Role of Complement C5a and Histones in Septic Cardiomyopathy

Abstract

Polymicrobial sepsis (after cecal ligation and puncture, CLP) causes robust complement activation with release of C5a. Many adverse events develop thereafter and will be discussed in this review article. Activation of complement system results in generation of C5a which interacts with its receptors (C5aR1, C5aR2). This leads to a series of harmful events, some of which are connected to the cardiomyopathy of sepsis, resulting in defective action potentials in cardiomyocytes (CMs), activation of the NLRP3 inflammasome in CMs and the appearance of extracellular histones, likely arising from activated neutrophils which form neutrophil extracellular traps (NETs). These events are associated with activation of mitogen-activated protein kinases (MAPKs) in CMs. The ensuing release of histones results in defective action potentials in CMs and reduced levels of [Ca²⁺]i-regulatory enzymes including sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA2) and Na⁺/Ca²⁺ exchanger (NCX) as well as Na⁺/K⁺-ATPase in CMs. There is also evidence that CLP causes release of IL-1β via activation of the NLRP3 inflammasome in CMs of septic hearts or in CMs incubated in vitro with C5a. Many of these events occur after in vivo or in vitro contact of CMs with histones. Together, these data emphasize the role of complement (C5a) and C5a receptors (C5aR1, C5aR2), as well as extracellular histones in events that lead to cardiac dysfunction of sepsis (septic cardiomyopathy).

1. Introduction

Sepsis in United States kills more than 200,000 individuals per year and epidemiological studies show an increased incidence of deaths from sepsis and septic shock over the past two decades in US (Kumar, et al. 2011; Martin, et al. 2003). Sepsis and septic shock are the most common cause of death among critically ill patients in intensive care units (ICUs) with estimated annual costs of care exceeding $17 billion (Angus, et al. 2001; Dombrovskiy, et al. 2007; Russell 2006). Despite nearly 100 clinical trials, there are no FDA-approved drugs for use in sepsis. We do not adequately understand the molecular events in sepsis in ways that have translational implications. Pathogenesis of sepsis is complex, with many immune and non-immune mediators involved. In 1904, William Osler noted, “It appears that patients are dying not from their infections but rather their reaction to them” (Cho and Choi 2014). Sepsis has traditionally been considered to be a result of an uncontrolled inflammatory response that results in organ dysfunction (Cho and Choi 2014). Four key areas in sepsis are endothelial dysfunction, coagulation abnormalities, alterations in cell function, and dysregulated cardiovascular responses (Evans 2018). It is also now realized that sepsis can cause progressive symptoms developing early as well as weeks, months or years after “recovery” from sepsis (Delano and Ward 2016). Aggressive early intervention for managing infection (bacterial cultures and antibiotics therapy), better managing resuscitation fluids and vasopressors, as well as improved mechanical ventilation that avoids barotrauma in patients with sepsis-related acute respiratory distress syndrome (ARDS) have been strongly recommended for improving survival of patients with sepsis (Howell and Davis 2017; Rhodes, et al. 2017). When patients develop septic shock, there is often relentless progression to multi-organ failure and a lethal outcome (Gustot 2011; Shapiro, et al. 2009).

Cardiac dysfunction due to sepsis, “septic cardiomyopathy”, can happen in up to 80% of patients with septic shock (Beraud, et al. 2014). This condition may lead to mortality rates of 70% in these patients (Romero-Bermejo, et al. 2011), even though it is generally agreed that, for patients who survive sepsis, the cardiac defects are reversible, typically within 7-10 days after onset of sepsis (Romero-Bermejo, et al. 2011). Sepsis-related cardiac dysfunction does not confer long-term harm to the heart (Parker, et al. 1984). Cardiac-dysfunction includes reduced systemic vascular resistance, elevated cardiac output, reduced ventricular function and dilatation of ventricles (Parrillo, et al. 1990). Regarding pathogenesis of these events, some time ago a myocardial depressant substance was found in blood of septic patients that had in vitro suppressive effects on cardiomyocyte (CM) contractility and relaxation (Parrillo, et al. 1985). Subsequently, it was determined that a variety of proinflammatory cytokines (IL-1β, IL-6, TNF) had similar effects on CMs (Carlson, et al. 2005; Kumar, et al. 1996). Our group found that rodent CMs exposed to C5a released cardio suppressive cytokines including IL-1β, IL-6, and TNF (Atefi, et al. 2011). The multiplicity of cytokines with these activities has made it unlikely that in vivo neutralization or use of knock-out (K.O.) mice would effectively protect the heart during sepsis. Other factors that can similarly affect CMs include pathogen-associated molecular patterns (PAMPs) released from bacteria and damage-associated molecular patterns (DAMPs) released from injured tissues or apoptotic cells (Rudiger and Singer 2013). LPS is a prominent example of a PAMP, while histones (discussed in section 3.4.) are examples of DAMPs released during sepsis.

2. Methods to assess heart function in sepsis

There are several non-invasive methods to assess heart function following sepsis including functional (Echocardiography; Echo-Doppler), electrical (electrocardiograms; ECGs) measurements and biomarkers present in the plasma. These techniques are still commonly used in septic patients in the hospital to assess cardiac function and monitor the hemodynamic condition (Barnaby, et al. 2018; Beraud, et al. 2014; Guerin and Vieillard-Baron 2016; Haileselassie, et al. 2016; Ozdemir, et al. 2016; Quinten, et al. 2017; S, et al. 2014; Shashikumar, et al. 2017).

