Chapter 3 RESULTS
3.1 Part 1
Mesenteric I/R triggers intestinal epithelial apoptosis that accompanies villous destruction
Compared with non-ischemic tissues in sham controls, rats subjected to mesenteric
I/R showed villous blunting and epithelial denudation in the jejunum (Figure. 1A).
Mucosal destruction was associated with increased number of TUNEL(+) epithelial
cells and enhanced immunoreactivity to cleaved caspase-3 at regions close to the
denuded villous tips in intestines of I/R rats (Figure. 1B and 1C). Decreased
immunoreactivity of PCNA was found in the intestinal crypts in I/R rats compared to
sham controls (Figure. 1D). The severity of mucosal damage caused by I/R was
quantified by histopathological scoring (Figure. 1E).
Enteral instillation of the pancaspase inhibitor ZVAD reduced the degree of
mucosal injury and epithelial apoptosis caused by I/R (Figure. 1A-C, and 1E). Normal
intestinal histology was seen in sham-operated rats enterally administered ZVAD
(Figure. 1A). The mucosal caspase-3 activity was significantly increased after ischemia
(Table 1). Luminal pretreatment with ZVAD inhibited both baseline and
ischemia-induced mucosal caspase-3 activity (Table 3-1). This parameter was not
measured post-reperfusion owing to the marked destruction of the villous structure.
To assess the changes in the tight junctional structure, the level of occludin was
evaluated by Western blot. Increased cleavage of occludin was seen in the intestinal
mucosa of ischemic intestines compared to that of sham operation (Figure. 2).
Pretreatment with ZVAD diminished the level of occludin cleavage caused by ischemic
challenge (Figure. 2). Occludin levels were not examined post-reperfusion due to severe
epithelial denudation.
Increased intestinal permeability caused by I/R is dependent on epithelial apoptosis
The intestinal permeability changes caused by I/R were first evaluated using an ex
vivo assay that measured the luminal-to-serosal flux of a macromolecular probe HRP in
Ussing chambers (Yu et al., 2003). The transmural HRP flux rate in the intestine of I/R
rats was twice that of sham controls. The increase of HRP flux in I/R tissues compared
to sham controls was evident at 30-60 and 60-90 minutes after luminal addition of HRP
to the chambers (Figure. 3A).
In consideration of the time frame required to detect increased HRP flux and the
extracorporeal setting of high oxygen and glucose supply needed to maintain the
viability of tissues ex vivo ⎯ which may produce artificial results of the extent of I/R
injury ⎯ we also measured gut permeability changes in vivo by using ligated loops
administered with fluorescein-conjugated dextran (Wang et al., 1998) or a contrast agent
gadodiamide for the newly-developed MRI-based assay (Hsiao et al., 2009). A
significant increase of the lumen-to-blood passage of FITC-dextran was seen in I/R rats
compared to sham controls (Figure. 3B).
The novel MRI-based intestinal permeability assay monitors the portal drainage of
an enterally-administered contrast agent (gadodiamide) by quantifying the signals in the
liver and kidney as the areas of interest (Hsiao et al., 2009). Representative abdominal
images of sham and I/R rats were taken at various time points after the start of
reperfusion (Figure. 3C). The signals in liver and kidney in I/R rats were brighter than
in sham controls (Figure. 3C). The signal-to-noise ratio (SNR) of the areas of interest
was quantified from the MR images (Figure. 3D). The liver SNR values in sham
controls was consistently low throughout these time points, whereas in I/R rats the liver
signals were significantly elevated over time and remained high up to 60 minutes
post-reperfusion (Figure. 3D). The signal intensity in liver was 14 times higher in I/R
rats (SNR = 4.97 ± 0.45) than in sham controls (SNR = 0.35 ± 0.07) as early as 5
minutes post-reperfusion (Figure. 3D). The kidney SNR values in I/R rats were
significantly higher than that of sham controls within 15 minutes post-reperfusion
(Figure. 3D). The plasma gadodiamide concentration was 68 times higher in I/R rats
than in sham controls (Figure. 3D).
Pretreatment with intraluminal ZVAD partially decreased the gut permeability rise
triggered by I/R, as evidenced by lower fluorescein intensity in plasma samples (Figure.
3B), as well as lower SNR in the liver, kidney, and plasma in I/R+Z rats (Figure. 3C and
3-3D). The gut permeability in sham+Z rats was comparable to that of sham controls
(data not shown).
I/R-induced enteric bacterial translocation and mucosal inflammation are diminished by pretreatment with a caspase inhibitor
The bacterial counts in liver tissue in I/R rats were significantly higher than sham
controls (Figure. 4A). A similar increase of bacterial CFUs was demonstrated in the
spleen upon I/R insult (Figure. 4B). Intraluminal pretreatment with ZVAD abolished the
increase in bacterial CFUs in the liver and spleen of I/R rats (Figure. 4A and 4B).
The intestinal MPO activity (Figure. 5) and the mucosal levels of TNFα and
MIP-1α (Figure. 6) were higher in ischemic intestines compared to those with sham
operation, suggesting activation of inflammatory cells. Enteral instillation of ZVAD
diminished the rise of MPO activity (Figure. 5) and the increase of TNFα and MIP-1α
production induced by ischemia (Figure. 6). The inflammatory parameters in sham+Z
rats were comparable to those of sham controls (data not shown).
