Prohibitin 1 Modulates Mitochondrial Stress-Related Autophagy in Human Colonic Epithelial Cells

Introduction Autophagy is an adaptive response to extracellular and intracellular stress by which cytoplasmic components and organelles, including damaged mitochondria, are degraded to promote cell survival and restore cell homeostasis. Certain genes involved in autophagy confer susceptibility to Crohn's disease. Reactive oxygen species and pro-inflammatory cytokines such as tumor necrosis factor α (TNFα), both of which are increased during active inflammatory bowel disease, promote cellular injury and autophagy via mitochondrial damage. Prohibitin (PHB), which plays a role in maintaining normal mitochondrial respiratory function, is decreased during active inflammatory bowel disease. Restoration of colonic epithelial PHB expression protects mice from experimental colitis and combats oxidative stress. In this study, we investigated the potential role of PHB in modulating mitochondrial stress-related autophagy in intestinal epithelial cells. Methods We measured autophagy activation in response to knockdown of PHB expression by RNA interference in Caco2-BBE and HCT116 WT and p53 null cells. The effect of exogenous PHB expression on TNFα- and IFNγ-induced autophagy was assessed. Autophagy was inhibited using Bafilomycin A1 or siATG16L1 during PHB knockdown and the affect on intracellular oxidative stress, mitochondrial membrane potential, and cell viability were determined. The requirement of intracellular ROS in siPHB-induced autophagy was assessed using the ROS scavenger N-acetyl-L-cysteine. Results TNFα and IFNγ-induced autophagy inversely correlated with PHB protein expression. Exogenous PHB expression reduced basal autophagy and TNFα-induced autophagy. Gene silencing of PHB in epithelial cells induces mitochondrial autophagy via increased intracellular ROS. Inhibition of autophagy during PHB knockdown exacerbates mitochondrial depolarization and reduces cell viability. Conclusions Decreased PHB levels coupled with dysfunctional autophagy renders intestinal epithelial cells susceptible to mitochondrial damage and cytotoxicity. Repletion of PHB may represent a therapeutic approach to combat oxidant and cytokine-induced mitochondrial damage in diseases such as inflammatory bowel disease.


