Figures

Figure 1 Mesenteric I/R triggers enterocytic apoptosis that accompanies villous destruction. The jejunal tissues of sham-operated and I/R rats were processed for (A) H&E staining, (B) TUNEL assay, and immunostaining of (C) cleaved caspase-3, and (D) PCNA. Magnification: 200× in (A); Magnification:

400× in (B-D). (A) In comparison to non-ischemic tissues in sham controls, mucosal injury including villous blunting and epithelial denudation were observed in I/R rats. Enteral instillation of 120 µM ZVAD (Z) or 25 mM glucose (G) alleviated I/R-induced morphologic injury (labeled I/R+Z and I/R+G, respectively). Normal mucosal histology was seen in sham+Z and sham+G rats. (B) Enterocytes with increased TUNEL(+) reaction were noted in regions close to the denuded villous apex after mesenteric I/R. TUNEL(+) epithelial cells (arrows) and villous denudation (asterisks) were observed in the intestines of I/R rats. Enteral instillation of ZVAD or glucose decreased I/R-induced epithelial apoptosis. (C) Increased immunoreactivity to cleaved caspase-3 was seen at villous tips in I/R rats compared to sham controls. No staining was observed in intestinal tissues in sham+Z, sham+G, I/R+Z and I/R+G rats. (D) Immunoreactivity to PCNA (arrows) was detected in intestinal crypts in sham, sham+Z, and sham+G rats, No staining was seen in intestinal tissues in I/R rats. Enteral instillation of glucose attenuated the I/R-induced decrease of PCNA immunoreactivity, whereas ZVAD had no effect. (E) Histopathological scores in jejunal tissues of each group of rats. *P<0.05 vs. sham, ^#P<0.05 vs. I/R. (n = 6–8 / group)

Figure 2

Figure 2 Increased cleavage of tight junctional occludin after ischemic challenge is alleviated by enteral instillation of ZVAD or glucose. Western blot images of occludin in jejunal mucosa of rats subjected to sham operation (Sham) and ischemic (I) challenge with or without enteral instillation of ZVAD (Z) or glucose (G). Representative images from 4 separate experiments. (n = 4 / group)

Figure 3

Figure 3 I/R-triggered increase of epithelial permeability is ameliorated by enteral instillation of ZVAD or glucose. (A) Increased mucosal-to-serosal flux of HRP was seen in I/R rats compared to sham controls. HRP in serosal buffer was measured by kinetic enzymatic assay at 0-30 (I), 30-60 (II) and 60-90 (III) minutes after addition of HRP to the luminal buffer. *P < 0.05 vs. sham. (B) The 4-kDa FITC dextran (FD4) concentration in plasma sample collected from sham, I/R, I/R+Z and I/R+G rats at 60 min post-reperfusion. A significant increase of the lumen-to-blood passage of FD4 was seen in I/R rats compared to sham controls, which was decreased by instillation of ZVAD or glucose. (C) Abdominal images of sham, I/R, I/R+Z and I/R+G rats were taken at 0, 15, 30, and 60 min post-reperfusion. The arrow, arrowhead, and asterisk indicate the location of the liver, kidney, and ligated jejunal sac, respectively. In all panels, “0 min” indicates images taken before luminal gadodiamide was instilled into the jejunal sacs. (D) SNRs in the liver, kidney, and plasma samples were higher in I/R rats than in sham controls. The I/R-induced increase of gut permeability was ameliorated by enteral instillation of ZVAD or glucose. *P < 0.05 vs. sham; ^#P < 0.05 vs. I/R. (n = 5–6 / group)

Figure 4

Figure 4 I/R-induced enteric bacterial translocation is diminished by ZVAD or glucose. Numbers of bacterial CFUs in liver (A) and spleen (B) of sham, I/R, I/R+Z, and I/R+G rats were normalized to tissue weight. Each data point in the figure represents the value from one animal. The median values of bacterial counts (shown as bars) were significantly higher in I/R rats than in sham controls. The increase in bacterial counts caused by I/R was reduced by luminal instillation of ZVAD or glucose. *P < 0.05 vs.

sham; ^#P < 0.05 vs. I/R. (n = 6–8 / group)

Figure 5

Figure 5 Increased MPO activity caused by I/R is reduced by enteral instillation of ZVAD or glucose. The jejunal tissues of sham, I/R, I/R+Z and I/R+G rats were processed for the measurement of MPO activity (see Methods). One unit of MPO activity was defined as the quantity of enzyme degrading 1 μmol of H2O2 per minute. *P < 0.05 vs. sham; ^#P < 0.05 vs. I/R. (n = 6–8 / group)

Figure 6

Figure 6 Ischemic challenge augments the production of proinflammatory cytokines in gut mucosa.

