Potential conflict of interest: Nothing to report.
Adaptive responses to sepsis are necessary to prevent organ failure and death. Cellular signaling responses that limit cell death and structural damage allow a cell to withstand insult from sepsis to prevent irreversible organ dysfunction. One such protective pathway to reduce hepatocellular injury is the up-regulation of heme oxygenase-1 (HO-1) signaling. HO-1 is up-regulated in the liver in response to multiple stressors, including sepsis and lipopolysaccharide (LPS), and has been shown to limit cell death. Another recently recognized rudimentary cellular response to injury is autophagy. The aim of these investigations was to test the hypothesis that HO-1 protects against hepatocyte cell death in experimental sepsis in vivo or LPS in vitro via induction of autophagy. These data demonstrate that both HO-1 and autophagy are up-regulated in the liver after cecal ligation and puncture (CLP) in C57BL/6 mice or in primary mouse hepatocytes after treatment with LPS (100 ng/mL). CLP or LPS results in minimal hepatocyte cell death. Pharmacological inhibition of HO-1 activity using tin protoporphyrin or knockdown of HO-1 prevents the induction of autophagic signaling in these models and results in increased hepatocellular injury, apoptosis, and death. Furthermore, inhibition of autophagy using 3-methyladenine or small interfering RNA specific to VPS34, a class III phosphoinositide 3-kinase that is an upstream regulator of autophagy, resulted in hepatocyte apoptosis in vivo or in vitro. LPS induced phosphorylation of p38 mitogen-activated protein kinase (p38 MAPK), in part, via HO-dependent signaling. Moreover, inhibition of p38 MAPK prevented CLP- or LPS-induced autophagy. Conclusion: Sepsis or LPS-induced autophagy protects against hepatocellular death, in part via an HO-1 p38 MAPK-dependent signaling. Further investigations are needed to elucidate how autophagic signaling prevents apoptosis and cell death. (HEPATOLOGY 2011;)
Sepsis is a systemic inflammatory response that occurs as a consequence of an infectious insult. It is a significant health problem, with a mortality rate of 30%-60%. The predominant cause of morbidity and mortality is the development of multiple system organ dysfunction with subsequent organ failure. The cause of early organ dysfunction in the setting of sepsis is secondary both to cellular activation by bacterial products, including lipopolysaccharide (LPS), elaborated inflammatory cytokines, as well as hemodynamic abnormalities, leading to decreased oxygen delivery. Interestingly, early organ dysfunction from sepsis usually is not associated with cell death. Several studies have illustrated that in response to infection and sepsis, cells will undergo a metabolic shutdown as an adaptive response to protect against tissue injury and long-term structural damage.1
Mitochondria are responsible for greater than 90% of the body's oxygen consumption through oxidative phosphorylation and adenosine triphosphate (ATP) production, with less than 2% of this consumption leading to the production of oxygen free radicals (reactive oxygen species [ROS]) (i.e., hydroxyl radical, superoxide, or hydrogen peroxide) under normal respiration.2 Under physiological conditions, these ROS play an important role in cell signaling, leading to the induction of adaptive cellular responses. However, continued or excessive production of ROS, as can occur in sepsis, can be deleterious to mitochondria and other organelles.3 If injured or dysfunctional organelles and proteins are not addressed by adaptive responses, cells will die. This death can potentiate cellular injury as well as result in structural and irreversible damage to the organ. Adaptive responses include processes that are aimed at dealing with damaged organelles and proteins, allowing cells and tissues to recover. One such adaptive response is autophagy.