2.1. Echo-Doppler studies:

Echocardiographic assessment and Doppler imaging technology can detect numerous systolic and diastolic dysfunctions in different geometries of the heart (left and right) in patients with severe sepsis or septic shock. This technique can detect left ventricular (LV) and/or right ventricular (RV) systolic or diastolic dysfunction in patients with sepsis. The hemodynamic status in patients with septic shock can be assessed using echocardiographic views that may reveal LV and RV dysfunction (Vieillard-Baron et al., 2003). Echocardiographic studies have reported RV dysfunction in 55% of septic patients and combined right and LV dysfunction in 53% of patients with sepsis/septic shock (Vallabhajosyula, et al. 2017a). There are also some sophisticated technologies for evaluation of cardiac systolic and diastolic functions in patients with sepsis; such has speckle-tracking echocardiography (STE). In a very recent review, STE was reported to be a useful predictor for mortality in patients with sepsis (Vallabhajosyula, et al. 2018).

Our observations of cardiac function (systolic and diastolic parameters) in septic mice by Echo-Doppler studies (using a Vevo 770 High-Resolution In Vivo Imaging System) have been quite informative regarding cardiac dysfunction in hearts of septic mice (Fattahi, et al. 2017; Kalbitz, et al. 2016a). Sepsis-induced heart dysfunction 8 hr after cecal ligation and puncture (CLP) has been documented in septic mice from our previous studies (Fattahi, et al. 2017; Kalbitz, et al. 2016a; Kalbitz, et al. 2016b; Kalbitz, et al. 2015) as well as reports from other groups (Alhamdi, et al. 2015; Boluyt, et al. 2004; Hollenberg, et al. 2001; Hoover, et al. 2015; Zanotti-Cavazzoni, et al. 2009). We described in details of Echo-Doppler measurements in septic mice in our previous publications (Fattahi, et al. 2017; Fattahi and Ward 2017; Kalbitz, et al. 2016a; Kalbitz, et al. 2016b; Kalbitz, et al. 2015).

2.2. ECG studies:

Septic patients may show different changes in their ECG including sinus bradycardia, prolonged PR interval, prolonged QT interval, and Osborn waves due to hypothermia following sepsis (Drake and Flowers 1980; Tisdale, et al. 2013; Varriale and Ramaprasad 1995). QRS attenuation and increases in QRS duration have also been seen in the ECG tracings in septic patients (Madias and Bazaz 2003; Rich, et al. 2002). Sepsis can also induce myocardial depression and ST-elevation similar to what has been found in humans with myocardial infarction. (Hibbert, et al. 2012; Y-Hassan, et al. 2014).

Evaluating patients with sepsis in the pediatric intensive care unit (PICU) using ECG has shown P waves (PWd), QT (QTd), T-peak to T-end (Tp-e) intervals and Tp-e/QT, Tp-e/QTc ratios than were significantly higher in septic patients compared to age- and sex-matched healthy children. The Tp-e/QT ratio was found to be a valuable tool in predicting mortality for the septic patients in the PICU (Ozdemir, et al. 2016). Some of the ECG changes in septic patients are seen in the septic mice including prolongation of the PR, prolongation of the QRS complex (Bustamante, et al. 2002; Hoover, et al. 2015) and prolongation of QTc intervals (Hoover, et al. 2015). Details of electrical method (ECG recordings) in mice have been described previously (Hoover, et al. 2015).

2.3. Biomarkers evaluations:

Many biological markers (biomarkers) in plasma are elevated during the pathophysiological process of severe sepsis or septic shock. These markers may be used for diagnosis, prognosis and monitoring of clinical responses in septic patients. Only a few of the biomarkers currently are used routinely in the clinical setting of sepsis (due to the lack of sufficient sensitivity and specificity). Troponin-T (TnT) is one of the best known traditional cardiac biomarkers, being evaluated in plasma from patients with sepsis or septic shock at time of admission (Hamilton, et al. 2012; Vallabhajosyula, et al. 2017b). TnT elevation has been shown to be associated with the higher risk of death in septic patients (Bessiere, et al. 2013; Fromm 2007; Gualandro, et al. 2014; Mantzouris, et al. 2013). Different studies showed the levels of TnT to be associated with in-hospital and short-term (Vallabhajosyula, et al. 2017b; Vasile, et al. 2013) and even long-term mortality (Vallabhajosyula, et al. 2017b). However, it has been reported (using univariate analysis data) that TnT levels per se do not predict mortality in patients with severe sepsis or septic shock. Echocardiographic variables (including LV diastolic dysfunction and RV systolic dysfunction; dilatation) have correlated with TnT levels in serum from septic patients (Landesberg, et al. 2014).

