Temporally distinct regulation of pathways contributing to cardiac proteostasis during the acute and recovery phases of sepsis

Abstract

Background

Cardiac dysfunction is a common manifestation of sepsis and is associated with early increases in inflammation and decreases in myocardial protein synthesis. However, little is known regarding the molecular mechanisms regulating protein homeostasis during the recovery phase after the removal of the septic nidus. Therefore, the purpose of this study was to investigate diverse signal transduction pathways that regulate myocardial protein synthesis and degradation.

Methods

Adult male C57BL/6 mice were used to identify potential mechanisms mediating the acute (24 h) effect of cecal ligation and puncture (CLP) as well as long-term changes that manifest during the chronic (10 d) recovery phase.

Results

Acutely, sepsis decreased cardiac protein synthesis that was associated with reduced phosphorylation of S6K1/S6 but not 4E-BP1. Sepsis also decreased proteasome activity, although with no change in MuRF1 and atrogin-1 mRNA expression. Sepsis acutely increased apoptosis (increased caspase-3 and PARP cleavage), autophagosome formation (increased LC3B-II), and canonical inflammasome activity (increased NLRP3, TMS1, cleaved caspase-1). In contrast, during the recovery phase, independent of a difference in food consumption, global protein synthesis was increased, the early repression in proteasome activity was restored to basal levels, while stimulation of apoptosis, autophagosome formation and the canonical inflammasome pathway had abated. However, during recovery there was a selective stimulation of the non-canonical inflammasome pathway as evidenced by activation of caspase-11 with cleavage of Gasdermin D.

Conclusion

These data demonstrate a temporally distinct homeostatic shift in the cardiac proteostatic response to acute infection and recovery.

Introduction

The immune and metabolic alterations produced acutely by sepsis affect essentially all tissues and organ systems. However, the sepsis-induced changes in myocardial function are particularly relevant in the progression to septic shock and the ultimate recovery of the patient from the septic episode (1). Between 40–60% of all septic patients manifest with some type of cardiac dysfunction, with in vivo and in vitro preclinical studies demonstrating that sepsis negatively impacst a number of determinants of cardiac contractility and function (2). Moreover, sepsis-induced myocardial dysfunction is a key determinant of increased mortality in this condition and the ability to recover from the initial insult (3).

Although the etiology of this cardiac dysfunction is likely multifactorial, inflammation-enhanced oxidative stress appears to represent a central physiological process (4). The early sepsis-induced increase in proinflammatory cytokines in blood and heart is also associated with a reduction in global protein synthesis in this tissue (5), although the underlying mechanism has not been determined. Moreover, acting through a cytokine-mediated mechanism(s), sepsis also impairs the ability of nutrient signals to increase myocardial protein synthesis (6) potentially impairing the recovery of this organ. In addition to the synthesis of new proteins, cellular protein homeostasis is governed by alterations in protein degradation via a number of mechanisms, including the ubiquitin-proteasome pathway (UPP), apoptosis and autophagy (7). Because of the low mitotic potential of cardiomyocytes, proteostasis is of particular importance in maintaining normal heart function.

Despite recognizing that recovery from sepsis occurs over weeks or months and impairs quality-adjusted survival in humans (8), the large majority of published studies focus on the acute (< 48 h post-infection) phase of sepsis. Hence, there is limited information pertaining to the temporal progression by which protein homeostasis in heart is restored and changes that may develop during convalescence. The current study uses a murine model of cecal ligation and puncture (CLP) to study the acute phase of sepsis (24 h) as well as a CLP model that incorporates surgical removal of the punctured cecum to eliminate the septic nidus and the administration of antibiotics to investigate the convalescence or recovery phase (10 days) of sepsis (9, 10). Therefore, to investigate the aforementioned gaps in knowledge, we assessed temporal changes in the sepsis-induced signal transduction pathways regulating myocardial protein synthesis (i.e., protein translation initiation and elongation), ubiquitin-proteasome degradation, autophagy and apoptosis.

Materials and Methods

Animal care

Adult male C57BL/6 mice at 10 weeks of age (26.9 ± 1.9 g; Charles River Laboratories, Wilmington, MA) were used in all experiments described below. Mice were individually housed in polycarbonate cages with corncob bedding throughout the 1-week acclimation period and the subsequent experimental protocol. Standard rodent chow (Teklad Globak 2019, Harlan Teklad, Boston, MA) and water were provided ad libitum, and mice were maintained in a controlled environment (23°C; 30–70% humidity; 12h:12h light:dark cycle) in the animal care facility at the Pennsylvania State University (PSU) College of Medicine that is specific-pathogen free. All experiments were approved by the Institutional Animal Care and Use Committee at the PSU College of Medicine (#46-946-17) and adhered to National Institutes of Health (NIH) guidelines.

