Tumor necrosis factor-alpha and apoptosis signal-regulating kinase 1 control reactive oxygen species release, mitochondrial autophagy, and c-Jun N-terminal kinase/p38 phosphorylation during necrotizing enterocolitis.

BACKGROUND
Oxidative stress and inflammation may contribute to the disruption of the protective gut barrier through various mechanisms; mitochondrial dysfunction resulting from inflammatory and oxidative injury may potentially be a significant source of apoptosis during necrotizing enterocolitis (NEC). Tumor necrosis factor (TNF)-alpha is thought to generate reactive oxygen species (ROS) and activate the apoptosis signal-regulating kinase 1 (ASK1)-c-Jun N-terminal kinase (JNK)/p38 pathway. Hence, the focus of our study was to examine the effects of TNF-alpha/ROS on mitochondrial function, ASK1-JNK/p38 cascade activation in intestinal epithelial cells during NEC.


RESULTS
We found (a) abundant tissue TNF-alpha and ASK1 expression throughout all layers of the intestine in neonates with NEC, suggesting that TNF-alpha/ASK1 may be a potential source (indicators) of intestinal injury in neonates with NEC; (b) TNF-alpha-induced rapid and transient activation of JNK/p38 apoptotic signaling in all cell lines suggests that this may be an important molecular characteristic of NEC; (c) TNF-alpha-induced rapid and transient ROS production in RIE-1 cells indicates that mitochondria are the predominant source of ROS, demonstrated by significantly attenuated response in mitochondrial DNA-depleted (RIE-1-rho) intestinal epithelial cells; (d) further studies with mitochondria-targeted antioxidant PBN supported our hypothesis that effective mitochondrial ROS trapping is protective against TNF-alpha/ROS-induced intestinal epithelial cell injury; (e) TNF-alpha induces significant mitochondrial dysfunction in intestinal epithelial cells, resulting in increased production of mtROS, drop in mitochondrial membrane potential (MMP) and decreased oxygen consumption; (f) although the significance of mitochondrial autophagy in NEC has not been unequivocally shown, our studies provide a strong preliminary indication that TNF-alpha/ROS-induced mitochondrial autophagy may play a role in NEC, and this process is a late phenomenon.


METHODS
Paraffin-embedded intestinal sections from neonates with NEC and non-inflammatory condition of the gastrointestinal tract undergoing bowel resections were analyzed for TNF-alpha and ASK1 expression. Rat (RIE-1) and mitochondrial DNA-depleted (RIE-1-rho) intestinal epithelial cells were used to determine the effects of TNF-alpha on mitochondrial function.


CONCLUSIONS
Our findings suggest that TNF-alpha induces significant mitochondrial dysfunction and activation of mitochondrial apoptotic responses, leading to intestinal epithelial cell apoptosis during NEC. Therapies directed against mitochondria/ROS may provide important therapeutic options, as well as ameliorate intestinal epithelial cell apoptosis during NEC.


Introduction
Necrotizing enterocolitis (NEC) is the most common gastrointestinal surgical emergency in premature low birth-weight neonates, where prematurity is the single most important risk factor. Although several contributing factors for NEC have been identified, such as ischemia, bacteria, cytokines and enteral feeding, the exact mechanisms for its pathogenesis remain elusive. The clinical presentation of NEC is often nonspecific and unpredictable. NEC frequently involves diffuse areas of bowel necrosis and perforation, necessitating emergency operation. The presence of pneumatosis intestinalis detected on plain abdominal radiographs remains as a pathognomonic clinical feature. Despite extensive research during the past two decades, the exact pathogenesis of NEC for premature neonates remains ill-defined and largely unknown, hence, limiting development of novel preventive strategies.