Introduction
Autophagy is an evolutionarily conserved catabolic pathway that degrades cytoplasmic components such as long-lived proteins, macromolecules and damaged organelles including the endoplasmic reticulum and mitochondria through lysosomal degradation [1,2,3]. Autophagy is an adaptive response to extracellular stress, such as starvation, or intracellular stress, including the accumulation of misfolded proteins, damaged organelles, or the invasion of microorganisms, intended to promote cell survival and restore cell homeostasis [4]. Either apoptosis or autophagic cell death can be initiated in irreversibly damaged cells [5]. Malfunctioning autophagy has been associated with multiple diseases such as cancer, neurodegeneration, autoimmune diseases and inflammatory diseases, including inflammatory bowel disease (IBD) [6].
The two common, but disparate, forms of IBD, Crohn's disease and ulcerative colitis, share related characteristics such as mucosal damage and diarrhea but have distinguishing clinical features. The etiopathogenesis of IBD remains unknown but is thought to involve a combination of genetic and non-genetic risk factors that regulate mucosal immune response, mucosal barrier function, and response to microbial factors [7]. Multiple epithelial molecules have been identified as mediators of IBD pathogenesis including those that control epithelial homeostasis [8]. Genome-wide association studies and meta-analysis have identified the autophagy genes ATG16L1, IRGM, and LRRK2 as candidate loci involved in genetic susceptibility to Crohn's disease [9,10,11,12]. Mutation or deletion of ATG16L1 results in increased proinflammatory cytokine production, increased susceptibility to experimental colitis, and reduced capability to eradicate invading bacteria, indicating the importance of autophagy in suppressing intestinal inflammation [13,14,15].
Multiple studies have reported mitochondrial dysfunction in Crohn's disease and ulcerative colitis [16,17,18,19] as well as the dextran sodium sulfate and 2,4,6-trinitrobenzene sulfonic acid models of colitis [20,21]. Mitochondria are important regulators of autophagy and apoptosis. During normal function of the mitochondrial respiratory chain, reactive oxygen species (ROS), which are partially reduced oxygen species such as superoxide radical (O 2 2 ), hydrogen peroxide (H 2 O 2 ), hydroxyl radical (?OH), and peroxynitrate (NOO 2 ), are generated at low levels. Production of ROS is increased when mitochondria are damaged [22]. IBD is associated with increased ROS and decreased antioxidant enzymes in the intestinal mucosa [23,24,25,26]. It is widely accepted that ROS produced as a by-product of respiration as well as exogenous ROS can induce autophagy via mitochondrial damage [27,28]. Mitochondria are the main source of ROS for regulation of autophagy [29]. In fact, exogenous ROS and the proinflammatory cytokine tumor necrosis factor a (TNFa), both of which are increased during IBD, promote cellular injury and autophagy via mitochondrial ROS generation [29,30,31]. Defects in autophagy result in the accumulation of intracellular ROS and deformed mitochondria [14,32].
Prohibitin 1 (PHB) is an evolutionarily conserved, multifunctional 32 kDa protein implicated in cellular processes including the regulation of proliferation, apoptosis, and transcription [33,34,35,36]. PHB is predominantly localized to the mitochondria in intestinal epithelial cells [37] and multiple studies have shown that PHB plays a role in maintaining normal mitochondrial function and morphology (reviewed in [38]). It has been shown that PHB interacts with complex I and subunits of cytochrome c oxidase of the respiratory chain and regulates their assembly [39,40]. Loss of PHB in mitochondria impairs function of the mitochondrial respiratory chain [39,40]. One obvious effect of respiratory chain dysfunction is increased oxidant production leading to oxidative stress, which can cause alterations in mitochondrial morphology and membrane potential [29].
Expression of PHB is decreased in mucosal biopsies from ulcerative colitis and Crohn's disease afflicted patients and in animal models of colitis [37,41]. Pro-inflammatory cytokines such as TNFa and oxidative stress induced by exogenous H 2 O 2 decrease expression of intestinal epithelial PHB in vivo and in vitro [37,42]. Restoration of colonic epithelial PHB expression using genetic manipulation (villin-PHB transgenic mice) or therapeutic delivery to the colon via nanoparticle or adenovirus protected mice from experimental colitis [43,44]. Our recent data suggest that epithelial PHB sustains anti-oxidant expression [44] and has anti-inflammatory properties such as reducing TNFa-stimulated NF-kB activation [42]. This is in agreement with emerging data that suggest a role of PHB in combating oxidative stress in multiple cells types [39,40,45,46]. In this study, we investigated the potential role of PHB in modulating mitochondrial stress-related autophagy in intestinal epithelial cells. Here, we show that TNFa and IFNc-induced autophagy inversely correlates with PHB protein expression and that gene silencing of PHB induces mitochondrial autophagy via increased intracellular ROS. Inhibition of autophagy during PHB knockdown exacerbates mitochondrial depolarization and reduces cell viability. These data suggest that decreased PHB levels coupled with dysfunctional autophagy renders intestinal epithelial cells susceptible to mitochondrial ROS and cytotoxicity.