Autophagy is a well-conserved, intracellular, catabolic process where proteins and organelles are isolated by a double-membrane vesicle (i.e., autophagosome) targeted to the lysosome and degraded into their subcomponents, which can then be recycled.4 Specifically, mitochondrial autophagy (or mitophagy) can consume damaged and dysfunctional mitochondria to limit further ROS production, prevent the release of cytochrome c and mitochondrial death signaling, and potentially contribute to the regulation of oxygen consumption.5 Based on this, we hypothesized that autophagy is a protective response in sepsis to limit cellular death. Furthermore, we hypothesized that autophagy is regulated by heme oxygenase-1 (HO-1), which is part of a vital cell-signaling pathway that occurs in response to cellular injury or stress.6 HO-1 has been recognized as a protein that is essential to limit inflammation and prevent cell death or apoptosis, but the mechanisms, including a link to autophagy, are not well defined.
ATP, adenosine triphosphate; CLP, cecal ligation and puncture; HO-1, heme oxygenase-1; LPS, lipopolysaccharide; p38 MAPK, p38 mitogen-activated protein kinase; PI3K, phosphoinositide 3-kinase; siRNA, small interfering RNA; SnPP, tin protoporphyrin-IX; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling.
Materials and Methods
Primary mouse hepatocytes were harvested from C57BL/6 mice as previously described.7 They were cultured in William E media supplemented with penicillin (100 U/mL), streptomycin (100 μg/mL), insulin (0.16 mL), HEPES buffer (7.5 mL) (Gibco), and 5% fetal bovine serum (Gibco) on either gel-coated plates for protein extraction or coverslips for immunohistochemistry. Cells were used on day 2 of harvest. HO was inhibited with tin protoporphryin-IX (SnPP) (50 μM; Frontier Scientific) a known, nonspecific inhibitor of HO, or HO-1–specific small interfering RNA (siRNA) (50 μM; Ambion). Autophagy was inhibited with 3-methyladenine (2 mM; Sigma), a chemical inhibitor of phosphoinositide 3-kinase (PI3K), or with VPS34 siRNA (50 μM; Ambion). The p38 mitogen-activated protein kinase (p38 MAPK) was inhibited with SB203580 (20 μM; Calbiochem). One hour after the above treatments, cells were stimulated with LPS (100 ng/mL) to simulate bacterial infection. Cells were maintained at 37°C and 5% CO2 for 0-24 hours. Notably, all cell-culture experiments using LC3 western blotting (a protein central to autophagosome formation) as an endpoint were performed with the addition of the protease inhibitors, E-64d (10 mg/mL; Enzo Life Sciences) and pepstatinA (10 mg/mL; Sigma). Cells were then either harvested for protein extraction or fixed to coverslips with paraformaldehyde for immunohistochemistry.
Cecal Ligation and Perforation.
Animal protocols were approved by the University of Pittsburgh Institutional Animal Care and Use Committee. Experiments were performed in adherence to the National Institutes of Health Guidelines on the Use of Laboratory Animals. Cecal ligation and perforation (CLP) was performed on C57BL/6 male mice (Jackson Laboratories, Bar Harbor, ME) 6-8 weeks in age and weighing 20-25 g. These animals were anesthetized with pentobarbital (70 mg/kg, intraperitoneal [IP]). A 1- to 2-cm midline laparotomy was performed, and the cecum was identified. The stool was then manipulated to the tip of the cecum and was subsequently ligated 1 cm from the tip with a 2-0 silk tie. The cecum was then perforated with a 22-gauge needle and returned into the abdomen. The muscle and skin were closed with a running 2-0 silk suture. Sham-operated animals underwent laparotomy and bowel manipulation without ligation or perforation. Tissue and blood collection occurred at either 8 or 20 hours post-CLP. No antibiotics were used, and animals had free access to food and water pre- and postoperatively.
HO was inhibited in vivo through an IP injection of SnPP (50 mg/kg) 1 hour before CLP or with the use of in vivo HO-1–specific siRNA (50 μM/kg) (Invitrogen). This was administered via hydrodynamic tail vein injection, where the siRNA was made to the correct concentration in 2 mL of lactated ringers and given 3 days before CLP. The rapid injection of this large volume creates significant pressure to help promote siRNA uptake. Scrambled siRNA (50 μM/kg) was used as a control again via hydrodynamic tail vein injection. Autophagy was inhibited through the use of in vivo siRNA against VPS34 (50 μM/kg; Invitrogen). p38 MAPK was inhibited in vivo through IP injections of SB203580 (10 mg/kg; Calbiochem) 1 hour before CLP.
