 -\- mice show a reduced capacity both of exercising and of acclimating to cold temperature [223].

The fetal gene program is mediated by an incompletely known but modified expression of various microRNA [224-226]: Sucharov described that changes in the expression level of miR 100 and 133b contribute to the regulation of the fetal gene program [225]. The reduced expression of SIRT1 underlies the depressed mitochondrial biogenesis and FAO in the chronically failing heart. SIRT1 seems to be expressed at higher levels in the first phases of hypertrophy, the first physiologic response to the overload of the heart, as a response to the stressing event inducing it but begins to be downregulated when PARP activation takes place, due to the increased oxidative stress [227], reducing NAD+ availability. The reduction in SIRT1, provided with an antiapoptotic function, and the compromised mitochondrial activity may contribute to the increased cellular death in chronically failing myocardium [228,229]. Overall we may conclude that some similarities exist between the metabolic and morphological alterations observed in sepsis and CHF.

The Role of Altered Lipid Metabolism in Septic Myocardial Dysfunction 347

to toxic lipid byproducts formation. The chronic accumulation of neutral lipids is probably not, per se, immediately harmful for the myocardial cell in CHF but represents a mechanism of defence against the excess of the FFAs, more metabolically reactive, reaching it [241]. In the long term however lipid store may behaves as a time bomb because when a critical point is reached, either by accumulation of excessive quantity of lipid or by some event triggering an increased FAO, the accumulation of toxic byproducts such as ceramides, diacylglycerol, fatty

The lipid accumulation in the heart during infections has been described since the '90 of the last century [242] however the only complete study in sepsis is the article by Rossi and colleagues [169]. They showed in septic myocardium an increased staining by the dye oil red O, specific for lipid, and attributed it to an altered metabolism of the myocardium.

These alterations are the morphological expression of the switch from lipid to glucose utilization in a setting in which lipolysis is increased and FAO reduced by PPAR

Very little is known about the significance of lipid accumulation in the septic myocardium; initially interpreted as a degenerative marker in infectious diseases it could have in the setting of the acute septic myocardiopathy a meaning considerably different from the

In conclusion notwithstanding similar morphological aspects and the involvement of many common biochemical pathways the cardiac dysfunction in sepsis and in CHF may have

In the recent years the strict relationships between myocardial contractility and metabolism have been ascertained and a metabolic basis for cardiac contractile dysfunction verified in many clinical and experimental settings. Several metabolic alterations have been showed in CHF and septic myocardial dysfunction, their relationship with cardiac contractile dysfunction is not completely understood. They are probably a marker of the complex transcriptional and posttranscriptional alterations induced by the inflammatory, mechanical and metabolic stress the heart undergoes in both conditions and particularly of the coordinated compromise of mitochondrial function and of lipid oxidation but they represent also a direct inducer and amplifier of cellular damage. The significance of the morphological and metabolic alterations could be remarkably different in the slowly-evolving setting of CHF and in the acute one of the septic myocardial dysfunction, moreover the malignancy of the noxae underlying similar microscopic aspects may be very dissimilar. To understand in deep this aspect more thorough studies are needed. We currently do not know if they may represent a target for therapy because of the uncertain efficacy of the drugs until now used for this purpose but perhaps the light shed on the pathophysiological mechanism may give

chronic lipid accumulation in CHF even though the microscopic aspect is similar.

Certainly a lipidomic approach to the problem may be of great help.

us an important help in the search for a future effective treatment.

acyl-CoAs and acyl-carnitine ensues damaging the myocardial cell.

downregulation and by acute mitochondrial dysfunction.

different meanings and consequences.

**9. Conclusions** 