PHB protein expression inversely correlates with cytokine-induced autophagy in cultured colonic epithelial cells
Our previous studies showed that TNFa reduces expression of PHB in intestinal epithelial cells in vivo and in vitro [42]. Proinflammatory cytokines such as TNFa and IFNc have been shown to induce autophagy in human intestinal epithelial cell lines [31,47]. Confluent monolayers of Caco2-BBE cells were treated with 10 ng/ml TNFa or 50 ng/ml IFNc alone or in combination for 18 hours. As expected, TNFa and IFNc increased two biochemical signs of autophagy: the conversion of LC3-I to LC3-II, indicated by normalizing LC3-II to LC3-I protein levels, and increased beclin-1 protein expression ( Figure 1A and 1B) [27,48]. Conversely, PHB protein levels in the same samples were decreased by TNFa and IFNc ( Figure 1A and 1B). The effect of TNFa and IFNc given in combination reflected that of cells treated with either cytokine alone and therefore, we did not pursue the effects of these cytokines in combination. It is widely accepted that the tumor suppressor p53 regulates autophagy [49]. Since Caco2-BBE cells have mutated p53 [50], we assessed the involvement of p53 in autophagy induction by TNFa and IFNc in wild-type (WT) and p53 null HCT116 colonic epithelial cells. HCT116 cells, including p53 null cells [51], also showed the conversion of LC3-I to LC3-II and increased beclin-1 protein expression suggesting that the effect of TNFa and IFNc to increase autophagy and decrease PHB protein expression is independent of p53 signaling ( Figure 1C and 1D).

Exogenous PHB expression reduces basal autophagy and TNFa-induced autophagy in intestinal epithelial cells
Since PHB expression inversely correlated with the induction of autophagy in colonic epithelial cells, we determined whether exogenous PHB expression could affect autophagy. Caco2-BBE cells stably overexpressing GFP-tagged PHB (pEGFPN1-PHB) show decreased LC3-II conversion from LC3-I and reduced beclin-1 protein expression (Figure 2A and 2B), suggesting a reduction in basal autophagy. Treatment with TNFa or IFNc increased LC3-II and beclin-1 protein abundance in empty vector expressing cells (Figure 2A and 2B), reflecting the same response as WT Caco2-BBEs in Figure 1A. PHB overexpressing cells showed decreased TNFa-induced LC3-I conversion to LC3-II and beclin-1 protein expression compared to vector-transfected cells, whereas PHB overexpression did not affect IFNc-induced LC3-II or beclin-1 protein expression (Figure 2A and 2B). Since PHB overexpression did not decrease IFNc-induced autophagy, this would suggest that the IFNc autophagy pathway is distinct from that of TNFa. An antibody specific to GFP was used to assess protein expression of GFP-PHB and GFP in PHB and vector overexpressing cells, respectively.

Knockdown of PHB induces autophagy
Since PHB protein expression inversely correlated with TNFaand IFNc-induced autophagy in Caco2-BBE cells, we next determined whether knockdown of PHB expression could induce autophagy. Caco2-BBE cells transfected with siPHB showed conversion of LC3-I into LC3-II ( Figure 3A) as well as increased beclin-1 protein expression ( Figure 3B) compared to cells transfected with a negative control siRNA (siNeg ctl). PHB knockdown stimulated the redristribution of GFP-LC3 fusion protein from a diffuse signal to cytoplasmic puncta indicative of autophagosomes ( Figure 3C). The mitochondrial stress proteins Cpn60, PKR, and ClpP showed no change in expression by Western blot upon PHB knockdown suggesting that inhibition of PHB does not induce the mitochondrial unfolded protein response ( Figure S1). The efficiency of siRNA knockdown was validated by western blot ( Figure 3A). Protein expression of PHB was reduced ,80% 96 hours after transfection.
The effect of PHB knockdown in Caco2-BBE cells was corroborated in HCT116 cells. Knockdown of PHB induced the conversion of LC3-I to LC3-II, increased beclin-1 protein expression ( Figure 4A and 4B) and increased the formation of cytoplasmic puncta by the GFP-LC3 fusion protein ( Figure 4C). Although knockdown of PHB in HCT116 cells (approximately 50% reduction compared to control; Figure 4A) was not as strong as in Caco2-BBE cells, the results on autophagy were similar to that in Caco2-BBE cells. p53 inhibition, including p53 knockout by homologous recombination, initially induces autophagy by ER stress followed by mitophagy (autophagy of mitochrondria) [52]. To determine whether the induction of autophagy by PHB knockdown is dependent on p53, markers of autophagy were assessed in p53 2/2 HCT116 cells transfected with siPHB. p53 2/2 cells showed increased LC3-II and GFP-LC3 puncta formation compared to WT cells ( Figure 4A and 4C) as previously described [52]. p53 2/2 cells also exhibited increased beclin-1 protein expression compared to WT cells ( Figure 4A and 4B). Knockdown of PHB caused a further increase in autophagy in p53 2/2 cells, suggesting that siPHB-mediated autophagy is independent of p53 signaling. ER stress is not further increased in p53 2/2 cells upon PHB knockdown ( Figure S2).

