Is Sepsis-Induced Organ Dysfunction an Adaptive Response?

Organ dysfunction is a hallmark of sepsis. Scientific investigation has focused on identifying potential causes and therapeutic targets of this component of the syndrome. Although various pathways and cellular systems are altered by sepsis and inflammation, to date no unifying or causative etiology has been uncovered. Historically, clinicians and investigators have viewed sepsis-induced organ dysfunction as a pathologic process that is deleterious to the survival of the host. Indeed, multiorgan dysfunction syndrome is the primary antecedent to sepsis-associated mortality. Recently, an interesting alternative hypothesis has been proposed: Does organ dysfunction during sepsis represent an adaptive prosurvival response? This concept is based on the observation that, despite physiologic and biochemical dysfunction, there is minimal evidence of cell death in affected organ systems, survivors rapidly recover organ function, and the downregulation of metabolism described during sepsis resembles a hibernating or suspended-animation state.

In nature, hibernation (torpor) is a protective adaptation to harsh environmental conditions and is a regulated, seasonal response largely coordinated by changes in mitochondrial respiration. This response allows hibernating mammals to reduce their metabolism to promote survival amid decreased substrate availability. Although a profound and prolonged metabolic downregulation can trigger death, the conservation of this physiologic response across models and species suggests that some degree of transient energetic swoon early in sepsis is likely to be adaptive. In this chapter, we review (1) the mechanisms that downregulate metabolism in sepsis, highlighting the role of mitochondria; (2) the evidence supporting the development of a hibernation-like state in during sepsis; and (3) the role for mitochondrial biogenesis to restore organ function and promote survival.

Mitochondria as the Mediator of Metabolic Downregulation in Sepsis

It has been proposed that an acquired defect in oxidative phosphorylation prevents cells from using molecular oxygen for adenosine triphosphate (ATP) production and potentially causes sepsis-induced organ dysfunction. Most energy production in vertebrate cells occurs in the mitochondria and is generated by aerobic respiration. This process, called “oxidative phosphorylation,” couples oxidation of NADH (nicotinamide adenine dinucleotide) and flavin adenine dinucleotide with phosphorylation of adenosine diphosphate (ADP) to form ATP. Oxidative phosphorylation is accomplished by a series of enzyme complexes termed the electron transport system (ETS). Located on the mitochondrial inner membrane, these enzymes use energy released during transfer of electrons between complexes to actively pump protons from the mitochondrial matrix into the intermembrane space. The resultant proton motive force is then used by ATP synthase (complex V) to synthesize ATP from ADP.

Each mitochondrion contains 2 to 10 copies of a circular, double-stranded DNA called mitochondrial DNA (mtDNA). mtDNA encodes key subunits of the ETS enzyme complexes, whereas structural subunits and the mitochondrial translational machinery primarily arise from nuclear genes. Thus, expression of genes that give rise to the protein complexes of the ETS is under dual control. An acquired defect in gene expression, protein translation, or functional activity of any of the ETS enzymes could impair oxidative phosphorylation and lead to sepsis-induced organ dysfunction.

Ultrastructural mitochondrial abnormalities have been recognized across organ systems in in vivo, ex vivo, and in vitro models of sepsis for over 30 years. For example, Crouser et al. demonstrated marked swelling and disruption of mitochondrial architecture in the liver 24 hours after cecal ligation and puncture (CLP) in a cat model. Similar morphologic abnormalities have been noted in mitochondria taken from heart, endothelial cells, intestinal epithelial cells, kidney, and skeletal muscle in animal models of sepsis. In human sepsis, heart and liver biopsies obtained immediately postmortem from adult nonsurvivors showed substantial accumulation of hydropic mitochondria.

Functional changes in mitochondrial respiration have been variably reported as increased, decreased, or unchanged in short-term sepsis models, although longer-term models more consistently demonstrate depressed mitochondrial function. Data about mitochondrial dysfunction in human sepsis remain relatively scant, although most—but not all—studies demonstrate decreased mitochondrial oxygen consumption in immune and nonimmune cells. When considering bioenergetic impairment in sepsis, investigators have most commonly focused on NADH:ubiquinone oxidoreductase (complex I) and cytochrome oxidase (complex IV). As the largest complex of the ETS (comprising 45 proteins), complex I is subject to impairment from changes in various protein subunits. Multiple studies have demonstrated decreased activity of ETS complex I in sepsis models and in humans. Complex IV is composed of only 13 subunits, many of which have been investigated in sepsis. Subunits 1, 2, and 3 make up the catalytic center and are encoded by mtDNA. The other 10 subunits arise from nuclear DNA. Subunit 1, the active site, houses the heme aa 3 binuclear center. Numerous studies have demonstrated abnormalities in expression and function of cytochrome oxidase during sepsis and in related models. For example, steady-state levels of cytochrome oxidase subunit I messenger RNA (mRNA) and protein are decreased in the murine heart after CLP and in endotoxin-stimulated macrophages. Injection of LPS into healthy human volunteers also resulted in widespread suppression of genes regulating mitochondrial energy production and protein synthesis. Reductions in ETS complex mRNA expression and protein translation result in reduced enzyme content and could affect the bioenergetic capacity of the cell.