Experimental design

Mice were randomly assigned to either a septic group or a control group that was time-matched. This study employed two different models, one designed to mimic the acute phase of peritonitis and the other the recovery phase. While these two models are distinct (e.g., differing in nutritional intake and antibiotic treatment), they have been purposely designed to recapitulate the two time points after sepsis that were being studied.

Polymicrobial peritonitis was induced using CLP and separate groups of mice were studied in the acute phase of sepsis (i.e., 24 h post-CLP) and during the recovery phase (i.e., day 10 post-CLP), as previously described (9, 10). Briefly, mice were anesthetized with isoflurane (3–4% induction with 2–3% maintenance; Vedco, St. Joseph, MO) in 100% oxygen, and a warming blanket used to maintain body temperature. Isoflurane was selected because of its rapid induction time and the need for short-term anesthesia (< 10 min) during the CLP procedure. The abdomen was shaved and a midline laparotomy was performed after cleaning with betadine. The cecum was ligated and punctured through-and-through using a 25-g needle. Cecal material was extruded from the puncture site to ensure patency. The cecum was returned to the abdominal cavity and the rectus abdominis was closed using 5-0 silk and the skin closed with metal wound clips. One mL of warm sterile saline (0.9%) containing 0.05 mg/kg of the analgesic buprenorphine (Reckitt Benckiser Pharmaceuticals, Richmond, VA) was administered subcutaneously 30 min prior to the start of surgery to all mice. Sham-control mice underwent the same surgical procedure except the cecum was not ligated or punctured. Mice were subsequently transferred to a warming pad until they regained consciousness. As septic mice consume little or no food during the first night after CLP (9), both control and septic mice were fasted to ensure the same basal nutritional state. Mice were euthanized at 24 h after laparotomy ± CLP and this is referred to as the “acute sepsis” group.

A separate cohort of mice was used to examine changes 10 days post-CLP, and this group is referred to as “sepsis-recovery”. Control mice had their surgery performed one day after septic mice. This staggered approach permitted us to determine the food consumed by septic mice each day and then provide control animals with the same amount of food. Hence, over the 10-day course of the study, septic and control mice had the same nutritional intake and are considered pair-fed. For these sepsis-recovery mice, antibiotic (0.5 mg meropenem; Fresenius Kabi, Lake Zurich, IL) and buprenorphine were injected subcutaneously (total volume 1 ml of saline) every 12 h starting 24 h post-CLP (i.e., day 1) and continued for the next 5 days. Mice were anesthetized at 48 hrs post-CLP and the original incision was reopened. The ischemic portion of the cecum distal to the original suture ligation was then resected and the abscess surrounding the cecum was excised. The peritoneal cavity was washed with warmed saline (5 ml) to remove the debris and abscess. The abdominal incision was again closed and resuscitation provided as above, and mice were observed through day 10. Control mice received the same volume of fluid resuscitation with buprenorphine but without antibiotic. It is noteworthy that the two control groups (e.g., acute sepsis and sepsis-recovery) used in this study are not equivalent because of differences in the nutritional state of mice and the time post-surgery when mice were euthanized. These differences may account for the observed differences between the two control groups for some of the endpoints determined (discussed later).

In general, mice were monitored approximately every 8 hours and were weighed daily. If mice were observed to be moribund, immobile, unable to consume food and/or water, or had a rectal temperature < 30°C for more than 8 hours, mice were euthanized by certified research personnel that had successfully completed appropriate institutional animal training. Mice not meeting these humane endpoints were deeply anesthetized with isoflurane and euthanized by cardiac excision and exsanguination. There were no deaths in the control group (n = 10) during the 10-day experimental protocol. For the septic mice (n = 30), there were no deaths within the first 24 hrs, 2 deaths within 48 hrs, 6 deaths within 72 hrs, 8 deaths within 96 hrs, and 13 deaths after 120 hrs. No mice died after day 5 and the overall 10-day mortality was 43%. We believe this model is more representative of the recovery phase than a chronic form of septic disease because the nidus of the infection (i.e., cecum) was removed, the peritoneal cavity irrigated, and a 5-day course of antibiotics administered. As a result, food consumption of sepsis-recovery mice on days 5–10 did not differ from food consumption determined under basal pre-surgical conditions, and there is no difference in body weight between time-matched control mice and sepsis-recovery mice on days 9 and 10. Furthermore, the plasma IL-1β concentration had returned to basal control levels, and concentrations of TNFα and IL-6 were reduced 97–99.5% from concentrations observed 24 h post-CLP (9, 10). Finally, data from preliminary studies indicated there were no additional deaths when septic mice were observed between 10 to 21 days post-CLP (unpublished observation).