Reactive oxygen species (ROS), generated as a result of ischemia-reperfusion injury to the gut, have been linked to the development of NEC in premature infants. 1,2 However, cytokines are also thought to play a role in ROS generation, contributing to severe gut inflammation and injury during NEC hallmarked by the exaggerated inflammatory responses by the premature immune system. 3,4 Tumor necrosis factor (TNF)α, a pro-inflammatory cytokine implicated in various inflammatory diseases of the small intestine, 5,6 is thought to contribute to the pathogenesis of NEC. Recent in vivo NEC studies have demonstrated a significant decrease in the severity and incidence of intestinal injury with anti-TNFα therapy. 7,8 TNFα-induced oxidative stress via mitochondrial ROS (mtROS) generation has recently emerged as a new mechanism of inducing cellular injury. Several studies have shown that mtROS generated by TNFα can oxidize the reduced thioredoxin-apoptosis signal-regulating kinase 1 (Trx(SH) 2 -ASK1) complex, [9][10][11][12][13] thus initiating its dissociation and inducing activation of ASK1 and its downstream stress signaling targets such as c-Jun N-terminal kinase (JNK) and p38 mitogen-activated protein kinase (MAPK) pathways. [14][15][16] ASK1 is a member of the MAPK family, and is an upstream activator of JNK and p38-MAPK signaling cascades. 17 ASK1 can be activated in response to TNFα/ROS and triggers various biological responses such as apoptosis, inflammation, differentiation and survival in various cell types. 13,[18][19][20][21][22] TNFα/mtROS-induced Trx oxidation results in the dissociation of the Trx(SH) 2 -ASK1 complex, release of ASK1 and activation of p38 MAPK and JNK stress response pathways. 16,[23][24][25][26] Thus, ASK1 is important in the regulation of ROS and inflammation-induced apoptotic signaling in injured cells; 16,27,28 however, its role in intestinal epithelial cells during oxidative injury is unknown.
Oxidative stress and inflammation may contribute to the disruption of the protective gut barrier through various mechanisms; however, mitochondrial dysfunction resulting from inflammatory and oxidative injury may potentially be a significant source of apoptosis during NEC. Hence, the focus of our study was to examine the effects of TNFα/ROS on mitochondrial function, and ASK1-JNK/p38 cascade activation in intestinal epithelial cells. evaluate mouse intestinal sections for evidence of autophagy. Initially, we treated RIE-1 cells with TNFα for various time points (15,30,60, 90 min and 24 hours), and labeled cells with organellespecific dyes, MitoTracker (mitochondria, red fluorescence) and LysoTracker (lysosomes, green fluorescence). Laser scanning confocal microscopy did not reveal significant mitochondrial autophagy at early time points (data not shown); however, when cells were treated with TNFα for 24 hours (h), and then labeled, we found significant co-localization of damaged mitochondria with lysosomes (yellow fluorescence) and typical morphologic changes (Fig. 2D) consistent with mitochondrial damage and autophagy of dysfunctional mitochondria. These findings suggest that TNFα-induced mitochondrial autophagy is a late phenomenon in contrast to more rapid and transient mitochondrial ROS production, MMPΔ, reduction in cellular respiration, activation of mitochondrial apoptotic signaling pathways and JNK/p38 stress pathways. Late autophagy may represent cellular inability to cope with overwhelming burden of damaged mitochondria. When taken together with RIE-1 cell death ELISA results (Fig.  3A) after TNFα treatment, these data suggest that mitochondrial autophagy may play a pro-apoptotic role during cytokineinduced injury in intestinal epithelial cells.
Cross-sectional views of mouse NEC intestinal villi showed ubiquitous autophagic vacuolization in intestinal epithelial cells in contrast with healthy controls (Fig. 2E). The degree of vacuolization is significant and may represent an adaptive function in stressed intestinal epithelial cells in vivo. Histological evaluation and electron scanning microscopy of human NEC sections would be helpful in examining autophagic vacuolization in intestinal epithelial cells.
TNFα-induced RIE-1 cell apoptosis is mitochondriadependent. Next, we sought to examine whether TNFα induces apoptosis in intestinal epithelial cells. We treated RIE-1 and RIE-1-ρ° cells with TNFα and DNA fragmentation was quantitated. TNFα-induced RIE-1 cell death was significantly transient drop in MMP, increased permeability and "leakiness" of the mitochondrial membrane which could result in the release of an apoptosis-activating molecule such as cytochrome c into the cytosol. MMP depolarization is an important early indicator of apoptotic signaling activation, and hence, transient and rapid MMPΔ in response to cytokine-induced injury demonstrates mitochondrial susceptibility in RIE-1 cells.