Cells were fixed on coverslips with paraformaldehyde for 15 minutes and then rinsed with cold phosphate-buffered saline (PBS). Slides were then stained for LC3 (Novus) to monitor autophagy.
Liver tissue from mice was removed after perfusion with cold PBS and paraformaldehyde. Tissue was then placed in paraformaldehyde for 1 hour, then switched to 30% sucrose solution for 12 hours. The tissue was then slowly frozen in 2-methylbutane. Sections were stained using antibodies against LC3, HO-1 (Enzo Life Sciences), VPS34 (Invitrogen), and phosphorylated p38 MAPK (Cell Signaling). Images were taken with a Zeiss 510 inverted confocal microscope.
Western Blot Analysis.
Primary mouse hepatocytes were washed with cold PBS, collected in lysis buffer, sonicated, spun (10,000g for 15 minutes), and then the supernatant was transferred to a new tube. Protein concentrations were determined using the bicinchoninic acid protein assay kit. Samples were then mixed with loading buffer and run on a 15% sodium dodecyl sulfate–polyacrylamide gel. This gel was then transferred to a polyvinylidene fluoride membrane at 250 mA for 2 hours. The membrane was blocked in 5% milk for 1 hour and then incubated in primary (LC3 or activated caspase-3; Cell Signaling Technology) in 1% milk or phosphorylated p38 MAPK in 5% bovine serum albumin overnight. Membranes were in TBS-Tween 20 (TBST) for 30 minutes, then placed in secondary antibody linked to horseradish peroxidase for 1 hour and washed for 1 hour in TBST before being developed using a chemiluminescence substance (Thermo Scientific).
For electron microscopy, mice were perfused with cold PBS, then with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.4) and processed for transmission electron microscopy (TEM) as described.8 After dehydration, thin sections were stained with uranyl acetate and lead citrate for observation under a JEM 1011CX electron microscope (JEOL, Peabody, MA). Images were acquired digitally from a randomly selected pool of 10-15 fields under each condition.
Fixed cells or tissue samples underwent terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining with the En Roche kit, per the manufacturer's protocol. Images were taken with a Zeiss 510 inverted confocal microscope.
Mitochondrial Membrane Potential.
Mitochondrial membrane potential was determined using the mitochondrial dye, JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolocarbocyanine iodide; Molecular Probes, Eugene, OR). In the cytosol and mitochondria at low membrane potential, the monomeric form of JC-1 fluoresces green (emission at 525 nm), whereas within the mitochondrial matrix at high membrane potentials, JC-1 forms aggregates that fluoresce red (emission at 590 nm). Samples were incubated with JC-1 at a final concentration of 1 μM at 37°C for the last 30 minutes of the experiment. Flow cytometry (Guava, Millipore) was used, and red and green fluorescence was determined. Results are expressed as the ratio of red:green fluorescence. Total cell counts were also obtained through the use of flow cytometry.
Cell Titer-Glo luminescent cell viability assay (Promega), per the manufacturer's instructions, was used for the quantification of ATP content. Luminescence was measured using the SoftMaxPro ATPase Assay program on a Synergy Mx (Biotek) plate reader.
Cecal Ligation and Puncture or LPS-Induced Autophagy in Liver Cells.