Cardiac ventricular tissue was excised at either 24 h or 10 d post-CLP. The heart was freeze-clamped to the temperature of liquid nitrogen and stored at −70°C until analyzed. Because of the extensive nature of the endpoints to be assessed (below), no tissue was processed for histological examination. This approach has the advantage of permitting systematic multiparameter analysis of mRNA and protein content and enzymatic activity from the same tissue that was necessary to assess molecular markers of myocardial protein balance.

Protein synthesis

In vivo protein synthesis was measured using the nonradioactive SUnSET method, as modified in our laboratory (9, 10). Puromycin (0.04 μmol/g body weight, dissolved in sterile saline) was administered by intraperitoneal injection 30 min prior to euthanasia (isoflurane as described above). At this relatively low-dose, puromycin is incorporated into elongating peptide chains and the relative global rate of protein synthesis is estimated by determining the synthesis of puromycin-labelled peptides by Western blotting (Kerafast, Boston, MA).

Western blot analysis

Briefly, a portion of heart was homogenized in ice-cold homogenizing buffer (in mM): 20 HEPES (pH 7.4), 2 EGTA, 0.2 EDTA, 100 KCl, 50 β-glycerophosphate, 50 NaF, 0.5 sodium orthovanadate, 1 Benzamidine, 0.1 PMSF and 1 DTT (9, 10). The protein concentration of each tissue homogenate was quantified (Bio-Rad Protein Assay, Hercules, CA) and SDS-PAGE was performed using equal amounts of total protein per sample. Additionally, each PDVF membrane was stained with Ponceau S (Aqua Solutions, Deer Park, TX) to verify equal protein loading. Blots were then blocked in 5% (w/v) nonfat dry milk and incubated overnight with primary antibody at 4°C. Antibodies (Cell Signaling, Beverly, MA, unless otherwise noted) included: (4E-BP1; Bethyl Laboratories, Montgomery, TX; A300–501), 4E-BP1(Ser65; #9451), S6K1 (Santa Cruz, Santa Cruz, CA; sc-230), ribosomal protein S6 kinase 1 (S6K1; Thr 389; #9205), rpS6 (#2217), rpS6 (Ser240/244; #2215), eukaryotic elongation factor (eEF2; #2332), eEF2 (Thr56; #2331), total eIF2Bε (#3595), p62 #5114), light chain 3B (LC3B; #3868), autophagy related protein (Atg) 12 (#4180), beclin (#3738), cleaved caspase-3 (#9661), cleaved nuclear poly (ADP-ribose) polymerase (PARP; #9542), activating transcription factor 4 (ATF4; courtesy Dr. Michael Killberg), c-Jun N-terminal kinase (JNK; #9252), JNK (Thr183/Tyr185; #9251), NLR pyrin domain containing 3 (NLRP3; #15101), adaptor protein containing a C-terminal caspase-recruitment domain (ASC; aka TMS1; #13833), caspase-1 (p10; Santa Cruz, CA; sc-514), Gasdermin (Santa Cruz; sc-393656), caspase 11 (#14340), calpain (Abcam, Cambridge, MA; ab28258), and calpastatin (Abcam; ab88065). Blots were then washed with an appropriate horseradish peroxidase conjugated secondary antibody and developed with enhanced chemiluminescence (ECL) reagents (Pierce Chemical, Rockford, IL) according to manufacturer’s instruction. Blots were imaged using the FluorChem (ProteinSimple, San Jose, CA) and densities in the linear range were quantified using Image J (NIH, Bethesda, MD). Each Western blot included samples from mice in the acute sepsis and sepsis-recovery groups as well as samples from the respective time-matched control mice.

Proteasome assay

A portion of heart was homogenized and reconstituted in cold assay buffer (Proteasome Activity Fluorometric Assay Kit; BioVision; Milpitas, CA) and the assay was performed exactly as described (9, 10). The protein content was assayed in a sample aliquot (BioRad; Hercules, CA). Each sample was measured in the presence and absence of the provided proteasome inhibitor to account for non-proteasomal degradation of the substrate. Proteasome activity (nmol/min/mg protein) was calculated by the change in the fluorescence signal where 1 unit of activity equals 1.0 nmol of the fluorophore 7-amino-4-methylcoumarin (AMC) per minute.