The oxygen consumption level in TNFα-treated RIE-1 cells was measured using a Clark-type electrode. TNFα treatment induced a significant decrease in oxygen consumption level of RIE-1 cells within the first minute of treatment with relatively depressed levels; this effect persisted for 5 min after TNFα treatment (Fig. 2C). This finding demonstrates that mitochondrial functional changes occur rather rapidly in response to TNFα, and that the mitochondrial oxygen consumption is rapidly decreased within the first minute of TNFα exposure. Taken together, these results demonstrate that TNFα induces significant mitochondrial dysfunction in intestinal epithelial cells, resulting in functional derangements such as increased production of mtROS, significant alteration in MMP and decreased oxygen consumption.
Organelle autophagy occurs as a result of cellular injury. Hence, we next examined the effects of TNFα treatment on mitochondrial autophagy in RIE-1 cells and sought to western blot analysis. The expression of mitochondrial apoptotic markers (apoptosis-inducing factor (AIF), APAF-1, cytochrome c) and ATP synthase-β, a marker for the activity of the electron transport chain, were increased in RIE-1 cells after TNFα treatment (Fig. 3C). This effect peaked at 15 min and returned to basal levels by 60 min. In contrast, the mtDNA-depleted RIE-1-ρ° cells displayed no significant alteration in expression levels of mitochondrial apoptotic markers, with the exception of a minimal increase in cytochrome c release at 15 min. This finding may represent either delayed protein degradation or altered nuclear encoding of mitochondrial proteins and enzymes that is unaffected by mtDNA silencing. decreased in TNFα-treated RIE-1-ρ° cells (Fig. 3A). These data imply that TNFα-exerted cellular injury mechanism(s) is predominantly mitochondria-dependent in RIE-1 cells.
When baseline levels of mitochondrial apoptotic markers such as Apoptotic protease activating factor 1 (APAF-1) and cytochrome c in RIE-1 and RIE-1-ρ° cells are compared, the mitochondrial expression of these apoptotic molecules is significantly reduced in mtDNA-silenced RIE-1-ρ° cell line (Fig. 3B). Hence, the effect of cytokine-induced injury may be dependent or independent of mitochondrial apoptotic arsenal. To test this hypothesis, we examined the effects of TNFα on mitochondrial apoptotic pathway activation in intestinal epithelial cells by

ROS trapping and ASK1 siRNA attenuate TNFα-induced apoptosis and JNK/p38 pathway activation in RIE-1 cells.
To determine the effectiveness of mitochondria-targeted potential therapies during TNFα/ROSinduced cell injury, we used ASK1 silencing via short interfering RNA (siRNA) treatment and spin-trapping compound, α-phenyl-N-tbutylnitrone (PBN). The dissociation of the Trx(SH) 2 -ASK1 complex, both cytosolic and mitochondrial, is a crucial step in activating the JNK-mediated apoptotic signaling cascade. 22-26 ASK1 inhibition can be protective during TNFα-induced cell injury. Hence, we targeted ASK1 with siRNA silencing method and examined activation of JNK and p38 apoptotic pathways, measured by phosphorylation levels of proteins following TNFα treatment (Fig. 4A). Protein analysis revealed that ASK1 silencing resulted in significant reduction of JNK and p38 phosphorylation levels in TNFα-treated cells. The differential phosphorylation of JNK isoforms was observed in RIE-1 cells, indicating a complex isoformspecific activation process induced by TNFα treatment in intestinal epithelial cells. These findings warrant future studies focusing on mitochondrial ASK1 targeting specifically and examining JNK isoform-specific activation in TNFα-treated intestinal epithelial cells.