C-reactive protein (CPR) and procalcitonin (PCT) are used worldwide routinely, although neither of them is specific for sepsis (Giannakopoulos, et al. 2017; Pierrakos and Vincent 2010; Prucha, et al. 2015). It was reported that CPR and PCT are often elevated in other inflammatory conditions without sepsis (Henriquez-Camacho and Losa 2014; Pierrakos and Vincent 2010). Some novel biomarkers include presepsin (sCD14-ST) (Endo, et al. 2012; Masson, et al. 2014; Novelli, et al. 2013; Shozushima, et al. 2011; Ulla, et al. 2013; Yaegashi, et al. 2005), CD64 (Bhandari 2014), soluble urokinase plasminogen activator receptor (suPAR) (Eugen-Olsen, et al. 2002; Koch, et al. 2011), proadrenomedullin (proADM) (Suberviola, et al. 2012) and the soluble triggering receptor expressed on myeloid cells (sTREM-1) (Adib-Conquy, et al. 2007; Ferat-Osorio, et al. 2008), Pentraxin 3 (PTX3) (Liu, et al. 2014), Creatine kinase (CK) (Hua, et al. 2012; Oliveira, et al. 2008; Zhang, et al. 2012), and CD73 (Bellingan, et al. 2014; Hasko, et al. 2011) were reported to be useful biomarkers for diagnosis and prognosis of sepsis.

Taken together, there is no reliable biomarker that is specific for sepsis. If novel reliable biomarkers with a high sensitivity and specificity are discovered in the future, they may be used in the clinical setting, probably as a useful tool to improve identification and subsequent treatment of septic patients.

3. Role of complement C5a in septic cardiomyopathy

It is known that sepsis causes robust complement activation, resulting in robust generation of C5a and it is also known that pathophysiology of polymicrobial sepsis in mice following CLP is C5a-dependent (Hoesel, et al. 2007b; Niederbichler, et al. 2006). C5a is a powerful proinflammatory mediator generated at the midpoint of the complement system, interacting with its receptors (C5aR1 and C5aR2) after onset of sepsis. It is reported that interaction of C5a with its receptors is linked to many of adverse harmful consequences of sepsis (Czermak, et al. 1999; Hoesel, et al. 2007b; Huber-Lang, et al. 2001a; Huber-Lang, et al. 2001b; Niederbichler, et al. 2006; Riedemann, et al. 2002; Riedemann, et al. 2003; Rittirsch, et al. 2008). C5a reacts with its receptors leading to the “cytokine storm”, apoptosis of lymphoid cells, loss of innate immune functions of neutrophils (PMNs, polymorphonuclear leukocytes), cardiomyopathy, disseminated intravascular coagulation, and complications associated with development of organ dysfunction (Hotchkiss, et al. 1999; Huber-Lang, et al. 2001b; Ward 2004; Ward 2010a; Ward 2010b). These events, often lead to lethality. We have found that C5a causes dysfunction in both contractility and relaxation in CMs (Niederbichler, et al. 2006) and that CMs exposed to C5a released proinflammatory cytokines such as IL-1β, IL-6, and TNF (Atefi, et al. 2011) which may contribute to myocardial depression during sepsis (Fernandes and de Assuncao 2012). We have shown that C5a is associated with septic cardiomyopathy (Hoesel, et al. 2007b; Kalbitz, et al. 2016b; Kalbitz, et al. 2015; Niederbichler, et al. 2006) and that interception of C5a improves survival, reducing cardiac dysfunction in septic mice (Huber-Lang, et al. 2001a; Kalbitz, et al. 2016b; Niederbichler, et al. 2006; Rittirsch, et al. 2008). Various types of antibodies and compounds such as peptides or nonpeptides have actively been developed. These substances act as inhibitors of complement components C5 and C5a and antagonists of the C5a receptor (Horiuchi and Tsukamoto 2016). Different reports demonstrated that infusion of a neutralizing antibody to C5a or C5a activity can be used as strategies to reduce cardiac dysfunction (Czermak, et al. 1999; Flierl, et al. 2009; Guo, et al. 2000; Guo, et al. 2006; Hoesel, et al. 2007a; Huber-Lang, et al. 2001a; Huber-Lang, et al. 2001b; Laudes, et al. 2002; Riedemann, et al. 2004; Sprong, et al. 2003). We found that C5a receptors are expressed on normal humans CMs (Fattahi and Ward 2017) and that C5a receptors are overexpressed on CMs following sepsis in mice (Atefi, et al. 2011). Our focus has been mainly on the early events developing after CLP, the time which the high levels of proinflammatory cytokines and extracellular histones are present in the plasma (Kalbitz, et al. 2016a; Kalbitz, et al. 2015). It is obvious that a better understanding of molecular events in sepsis is imperative. It is known that complement activation plays a deleterious role in the harmful effects of sepsis. Following is the list of the events occurring during sepsis which result in septic cardiomyopathy development. All of these events are dependent on C5a.