RNA extraction and real-time quantitative PCR

Total RNA was extracted using Tri-reagent (Molecular Research Center, Inc., Cincinnati, OH) and RNeasy mini kit (Qiagen, Valencia, CA) following manufacturers’ protocols. On-column DNase I treatment was used to remove residual DNA contamination. RNA was eluted from the column with RNase-free water and an aliquot was used for quantitation (NanoDrop 2000, Thermo Fisher Scientific, Waltham, MA). RNA quality was analyzed on a 1% agarose gel. Total RNA (1 μg) was reverse transcribed using superscript III reverse transcriptase (Invitrogen, Carlsbad, CA). Real-time quantitative PCR was performed using 25 ng of cDNA in a StepOnePlus system using TaqMan gene expression assays (Applied Biosystems, Foster City, CA) using primers as described previously by our laboratory (9, 10). The comparative quantitation method 2-ΔΔCt was used in presenting gene expression of target genes in reference to the endogenous control. Samples for determination of cardiac mRNA content in mice from both acute and recovery sepsis were run simultaneously.

Plasma CK-MB and cTnI

The plasma concentration of the MB fraction of creatine kinase (CK-MB) was determined on a VITROS 5600 automated biochemical analyzer (Ortho-Clinical Diagnostics, New York) and the plasma analysis for cardiac troponin I (cTnl) was performed by ELISA (Life Diagnostics, USA).

Data and statistical analysis

Data are presented as the mean ± standard error of mean (SEM) with the number of mice in each group presented in the figure legends. Unless otherwise noted, statistical analysis of the data were analyzed on commercial statistical software (SigmaPlot, Systat, San Jose, CA) using a two-way ANOVA (sepsis × time) with Student-Newman-Keuls (SNK) post hoc test. Difference were considered significant when P < 0.05 and such differences are graphically depicted by letters (e.g., a,b,c) above bars on each graph. There are no differences between group means when bars have the same letter, whereas bars with different letters indicate a statistically significant difference.

Results

Protein synthesis and mTOR signaling

Cardiac protein synthesis was reduced by 45% during the acute phase of sepsis but increased 30% above time-matched pair-fed control values during the recovery phase (Fig 1). As the mammalian or mechanistic target of rapamycin complex 1 (mTORC1) integrates various hormonal, nutrient and energy signals (11), we determined the activity of this kinase by assessing the extent of phosphorylation of two immediate downstream substrates S6K1 and 4E-BP1 (11) (Fig 2). S6K1 phosphorylation was reduced during the acute phase of sepsis and increased ~2-fold above control levels during recovery (Fig 2A, 2F). A comparable temporal change was seen with S240/244-phosphorylated S6 (Fig 2B, 2F) and eIF4B (data not shown), both authentic S6K1 substrates. In contrast, there was no consistent sepsis-induced change in S65-phosphorylated 4E-BP1 (Fig 2C, 2F).

Fig 1. Temporal changes in cardiac protein synthesis during the acute and recovery phases of sepsis.

Panel A, global protein synthesis was determined in cardiac ventricular tissue by measuring puromycin incorporation into protein, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Panel A, representative Western blot of control (C) and septic (S) mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Panel B, Representative Ponceau S blot demonstrating equal protein loading. Panel C, values in bar graph represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Values with different letters (a,b,c) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Fig 2. Temporal changes in proteins regulating protein synthesis in heart via translation initiation or elongation during the acute and recovery phases of sepsis.

Panels A–E, bar graphs in which values represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to the total amount of the respective protein or to tubulin, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Panel F, representative Western blots of control and septic mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Values with different letters (a,b,c) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Global protein synthesis can also be regulated by changes in the eIF2/eIF2B system (11) and therefore we assessed the relative amount of the catalytic epsilon (ε)-subunit for the eIF2B complex. Our data (Fig 2D) indicate the amount of total eIF2Bε was increased similarly during both the acute and recovery phases of sepsis. Protein synthesis can also be regulated at the level of peptide-chain elongation where specifically the up-regulation of T56-phosphorylated eEF2 inhibits translation elongation. The extent of eEF2 phosphorylation in heart during acute sepsis was increased, whereas there was no difference in eEF2 phosphorylation between control and septic mice during sepsis-recovery (Fig 2E, 2F).