(A) Mucosal levels of TNFα, which increased after ischemic challenge, were attenuated by luminal instillation of ZVAD or glucose. (B) The increase of mucosal MIP-1α levels induced by ischemic challenge was ameliorated by ZVAD or glucose. *P < 0.05 vs. sham; ^#P < 0.05 vs. I. (n = 6 / group)

Figure 7

Figure 7 Blockage of phloridzin-sensitive SGLT1 inhibits the glucose protection against I/R-increased epithelial permeability and bacterial translocation. (A) Administration of phloridzin (PHZ) increased the liver SNR in I/R+G rats to a comparable level to I/R rats in a dose-dependent manner.

The liver SNR values were acquired at 60 minutes post-reperfusion. Phloretin (PHT) had no effect on the liver SNR values in I/R+G rats. * P < 0.05 vs. 0 mM phloridzin. (B) Glucose-mediated reduction of BT to liver and spleen was abolished by phloridzin (PHZ; 2.5 mM), but not phloretin (PHT; 2.5 mM). The bar represents the median of bacterial counts in each group. * P < 0.05 vs. I/R+G. (C) Immunofluorescent staining demonstrated the apical expression of SGLT1 (green color) on jejunal villi of sham, sham+G, and I/R+G rats. Lack of SGLT1 staining was correlated with epithelial sloughing in the intestine in I/R rats.

The cell nuclei are stained blue. (n = 4-6 / group)

Figure 8

Figure 8 Anti-apoptotic PI3K signaling is involved in the mechanism of glucose protection against I/R-induced intestinal permeability rise. (A) Inhibition of PI3K by LY294002 (LY) or wortmannin (W) blocked the glucose rescue from cell apoptosis caused by I/R. TUNEL(+) epithelial cells (arrows) and disruption of epithelial continuity (arrowheads) were noted in I/R+G+LY and I/R+G+W rats.

Magnification: 400×. (B) The plasma gadodiamide concentrations in I/R+G+LY and I/R+G+W rats were higher than that of I/R+G rats, indicating that PI3K inhibitors partially diminished the glucose protection against I/R-induced increase of epithelial permeability. *P < 0.05 vs. I/R+G. (n = 6 / group)

Figure 9

Figure 9 I/R-decreased mucosal Akt activity is inhibited by enteral instillation of glucose. Mucosal lysates from rat jejunal tissues were collected for the measurement of Akt activity. Phospho-Akt protein was immunoprecipitated from the mucosal lysates by anti-phospho-Akt antibody, and exogenous GSK3 served as a substrate for Akt kinase reaction in vitro. The level of phospho-GSK3 was examined by Western blot. (A) Decreased Akt activity was seen in the intestinal mucosa of I/R rats compared to sham controls. Luminal glucose increased the mucosal Akt activity in both sham and I/R rats. (B) Administration of a PI3K inhibitor LY294002 (LY) decreased the mucosal Akt activity in I/R+G rats. *P

< 0.05 vs. sham; ^#P < 0.05 vs. I/R. (n = 6 / group) (C) Representative images of jejunal tissues stained for Akt (green color); cell nuclei are shown in blue. (a) Expression of cytosolic Akt was found in jejunal epithelial cells in sham controls. (b) Staining of Akt was noticed on brush border (arrows) and subcellular organelles (arrowheads) in villous epithelial cells in sham+G rats. (c) Loss of Akt staining was correlated with epithelial sloughing in I/R rats. (d) Histological improvement and staining of Akt on brush border (arrows) and subcellular organelles (arrowheads) in enterocytes were seen in I/R+G rats.

Figure 10

Figure 10 Enteral instillation of glucose induced the phosphorylation of Akt and downstream targets such as mTOR, Bad and FoxO1/3a but not IκBα in ischemic intestines. Western blot showing the levels of phosphorylated Akt, IκBα, mTOR, Bad and FoxO1/3a in jejunal mucosa of sham and ischemic rats after enteral instillation of glucose. Blots from 3 independent experiments were quantified by densitometry. (A) Decreased levels of phosphorylated Akt were seen in ischemic tissues compared to sham controls. Luminal glucose increased the phosphorylation of Akt in both sham and ischemic tissues.