3.1. C5a effects on Ca²⁺ homeostasis in cardiomyocytes

In earlier reports we found that C5a affects intracellular calcium ([Ca²⁺]i) homeostasis and electrophysiological functions of CMs (Kalbitz et al., 2016)(Kalbitz, et al. 2016b). Here we provide more details of how these events develop in CMs. Details of CMs isolation are described in our recent publications (Fattahi, et al. 2017; Kalbitz, et al. 2016a; Kalbitz, et al. 2016b; Kalbitz, et al. 2015). In CMs, phospholamban (PLB), located on membrane of sarcoplasmic reticulum, plays a major role in [Ca²⁺]i homeostasis and in regulation of cardiac function (Li, et al. 2013; Luo, et al. 1996; Simmerman and Jones 1998). Phosphorylation of PLB (phospho-PLB) occurs during development of septic cardiomyopathy. We hypothesized that phosphorylation of PLB following C5a ligation with C5a receptors on CMs may underlie the defects in cellular calcium regulation in CMs. Western blotting using PLB and phospho-PLB specific antibodies revealed that C5a over one hr induced progressive phosphorylation of PLB in a dose dependent manner (Figure 1a). Phosphorylation of PLB enhanced cardiac contractility by increasing sarcoplasmic reticulum Ca²⁺ levels and [Ca²⁺]i transients (Chu, et al. 2000; Wegener, et al. 1989). We next determined the effects of C5a on [Ca²⁺]i + transient amplitudes in rat CMs (0.5 Hz pacing, Figure 1b-e). Consistent with elevated phospho-PLB, the [Ca²⁺]i transient amplitude was also elevated by C5a in a dose dependent manner, peaking at approximately 150 ng/ml C5a (Figure 1b). Furthermore, the amplitude of spontaneous Ca²⁺ releases in the diastolic interval as observed in Figure 1d-e increased with C5a dose (Figure 1c). In frames d-f of Figure 1, we determined the effects of C5a on CM responses to progressively elevated stimulation frequency, up to 5 Hz. At faster pacing frequencies, C5a caused sustained elevations of cytosolic Ca²⁺ when compared to the absence of C5a (frames f and g) and reduced the transient amplitude. The mechanism to explain the frequency-dependent effects of C5aR1 activation is unclear but may be attributed to the prolongation of the Ca²⁺ transient duration (Figure 1h) in CMs exposed to C5a. The slower removal of Ca²⁺ from the cytosol, especially at faster pacing rates, resulted in accumulation of cytosolic Ca²⁺ which may cause diastolic dysfunction. Prolongation of the Ca²⁺ transient duration is puzzling in the presence of increased phospho-PLB (Figure 1a) and suggests that other mechanisms that affect [Ca²⁺]i homeostasis may also be altered by C5a engagement of its receptors

Figure 1: Quantification of C5a effects on CM Ca²⁺ homeostasis.

a. Western blotting of lysates from rat CMs indicates that C5a, in a dose-dependent manner, induced phosphorylation of phospholamban (PLB) after a one-hour period at 37°C. b. C5a application (up to 340 ng/ml) to CMs increased the intracellular Ca²⁺ amplitude (0.5 Hz pacing). c. The amplitude of spontaneous Ca²⁺ release and waves also increased with C5a dose. For frames B and C, N = 3, n = 7-27 per dose (N = number of separate experiments; n = number of CMs used in each study). d. Control frequency responses of intracellular Ca²⁺ transients. e. Frequency responses of intracellular Ca²⁺ transients following C5a exposure (45 ng/mL). f. Presence of C5a (45 ng/mL) caused elevations of intracellular Ca²⁺ as indicated by higher baseline fluorescence with increasing frequency (black circles) as opposed to no C5a (open circles). g. Relative increase (%) of basal Ca²⁺ levels after C5a addition (45 ng/mL). The greatest elevations of resting Ca²⁺ occurred at higher pacing frequencies. h. The duration of electrically paced intracellular Ca²⁺ transients was greater following C5a addition (45 ng/mL) (black bars) over a range of pacing frequencies (0.5-3Hz) as compared to absence of C5a (white bars). * p<0.05, n = 2, n = 5.

3.2. Sepsis-induced [Ca²⁺]i increases in CMs in a C5a-dependent manner

We have shown in our previous studies the importance of [Ca²⁺]i flux during sepsis and found the presence of defective current densities for Ik1, L-type calcium channel, and Na⁺/Ca²⁺ exchanger (NCX) in CMs from septic mice (Kalbitz, et al. 2016a). We also showed effects of C5a on [Ca²⁺]i homeostasis and electrophysiological functions in single CMs (Kalbitz, et al. 2016b). In this report, we used rat and mouse CMs to study the effect of C5a on [Ca²⁺]i influx from CMs, either after incubation of rat CMs with recombinant rat C5a, rrC5a (in vitro studies) or mouse CMs after CLP to see effect of C5a receptor (in vivo studies). The methodology was described in our previous report in which we measured [Ca²⁺]i, using calcium indicator fluo-3 AM, in CMs after exposure to histones or after CLP (Kalbitz, et al. 2015). For the new studies we measured [Ca²⁺]i in rat or mouse CMs by flow cytometry using calcium indicator fluo-3 AM labeled CMs. In Figure 2, we found increased [Ca²⁺]i influx in rat CMs after exposing them to rrC5a in vitro (a-c). We found that rat CMs exposed to rrC5a (60 min at 37°C) developed dose-dependent increases in [Ca²⁺]i (a). Frame b shows a typical set of tracings for CMs exposed only to buffer (control buffer) or to C5a (1 μg/ml at 37°C), as measured by flow cytometry. Frame c shows increases in [Ca²⁺]i in CMs as a function of time of exposure (at 37°C) to C5a. Frames d and e are [Ca²⁺]i measurements in mouse CMs after CLP as a function of time (frame d) and evidence that the absence of either C5a receptor prevented CLP-induced increases in mouse CMs (frame e). The increases in cytosolic [Ca²⁺]i were associated with both systolic and diastolic cardiac dysfunction (Kalbitz, et al. 2016b; Kalbitz, et al. 2015). These data showing increased [Ca²⁺]i by C5a are in line with the data for ROS in which we found ability of rrC5a to cause increased reactive oxygen species (ROS) in isolated rat CMs (Kalbitz, et al. 2016b). Use of wild type (Wt) mice and induction of sepsis resulted in increased ROS (Kalbitz, et al. 2016b) in mouse CMs. We also found that ROS release by C5a also required C5a receptors (C5aR1 and C5aR2) (Kalbitz, et al. 2016b) and that the response was significantly attenuated in the lack of either C5aR1 or C5aR2.