Application of spin-trapping of RIE-1 cells showed significant attenuation in TNFαinduced RIE-1 cell death (Fig. 4B), thus demonstrating a protective effect of ROS-trapping by PBN in TNFα/ROS-damaged intestinal epithelial cells. Protein analysis of cell lysates revealed significantly attenuated levels of proapoptotic cytochrome c, and marked reduction  Mitochondrial dysfunction is found in many disease processes, including fulminant hepatic failure in neonates. 34,35 Susceptibility of premature neonatal intestine to significant oxidant injury during NEC as a result of mitochondrial dysfunction has not been explored. We have attempted to elucidate some of the early mitochondrial functional derangements in RIE-1 cells with TNFα treatment, and have shown that mitochondrial integrity during inflammation is compromised, leading to early activation of pro-apoptotic and mitochondria-dependent signaling in vitro (Fig. 5).
It is unclear whether cellular autophagy is a pro-survival or pro-apoptotic cellular mechanism. It serves an important purpose in disease processes requiring the maintenance of healthy population of mitochondria. 36,37 Our findings show that mitochondrial damage due to TNFα/ROS elicits late mitochondrial autophagy after TNFα exposure in vitro, suggesting either slow mitochondrial turnover or delayed mitochondrial biogenesis in injured RIE-1 cells. These results suggest that TNFα affects mitochondrial homeostasis and further studies are necessary to gain insight into intestinal mitochondrial dysfunction during NEC. Our previous findings of oxidative stress-induced intestinal cell death and MMP collapse 38,39 along with the rapid, transient inflammation-induced mtROS production, early mitochondrial pro-apoptotic response and ASK1-JNK/p38 stress pathway activation in intestinal epithelial cells in the current study, suggest a strong role for ROS-mediated mechanism(s) of cellular damage during NEC.
The selective phosphorylation of JNK isoforms, p56 and p45, in response to TNFα suggests that there may be a differential pathway of activation of downstream signaling proteins in rat intestinal epithelial cells in vitro. For example, although both JNK isoforms are phosphorylated during TNFα stimulation, the predominantly phosphorylated JNK isoform is p45. In contrast, pretreatment with the PBN scavenger specifically decreases p54 phosphorylation and not phosphorylation of p45 JNK isoform. Furthermore, ASK1 silencing with siRNA leads to suppressed phosphorylation levels of both JNK isoforms. These findings suggest a complex and possibly selective (cytosolic or mitochondrial only) responses to the TNFα-mediated activation of ASK1-JNK/p38 signaling pathways in intestinal epithelial cells in vitro.
Regulation of caspase-dependent apoptotic signaling via cytochrome c and APAF-1 release from the mitochondrial matrix appears to be a specific response to TNFα treatment as compared to caspase-independent apoptotic signaling. Though both apoptotic signaling pathways are activated by TNFα, these findings support our hypothesis that injured mitochondria are a significant source of apoptotic signaling in intestinal epithelial cells during TNFα/ROS-induced injury, and can lead to potentially detrimental compromise of the gut mucosal barrier integrity.
Previous in vivo study by Halpern et al. had already demonstrated significant attenuation of NEC in neonatal rat, both in severity and incidence of intestinal injury, with anti-TNFα treatment. In our study, we examined the in vitro effect of TNFα on molecular signaling mechanisms and mitochondrial functional changes in intestinal epithelial cells.

Discussion
Cytokines are thought to play a central role in gut inflammation and injury during NEC by inducing exaggerated inflammatory responses, leading to significant intestinal injury in premature neonates. 3,4 Previous studies have demonstrated that activation of inflammatory mediators such as TNFα, IL-1, NFκB, toll-like receptors (TLR), IL-8 and inducible NO synthase (iNOS) may play a significant role in the pathogenesis of NEC. 7,29,30 Recently, we demonstrated an anti-inflammatory action of peroxisome proliferator-activated receptor (PPAR)γ using in vivo model of NEC, and inhibition of NFκB pathway, a critical transcription factor for the activation of inflammatory mediators and cytokines. 31 De Plaen et al. also have demonstrated that inhibition of NFκB pathway during NEC ameliorates bowel injury and improves survival in vivo. 29 Recent studies have also focused on molecular signaling mechanism(s) within the TLR signaling pathway 32 and cyclooxygenase-2 (COX-2) in intestinal homeostasis and inflammation associated with NEC. 33 Modulating early cellular inflammatory pathways during NEC may improve overall survival for neonates.