(B) Ischemic challenge augmented the phosphorylation levels of IκBα, which was decreased by glucose.

(C) Luminal glucose increased the levels of phosphorylated mTOR, Bad and FoxO1/3a in jejunal mucosa in ischemic rats. *P < 0.05 vs. sham; ^#P < 0.05 vs. I. (n = 6 / group)

Figure 11

Figure 11 Schema of SGLT1-mediated anti-apoptotic PI3K/Akt signaling pathway attenuated I/R-induced epithelial apoptosis, mucosal damages and barrier defects.

Figure 12

Figure 12 Necrotic death was triggered by hypoxic challenge in human colonic carcinoma cells. (A) Caco-2 cells were exposed to normoxia (Nx) or hypoxia (Hx) in glucose-free media (Φ) for various time points. Increased LDH activity was found in the cell media of Hx+Φ, but not Nx+Φ cells, in a time-dependent manner. *P<0.05 vs. Nx+Φ (n=6/group). (B) Representative time-lapse images showing morphological changes in Hx+Φ cells for 24 hrs. Cytosolic vacuolation (→) and cell detachment (#) were noted in Hx+Φ cells. (C) Representative time-lapse images showing temporal alterations of mitochondrial transmembrane potential in Hx+Φ, but not Nx+Φ cells. The aggregated form of JC-1 (J-aggregate; red fluorescence) accumulated in functional mitochondria in normoxic cells throughout each time point. In hypoxic cells, a transient increase in red fluorescence intensity was seen after 4-8 hrs followed by a decline at later time points (16-24 hrs) associated with an increase in green fluorescence (the monomer form of JC-1; J-monomer) in the cytoplasm. (D) The ratio of J-aggregate to monomer was quantified in Nx+Φ and Hx+Φ cells at various time points. In contrast to normoxic cells, hypoxic challenge induces transient hyperpolarization and a final collapse of the mitochondrial transmembrane potential. *P<0.05 vs.

Nx+Φ at individual time points. (n=8/group) (E) Hypoxic challenge decreased the transepithelial resistance (TER) of cells compared to normoxic conditions. Data are presented as the absolute TER value at various time points. *P<0.05 vs. Nx+Φ. (n=6/group) (F) Hypoxic cells displayed heightened apical-to-basolateral flux of dextran probe in a time-dependent manner. *P<0.05 vs. Nx+Φ. (n=6/group) (G) Representative images of tight junction ZO-1 staining in cells exposed to normoxia and hypoxia for 16 hrs. Tight junction disruption and cell detachment (asterisks) were observed in hypoxic cells.

(n=6/group)

Figure 13

Figure 13 No sign of cell apoptosis was seen after hypoxic challenge. Caco-2 cells were exposed to normoxia (Nx) or hypoxia (Hx) in glucose-free media (Φ) for various time points. (A) No increase in oligonucleosome formation (an indicator of DNA fragmentation) was seen in cells exposed to hypoxia in glucose-free media for (a) 8 hrs and (b) 16 hrs. (B) The caspase-3 activity levels were comparable between hypoxic and normoxic cells at the 8-hr time point. (n=6/group)

Figure 14

Figure 14 Hypoxic challenge induced LDH leakage and plasma membrane disruption in human colonic carcinoma HT29 cells. (A) HT29 cells were exposed to normoxia (Nx) or hypoxia (Hx) in glucose-free media (Φ) for 16 hrs. Pretreatment with necrostatin-1 (a specific RIP1 inhibitor) decreased the hypoxia-induced LDH leakage. *P<0.05 vs. Nx. (n=6/group). (B) Representative photoimages of cells hypoxia exposed for 0, 8, and 16 hrs. Cytosolic vacuolation (→) and cell detachment (#) were noted in Hx+Φ cells. (n=6/group).

Figure 15

Figure 15 Hypoxia-induced necrotic cell death is dependent on RIP signaling pathways. (A) Pretreatment with necrostatin-1 (a specific RIP1 inhibitor) decreased the hypoxia-induced LDH leakage in a dose-dependent manner. *P<0.05 vs. respective Nx+Φ groups; ^#P<0.05 vs. ‘0 mM’ in Hx+Φ cells.