C57BL/6 mice were randomized to sham operation or cecal ligation and puncture. Mice were sacrificed 8 or 20 hours after this insult, and liver tissue was collected. Induction of autophagy was determined using western blotting, immunohistochemistry, and TEM. These data demonstrate that cecal ligation and puncture increases hepatic autophagosome formation at 8 hours by electron microscopy (Fig. 1A). Additionally, these micrographs demonstrate evidence of mitophagy and, further, increased LC3 staining and punctae formation, as shown by western blotting and fluorescent immunohistochemistry (Fig. 1B,C). LC3 is a terminal autophagic protein that, upon the induction of autophagy, becomes membrane bound to the autophagosome and is a classic marker used for monitoring changes in autophagy. Increased autophagy is seen at both early and later time points (data are shown for 8 hours). LPS treatment of primary mouse hepatocytes (100 ng/mL) in vitro also resulted in a time-dependent increase in LC3 protein expression and punctae formation, as shown by western blotting and immunohistochemistry (Fig. 2A,B). Together, these data suggest that autophagy is part of the adaptive stress response to infection or LPS.
LPS or CLP Transiently Decreased Cellular ATP and Mitochondrial Membrane Potential, but Did Not Induce Cell Death.
The influence of experimental sepsis on hepatic cell death was investigated. LPS treatment of primary mouse hepatocytes in vitro resulted in a transient decrease in mitochondrial membrane potential and cellular ATP levels (Fig. 3A,B). These values were maximally decreased 4 hours after LPS and normalized by 24 hours. There was no evidence of hepatocyte cell death, as measured by cell counts (Fig. 3C) as well as crystal violet or TUNEL staining (data not shown). Consistent with previous studies, experimental sepsis does not result in significant liver cell death at the time points examined, as determined by TUNEL staining9, 10 (Fig. 4C).
Sepsis-Induced Autophagy Is Dependent on HO-1.
HO-1 is up-regulated in response to both heme and nonheme stress in cells and when HO-1 is knocked down or activity is inhibited; tissues, including the liver, demonstrate increased injury in response to insults, such as ischemia/reperfusion, hemorrhage, and immune-mediated hepatitis.7, 11 In addition, HO-1 is a key protein in the adaptive response to infection. Mice deficient in HO-1 (hmox1−/−) have an increased susceptibility to infection.12 Consistent with previous findings, hepatic HO-1 is up-regulated in response to cecal ligation and puncture, as determined by real-time polymerase chain reaction (RT-PCR) and western blotting (Fig. 4A,B). As demonstrated above, CLP results in increased autophagy, as demonstrated by immunohistochemistry for LC3. Inhibition of HO activity using SnPP resulted in decreased LC3 staining by immunohistochemistry (Fig. 4C). Furthermore, inhibition of HO activity increased cell death in the liver, as demonstrated by increased TUNEL staining (Fig. 4C). VPS34 is a class III PI3K that is important in promoting autophagic signaling. The influence of HO-1 on sepsis-induced autophagy was also determined using HO-1–specific siRNA. Knockdown of HO-1 inhibited CLP-induced up-regulation of LC3, as well as prevented up-regulation of the proximal autophagy-inducing protein, VPS34 (Fig. 4D).
The role of HO-1 in the induction of autophagy and protection against cell death was further investigated in LPS-treated hepatocytes. LPS-induced autophagy was diminished by inhibition of HO activity using SnPP or knockdown of HO-1 with HO-1–specific siRNA (Fig. 5A,B). Additionally, inhibition of HO activity resulted in increased hepatocyte cell death and apoptosis with SnPP treatment (Fig. 5C,D). Moreover, overexpression of HO-1 by adenoviral gene transfer in hepatocytes increases autophagy, compared to control adenoviral transfection; however, in LPS-treated hepatocytes, there is no additional gain of function with this gene transfer (data not shown). This may be secondary to the finding that HO-1 levels are substantially increased in LPS-treated cells.
Inhibition of Autophagy Increases Liver Injury and Cell Death from Sepsis.