Figure 2. C5a and Sepsis-Induced [Ca²⁺]i increases in CMs are C5a Receptor-dependent.

[Ca²⁺]i in rat CMs exposed in vitro to C5a (frame a, b, and c) or after CLP (frame d and e), as determined by flow cytometry using calcium indicator fluo-3AM. a. Representative cell plots of unstimulated (using control buffer) (left line) by flow cytometry and C5a-treated (1 μg/ml, 60 min at 37°C, right line) CMs. b. [Ca²⁺]i responses of rat CMs to increasing amounts of rrC5a after 60 min at 37°C. c. Increased [Ca²⁺]i in CMs exposed in vitro to rrC5a (1 μg/ml) for the indicated periods of time. d. [Ca²⁺]i in CMs isolated from rats 8, 16, 24 and 48 hr after CLP. e. Altered levels of [Ca²⁺]i in CMs isolated after 16 hr CLP in Wt, C5aR1^−/− and C5aR2^−/− mice. For each bar, n > 8 from different CM preparations. *p <0.05.

3.3. Reduced SERCA2, NCX and Na⁺/K⁺-ATPase in CMs during sepsis are C5a-dependent

One of the known events happening during septic cardiomyopathy is impaired calcium handling in CMs (Li, et al. 2013). We have recently shown that, when mice developed CLP-induced sepsis, there was a trio of homeostatic enzymes in heart that were impaired during sepsis (Fattahi and Ward 2017; Kalbitz, et al. 2016a; Kalbitz, et al. 2016b). These enzymes included: Na⁺/K⁺-ATPase, which is critical for action potentials, together with two [Ca²⁺]i regulatory enzymes: sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA2) and the NCX. Reductions of NCX and SERCA2 during sepsis impaired cytosolic clearance of [Ca²⁺]i after CLP, indicating that these enzymes play an important homeostatic role in regulating cytosolic [Ca²⁺]i in CMs (Zhang, et al. 2004). In CMs after systole, SERCA2 and NCX rapidly reduced [Ca²⁺]i to presystolic levels. During polymicrobial sepsis, it was shown that action potentials were faulty, together with the inability to regulate the increased [Ca²⁺]i developing during diastole. We have shown that the enzymatic activities of these three enzymes that maintain homeostasis as well as their protein levels (as defined by ELISA, Western blot and mRNA data), were strikingly reduced after sepsis in heart (Fattahi and Ward 2017; Kalbitz, et al. 2016b). Using immunofluorescent (IF) staining to detect the expression of plasma membrane Na⁺/K⁺-ATPase, we were able to detect the Na⁺/K⁺-ATPase of the plasma membrane. Figure 3 shows expression of Na⁺/K⁺-ATPase of the plasma membrane on CMs from control mice (frame a) and in frozen heart tissue (frame b) from healthy mouse using IF staining. IF staining was performed as described in our previous report (Kalbitz, et al. 2016b). Anti-alpha 1 Sodium Potassium ATPase antibody (ab7671, Abcam) was used as primary antibody and goat anti-mouse IgG-AlexaFluor488 (Jackson ImmunoResearch) was considered as the secondary antibody for IF labeling. Our recent data with IF staining showed evidence of reduced levels of Na⁺/K⁺-ATPase of the CM in the septic mouse 8 hr after CLP (Kalbitz, et al. 2016b). These reductions occurred as early as 8 hr after sepsis, at a time that Echo-Doppler measurements also showed defective cardiac function, all being the hallmark of the “cardiomyopathy of sepsis”. Our data were in line with the earlier reports showing reduced levels of SERCA (He, et al. 2007; Ren, et al. 2002; Wu, et al. 2004; Wu, et al. 2016; Wu, et al. 2001; Zhu, et al. 2005) and NCX (Hobai, et al. 2015a; Hobai, et al. 2015b) on CMs of septic mice. The cardiomyopathy was reversible, with the various parameters of responses returning to normal limits around 48-72 hr after sepsis (Kalbitz, et al. 2016a; Kalbitz, et al. 2015). We have found that these events (reductions in homeostatic enzymes) were C5a receptor-dependent as these decreases during sepsis were significantly diminished in C5aR1 or C5aR2 K.O. mice (Kalbitz, et al. 2016b).