(n=6/group) (B) Knockdown of RIP1 by siRNA reduced LDH leakage in hypoxic cells. No effect was seen by negative control (CON) siRNA. Knockdown efficiency of transfected cells was confirmed by Western blots. *P<0.05 vs. respective Nx+Φ groups. ^#P<0.05 vs. CON. (n=3/group) (C) Representative images showing that necrostatin-1 inhibited morphological damage and cell detachment caused by 8-hr and 24-hr hypoxia. (D) Representative images showing transient mitochondrial hyperpolarization (an increase in red fluorescence intensity after 8 hrs) and a final collapse of transmembrane potential (an increase in green fluorescence intensity after 24 hrs) in hypoxic cells treated with necrostatin-1. The results suggest that RIP1 may not be upstream of mitochondrial dysfunctions. (n=4/group)

Figure 16

Figure 16 Supplementation of glucose prevented hypoxia-induced necropoptosis in colonic cancer cells. Cells exposed to normoxia (Nx) and hypoxia (Hx) were given glucose (0 or 25 mM) to evaluate death resistance. (A) Immunoprecipitation blots showing the formation of RIP1-RIP3 complex and phosphorylation of RIP1 in hypoxic cells without supplementation (labeled as ‘Φ’). No sign of RIP signaling was seen in hypoxic cells added 25 mM of glucose (labeled as ‘G’), and normoxic counterparts with or without glucose. The experiments were repeated twice and similar results were obtained.

(n=3/group). (B) The hypoxia-induced LDH leakage was decreased by glucose in a dose-dependent manner. (C) The changes in mitochondrial transmembrane potential caused by hypoxic challenge were inhibited by glucose addition. (D) The TER drop caused by hypoxia was reversed in cells treated with glucose compared to those without. (E) Addition of glucose prevented the increased dextran permeability caused by hypoxia. (F) Representative images of tight junction ZO-1 staining in cells exposed to hypoxia for 16 hrs with or without glucose. Hypoxia-induced tight junctional disruption and cell detachment (asterisks) was prevented by glucose. (B-E) *P<0.05 vs. respective Nx groups; ^#P<0.05 vs. ‘Hx+Φ’.

(n=6/group).

Figure 17

Figure 17 Addition of sugar analogs or amino acid did not reduce hypoxia-induced necrosis. The inhibition of hypoxia-induced LDH leakage was not seen by addition of 3-OMG (O), mannitol (M), or glutamate (GM). P<0.05 vs. Nx+Φ. (n=8/group).

Figure 18

Figure 18 Hypoxic challenge induced HIF1α activation and hypoxia-targeted gene (GLUT-1 and -4) expression in the presence of glucose. (A) Representative immunofluorescence staining of HIF1α in hypoxic cells given glucose for 0, 2 and 4 hrs. Superimposed images of HIF1α staining (red) merged with cell nucleus (blue) showed cytoplasmic distribution of HIF1α under normoxic conditions and nuclear translocation of HIF-1α after hypoxic challenge for 2 and 4 hrs (n=6/group). (B) The mRNA levels of GLUT-1, -2, -3 and -4 following hypoxic challenge as determined by semiquantitative PCR analysis. (C) Representative quantification of mRNA expression of GLUT-1 and -4 in hypoxic cells by real-time PCR.

(D) The protein levels of GLUT-1, -2, -3 and -4 after hypoxic challenge as determined by Western blots.

(B and D) *P<0.05 vs. ‘0 hr’. (n=4/group)

Figure 19

Figure 19 Resistance to hypoxia-induced necrosis is conferred by glycolytic pyruvate scavenging of mitochondrial superoxide in HT29 cells. (A) Various concentrations (0, 1, 5, and 25 mM) of glucose were added to hypoxic cells, and the LDH activity in the cell media was measured. *P<0.05 vs. Nx at 0 mM. (B) Cells were normoxia and hypoxia exposed in the absence (Φ) or presence of ethyl pyruvate (pyr;

25 mM), and the LDH activity in cell media was measured. *P<0.05 vs. respective Nx+Φ groups.