Proteolysis

Protein balance can also be modulated by a change in protein degradation, and the UPP appears primarily responsible for catabolic condition-induced enhanced rates of protein breakdown (12). Therefore, we measured the in vitro protease activity of the 20S proteasome core that catalyzes the final step of the UPP. In the acute phase, the 20S proteasome activity was reduced 50% in hearts from septic mice, compared to control values; however, this decrement was not seen during the recovery phase (Fig 3A). The mRNA content for the muscle-specific ubiquitin-E3 ligases MuRF1 and atrogin-1 was also determined, but no time- or sepsis-induced changes in either ligase was detected (Fig 3B, 3C). Calpains can modify cytoskeletal proteins, among others, by selective and limited proteolytic degradation. Western blot analysis of calpain I indicated there was a main effect of sepsis, but only the increase in calpain I observed during the recovery period was statistically significant (Fig 3D, 3G). We also assessed calpastatin which is an endogenous inhibitor of calpain. Although we did not detect high molecular weight proteins (e.g., 100–120 kDa) as reported by others (13), a major band was visualized at ~75 kDa as well as multiple bands at lower molecular weights (e.g., ~50 and 25 kDa) that represent degraded calpastatin. For the 75 kDa form of calpastatin, there was a main time effect with levels being lower in the recovery compared to the acute phase (Fig 3E, 3G). Furthermore, sepsis increased the 75 kDa form of calpastatin during the recovery phase, compared to time-matched control values. In contrast, when the major degradative products of calpastatin (~50 and 25 kDa) were quantitated there was an 8-fold elevation in hearts from septic mice during recovery, compared to control values (Fig 3F, 3G).

Fig 3. Temporal changes in proteins regulating protein degradation in heart during the acute and recovery phases of sepsis.

Panel A, In vitro-determined 20S proteasome activity in heart; Panels B and C, mRNA content of the muscle-specific E3 ligases MuRF1 and atrogin-1 normalized to GAPDH. Panels D and E, protein content for calpain I and calpastatin where data were normalized to tubulin. For all bar graphs, values represent mean ± SEM from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to control values during the acute phase. Values with different letters (a,b,c) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Apoptosis

Programed cell death is enhanced in sepsis and can decrease cell number in several tissues (14). A predominant apoptotic signaling cascade involves the activation of caspase-3 and the subsequent cleavage of PARP. Acute sepsis more than doubled the amount of cleaved caspase-3 and cleaved PARP (Fig 4A–4B); however, there was no difference between septic and control mice during the recovery phase. Likewise, activation of JNK via its phosphorylation increases transcription of a number of pro-apoptotic genes. JNK phosphorylation on T183/Y185 was increased 2-fold during acute sepsis, while there was no difference in JNK phosphorylation between groups during sepsis recovery (Fig 4C). Finally, the unfolded protein response can lead to apoptosis via activation of the ATF/CHOP pathway. In contrast to the early sepsis-induced increased in caspase-3, cleaved PARP and phosphorylated JNK, ATF4 did not differ in heart between septic and control mice at either time point assessed (Fig 4D).

Fig 4. Temporal changes in proteins regulating apoptosis in heart during the acute and recovery phases of sepsis.

Panels A–D, bar graphs in which values represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to the total amount of the respective protein or to tubulin, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Panel E, representative Western blots of control and septic mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Values with different letters (a,b,c) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Autophagy

Autophagy is a proteolytic degradative process that removes cytosolic components and its dysregulation can produce cardiomyocyte death and cardiomyopathy (14). During autophagy the cytosolic form of LC3 (LC3-I) is conjugated to phosphatidylethanolamine to form LC3-II. Hence, the increase in myocardial LC3-II during the acute phase is suggestive of enhance autophagy (Fig 5A, 5E). However, this sepsis-induced increase was not seen during sepsis-recovery. Autophagy is a multiple step process requiring a number of key intracellular proteins, including Atg5/12, p62 and Beclin. In this regard, an increase in Atg5/12 and Beclin and decrease in p62 are often associated with an increase in autophagy. In contrast to the early increase in LC3B-II described above, the relative content for each of these regulatory proteins did not differ between control and septic mice at either time point examined (Fig 5B–5E). However, for the Atg5/12 complex and p62 there was a time effect as the relative content of these two proteins was lower in the recovery phase compared to 24-h post-CLP.

Fig 5. Temporal changes in proteins regulating autophagy in heart during the acute and recovery phases of sepsis.

Panels A–D, bar graphs in which values represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to tubulin, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Panel E, representative Western blots of control and septic mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Values with different letters (a,b) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Inflammasome

The inflammasome is a multiprotein complex that when activated by either the canonical or non-canonical pathway triggers the inflammatory process and can lead to cell death (15). Protein expression for components associated with the canonical inflammasome pathway, such as TMS1, NLRP3 and cleaved caspase 1, were increased in heart during the acute phase of sepsis but not during the recovery phase (Fig 6). We were unable to detect cleavage of pro-IL-1β to the mature protein (data not shown), the ultimate end product of increased caspase 1 activity. In contrast, expression levels of caspase 11 together with total and cleaved Gasdermin D, the latter being a marker of non-canonical inflammasome pathway activation, were consistently and markedly increased during the recovery phase but not at 24-h post-CLP (Fig 7).