PHB knockdown increases intracellular reactive oxygen species and induces mitochondrial depolarization
Autophagy is induced by multiple stimuli including damage to organelles such as the endoplasmic reticulum and mitochondria. Given that the predominant subcellular localization of PHB is in the mitochondria in most cell types studied to date, emerging data suggest that PHB plays a role in stabilizing components of the electron transport chain [38]. Depletion of PHB function has been shown to cause electron transport chain disruption and increased intracellular oxidative stress [39,40]. As an indicator of intracellular ROS, Caco2-BBE cells transfected with siPHB or siNeg ctl were loaded with the oxidation-sensitive dye H 2 DCF-DA and assayed for DCF fluorescence 10, 20 and 30 minutes after loading using a plate reader. PHB knockdown caused a 30% increase in intracellular ROS compared to control cells at each time point measured ( Figure 5A; No tx). Recent data has shown that specific gene mutations or deficiency in autophagy is linked to susceptibility to inflammatory bowel disease [9,10,11,12]. To mimic this situation, autophagy was inhibited using 100 nM Baf A for 12 hours or RNAi against the autophagy gene ATG16L1. Efficiency of ATG16L1 knockdown is shown in Figure 5D. Inhibition of autophagy increased intracellular ROS in both control and siPHB-transfected cells; the magnitude, however, was significantly greater in siPHBtransfected cells at all time points ( Figure 5A; Baf A and siATG16L). Caco2-BBE cells stably overexpressing PHB show reduced DCF fluorescence compared to vector-transfected cells ( Figure 5B; No tx). In the setting of PHB overexpression coupled with autophagy inhibition (Baf A or siATG16L1), ROS levels are not further increased and in fact are decreased during loss of ATG16L1. These results suggest that PHB expression regulates intracellular ROS and that autophagy prevents the accumulation of ROS.
To determine whether the increase in intracellular ROS is associated with altered mitochondrial integrity, mitochondrial membrane potential (MMP) was assessed using JC-1 dye. As shown in Figure 5C, the percentage of Caco2-BBE cells testing positive for depolarized mitochondria increased from ,29% of control cells to ,41% of cells with PHB knockdown in the absence of inhibitors of autophagy (No tx, left panels), suggesting that loss of PHB negatively affects mitochondrial integrity. Inhibition of autophagy using Baf A or RNAi against ATG16L1 increased the percentage of control cells ( Figure 5C, top panels) with mitochondrial membrane depolarization from ,29% to ,36% (Baf A) and ,51% (siATG16L1). In cells with silenced PHB, treatment with Baf A increased the percentage of cells with mitochondrial membrane depolarization from ,41% to ,51% and siATG16L1 increased the percentage to ,56% ( Figure 5C, bottom panels). These results suggest that loss of PHB function induces mitochondrial dysfunction and that inhibition of autophagy increases the number of cells with altered mitochondrial function within the cell population. Silencing of PHB expression reduces cell viability Cell cytoxicity was assessed by measuring the release of lactate dehydrogenase. Across all treatments, Caco2-BBE cells transfected with siPHB were less viable than control cells ( Figure 6A). Inhibition of autophagy using Baf A further decreased viability in siPHB-transfected cells compared to control cells. RNAi against the autophagy gene ATG16L1 reduced viability in both siPHB and siNeg ctl cells, but the effect was greater in siPHB cells. These results suggest that PHB knockdown reduces cell viability and that autophagy acts to promote cell survival. The effect of PHB knockdown to reduce cell viability was exacerbated by the pro-inflammatory cytokines TNFa and IFNc ( Figure 6A). IFNc reduced viability in cells with PHB knockdown versus control cells, but in combination with autophagy inhibition IFNc had no further effect than inhibition of autophagy alone. In contrast, TNFa dramatically reduced cell viability in siPHB-transfected cells with inhibited autophagy (Figure 6A), supporting the concept that the IFNc autophagy pathway is distinct from that of TNFa. These results suggest that autophagy is important for cell survival when TNFa levels are high and when PHB expression is reduced, a scenario present during intestinal inflammation.
To determine whether cell death induced by PHB knockdown involves apoptosis, cleaved Caspase-3 protein expression and TUNEL staining were determined. Levels of cleaved Caspase-3 protein were increased during PHB knockdown and further increased by autophagy inhibition and TNFa treatment ( Figure 6B). The percent TUNEL positive cells was increased during PHB knockdown and further increased by autophagy inhibition and TNFa treatment ( Figure 6C), suggesting that cell death is at least partially due to apoptosis. To determine whether increased intracellular ROS is associated with increased cell death during autophagy inhibition and TNFa treatment, DCF fluorescence was measured in cells transfected with or without siPHB and treated with TNFa and autophagy inhibitors. Figure 6D shows that knockdown of PHB increased ROS levels for all treatment groups compared to control. TNFa treatment during inhibition of autophagy causes a further increase in intracellular ROS, which is further exaggerated in cells with PHB knockdown.
Treatment with NAC, a ROS scavenger, prevents siPHBinduced mitochondrial stress-related autophagy To determine whether mitochondria are indeed recycled during siPHB-induced autophagy, siPHB and GFP-LC3 co-transfected cells were labeled with MitoTracker, a mitochondria-specific dye (red fluorescence). Laser scanning confocal microscopy revealed significant co-localization of mitochondria with GFP autophagosomes in cells with silenced PHB compared to control cells ( Figure 7A, upper panels versus middle panels), suggesting mitophagy. To determine whether increased intracellular ROS is the mechanism by which PHB knockdown induces autophagy, cells were treated with 1.0 mM NAC for 24 hours prior to collection. The addition of NAC prevented the increase in GFP-LC3 puncta, indicative of autophagosomes, by PHB knockdown and prevented the co-localization of GFP-LC3 with mitochondria ( Figure 7A, lower panels).
Caco2-BBE cells were transfected with siPHB for 96 hours and treated with 1.0 mM or 10.0 mM NAC for 24 hours starting at 72 hours post-transfection. PHB knockdown stimulated the conversion of LC3I to LC3II compared to siNeg ctl cells ( Figure 7B), indicating autophagy, as shown in Figure 3A. The addition of NAC prevented siPHB-induced conversion of LC3I to LC3II in a dose-dependent manner ( Figure 7B). These results suggest that silencing of PHB expression increases intracellular oxidative stress, which subsequently induces mitophagy. DCF fluorescence was measured to ensure that NAC treatment reduced intracellar ROS ( Figure 7C).