Figure 3. Na⁺/K⁺-ATPase staining (plasma membrane marker) on mouse cardiomyocyte (a) or on frozen heart tissue (b) from control mouse.

Green: Plasma membrane (Na⁺/K⁺-ATPase), Blue: Nuclei (DAPI)

3.4. C5a causes NET formation and release of extracellular histones from PMNs

Robust activation of complement C5a during sepsis was associated with progressive impairment of the innate immune system in which C5a reacted with its receptors (C5aR1 and C5aR2) on PMNs (Ward 2004; Ward 2010a; Ward and Gao 2009). We described PMN dysfunction during sepsis, including defects in cytokine production and the respiratory burst, that were dependent on C5a and its receptors (Huber-Lang, et al. 2002). We showed C5a-C5a receptor signaling plays an essential role in neutrophil innate immunity during sepsis. Blockade of either the ligand or the receptor improved survival rates in experimental sepsis (Guo, et al. 2003). As a recent discovery related to the effects of C5a on PMNs, C5a interacted with its receptors on phagocytes (mainly on PMNs) resulting in PMN activation to form neutrophil extracellular traps (NETs). In these events, extracellular histones were released along with products from PMN granules (Fattahi, et al. 2015; Fuchs, et al. 2010). Histones acted as DAMPs when they were released into the extracellular space (Chen, et al. 2014). It has been shown that histones are released in response to inflammatory challenge, contributing to endothelial dysfunction, organ failure and death during sepsis (Kalbitz, et al. 2015; Xu, et al. 2009). Extracellular histones appeared to function as major mediators responsible for harmful effects of sepsis (Alhamdi, et al. 2015; Ekaney, et al. 2014; Kalbitz, et al. 2015; Wang, et al. 2015; Xu, et al. 2009). There was evidence of increased extracellular histones in plasma during sepsis which we found that ELISA kit can detect all of 4 individual histones (H1, H2A, H2B, and H4). Histones have been postulated to be a major cause of death during sepsis (Alhamdi, et al. 2015; Chaput and Zychlinsky 2009; Ekaney, et al. 2014; Kalbitz, et al. 2015; Nakahara, et al. 2013; Xu, et al. 2009). In “sterile sepsis” conditions such as ischemia/reperfusion, hemorrhagic shock or polytrauma, histones also appeared in the blood and may induce events in organs similar to those developing during infectious sepsis (Abrams, et al. 2013; Allam, et al. 2012). Extracellular histones play important roles in pathogenesis of sepsis due to development of apoptosis in a variety of cells and proinflammatory effects, which can induce cell damage or death (Xu, et al. 2015). Histones are released into the extracellular space in three ways, either freely, as a DNA-bound nucleosome, or as a component of NETs. All three types have been detected in serum after trauma and sepsis (Alhamdi, et al. 2015; Garcia-Gimenez, et al. 2017; Huber-Lang, et al. 2018; Kalbitz, et al. 2015; van der Poll, et al. 2017; Xu, et al. 2009). When in the extracellular space, histones act as DAMP proteins driving immune activation and cytotoxicity through interaction with toll-like receptors (TLRs), complement and cell membranes phospholipids (Marsman, et al. 2016; Silk, et al. 2017; Xu, et al. 2011). We have found the role of extracellular histones in polymicrobial sepsis to be very harmful for cells as there was evidence of damage to mouse hearts perfused with histones or isolated CMs incubated with histones, based on ECG tracings (Kalbitz, et al. 2015). A few studies (including our own) have shown that neutralizing antibodies to histones sharply reduced sepsis-induced organ dysfunction as well as greatly improving survival in septic mice (Kalbitz, et al. 2015; Xu, et al. 2009). Histone appearance in septic plasma was complement C5a- and PMN-dependent and that both C5a receptors (C5aR1 and C5aR2) were required for histone presence in septic plasma (Kalbitz, et al. 2015). When we used neutralizing histone antibody (BWA3) in our septic mice, there was remarkably reduced heart dysfunction (as defined by Echo-Doppler parameters) 8 hr after CLP-induce sepsis (Kalbitz, et al. 2015). We also showed increased levels of [Ca²⁺]i in mouse CMs exposed to purified histones (from calf thymus), caused loss of homeostasis, especially in septic mice (Kalbitz, et al. 2015). A recent paper has described appearance of extracellular histones (including H2B and H3) in plasma from septic humans (Garcia-Gimenez, et al. 2017; van der Poll, et al. 2017). Interestingly, the plasma levels of histones were remarkably similar to the levels we have found in mice with polymicrobial sepsis (Kalbitz, et al. 2015).

