Sepsis protects the myocardium and other organs from subsequent ischaemic/reperfusion injury via a MAPK-dependent mechanism

Sepsis remains a major cause of mortality, and there is recent data to suggest that the incidence of sepsis is increasing [1]. Treatment options remain limited to early antibiotic therapy, resuscitation and aggressive support of failing organ systems. Multiorgan dysfunction syndrome (MODS), the most severe end of the spectrum of sepsis syndrome, may involve the myocardium [2,3], and this has clear prognostic implications, with some reports pointing to a direct correlation between the onset of myocardial dysfunction and mortality in septic shock [4].

The pathophysiology of septic cardiomyopathy has not been fully elucidated. Although apoptosis is postulated to play an important role in the pathophysiology of septic cardiomyopathy [5,6], post-mortem studies have demonstrated relatively little cell death, even in cases of undoubted severe septic cardiomyopathy [7,8]. This has led to the suggestion that activation of pro-survival pathways may occur in association with pro-apoptotic pathway activation and that septic cardiomyopathy, although ultimately dysfunctional for the organism, is a reactive process of metabolic downregulation, serving a cellular preservation function broadly similar to preconditioning [6].

Preconditioning is the phenomenon whereby prior exposure to a stimulus, often one normally held to be injurious, triggers phenotypic changes that confer resistance to subsequent insults. It was first described by Murry et al. in 1986 [9], following the discovery that brief episodes of non-lethal myocardial ischaemia, followed by brief periods of reperfusion, provided protection against a subsequent more prolonged period of ischaemia/reperfusion. The concept of endotoxin-induced increased tissue tolerance first emerged in the 1960s [10]. These early descriptions of ‘endotoxin tolerance’ in animal models were followed by discoveries that a similar phenomenon occurs in humans [11]. Sepsis preconditioning of cardiac tissue may also occur in response to brief, non-fatal episodes of sepsis and bears undoubted observational similarities to ischaemic preconditioning, being inhibited, for example, by cycloheximide, a well-known blocker of ‘late ischaemic preconditioning’ [12]. The mechanisms whereby sepsis confers protection to the heart against subsequent ischaemia/reperfusion injury, however, are unknown, and whether this phenomenon most closely approximates myocardial stunning, preconditioning or other unrelated effects remains to be firmly established.

In the current study, we hypothesised that sepsis would protect the myocardium from subsequent ischaemia/reperfusion (IR) injury. We further hypothesised that the mitogen-activated protein kinases (MAPK) and nuclear factor-kappaB (NF-κB) pathways mediated this protective effect.

Previous investigators have reported that exposure of hearts to sublethal dosing of endotoxin confers protective effects against myocardial ischaemia/reperfusion injury in a variety of experimental models [13-15], including the Langendorff mode [16-19]. However, data in the literature examining the effects of polymicrobial sepsis, such as that induced by caecal ligation and operation, is less frequently encountered. It should be noted that inherent difficulties exist with experimental models utilising exposure of animal species to endotoxin. Although a reproducible experimental model that is easy to establish, endotoxaemia is not a perfect model of sepsis, with differing cytokine kinetic responses, duration of sepsis and time to death [20]. Making a conceptual advance on LPS exposure as a model of sepsis, we utilised the CLP procedure to induce polymicrobial sepsis. CLP is an in vivo model of sepsis that has been shown to be valid and reproducible [21] and is a more accurate replication of systemic inflammatory response to an infectious stimulus [22], more loosely mimicking an important clinically encountered scenario and is considered by some to be an essential preclinical model for sepsis studies [23].

The results of our experiments demonstrate that septic animals sustained marked sepsis-induced myocardial dysfunction, as evidenced by significantly reduced baseline (d)LVP and LVP_max compared to control animals, p = 0.0001 and p = 0.00003, respectively (Figures 1 and 2). Following ischaemia and reperfusion, CLP hearts showed more rapid recovery of functional measures, despite significantly reduced baseline function, and infarct size was significantly reduced in septic hearts (Figure 3). These results strongly support a preconditioning effect of polymicrobial sepsis.

Episodes of fibrillation were seen more frequently in the control group compared to the sepsis group (28% vs. 14%). This likely reflects the severity of IR injury, with protection conveyed in the CLP group conveyed by sepsis preconditioning. The use of lignocaine to prevent or treat reperfusion arrhythmias in the Langendorff model of the isolated heart has been well described [16]. Episodes of fibrillation were promptly treated and therefore unlikely to affect functional outcomes or infarct size.

Studies have shown sepsis preconditioning to demonstrate putative similarities to ischaemic preconditioning, particularly so-called ‘late preconditioning’ [12,24]. For example, with induction of gram-negative sepsis, myocardial protective effects are demonstrable at 6 h [18,25], and the myocardial protective effects of LPS exposure were reported to appear after 12 to 24 h [12]. As mentioned previously, the administration of the drug cycloheximide was found to inhibit protection, an effect demonstrable with late ischaemic preconditioning [26]. Additionally, there is strong evidence from multiple sources that the protective effects of LPS utilise signaling pathways known to be of central importance in ischaemic preconditioning, namely NOS2 and Akt [24]. Despite these similarities and intense research in the area of ischaemic preconditioning, the signaling pathway through which sepsis protects the heart from IR injury remains to be fully elucidated.