Discussion
In addition to the emerging role of autophagy in controlling bacterial invasion [15,53,54], autophagy recycling represents the final tier of mitochondrial quality control. Multiple studies have reported mitochondrial dysfunction in Crohn's disease and ulcerative colitis [16,17,18,19] and mouse models of colitis [20,21]. Mitochondria are important regulators of autophagy and apoptosis. Exogenous ROS and cytokines such as TNFa, both of which are increased during IBD, promote cellular injury and autophagy via mitochondrial ROS generation [29,30,31]. We show here that PHB modulates autophagy in intestinal and colonic epithelial cells. Reduced PHB protein expression by RNA interference induces autophagy of mitochondria and treatment with NAC, a ROS scavenger, prevents siPHB-induced autophagy. These results suggest that loss of PHB expression induces autophagy via increased intracellular ROS. PHB knockdown induces mitochondrial membrane depolarization, suggesting mitochondrial damage, and increased intracellular ROS which is likely generated via dysfunctional respiration, all of which are exacerbated by inhibition of autophagy. Therefore, autophagy plays a protective role during conditions when PHB expression is low in intestinal epithelial cells.
We have previously shown that expression of PHB protein is reduced in mucosa during active inflammatory bowel disease and in Caco2-BBE cells treated with TNFa or exogenous oxidants [37,42]. It has been demonstrated that TNFa and IFNc induce autophagy in intestinal epithelium [31,55] and that autophagy can attenuate inflammatory responses, thereby maintaining intestinal homeostasis [56]. PHB protein levels inversely correlated with TNFa and IFNc-induced autophagy in Caco2-BBE and HCT116 cells. Furthermore, exogenous PHB expression reduced basal autophagy and TNFa-induced autophagy, suggesting that expression of PHB modulates the cellular autophagic response. One potential mechanism whereby PHB regulates beclin-1, the first mammalian gene shown to induce autophagy, is through the transcription factor E2F1. The beclin-1 promoter contains a putative E2F1 binding site [57] and PHB has been shown to regulate E2F1 activity [35]. Future studies will investigate this possibility. It is compelling to speculate that reduced PHB expression coupled with dysfunctional autophagy, an emerging susceptibility trait in Crohn's disease [13,14,15], could render epithelial cells unable to recycle damaged mitochondria and thus they succumb to cell death. It has been shown that when less than 66% of the mitochondria within a cell are damaged, autophagy predominates to restore cell homeostasis, while apoptosis is triggered when the percentage increases to involve most of the mitochondrial population [58,59]. The likely reason cell viability is decreased during PHB knockdown coupled with autophagy inhibition or TNFa treatment is that the threshold of damaged mitochondria exceeds the capacity of the cell.
Our finding that decreased PHB levels coupled with inhibited autophagy increased cell death has important clinical implications. Patients with inflammatory bowel disease exhibit decreased expression of PHB in intestinal epithelial cells [37,41]. Dysfuntion in autophagy genes such as ATG16L1, IRGM, and LRRK2, are emerging as potential susceptibility traits in patients with inflammatory bowel disease [9,10,11,12]. We speculate that decreased expression of PHB during active inflammatory bowel disease is a signal to the epithelial cell that there is inflammatory stress and that autophagy is subsequently induced to maintain cell viability and return homeostasis. Thus, these findings support an important role of autophagy in intestinal health and lend further insight into the mechanisms of dysfunctional autophagy via PHB in inflammatory bowel diseases. In the setting of PHB overexpression coupled with autophagy inhibition, ROS levels were not further increased and in fact are decreased during loss of ATG16L1. These data support the therapeutic concept of repletion of PHB levels in the setting of dysfunctional autophagy.