Changes in mRNA and protein levels of key enzyme complex subunits are only functionally significant if they lead to or contribute to enzyme dysfunction. To this point, myocardial cytochrome oxidase activity decreased to 51% of baseline in baboons after Escherichia coli infusion. In murine sepsis, myocardial cytochrome oxidase inhibition was reported after CLP. This inhibition was initially competitive but later became noncompetitive. This change occurred at a time when cardiac function was markedly impaired and when mortality was high. Cytochrome oxidase dysfunction also has been shown in septic liver and in the medulla of the endotoxemic rat. Furthermore, reduced state 3 mitochondrial oxygen consumption has been demonstrated in the neonatal rat heart, feline liver, and rat diaphragm during endotoxemia.

Complex IV contains two heme subgroups (cytochrome a and a3 ) that assist in the transfer of electrons and reduction of oxygen to water. A reduced cytochrome aa 3 redox state in the absence of tissue hypoxia indicates a defect in mitochondrial oxygen use and suggests impaired oxidative phosphorylation. Several investigators have demonstrated reduced redox status during endotoxemia and gram-negative bacteremia in the heart, brain, skeletal muscle, and intestine in various animals. In addition, diminished heme aa 3 content in heart and skeletal muscle has also been shown in experimental sepsis.

Bioenergetic failure as a potential cause of sepsis-induced organ failure is not a new concept. With regard to sepsis-associated myocardial depression, early investigation extensively evaluated oxygen delivery, global myocardial perfusion, and high-energy phosphate levels. These studies clearly demonstrated that coronary blood flow and global cardiac perfusion were maintained and often increased during sepsis. In addition, there is evidence to suggest that tissue oxygen tension was unchanged in the dysfunctional septic heart. These findings argue strongly against decreased oxygen availability as a cause of myocardial depression in sepsis and support a defect in oxygen utilization. Organ-specific impairment of oxygen utilization is further supported by a progressive decline in whole-body oxygen consumption and resting metabolic rate with increasing severity of sepsis. Although other studies have reported oxygen consumption to be unchanged or increased in sepsis, it has been postulated that this may be due to an uncoupling of mitochondrial respiration leading to inefficient electron flux and heat generation.

However, the literature is less clear regarding ATP availability. In many studies, preserved ATP levels were demonstrated in dysfunctional septic myocardium. Other investigations reported decreased high-energy phosphates in experimental sepsis and endotoxemia. In a study of 28 adults with severe sepsis, 12 (43%) of whom died of sepsis-related multiple organ dysfunction syndrome, nonsurvivors were distinguished from sepsis survivors and nonseptic controls by lower levels of ATP in skeletal muscle. However, even preservation of ATP does not imply an absence of mitochondrial dysfunction in sepsis. During reduced oxygen delivery and cellular hypoxia, cells can adapt to maintain viability by downregulating oxygen consumption, energy requirements, and ATP demand. Thus, although ATP content may remain unchanged, ATP use can be decreased dramatically. In the heart, this response is called myocardial hibernation and classically occurs during myocardial ischemia. This adaptive, prosurvival response results in cardiomyocyte hypocontractility with preserved cellular ATP. If cellular metabolic activity continued unchanged despite mitochondrial dysfunction, then ATP levels would inevitably diminish and cell death pathways would be activated. Because cell death does not appear to be a primary feature of sepsis-induced organ dysfunction, it follows that cells may instead adapt to cope with the falling energy supply. Thus, finding preserved ATP during sepsis reveals little about the integrity of oxidative phosphorylation and may support the notion of a similar prosurvival response.