(n=8/group). (C) Representative immunofluorescence staining of HIF1α in HT29 cells exposed to normoxia or hypoxia with glucose. Superimposed images of HIF1α staining (red) merged with cell nucleus (blue) showed cytoplasmic distribution of HIF1α under normoxic conditions and nuclear translocation of HIF1α after hypoxia (n=6/group). (D) The protein levels of GLUT-1, -2, -3 and -4 following exposure to normoxia or hypoxia. (n=3-4/group) (E) Addition of glucose decreased mitochondrial ROS levels in hypoxic HT29 cells. Replacing glucose with a pyruvate derivative (pyr) also reduced mitochondrial ROS in hypoxic cells. *P<0.05 vs. respective Nx groups; ^#P<0.05 vs. Hx+Φ (n=8/group).

Figure 20

Figure 20 Anti-necrotic resistance of human colonic carcinoma cells was associated with anaerobic glycolytic metabolism. Cells were pretreated with vehicle (veh), iodoacetate (IA, 1 mM; a glycolytic inhibitor to GPD), or UK5099 (UK, 10 μM; a MPC inhibitor) prior to hypoxic challenge in the presence of glucose. The LDH activity in cell media (A), and the intracellular levels of ATP (B), pyruvate (C), and lactate (D) were measured. *P<0.05 vs. respective Nx groups; ^#P<0.05 vs. ‘veh’. (n=8/group)

Figure 21

Figure 21 Pyruvate inhibited hypoxia-induced necrotic death in an energy-independent mechanism.

Colonic cancer cells were normoxia and hypoxia exposed in the absence (Φ) or presence of ethyl pyruvate (pyr; 25 mM). Addition of pyruvate derivative prevented the LDH leakage (A), RIP1/3 complex formation (B), and morphological damage (C) caused by hypoxia. However, the cellular ATP levels (D), barrier function (E), and mitochondrial transmembrane potential (F) were not restored by pyruvate. (A, D, E, and F) *P<0.05 vs. respective Nx groups; ^#P<0.05 vs. ‘Hx+Φ’. (A and D, n=8/group; B, C, E, and F, n=4/group)

Figure 22

Figure 22 Pyruvate-mediated mitochondrial superoxide scavenging plays a critical role in resistance to hypoxia-induced necroptosis. (A) Increased MitoSOX fluorescence units were seen in hypoxic cells at the 8-hr time point, suggesting mitochondrial superoxide production upon hypoxic stress. The generation of mitochondrial ROS may be partially prevented by antioxidants (200 μM BHA or 1 mM apocynin), whereas vehicle control (veh) had no effect. *P<0.05 vs. respective Nx groups; ^#P<0.05 vs.

‘veh’ (n=8/group). (B) The hypoxia-induced LDH leakage was abolished by pretreatment with antioxidants, suggesting that cell necrosis was dependent on mitochondrial ROS production. *P<0.05 vs.

respective Nx groups; ^#P<0.05 vs. ‘veh’ (n=8/group). (C) Immunoprecipitation blots showing the formation of RIP1/3 complex in hypoxic cells, which was inhibited by BHA. The experiments were repeated twice and similar results were obtained. (n=3/group). (D) Supplementation of glucose decreased mitochondrial ROS levels in hypoxic cells, which was reversed by pretreatment with iodoacetate (IA, 1 mM), but not UK5099 (UK, 10 μM) or vehicle (veh). Replacing glucose with a pyruvate derivative (pyr) also reduced mitochondrial ROS in hypoxic cells. *P<0.05 vs. respective Nx groups; ^#P<0.05 vs. ‘veh’.

(n=8/group)

Figure 23

Figure 23 Supplementation of glucose did not alter the activities of redox enzymes in normoxic and hypoxic cells. The activities of (A) catalase, (B) superoxide dismutase (SOD), (C) glutathione reductase (GR), or (D) glutathione-S-transferase (GST), were measured in cells exposed to normoxia (Nx) and hypoxia (Hx) for 8 hrs in media containing 0 mM (Φ) or 25 mM of glucose (G). All of the enzyme activities were comparable between ‘Hx+G’ and ‘Hx+Φ’ cells. These results suggest that the cytoprotective effect exerted by glucose was not through redox mechanism. The GR activity in ‘Hx+Φ’

cells was lower than ‘Nx+Φ’ groups, suggesting that hypoxic challenge may decrease the GR activity in the absence of glucose. *P<0.05 vs. Nx+Φ. (n=8/group).

Figure 24

Figure 24 Schema of the resistance to necroptotic cell death induced by hypoxia in colorectal cancer is conferred by glycolytic pyruvate scavenging of mitochondrial superoxide.