Basal autophagy was reduced by PHB overexpression in Caco2-BBE cells. We have shown previously that exogenous PHB expression in Caco2-BBE cells induced the expression of glutathione-S-transferase p, an antioxidant enzyme that catalyzes the conjugation of electrophiles to GSH [37], and modulates the activity of nuclear factor erythroid 2-related factor 2 (Nrf2), a transcriptional regulator of antioxidant response [42]. PHB overexpression may reduce basal autophagy through an increased antioxidant response on ROS production during normal physiological respiration. This issue warrants further investigation.
Although PHB protein levels inversely correlated with IFNcinduced autophagy, overexpression of PHB in Caco2-BBE cells did not affect IFNc-induced autophagy. This is in contrast to the response to TNFa. NF-kB, a major downstream signaling pathway of TNFa, has been implicated in modulating autophagy [60]. Our previous work demonstrated that exogenous PHB expression reduced basal and TNFa-stimulated NF-kB activation [42]. It is possible that this effect on NF-kB reduces TNFainduced autophagy as compared to IFNc since IFNc is not a major activator of NF-kB. Furthermore, TNFa promotes autophagy via mitochondrial ROS generation [31]. TNFa treatment during inhibition of autophagy exacerbated intracellular ROS levels in cells with knockdown of PHB. Since PHB overexpression did not decrease IFNc induced autophagy, this would suggest that the IFNc autophagy pathway is distinct from that of TNFa. Exogenous PHB expression can likely prevent TNFa-induced autophagy since PHB overexpression reduces ROS generation. The mechanism whereby PHB overexpression dampens basal and TNFa-induced, but not IFNc-induced, autophagy requires further investigation.
It is widely accepted that the tumor suppressor p53 regulates autophagy depending upon is subcellular localization [49]. Normal levels of p53 maintain a tonic inhibition of autophagy; autophagy is induced via a reduction in cytoplasmic p53 levels [61]. p53 can activate genes that induce autophagy including damage-regulated autophagy modulator (DRAM) and sestrins 1 and 2 [62]. Since Caco2-BBE cells have mutated p53, we included WT and p53 null HCT116 colonic epithelial cells to assess the involvement of p53 in autophagy mediated by PHB knockdown. PHB levels inversely correlated with TNFa-or IFNc-induced autophagy regardless of p53 status. As previously described [52], p53 2/2 HCT116 cells exhibited increased autophagy compared to WT cells. p53 knockout initially induces autophagy by ER stress followed by mitophagy [52]. Knockdown of PHB caused a further increase in autophagy in p53 2/2 cells, suggesting that autophagy mediated by PHB knockdown is independent of p53. ER stress markers indicate that ER stress was not further increased upon PHB knockdown. We speculate that p53 null cells with PHB knockdown likely show more autophagy due to a combination of ER stress-induced autophagy (due to p53 knockout) and mitophagy (due to loss of PHB expression and p53 knockout).
In conclusion, we demonstrate that the mitochondrial protein PHB modulates autophagy in intestinal epithelial cells via intracellular ROS signaling. Decreased PHB expression coupled with inhibition of autophagy, renders intestinal epithelial cells unable to maintain cell homeostasis and susceptible to mitochondrial damage and cytotoxicity. These findings have elucidated a molecular pathway whereby increased ROS by decreased PHB may enhance inflammation in patients with inflammatory bowel disease.