Previously, we investigated the effects of ROS, generated as a result of ischemia-reperfusion injury to the gut, and activation of apoptotic and survival signaling during NEC. The aim of the present study was to examine the effects of pro-inflammatory TNFα on mitochondrial dysfunction in intestinal epithelial cells during NEC, since premature neonatal gut is thought to be more susceptible to inflammatory cascade activation and oxidative injury.
In this study, we observe significant intestinal expression levels of pro-inflammatory TNFα and apoptotic ASK1 molecules throughout all layers of the intestine. This indicates that intestinal inflammatory process during NEC is in fact transmural. Tissue ASK1 levels also share similar pattern of apoptotic activation and injury in premature neonatal gut when compared with TNFα. One may argue that the human tissue expression levels may not fully reflect the exact extent of apoptotic signaling, and likely represent a late stage of the disease. Early NEC tissue analysis is a limiting factor in understanding early intestinal responses and expression levels of TNFα and ASK1.
Complex relationship between inflammation, oxidative stress and activation of apoptotic signaling in intestinal epithelial cells may largely depend on early mitochondrial responses during cellular injury. We demonstrate significant rise in mtROS production, alteration of mitochondrial function, activation of ASK1-JNK/ p38 stress signaling and mitochondrial apoptotic cascade in rat intestinal epithelial cells with TNFα stimulation. Neonatal tissue staining together with our in vitro data demonstrate that inflammation and oxidative stress can induce significant mitochondrial deregulation and activation of mitochondria-selective apoptotic response, and may possibly lead to intestinal epithelial cell death in premature neonatal gut during NEC.
are the main source of intracellular ROS during TNFα exposure in intestinal epithelial cells; and (iii) mitochondria are susceptible to TNFα injury; (iv) activation of ASK1-JNK/p38 and mitochondrial apoptotic pathways occurs during inflammation-mediated ROS injury in intestinal epithelial cells, suggesting a central role for mitochondrial dysfunction during TNFα-induced oxidative stress. Therapies directed against mitochondria/ROS may provide important therapeutic options, as well as ameliorate intestinal epithelial cell apoptosis during NEC.

Human intestinal NEC sections.
Paraffin-embedded intestinal sections from 20 neonates with NEC and 3 neonates with noninflammatory condition of the gastrointestinal (GI) tract (intestinal atresia; control) undergoing bowel resection were analyzed. Intestinal tissues were fixed and paraffin-embedded for further analysis. Control and NEC sections (5 μm) were prepared for immunohistochemical analysis. Sections were incubated with rabbit anti-TNFα (1:200) and anti-ASK1 (1:100) antibodies overnight at 4°C, then incubated with an anti-rabbit secondary antibody and stained with DAB chromogen (Dako Cytomation EnVision ® + System-HRP (DAB) kit, Carpinteria, CA). Slides were washed, counterstained with hematoxylin, dehydrated and Based on our findings, we propose that developing therapies specifically directed against mitochondria may be beneficial in reducing activation of apoptotic signaling cascades as a result of cytokine-mediated oxidative stress in intestinal epithelial cells during NEC. Increasing ROS production by the electron transport chain of dysfunctional mitochondria can be attenuated by various ROS scavenging compounds. Our successful use of the spin-trapping compound, PBN, which protects RIE-1 cells from TNFα/ROS-mediated death and attenuates activation of ASK1-JNK/p38 apoptotic pathway signaling, suggests that ROS scavengers may be beneficial.
PBN is widely used for ROS scavenging, and most importantly, has been shown to reverse the age-related oxidative changes and to reduce oxidative damage from ischemia/reperfusion injury. [40][41][42][43][44] The antioxidant activity of PBN protects biologically important molecules from oxidative damage. Although this effect has not been demonstrated in NEC intestinal tissues, increasing scientific evidence of protective effects of free radical scavengers supports their possible use in clinical application, specifically in conditions requiring the targeting of diseases of mitochondrial dysfunction.