Fig 6. Temporal changes in proteins regulating canonical inflammasome activity in heart during the acute and recovery phases of sepsis.

Panel A, representative Western blots of control and septic mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Panels B–D, bar graphs in which values represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to tubulin, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Values with different letters (a,b) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Fig 7. Temporal changes in proteins regulating non-canonical inflammasome activity in heart during the acute and recovery phases of sepsis.

Panel A, representative Western blots of control and septic mice during the acute phase (24 h) and the recovery phase (10 day) of sepsis produced by cecal ligation and puncture. Samples from all four experimental groups were run on the same gel and the white bar indicates that intervening lanes of the gel have been removed for presentation purposes. Panels B–D, bar graphs in which values represent mean ± SEM for Western blot data from all 4 groups (n = 10, 14, 10, 14, respectively). Data were normalized to tubulin, and the data were expressed as the percentage of the pair-fed control value during the acute phase. Values with different letters (a,b) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05).

Cardiac inflammation and damage

We also assessed various markers of inflammation in heart tissue. The mRNA content for IL-1β, IL-6 and TNFα was increased at 24 h post-CLP, but the expression of these cytokines did not differ from control values during sepsis-recovery (Table 1). Finally, the plasma concentrations for CK-MB (primarily found in the heart) and cTnI (a sensitive and specific indicate of myocardial damage) were both increased at the 24-h time point but were not elevated during sepsis-recovery (Table 1). These data collectively suggest the early sepsis-induced increase in cardiac inflammation and damage have largely resolved by day 10.

Table 1.

Cardiac inflammation and damage markers during acute sepsis and recovery

	Acute Phase (24 h)		Recovery Phase (10 days)
	Control	Sepsis	Control	Sepsis
Cardiac mRNA, AU/GAPDH
IL-1β	1.00 ± 0.08a	5.87 ± 2.13b	1.05 ± 0.13a	1.16 ± 0.15a
IL-6	1.00 ± 0.15a	25.7 ± 7.4b	1.21 ± 0.14a	1.54 ± 0.47a
TNFα	1.00 ± 0.11a	4.33 ± 1.2b	0.97 ± 0.10a	1.11 ± 0.22a
Plasma
CK-MB, U/L	145 ± 27a	985 ± 177b	132 ± 38a	166 ± 41a
cTnI, ng/ml	1.1 ± 0.2a	3.3 ± 0.5b	0.8 ± 0.3a	1.1 ± 0.3a

Values represent mean ± SEM for data from all 4 groups (n = 10, 14, 10, 14, respectively). Values with different letters (a,b) are statistically different from each other by 2-way ANOVA with follow-up SNK (P < 0.05). IL-1β, interleukin-1β; TNFα, tumor necrosis factor-α; CK-MB, creatine kinase MB fraction, cTnl, cardiac troponin I.

Discussion

Our data demonstrate that sepsis acutely modulates specific protein synthetic and degradative pathways in whole cardiac tissue and suggests compensatory mechanisms are invoked to restore proteostasis during the recovery phase (Fig 8). On the synthetic side of the protein balance equation, sepsis acutely decreased global protein synthesis. As mTORC1 is a key regulator of protein synthesis (11), we assessed authentic downstream substrates for this kinase. In this regard, the decrease in protein synthesis was associated with a selective reduction in the activity of S6K1 in the absence of a detectable decrement in 4E-BP1 phosphorylation. Moreover, as eEF2 phosphorylation was increased by sepsis, it appears that a sepsis-induced decrease in both translation initiation and elongation may contribute to the early decrease in protein synthesis. It is noteworthy that these results differ from those previously reported where sepsis did not acutely alter cardiac protein synthesis (16, 17). Possible explanations for this difference might include the species used (mouse vs rat) or, of potentially greater importance, the type and/or severity of the sepsis model (CLP vs intraabdominal inoculation of purified Escherichia coli and Bacteroides fragilis). In contrast, the recovery phase was characterized by an up-regulation of cardiac protein synthesis that occurred with a coordinated increase in S6K1/S6 phosphorylation and the return of eEF2 phosphorylation to baseline. This temporal progression of sepsis-induced changes in cardiac protein synthesis is comparable to that reported for skeletal muscle (9).