To determine whether autophagic signaling contributes to protection against cell death, autophagy was suppressed in vivo by siRNA specific for VPS34. CLP-induced VPS34 was shown to be dependent on HO-1 (Fig. 4D). C57BL/6 mice were treated with scrambled or VPS34-specific siRNA by hydrodynamic tail vein injection 72 hours before cecal ligation and puncture. Adequate knockdown of VPS34 in the liver was determined by RT-PCR of liver homogenates and immunohistochemistry (Fig. 6A,B). The influence of autophagy was measured by TEM, looking for autophagosomes (Fig. 6C), and immunohistochemistry, in a search for LC3 (data not shown). CLP increased hepatic autophagic signaling in control scrambled siRNA-treated mice; however, VPS34 siRNA-treated mice demonstrated minimal induction of autophagy (Fig. 6C). Moreover, inhibition of autophagy with VPS34 siRNA resulted in increased cell death and tissue injury, as measured by alanine aminotransferase (Fig. 6D,E). These findings were recapitulated in LPS-treated hepatocytes, which demonstrated decreased autophagy and increased cell death after treatment with VPS34 siRNA or the pharmacological inhibitor 3-MA.
HO-1 Induces Autophagy Through Activation of p38 MAPK.
The mechanism by which HO-1 induces autophagy was next investigated. These studies focused on the MAPKs, which have been shown to modulate the induction of autophagic signaling. LPS-treated hepatocytes increased the phosphorylation of p38 MAPK (Fig. 7A), but not p42/44 MAPK or JNK (data not shown). HO inhibition with SnPP decreased LPS-induced phosphorylation of p38 MAPK. Additionally, pharmacological inhibition of p38 MAPK with SB203580 (10 μM) prevented LPS-induced autophagic signaling, as demonstrated by LC3 punctae formation (Fig. 7B) and western blotting (Fig. 7C). Similar results were found in vivo. CLP increased HO-1 protein expression and phosphorylation of p38 MAPK. Inhibition of HO-1 with in vivo–specific siRNA decreased CLP-induced phosphorylation of p38 MAPK (Fig. 7D), Thus, knockdown of HO-1 or inhibition of HO activity decreases sepsis or LPS-induced p38 MAPK, VPS34, and LC3. The role of p38 MAPK in the induction of VPS34 and autophagy was next investigated. Mice were treated with SB203580 before CLP to inhibit the phosphorylation of p38 MAPK. SB203580 prevented CLP-induced VPS34 and LC3 (Fig. 8).
In summary, cecal ligation and puncture or in vitro LPS treatment induced autophagic signaling modulated by HO-1. Inhibition of HO-1 activity or knockdown of HO-1 resulted in decreased autophagy and increased hepatic apoptosis and cell death. Likewise, inhibition of autophagy by knockdown of VPS34 or pharmacological inhibition of PI3K (class III) resulted in increased cell death. Additionally, these data suggest that sepsis or LPS induces autophagy via an HO-1, p38 MAPK, VPS34 signaling pathway.
Several previous investigations have suggested that autophagy serves as an adaptive response to cellular stress that avoids cell death.13 Studies have shown that nutrient deprivation in HeLa or HCT116 cancer cells induces autophagy, and inhibition of this pathway results in apoptotic signaling.14, 15 Additionally, knockout of the autophagic protein, ATG5, can result in increased apoptosis in neurons or T cells.16, 17 The mechanism through which autophagy decreases cell death has not been delineated in these previous studies. Several possible mechanisms include bioenergetic failure, increased oxidative stress, and/or failure to remove damaged organelles and misfolded proteins in the setting of inhibition of autophagy. Our findings suggest this to be relevant in LPS-treated hepatocytes.
These data demonstrate a lack of cell death in early sepsis and are consistent with findings of previous studies.1 Multiple investigations have demonstrated, both in animal models and from human tissues, that early sepsis-induced organ dysfunction is not associated with cell death and structural damage.9 Brealey et al. illustrated that in sepsis, there is a decrease in metabolic rate and energy production that leads to organ dysfunction, and hypothesize that this is an adaptive response to increase the chance of survival of the cell and thus organs in the setting of the insult of sepsis. By decreasing metabolic demands and “hibernating,” the organism has time to clear the infectious insult and overcome the massive inflammatory response without sustaining significant cellular death, allowing for organ recovery.10 These data in this article that illustrate sepsis-induced autophagy and mitophagy support this hypothesis, in that autophagy may serve as an important mechanism to rid the cell of damaged organelles, including mitochondria. Damaged or dysfunctional mitochondria that accumulate in a cell can lead to cell death by a number of potential pathways, including the mitochondrial pathway of apoptosis. Our data demonstrate in vitro that LPS treatment of hepatocytes causes a decrease in mitochondrial membrane potential, and, over time, there is a normalization of mitochondrial membrane potential. Likely, autophagy contributes significantly to a process of mitochondrial homeostasis, which would also involve the processes of mitochondrial fission or fusion as well as biogenesis.