PHB knockdown and overexpression
Cells were transiently transfected with Stealth RNAi TM against PHB1  or Stealth RNAi TM siRNA Negative Control Med GC (Invitrogen, Carlsbad, CA) at 20 mm concentration. Caco2-BBEs were transfected using AmaxaH Cell Line Optimization NucleofectorH Kit T (Lonza, Basel, Switzerland), while HCT116 were transfected with LipofectAmine 2000 (Invitrogen). Cells were transfected with siRNA for 96 hours ( Figure S1A-D). For PHB overexpression studies, Caco2-BBE cells were transiently transfected with either pEGFPN1 expression vector (V) or pEGFPN1-PHB (P) for 72 hours (generation of the pEGFPN1-PHB construct is described below).

Autophagy activation and inhibition
Serum deprived cells were treated with 10 ng/ml recombinant human TNFa (R&D Systems, Minneapolis, MN) or 50 ng/ml recombinant human IFNc (eBioscience, San Diego, CA) for 18 hours. When treating Caco2-BBE cells, TNFa was administered to the basolateral chamber, while IFNc was administered to the apical and basolateral chambers. To inhibit autophagy, cells were treated with 100 nM Bafilomycin A 1 (Baf A; Sigma-Aldrich, St. Louis, MO) 24 h prior to collection or co-transfected with 20 mm siATG16L1 (59-AUUACUGCCAGAUAGGGAACC-CUUG-39). Efficiency of siRNA knockdown was assessed by Western blotting ( Figure 5D). Cells were treated with 1.0 or 10.0 mM N-acetyl-L-cysteine (Sigma-Aldrich), a ROS scavenger, for 24 hours prior to collection.