Fig 8. Schematic diagram of changes in proteostasis in the heart during the acute (top panel) and recovery (bottom panel) phases of sepsis produced by cecal ligation and puncture.

Thick solid lines and arrows indicate a stimulation of the respective pathway, whereas a thick dotted line indicates stimulation of an inhibitory pathway; thin lines indicate no detectable change in the activation state of the pathway. Abbreviations: 4E-BP1; eukaryotic initiation factor 4E binding protein-1; mTORC1, mammalian target of rapamycin complex 1; S6, ribosomal protein S6; S6K1, ribosome protein S6 kinase 1; UPP, ubiquitin proteasome pathway; eEF2, eukaryotic elongation factor; LC3B, light chain 3B; Atg, autophagy related protein; PARP, nuclear poly (ADP-ribose) polymerase; ATF4, activating transcription factor 4; JNK, c-Jun N-terminal kinase; NLRP3, NLR pyrin domain containing 3, TMS1, (aka ASC; adaptor protein containing a C-terminal caspase-recruitment domain; IL, interleukin; S, serine; T, threonine.

Alterations in protein breakdown also impact tissue protein balance and sepsis has consistently been shown to increase the UPP activity in skeletal muscle from rodents and humans (18, 19). In contrast, there is a paucity of data pertaining to sepsis-induced changes in UPP activity in heart. In contradistinction to skeletal muscle, we show that sepsis acutely decreases 20S proteasome activity, a response comparable to the decrease in cardiac 20S B1 and B2 as well as 26S B5 activity seen 8 hrs after injection of lipopolysaccharide (LPS) (20). Whereas the sepsis-induced increase in proteasome activity in skeletal muscle is associated with a coordinated increase in the expression of muscle-specific atrogenes (21), we detected no alterations in myocardial MuRF1 or atrogin-1 mRNA content during acute sepsis suggesting an alternative mechanism is operational in cardiac tissue. A similar disconnect between these E3 ligases and UPP activity has been previously reported in skeletal muscle (22).

The activation of calpains can be an important regulator of skeletal muscle breakdown allowing for the release of actin and myosin that are then subsequently degraded by the UPP (23). Others have reported that an acute sepsis-induced increase in cardiac calpain-1 mRNA and protein content was associated with a reduction in the contractile proteins α-actin and heavy-chain cardiac myosin (24). However, no such change in calpain I was detected in the acute phase of the current study. This cardiac response differs from the sustained sepsis-induced increase in the UPP in skeletal muscle that leads to the characteristic atrophic response in this tissue (11). Overall, it appears that myocardial protein content and mass are preserved during early sepsis by the coordinated and proportional change in protein synthesis and protein breakdown.

Apoptotic death of cardiomyocytes is potentially detrimental as the heart has a limited regenerative capacity and cell loss, by any mechanism, can impair myocardial contractility. Several independent laboratories have provided relatively consistent data indicating that sepsis acutely increases cardiac apoptosis as illustrated by the increased presence of TUNEL (Terminal deoxynucleotidyl transferase dUTP nick end labeling)-positive cells as well as increased protein expression of BAX (bcl-2-like protein 4), cleaved caspase-3 and cleaved PARP (25–28). Furthermore, chemical inhibition of PARP cleavage using 3-aminobenzamide has been reported to prevent apoptosis and mitochondrial damage due to sepsis (28). Sepsis also increases JNK phosphorylation which is mediated by IL-1β and TNFα, both of which are increased in the blood and heart of septic animals and can induce apoptosis and contractile dysfunction (25, 29). Our data obtained 24 h post-CLP are entirely consistent with these early reports. Moreover, our data suggest that the early sepsis-induced increase in apoptosis is ATF4-independent. In contrast, the initial increase in markers of cardiac apoptosis is transient and abates during the recovery phase of sepsis as the relative abundance of cleaved caspase-3 and PARP, and the phosphorylation of JNK did not differ between septic and control mice at day 10. While the early increase in apoptosis may impair contractile function, apoptotic cell death may limit inflammation thereby minimizing further release of damage-associated molecular patterns and dampening the immune response in heart. Hence, the return of apoptosis to basal control levels during recovery may suggest an amelioration of the inflammatory state, which was confirmed by the return of IL-1β, IL-6 and TNFα mRNA content to control levels.