The regulation of autophagy via HO-1 signaling is a novel finding. HO-1 is an important molecule in the liver and other organs and is involved in restoring cellular homeostasis in response to multiple insults, including infection and sepsis.6 This can occur via multiple parallel pathways. HO-1, which is up-regulated in sepsis, is an adaptive response to metabolize free intracellular heme released by injured cells. One could hypothesize that a cellular response to increased intracellular heme, which is associated with protein breakdown and intracellular stresses, would also require other intracellular degradative pathways, such as autophagy, to process the nonheme “waste” and injured organelles at the same time. Thus, up-regulation of autophagic signaling with HO-1 would be necessary. HOs may also directly regulate aerobic respiration through the production of carbon monoxide (CO). We and others have shown that HO-1/CO can increase the production of hepatic mitochondrial ROS via inhibition of cytochrome c oxidase to initiate adaptive signaling to prevent cell death.18-20 Additionally, we have shown, in LPS-treated macrophages, that CO increases mitochondrial ROS to increase the phosphorylation of p38 MAPK.21 The findings in this study are consistent with such signaling pathways, in that HO-1 modulates the phosphorylation of p38 MAPK to induce autophagic signaling. Other additional potential signaling mechanisms include the direct effect of HO-1 on activation of class III PI3Ks to promote autophagic signaling. We and others have shown, in hepatocytes, that HO-1 or CO can activate PI3Ks.22 These mechanisms of action require further investigation.
Furthermore, HO-1 signaling is known to inhibit apoptosis.21 The mechanisms in which apoptosis are inhibited by HO-1 signaling has not been clearly elucidated. Brouard et al. demonstrated that the product of HO, CO, is able to inhibit tumor necrosis factor-alpha–induced apoptosis in endothelial cells through the activation of p38 MAPK.23 Our previous work demonstrated that HO or CO could prevent the spontaneous apoptosis of hepatocytes via PI3K signaling to influence nuclear factor-κB.22 The influence of HO-1 as a key inducer of autophagic signaling as part of an adaptive response to stress, thereby preventing accumulation of damaged and dysfunctional mitochondria to prevent apoptosis, is suggested in this article. Interestingly, with increased autophagy and mitophagy, these data demonstrate that bioenergetic failure and cell death are prevented. This suggests that there are compensatory mechanisms that take place, such as increased anaerobic respiration, a compensatory increase in oxidative phosphorylation by uninjured mitochondria, or restoration of a healthy mitochondrial population by mitochondrial fission/fusion or biogenesis. These data support the hypothesis that HO-1 acts as a central molecule to influence cellular “decision” between autophagy and apoptosis.
Whether activation of autophagy directly decreases apoptosis, or whether the divergence occurs more proximally and signaling proceeds down an autophagic versus an apoptotic pathway, is yet unknown. We hypothesize that autophagy is a necessary process to remove damaged organelles, such as mitochondria. If this process fails or becomes overwhelmed, these damaged organelles trigger an apoptotic death. This may occur via the release of cytochrome c from mitochondria, triggering the “mitochondrial” pathway of apoptosis or via other pathways. This remains to be elucidated.
Therefore, autophagy is a protective molecular pathway in the setting of sepsis, and understanding its regulation is important to further understand the pathophysiology of sepsis and make advances that could decrease the morbidity and mortality from this disease process.