Generation of stably-transfected Caco2-BBE cell expressing pEGFPN1-PHB
A single PHB PCR product corresponding to the entire coding region of PHB (818 bp) was generated from Caco2-BBE cells using an antisense primer with a mutated PHB stop codon, underlined (59-AATTGGATCCCCTCCCTGGGGCAGCTGGA-39). The PCR product was ligated into pEGFPN1 vector (Clontech, Palo Alto, CA) using the Quick Ligation Kit (New England Biolabs, Ipswich, MA) and sequenced. Caco2-BBE cells were transfected with pEGFPN1-PHB or empty pEGFPN1 vector using Lipofectamine 2000 (Invitrogen) and the transfected clones were selected under 0.12% geneticin (Sigma-Aldrich), and fluorescent cells were isolated using flow cytometry (fluorescence-activated cell sorting).

Confocal microscopy
Caco2-BBE cells grown to confluency on filters were transfected with pSelect-NGFP-LC3 for 96 hours and incubated with 100 nM MitoTracker dye (Invitrogen) for 15 minutes. A subset of cells was treated with 1.0 mM NAC for 24 hours prior to collection. Cells were fixed in buffered 4% paraformaldehyde for 20 minutes, washed with PBS, and counterstained with DAPI to visualize nuclei. Samples were mounted in p-phenylenediamine/glycerol (1:1) and analyzed by confocal microscopy.

DCF assay
As a measure of intracellular ROS, conversion of the nonionic, nonpolar 29, 79 -dichlorodihydrofluorescein diacetate (H 2 DCFDA: Invitrogen) to fluorescent 29, 79 -dichlorofluorescein (DCF) was measured. Caco2-BBE cells were loaded with 10 mM H 2 DCFDA for 10 minutes and fluorescence was quantitated 10, 20 and 30 minutes later according to the manufacturer's protocol using a plate reader.

Cytotoxicity test
Lactate dehydrogenase (LDH) cytotoxicity detection kit (Clontech, Mountain View, CA) was used to measure cell viability. LDH determines the secretion of LDH into the culture medium from dead or membrane-damaged cells. Caco2-BBE cells transfected with siRNA for 96 hours and were treated with TNFa, IFNc, or Baf A alone or in combination for 18 hours prior to collection. An aliquot of 100 ml of culture media was added to 100 ml of LDH reagent and % cytotoxicity and % viable cells were measured according to the manufacturer's protocol using a plate reader.
Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) staining TUNEL staining was performed on confluent Caco2-BBE cells as described by the manufacturer's protocol (Roche, Indianapolis, IN).

FACS analysis of mitochondrial membrane potential
Mitoprobe TM JC-1 assay kit was used (Molecular Probes, Eugene, OR) to detect mitochondrial membrane depolarization.
JC-1 is a cationic dye which gives orange emission upon aggregates in normal polarized mitochondria and in monomeric form gives green fluorescence in the depolarized mitochondria. Briefly, 4610 5 Caco2-BBE cells transfected with siNeg Ctl or siPHB for 96 hours were trypsinized, washed and resuspended in a final volume of 1 ml of warm media. Cells were stained with 2 mm JC-1 dye for 15 minutes at 37uC protected from light. After incubation, cells were washed and resuspended in 500 ml of phosphate buffered saline and were analyzed on Becton Dickinson FACS Canto II flow cytometer.

Statistical analysis
Values are expressed as mean 6 SEM. Comparisons between TNFa or IFNc treatment or siPHB versus control were analyzed by unpaired Student's t-test. Comparisons between PHB RNAi and autophagy inhibition were analyzed by two-way analysis of variance to test for a significant interaction between PHB knockdown and impaired autophagy. Subsequent pair wise comparisons used Bonferroni post-hoc tests to test for significant differences between two particular groups. p,0.05 was considered statistically significant in all analyses.