4. Sepsis causes NLRP3 inflammasome activation in CMs in a C5a-dependent manner

The inflammasome is a multiprotein complex that mediates the activation of proteins which are involved in the maturation and secretion of pro-inflammatory cytokines including IL-1β and IL-18 (Franchi, et al. 2009; Martinon, et al. 2002) and induction of ‘pyroptosis’ (cell death induced by bacterial products) (Fink and Cookson 2005). Caspase-1 (“IL-1 converting enzyme”) promotes secretion of these proinflammatory cytokines (Franchi, et al. 2009). NLRP3 (nucleotide-binding oligomerization domain–like receptor with pyrin domain), one of the NLR-subset inflammasomes, was introduced as an important mediator pathway of the innate immune system over the past decade (Martinon, et al. 2002). NLRP3 and the adaptor ASC (apoptosis-associated speck-like protein containing a caspase recruitment domain) are critical components of the inflammasome complex that links microbial and endogenous ‘danger’ signals to caspase-1 activation (Cerretti, et al. 1992; Mariathasan, et al. 2004; Martinon, et al. 2002; Thornberry, et al. 1992). PAMPs can activate caspase-1 via NLRP3 inflammasome (Kim, et al. 2010; Munoz-Planillo, et al. 2009). Several diseases have been associated with dysregulated activation of caspase-1 and secretion of IL-1β (Franchi, et al. 2009). There is also evidence of involvement of ASC in endotoxemia in mice (Mariathasan, et al. 2004). Several studies showed activation of NLRP3 inflammasome in different cardiac diseases or myocardial injuries (Bracey, et al. 2013; Pomerantz, et al. 2001), including our study on septic cardiomyopathy (Kalbitz, et al. 2016a). The studies have also shown the protective effects of blocking caspase-1 (Bracey, et al. 2013; Hwang, et al. 2001; Pomerantz, et al. 2001) or IL-1β (Abbate, et al. 2010) in myocardial ischemia. Using NLRP3 or Caspase-1 K.O. mice we showed remarkably decreased levels of extracellular histones in LPS-induced acute lung injury model (Grailer, et al. 2014). We also showed that extracellular histones could result in NLRP3 inflammasome activation (Grailer, et al. 2014). We have found evidence of the NLRP3 activation in heart during sepsis (Kalbitz, et al. 2016a). As shown in Figure 4a, lysates from three different normal Wt hearts contained components of the inflammasome (NLRP3, ASC, caspase 1 and IL-1β). In frame b and c mRNA studies were performed on mouse CMs using RT-PCR the NLRP3 and IL-1B primers. In frame b, control (con) CMs had mRNA for NLRP3 arbitrarily set at a value of 1, while addition of C5a or LPS caused 1.8 fold and 2.7 fold increases, respectively, in mRNA. Frame c shows that 16 hr after CLP hearts had a great increase (approximately 50 fold) of IL-1β mRNA. Details of the RT-PCR technology and sequence of the primers for NLRP3 and IL-1β were described in our previous report (Kalbitz, et al. 2016a). In frame d, frozen sections from LVs 12 hr after CLP stained for IL-1β showed upregulation of IL-1β consistent with the mRNA data in mouse heart after CLP (frame c). Control ventricle (not shown) did not show any staining for IL-1β. For IL-1β staining, rabbit anti-mouse antibody was used as primary (Abcam, Inc., Cambridge, MA, USA) and donkey anti-rabbit (AF-488) as secondary antibody (Jackson ImmunoResearch Laboratories). Details of the IF staining was described in our previous report (Kalbitz, et al. 2016a). The data in Figures 4 suggest that CMs contain the NLRP3 inflammasome. We showed that the absence of NLRP3 (NLRP3^−/− mouse) has a protective effect against cardiac dysfunction as well as cytokine production during sepsis (Kalbitz, et al. 2016a). We also found reduced levels of histones release in plasma in NLRP3 and Caspase-1 K.O. mice compared to the Wt after CLP (Kalbitz, et al. 2015). Collectively, inflammasome can be a good therapeutic target for reducing the harmful effect of the sepsis leading to septic cardiomyopathy.

Figure 4. Constitutive presence of components of the NLRP3 inflammasome and their upregulation in vitro and after CLP.

In frame a. three normal mouse hearts were homogenized and lysed (TRIzol), and then subjected to RT-PCR mRNAs for GADPH, NLRP3, ASC, caspase 1 and IL-1β. As is evident, all components were detected (no reverse transcriptase, no RT). b. In vitro, mouse CMs were incubated with C5a or LPS (4 hr at 37°C). As is evident, there was upregulation o NLRP3 in CMs incubated with either agonist. c. Mouse CMs in sham mice or in CLP mice (16 hr), the latter showing a very large increase (>50 fold) in mRNA for IL-1β. For each bar, n = 5. d. Frozen ventricles 12 hr after CLP stained for IL-1β.