Autophagy is a type of programmed cell death involving lysosomal degradation and recycling of cellular proteins and organelles. This process can often be cytoprotective serving a pro-survival function in catabolic conditions (14), and inhibition of autophagy increases sepsis mortality (30). Sepsis acutely increases formation of autophagic vacuoles as assessed by electron microscopy (31) and increases LC3-II stained punctate cells in the left ventricle and LC3-II in whole cardiac tissue (14). Increases in LC3-II can indicate either an increase in autophagosome biogenesis or a block in lysosomal degradation. Previous work revealed that both Ras-related protein Rab7 and lysosomal associated membrane protein (LAMP)-1 were decreased 4–24 h after CLP. Moreover, acute sepsis decreased the colocalization of LC3 with LAMP1 suggesting inhibition of autophagosome-lysosome fusion. Finally, degradation of autophagosomes appears decreased as evidenced by an increased number of larger autophagosomes containing mitochondria and partial digestion of autophagosome contents after CLP (14). While our data do confirm an early sepsis-induced increase in LC3-II, we did not detect concomitant changes in beclin and the Atg5–12 complex, or a decrease in p62 content which would collectively imply an increased in autophagy. Hence, while our data suggest that sepsis increases autophagosome formation, they do not provide unequivocal evidence for an overall change in autophagic flux. As with the above mentioned changes in apoptosis, the sepsis-induced increase in LC3-II was not sustained having returned to basal control levels during the recovery phase.

Pyroptosis is a distinct form of cell death defined as being caspase-1-dependent and is enhanced by numerous environmental prompts, including sepsis (32). Caspase-1 plays little role in apoptosis and conversely, the apoptotic caspases, including caspase-3, do not appear to be predominantly involved in pyroptosis (32). What is referred to as the canonical inflammasome pathway includes the NLRP3 complex consisting of NLRP3, the adapter protein TMS1 (ASC1) and caspase-1. Association of caspase-1 with this multi-protein complex allows its processing and activation that in turn facilitates the cleavage of inactive precursors of IL1β and IL-18 to form the mature cytokines (32), and sepsis has been reported to increase IL-1β in muscle and blood (6). The relative importance of this pathway in sepsis is evidenced by the smaller decrement in cardiac contractile function and improved survival of septic mice with a gene deletion in NLRP3 or caspase-1 (15, 33). Moreover, pretreatment of rats with an IL-1 receptor antagonist prevents the sepsis-induced impairment in nutrient-stimulated myocardial protein synthesis (17). However, activation of caspase-1 appears to have other detrimental effects as neutralization of IL-1 and IL-18 during sepsis does not completely mimic the phenotype seen in caspase-1 deficient septic mice (33). Our data from septic mice in the acute phase confirm these early observations and extend them by demonstrating that, similar to the activation of apoptosis and autophagosome formation, these sepsis-induced changes were transient had resolved during the recovery phase.

A non-canonical NLRP3 pathway has also been described whereby caspase-11 functions as a pattern recognition receptor and is activated by cytosolic LPS (34). Caspase-11 activation results in proteolytic cleavage of the cytosolic protein Gasdermin D (35). The newly generated N-terminal fragment of Gasdermin D associates with the plasma membrane thereby enhancing pore formation and ultimately causing cell death. Activation of this pathway was seen only during the recovery phase of sepsis as evidenced by the up-regulation of caspase-11 and Gasdermin D cleavage. Hence, sepsis regulated both the canonical and non-canonical NLRP3 pathways in a temporally distinct manner, a pattern not previously reported in heart or other tissues. As the caspase-11/Gasdermin D pathway is also necessary for the secretion of HMGB1 and IL-1α (35), our data suggest there is potential for continued release of selected inflammatory cytokines and DAMPs that may slow recovery. The physiological function of the up-regulation of this non-canonical pathway remains to be determined.

The primary limitation of the current study is the lack of

In this study, we have identified the signal transduction pathways that are up-regulated (e.g., apoptosis, autophagosome formation, canonical NLRP3 pathways) or down-regulated (protein synthesis and proteasome activity) in heart during the acute phase of CLP-induced sepsis in adult male mice. While apoptosis, autophagy and UPP activation returned to baseline control levels during recovery, our work reveals that myocardial protein synthesis is increased and pyroptosis via the non-canonical pathway is activated during convalescence. Hence, there are changes in proteostasis that persist long after the initial septic insult has resolved that may potentially impact heart structure and function. However, we are unable to directly determine the physiological importance of the sepsis-induced changes in signal transduction as the current study was limited by the lack of histological evidence and, more importantly, by any direct assessment of cardiac function. Future studies will be required to characterize both sepsis-induced changes in cardiac function and various aspects of proteostasis to help identify potential mechanisms by which sepsis leads to frailty and deconditioning due to weakening of cardiac reserve in the human population.