Development of a Hibernation-Like State in Sepsis

Metabolic downregulation is a crucial response that facilitates tolerance to a lack of energetic substrates during harsh environmental conditions and promotes survival during true hibernation. The hibernating state prevents a cellular bioenergetic crisis by reducing demand for ATP when substrate and/or oxygen supply are low and decreases mitochondrial oxidative stress. Central to this response are reduced oxygen consumption and cytochrome oxidase activity. In the hibernating frog, whole-body oxygen consumption decreases by 50% in normoxic 3° C water. Whole-body oxygen consumption and respiration of isolated skeletal muscle mitochondria decreased further when hibernating frogs were placed in hypoxic cold water. Furthermore, cytochrome oxidase activity in frog skeletal muscle progressively decreases during different stages of hibernation. In the hibernating ground squirrel, state 3 respiration decreased by approximately 70% in liver mitochondria. In squirrels that fail to hibernate, though, the changes observed in kidney cytochrome oxidase within their hibernating counterparts do not occur. Thus, it is clear that metabolic downregulation, in part because of reversible cytochrome oxidase inhibition and reduced activity, is key to initiating and maintaining the hibernating phenotype in various species. Importantly, the reduction in cytochrome oxidase activity and mitochondrial respiration characteristic of hibernation are similar to the changes seen during early sepsis.

Pharmacologic inhibition of cytochrome oxidase has been shown to induce a hibernation-like or suspended-animation state. Reversible inhibition of cytochrome oxidase with carbon monoxide (CO) arrests embryogenesis in Caenorhabditis elegans embryos yet preserves their viability in hypoxic conditions. In addition, noncompetitive cytochrome oxidase inhibition with inhaled hydrogen sulfide (H ₂S) induces a suspended-animation state in nonhibernating mice. On exposure to H ₂S, mice dramatically reduced their core body temperature and metabolic rate in a dose-dependent and reversible manner. At the cellular level, noncompetitive inhibition of cytochrome oxidase with sodium azide causes a rapid and reversible reduction in cardiomyocyte contraction and metabolic demand, mimicking myocardial hibernation.

Similar to hibernation and exposure to certain compounds, cytochrome oxidase inhibition is well described during sepsis. For example, in the heart, cytochrome oxidase was competitively inhibited during the early phase of sepsis and progressed to become noncompetitively inhibited during the late, hypodynamic phase. This specific pattern of enzyme inhibition is known to induce metabolic downregulation and a suspended-animation state. In addition, these biochemical changes coincide with the time course and progression of myocardial depression in humans, in which an early depression of cardiac contractility is compensated for by ventricular dilation and increased stroke volume with subsequent diastolic dysfunction and reduced cardiac output over time. The decrease in cardiac contractility seen early in sepsis is similar to the reduced systolic function characteristic of hibernation in grizzly bears and marmots and may reflect the decreased total-body oxygen requirement to produce ATP observed in sepsis and hibernation. However, the interval development of reduced diastolic relaxation later in sepsis is not observed in hibernating animals and may underlie a transition point during which metabolic downregulation becomes maladaptive and pathologic in sepsis compared with natural hibernation. When sepsis-induced mitochondrial and cellular defects become irreversible, as is the case with cytochrome oxidase inhibition over time, organ dysfunction may also become irreversible and lead to death. If this is true, then the challenges for clinicians will be to differentiate reversible adaptive organ “hibernation” from pathologic organ “failure,” to recognize when this switch has occurred, and to intervene to prevent the alteration.

Several different mediators may be responsible for metabolic downregulation and mitochondrial dysfunction in sepsis. The most likely offenders include nitric oxide (NO), CO, H ₂S, peroxynitrite, and reactive oxygen species. Certainly, all of these agents are endogenously produced in various tissues during sepsis, largely in response to an upregulation of the proinflammatory cytokines tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), high-mobility group protein-1, and others. The high levels of NO observed in sepsis may assist antimicrobial defense, but they are also cytotoxic to host cell oxidative phosphorylation at several postulated sites. Notably, the impairment in cytochrome oxidase activity during sepsis mimics that of true hibernation, suggesting that, at least early on, such impairment may be an adaptive response to an inflammatory insult. As discussed later, the failure to restore mitochondrial function and recover organ function after acute inflammation separates sepsis-induced organ dysfunction from true hibernation.

Mitochondrial Biogenesis to Restore Organ Function and Promote Survival

From an evolutionary perspective, it may be adaptive to reduce metabolic demands when oxygen and substrate availability are low to protect cells from a bioenergetic crisis and limit exposure to oxidative stress. Such a shutdown is manifest clinically as organ dysfunction. As with recovery from hibernation, though, once the infectious/inflammatory stimulus has abated, functional mitochondria are needed to restore cellular energy supply. All cells that undergo oxidative phosphorylation have robust quality control mechanisms to ensure a full complement of healthy mitochondria. Cells optimize the overall mitochondrial number, distribution, and function through a network of interrelated processes of biogenesis, fission, fusion, and mitophagy.