5. Sepsis induces MAPKs activation in CMs in a C5a-dependent manner

Mitogen-activated protein kinase (MAPK) family includes extracellular signal-regulated kinase (ERK), p38 and c-Jun NH2-terminal kinase (JNK). MAPK signaling is involved in a variety of cellular activities including proliferation, differentiation, survival, and death. Deviation from the strict control of MAPK signaling pathways has been implicated in the of many human diseases (Kim and Choi 2010). Myocardial ischemia/reperfusion (I/R) was reported to activate the MAPKs pathways including ERK1/2, JNK1/2, p38; and other primary protein kinases like the cell survival kinase, Akt; and the sodium-hydrogen exchanger (NHE) kinase, p90RSK (Armstrong 2004). There are several reports of MAPK signaling in the myocardial I/R injury and attenuation of myocardial injury by modulating MAPKs activation (Gargiulo 2014; Jeong, et al. 2012; Kim, et al. 2012; Suchal, et al. 2016; Suchal, et al. 2017; Yu, et al. 2015). For instance, there is evidence that antioxidant and anti-inflammatory treatment is involved in MAPK signaling pathway (Suchal, et al. 2016). Using this protocol, there was a significant inhibition of active JNK and p38 proteins, associated with attenuation of septic cardiopathy or after I/R injury of the heart (Suchal, et al. 2016). However, there are limited data on the role of MAPKs in cardiac dysfunction during sepsis. Sepsis has been shown to induce upregulation of p38-(phosphorylation) in adult rat ventricular myocyte (Gupta, et al. 2005). Inhibition of TNF-α/p38-MAPK/caspase-3 signaling pathway showed protection effect in myocardial injury in septic rats (Zhang, et al. 2016). In our recent report, using flow cytometry techniques, we showed the activation (phosphorylation) of the MAPK proteins including JNK1/2 and ERK1/2, as well as Akt in the isolated CMs from septic rats at different time points after CLP. In addition, we could detect phosphorylation of MAPKs and Akt in rat CMs as a function of time after in vitro exposure to rrC5a (Fattahi, et al. 2017). The septic mice lacking C5a receptors, C5aR1 or C5aR2 K.O., showed decreased levels of phosphorylated MAPKs (p38, JNK1/2 and ERK1/2) and phosphorylated Akt on the heart from the septic mice (Fattahi, et al. 2017). It was of interest that selective inhibitor of p38 MAPK; water-soluble salt of SB 203580 could remarkably attenuate cardiac dysfunction in septic mice as measured by ECHO-Doppler and also by IF staining of the heart from the septic mice (Fattahi, et al. 2017). These data suggest that cardiac dysfunction following sepsis appears to be linked to activation of MAPKs and Akt in heart which is C5a- or C5a receptor-dependent. Moreover, using selective p38 inhibitor can reduce these activation responses in the heart.

Conclusions

In this review, we have described in polymicrobial sepsis (following CLP) the appearance of extracellular histones in plasma in a C5a and in an NLRP3 inflammasome-dependent manner. Sepsis also caused MAPKs activation in CMs in a C5a-dependent manner. Collectively these events led to functional abnormalities associated with the “cardiomyopathy of sepsis”.

Figure 5 summarizes our current data, linking C5a and C5a receptors and the role of extracellular histones to events that compromise the heart function during sepsis. In the proposed cascade, CLP caused robust complement activation with release of C5a, interacting with its receptors (C5aR1 and C5aR2), resulting in activation of the NLRP3 inflammasome and MAPKs in CMs and the appearance of extracellular histones, likely arising from activated neutrophils, which form NETs and resulting in the release of histones. These events induced reductions in key Ca²⁺ regulatory proteins (SERCA2 and NCX) as well as Na⁺/K⁺-ATPase in CMs, leading to inability to clear cytosolic Ca²⁺ after systole, defective action potentials, buildup of ROS and [Ca²⁺]i and release of cytokines, ultimately leading to the “cardiomyopathy of sepsis”. Our studies show evidence of involving complement C5a and its receptors, as well as extracellular histones in events that lead to cardiac dysfunction during sepsis (septic cardiomyopathy).

Figure 5. Proposed cascade of events after CLP leading to “cardiomyopathy of sepsis”.

Briefly, complement C5a and extracellular histones induce reductions in key Ca²⁺ regulatory proteins as well as Na⁺/K⁺-ATPase protein in CMs leading to inability to clear cytosolic Ca²⁺, together with activation NLRP3 inflammasome and MAPKs resulting in the cardiomyopathy of sepsis.

As details of these pathways emerge, it might be possible to develop therapeutic strategies for blockade of the following targets: extracellular histones, the cardiac NLRP3 inflammasome, or MAPK proteins. Such interventions as well as blockade of C5a or its receptors would be expected to diminish the cardiomyopathy developing in sepsis. The neutralizing mAb to C5a appears to be safe for use in humans (Hammerschmidt, et al. 1980; Stevens, et al. 1986). Since a significant number of septic patients develop cardiac failure which may be lethal, our data suggest that early interventions in sepsis such as low molecular weight inhibitors of C5a may prevent the sequence of events that leads to death because of impaired cardiac function. It should be noted that, with modern interventions that do not cause baro-trauma in lungs, the majority of septic patients survive and after recovery do not show evidence of cardiac dysfunction. We need to increase our knowledge in the field of sepsis to better understand molecular pathways responsible for cardiac dysfunction developing after sepsis. Such information would help us to target specific molecules or proteins to be used for potential treatment in patients with septic cardiomyopathy

Highlights.

• Cardiomyopathy is a common complication of sepsis which can lead to high mortality rates (~ 70%).

• Cardiac dysfunction of sepsis disappears after recovery from sepsis, indicating cardiac dysfunction in sepsis is reversible.

• Cardiomyopathy is complement C5a and its receptors –dependent which develop during early sepsis.

• Extracellular histones as well as complement C5a, after interacting with its receptors (C5aR1, C5aR2), cause defective action potentials, buildup of ROS and [Ca²⁺]i and release of cytokines, ultimately leading to the “cardiomyopathy of sepsis”

• Activation of NLRP3 inflammasome and MAPKs in CMs are also involved in the process of septic cardiomyopathy.