In conclusion, we have demonstrated that: (i) the pro-inflammatory cytokine, TNFα, is abundant in neonatal NEC intestinal sections and can induce significant functional mitochondrial deregulation in intestinal epithelial cells in vitro; (ii) mitochondria Western blot analysis. Cell lysates were clarified with centrifugation (13,200 rpm, 20 min at 4°C) and stored at -80°C. Protein concentrations were determined using the method described by Bradford. 45 Equal amounts of total protein (20-30 μg) were loaded onto NUPAGE 4-12% Bis-Tris Gel and transferred to PVDF membranes, incubated in a blocking solution for 1 h (Tris-buffered saline containing 5% nonfat dried milk and 0.1% Tween 20), incubated with primary antibody overnight at 4°C, and then incubated with horseradish peroxidase-conjugated secondary antibody. Anti-β-actin antibody (1:5,000), total JNK (1:1,000) and p38 (1:1,000) were used for protein loading control. All primary antibodies were used in concentration of 1:1,000 to probe membranes. The immune complexes were visualized by ECL Plus (Amersham Biosciences, Piscataway, NJ). Quantitative densitometric analyses of all western blot bands (data not shown) were performed using ImageJ (Image Processing and Analysis in Java software, National Institutes of Health, MD).
ROS trapping with α-phenyl-N-t-butylnitrone (PBN) in RIE-1 cells. PBN is one of the most widely used spin-trapping compounds that target mitochondrial ROS to reverse age-related oxidative changes and to alleviate oxidative damage from ischemia/reperfusion injury. 40-44 RIE-1 (2 x 10 4 ) cells were incubated with 0.5 mM PBN for 2 h at 37°C. To control for the effects of DMSO, a control group of cells were incubated in fresh media with DMSO for 2 h. After incubation, cells were treated with TNFα (10 ng/mL) for 15 min, and protein was harvested for western blot analysis.
ASK1 siRNA transfection of RIE-1 cells. Rat SMART pool ASK1 and non-targeting control (NTC) siRNA duplexes (Dharmacon, Lafayette, CO) were used for transfection by electroporation (400 V/500 μF) in RIE-1 cells. Cells were maintained in medium for 48-72 h, then treated with TNFα (10 ng/ mL) for various time points. Extracted protein was analyzed by western blotting for phospho-JNK and phospho-p38 expression levels.
JC-1 assay for detection of MMP changes. To determine the effects of TNFα on MMP, we used MitoProbe JC-1 Assay kit (Molecular Probes, Eugene, OR). The collapse in the electrochemical gradient across the mitochondrial membrane was measured using JC-1 fluorescent cationic dye. RIE-1 cells (1 x 10 6 ) were treated with TNFα (10 ng/mL), washed with PBS and incubated with 2 μM JC-1 for 15 min at 37°C in darkness. Cells were washed again and analyzed on a FACScan flow cytometer. cover slipped. For negative control, sections were stained with rabbit IgG (not shown).
In vivo murine NEC model. Timed pregnant Swiss-Webster mice were purchased (Charles River Labs, Pontage, MI) and pup littermates were randomized to either control or NEC group. All mice in NEC groups were housed in a water bath at 37°C. They were hand-fed KMR liquid milk replacer formula (0.3 cc/g/day; q 3 h) using an animal feeding needle (24 g, ballpoint; Popper & Sons, New Hyde Park, NY). Control mice were maternally reared. To induce NEC, pups were stressed twice daily with hypoxia by placing them in a plexi-glass chamber, breathing 5% oxygen for 10 min. Each mouse was monitored daily for the clinical severity of NEC by assessing the level of activity, oral intake, weight change and abdominal exam findings. Mice that developed abdominal distention, respiratory distress and lethargy during the first 96 h of the experiment were sacrificed. After 96 h, all surviving mice were sacrificed and distal ileum was harvested for analysis. Segments of ileum were fixed in formalin and stored in 70% ethanol for paraffin embedding. Intestinal sections were stained with hematoxylin and eosin for analysis.
Cell lines and culture techniques. RIE-1 cells (a gift from Dr. Kenneth D. Brown; Cambridge Research Station, Cambridge, UK) were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 5% fetal bovine serum. Cells were maintained at 37°C under an atmosphere containing 5% CO 2 . Tissue culture media and reagents were obtained from Invitrogen (Carlsbad, CA).