Mitochondrial biogenesis is the process of synthesizing new functional mitochondria and can be induced by exercise, fasting, exposure to cold temperatures, oxidative stress, and inflammation. Depending on the stimulus, mitochondrial biogenesis is executed through several signaling pathways that converge on a common set of coactivators and transcription factors. The nuclear-encoded factors include peroxisome proliferator-activated receptor γ-1 coactivator-α (PGC-1α), nuclear respiratory factors (NRF-1 and NRF-2), nuclear factor erythroid-2-related factor 2 (Nrf2), and mitochondrial transcription factor A (TFAM). These proteins either increase transcription of nuclear-encoded mitochondrial proteins or are imported into the mitochondria to directly upregulate expression of mtDNA to promote ETS complex assembly.

Common inflammatory mediators of the innate immune response, including TNF-α, IL-6, and interferon-γ (IFN-γ), can activate one or several well-defined pathways of mitochondrial biogenesis. These cytokines/chemokines are produced in response to the presence of microbial antigens (pathogen-associated molecular patterns) that are sensed by cellular pattern recognition receptors such as toll-like receptors (TLRs). Subsequent activation of the nuclear factor-κB (NF-κB), mitogen-activated protein kinases, and protein kinase B (Akt) pathways lead to increased expression of the factors regulating mitochondrial biosynthesis. Upregulation of NO also stimulates mitochondrial biogenesis through increased PGC-1α activity. Finally, stimulation of heme metabolism in sepsis due to hypoxia and inflammation leads to a heme oxygenase-1 (HO-1)–mediated production of CO that increases Nrf2 activation, thereby further increasing mitochondrial biogenesis. Notably, despite elevated blood levels of cytokines, intracellular signaling may be impaired in severe sepsis and may be a mechanism leading to insufficient mitochondrial recovery in nonsurvivors.

The inhibition of oxidative phosphorylation and mitochondrial damage in sepsis also provide potent stimuli for mitochondrial quality control mechanisms. For example, an increase in the adenosine monophosphate/ATP or oxidized NAD (NAD+)/NADH ratios induces PGC-1α through several pathways. Mitochondrial damage due to oxidative stress results in mtDNA translocation to the cytoplasm, which upregulates NF-κB through TLR-9 signaling and acts as a danger-associated molecular pathogen to further promote inflammatory mediates that affect mitochondrial biogenesis. In this regard, removal of dysfunctional mitochondria—termed mitophagy —has also been shown to be protective in sepsis. For example, in the liver and kidney of hyperglycemic critically ill rabbits, biochemical markers indicating insufficient mitophagy were more pronounced in nonsurviving animals. Interestingly, after 3 and 7 days of illness, mitophagy was better preserved in animals treated with insulin to preserve normoglycemia, which correlated with improved mitochondrial function and less organ damage. Moreover, stimulation of mitophagy in the kidney with rapamycin correlated with protection of renal function in this study.

Data from animal studies and septic patients provide key evidence that mitochondrial recovery is predictive of or associated with recovery of organ function and survival. Haden et al. showed that mitochondrial biogenesis is evident over 1 to 3 days after a nonlethal exposure to Staphylococcus aureus in a rodent model, with a subsequent recovery of oxidative phosphorylation. The same group further showed that sepsis survival could be improved in rodents treated with daily exposure to a low dose of CO and demonstrated a mechanistic link to induction of HO-1, Nrf2, and Akt signaling. In humans, Carre demonstrated a significant association in the upregulation of the mitochondrial biogenesis factors PGC-1α and NRF-1 with survival in sepsis.

Evidence of mitochondrial recovery—and biogenesis in particular—is also present in hibernating animals. Ground squirrels in torpor exhibited a shift to slow-twitch type I muscle fibers that was accompanied by activation of PGC-1α and enhanced mitochondrial abundance and metabolism. Pharmacologic agents that induce mitochondrial biogenesis, such as pioglitazone, resveratrol, and recombinant human TFAM, may hold promise as a novel therapeutic strategy in sepsis-induced organ dysfunction that fails to recover after an initial “hibernation-like” phase.

Recent evidence clearly suggests that the resolution of inflammation is an active process driven by a group of specialized and unique mediators. These compounds are derived from polyunsaturated fatty acids that include lipoxins, E- and D-series resolvins, protectins, and maresins. These lipids, alone or in combination, suppress activated leukocytes and macrophage activity, inhibit proinflammatory cytokine production, attenuate inappropriate inflammatory responses, enhance bacterial clearance, and improve survival. Their importance to the current discussion lies in recent demonstration of a biosynthesis pathway for these lipid mediators in mitochondria that is activated after tissue injury. Although the role of these proresolving mediators in sepsis is unknown, the presence of a responsive pathway within mitochondria suggests their potential importance and identifies an area for future investigation.

Only gold members can continue reading. Log In or Register to continue