To investigate the in vitro effects of endotoxin exposure, we examined cell viability as measured by the MTT assay. The MTT assay provides an estimate of mitochondrial activity from which cell viability is deduced. In the presence of functioning mitochondria, MTT is reduced from a colourless compound to a blue product (MTT-formazan). The assay involves spectrophotometrical measurement of the concentration of the blue product. This is then calibrated post hoc to provide an indirect measure of mitochondrial function.

Lipopolysaccharide exerts its effects through TLR4-dependent NF-κB activation, leading ultimately to pro-inflammatory cytokine release. LPS is known to induce both pro-apoptotic and anti-apoptotic pathways, inducing apoptosis in endothelial cells [26] and hepatocytes [27] while promoting survival in monocytes [28], neutrophils [29], macrophages [30] and cardiomyocytes [4,12]. Although there is evidence to support a protective role played by LPS against myocardial ischaemia and reperfusion injury [13,16,19], less is known about the effects of LPS on apoptosis of cardiomyocytes. The current study in which LPS pretreatment of cardiomyocytes conferred protection against hypoxia-induced cell death (Figure 4) supports the findings of Chao et al. [4] who found reduced apoptosis in cardiomyocytes treated with LPS and subjected to hypoxia and serum deprivation, supporting a preconditioning effect of LPS in in vitro models of hypoxia. Limited studies exist outlining the effect of endotoxin preconditioning in other organ systems, and our findings in the current study speak to the fact that these findings can be generalised to other tissues and are not specific to cardiomyocytes through our demonstration of endotoxin preconditioning in alveolar epithelial (Figure 5) and human kidney (Figure 6) cells.

Due to the known similarities between ischaemic and sepsis preconditioning, and the important roles that stress kinases and NF-κB activation are known to play in the signaling of ischaemic preconditioning [31,32], we utilised blockers of the stress kinases, and ammonium PDTC, a direct inhibitor of the NF-κB pathway, to further examine potential signaling pathways involved in sepsis preconditioning.

The MAPK are a family of kinases that have been shown to be activated following multiple extracellular stimuli [33,34], with effects controlling cellular growth and differentiation [35-37]. Three MAPK have been identified in the cardiomyocyte: ERK, p38 and JNK. These are known to play key roles in cellular signal transmission between the cell surface and nuclei [38]. Although the signaling pathways through which sepsis confers myocardial protective effects have as yet to be elucidated, an important role has been proposed for the MAPK JNK, with additional reports of increased TNF-α-induced apoptosis associated with JNK inhibition [39]. This is supported by the findings of the current study, which demonstrates that the addition of JNKi to cardiomyocytes incubated under hypoxic conditions attenuated the protective effects against cell death associated with LPS (Figure 4). p38 MAPK is undoubtedly activated by LPS, although studies have revealed the conflicting effects of attenuated [39] and enhanced [40] apoptosis consequent to this. The results of the current study support a role of increased cell survival, as evidenced by attenuation of LPS-induced protective effects against hypoxia in cardiomyocytes (Figure 4), a finding that warrants further investigation. Although a role has been suggested by ERK 1/2 MAPK in the sepsis preconditioning signaling pathway [4], incubation of cardiomyocytes with inhibitors of ERK 1/2 had no demonstrable effect on endotoxin protection.

Studies have demonstrated an important role of NF-κB activation in delayed ischaemic preconditioning in vivo [41], and there is some evidence of a role played in signaling pathways of sepsis preconditioning [42]. In the current study, inhibition of NF-κB through the use of cellular incubation with PDTC attenuated endotoxin preconditioning, providing support to our hypothesis that this mediator may play an important role in sepsis preconditioning.

HMGB1, a so-called endogenous danger signaling molecule, is actively released by activated monocytes and macrophages and is a key late mediator of the inflammatory response to sepsis [43-45], with levels corresponding to the onset of mortality [43,46]. There is evidence that HMGB1 may also have preconditioning effects [47], and for this reason, we utilised anti-HMGB1 antibodies to ascertain whether HMGB1 may play a role in the signaling pathway of endotoxin preconditioning. Our results determined no demonstrable effects with HMGB1 inhibition.

These studies provide further confirmation of a preconditioning effect of sepsis in the intact heart subjected to global ischaemia/reperfusion. In addition, we have provided evidence of endotoxin preconditioning in cultured cardiomycocytes and further expanded on this to other organ systems through demonstration of effects in pulmonary alveolar cells and human kidney cells. Our investigation of the signaling pathway involved in these effects were reversed by NF-κB, p38 and JNK inhibition, highlighting potential important roles for these mediators in sepsis preconditioning. Taken together, these findings suggest a phenomenon with a high degree of relative organ specificity and requiring a complex signaling pathway.