To determine whether mitochondria are the main source of TNFα-induced intracellular ROS production, and that activation of apoptotic pathways is mitochondrial ROS-dependent, we established a RIE-1-ρ° from the parent RIE-1 cell line. Silencing of mitochondrial function was achieved by maintaining RIE-1 cells in isolation medium supplemented with 5% fetal bovine serum, ethidium bromide, EtBr; 0.1 μg/mL, uridine (50 μg/ mL), pyruvate (100 μg/mL) and glucose (4.5 g/L). To confirm successful mtDNA depletion in RIE-1-ρ° cell population, cells were evaluated with FACS cell sorting method for EtBr binding after 14-16 passages and cell lysates were analyzed for decreased mitochondrial cytochrome oxidase subunit 1 protein level using western blot analysis (data not shown).
Mitochondrial isolation. To determine the baseline expression of mitochondrial apoptotic markers in RIE-1 and RIE-1-ρ° (2 x 10 7 ) cells, mitochondria were isolated using a mitochondrial isolation kit according to a manufacturer's protocol (Pierce, Rockford, IL). Mitochondrial lysates were analyzed for mitochondrial apoptotic markers (APAF-1, cytochrome c) by western blot.
Cell death detection ELISA. RIE-1 and RIE-1-ρ° cells were plated onto 24-well plates for 24 h prior to TNFα treatment. Cells were incubated with TNFα (10 ng/mL). To determine the effects of ROS scavenging on intestinal epithelial cell survival, RIE-1 cells were first pretreated with α-phenyl-N-t-butylnitrone (PBN; 0.5 mM) for 2 h, and then incubated with TNFα (10 ng/mL). DNA fragmentation was evaluated using a Cell Death Detection 1 ml of suspended cells was placed in the respirometry chamber, and oxygen consumption was monitored for 3 min to establish a basal consumption rate. TNFα (10 ng/mL) was added, and O 2 consumption was monitored for approximately 5 min. Potassium cyanide (KCN; 2 mM) was added to another aliquot of cells to determine the rate of non-mitochondrial respiration. Rates of O 2 consumption were calculated over a linear, one minute interval. All experiments were repeated three times.
Statistical analysis. Minimum of three sets of experiments were performed to reproduce the results. The data were analyzed separately for each set using Kruskal-Wallis test. For oxygen consumption study, the effect of TNFα was assessed using one-sample t test against 100%. All effects and interactions were assessed at the 0.05 level of significance. Statistical computations were carried out using SAS 9.1 ® .
Confocal microscopy for ASK1 expression and mitochondrial autophagy. To examine the effects of TNFα on ASK1 expression and mitochondrial autophagy, RIE-1 cells (2 x 10 4 ) were grown in glass chambers overnight, and then treated with TNFα. Cells were first incubated with PBS containing 1% bovine serum albumin, and then incubated with ASK1 antibody and AlexaFluor ® 647-labeled goat anti-rabbit IgG to determine ASK1 expression. For mitochondrial autophagy, treated RIE-1 cells were incubated with organelle-specific dyes from MitoTracker and LysoTracker labeling kits according to the manufacturer's instructions (Molecular Probes, Eugene, OR). All cells were counterstained with Hoechst 33342 nuclear stain. Zeiss LSM 510 UV Meta laser scanning confocal microscopy was used to visualize mitochondrial and lysosomal co-localization. Data was analyzed using Zeiss LSM 5 Image software.
Mitochondrial oxygen consumption (QO 2 ). Mitochondrial oxygen consumption was measured using a Strathkelvin Mitocell S200 micro-respirometry system (Strathkelvin; UK), which utilizes a Clark-type oxygen electrode. Respirometry was adapted from Mozo et al. 46 Chamber temperature was maintained at 37°C. The respirometer was calibrated prior to experiments using O 2 -saturated water followed by sodium dithionite to provide 100% O 2 and 0% O 2 baselines, respectively. RIE-1 cells (1 x 10 7 /ml) were trypsinized, pelleted by centrifugation, and resuspended in culture media lacking fetal bovine serum. Briefly,