Luminal glucose decreased I/R-induced intestinal pathology Mucosal pathology and epithelial apoptosis
Enteral instillation of glucose significantly alleviated I/R-induced mucosal injury,
whereby the jejunal villi showed better structure and were covered by intact epithelial
layers without cell apoptosis, and the crypt regions showed PCNA immunoreactivity
(Figure. 1A-D). Normal intestinal histology was seen in sham+G rats (Fig. 1A).
Luminal glucose also reduced the increase in mucosal caspase-3 activity caused by
intestinal ischemia (Table 1).
Tight junctional integrity and epithelial permeability
The effect of glucose on intestinal barrier function was examined further. Enteral
instillation of glucose reduced the level of occludin cleavage in ischemic guts (Figure.
2).
Moreover, the presence of luminal glucose during I/R challenge diminished the
lumen-to-blood passage of FITC-dextran (Figure. 3B), as well as the gadodiamide
signals and SNR in the liver, kidney, and plasma samples (Figure. 3C and 3D).
Bacterial translocation and mucosal inflammation
The I/R-triggered increase of BT was abolished by luminal glucose. The bacterial
counts in liver and spleen were significantly lower in I/R+G rats than in I/R rats (Figure.
4A and 4B). The bacterial counts in sham+G rats were comparable to those of sham
controls (data not shown).
The intestinal MPO activity in I/R+G rats was decreased compared with I/R rats
(Figure. 5). Reduced mucosal levels of TNFα and MIP-1α were seen in ischemic
intestines instilled with enteric glucose (Figure. 6A and 6B). The intestinal
inflammatory parameters in sham+G rats were comparable to those of sham controls
(data not shown).
Phloridzin blockage of SGLT1 sugar uptake nullifies glucose protection in a dose-dependent manner
To verify the role of SGLT1 in the protective mechanism, pharmacological
inhibitors of specific transporters were instilled into the ligated sac in the presence of
glucose and the gut permeability changes were measured by MRI-based assay. We
found that luminal pretreatment with phloridzin (a specific SGLT1 inhibitor; 0.5-2.5
mM) dose-dependently increased the liver SNR values of I/R+G rats to levels
comparable to those of I/R rats (Figure. 7A). Phloridzin (2.5 mM) also inhibited the
glucose-mediated reduction of BT (Figure. 7B). On the other hand, pretreatment with
phloretin (an inhibitor of GLUT2; 2.5 mM) did not diminish the protective effect of
glucose on gut permeability (Figure. 7A) and BT (Figure. 7B). Moreover, apical
expression of SGLT1 was confirmed in the jejunal epithelium in sham controls (Figure.
7C). A lack of SGLT1 staining accompanied the epithelial sloughing seen in I/R rat
intestines; the presence of luminal glucose abolished this decrease (Figure. 7C).
PI3K/Akt signaling are involved in the glucose-mediated cell survival mechanism
To verify the involvement of PI3K/Akt signals in the glucose-mediated
cytoprotective mechanism, I/R+G rats were administered LY294002 (LY) or
wortmannin (W), which partially eliminated the glucose protection against I/R-induced
cell apoptosis and villous destruction (Figure. 8A) as well as permeability rise (Figure.
8B).
The activation status of Akt in gut mucosa was investigated by measuring the
kinase reaction of immunoprecipitated phospho-Akt to phosphorylate exogenous GSK3
in an in vitro assay. Decreased Akt activity in the intestinal mucosa of I/R rats was
evidenced by the lower levels of phosphorylated GSK3 in I/R samples than in sham
groups (Figure. 9A). Enteral instillation of glucose increased the mucosal Akt activity in
both sham and I/R tissues (Figure. 9A). The GSK3 phosphorylation levels in samples
from I/R+G+LY rats was lower than I/R+G rats (Figure. 9B), indicating that specific
inhibition of PI3K by LY294002 partly diminished the glucose-mediated activation of
Akt. Furthermore, immunofluorescent staining demonstrated the cytosolic expression of
Akt in jejunal epithelial cells in sham controls (Figure. 9C-a). Enteral instillation of
glucose induced the translocation of cytosolic Akt to the brush border and subcellular
organelles of enterocytes in sham+G rats (Figure. 9C-b). The loss of Akt expression was
correlated with the sloughing of intestinal epithelium in I/R rats (Figure. 9C-c), in which
these changes were attenuated by the addition of luminal glucose (Figure. 9C-d). The
phenomenon of Akt translocation to the brush border and to subcellular organelles in
epithelial cells was also seen in I/R+G rats (Figure. 9C-d).
The phosphorylation levels of Akt and downstream signals such as IκB, mTOR, Bad,
and FoxO1/3a in the mucosa of ischemic tissues were investigated by Western blot.
These parameters were not measured post-reperfusion due to severe mucosal denudation.
A significant decrease in phosphorylated Akt level was seen in ischemic tissues
compared to sham controls (Figure. 10A). Increased phosphorylation of Akt was seen
after enteral instillation of glucose in both sham and ischemic tissues (Figure. 10A).
Recent data indicate a link between Akt and IκBα/NFκB signals in promotion of cell
survival and resistance to apoptosis in enterocytes (Dan et al., 2008; Bai et al., 2009).
On the other hand, IκBα/NFκB signals also play key roles in proinflammatory cytokine
production in monocytes/macrophages and intestinal epithelial cells (Funda et al., 2001;
Suzuki et al., 2003; Selvaraj et al., 2005; Murphy et al., 2010). Our data showed that the
mucosal level of phospho-IκBα was significantly increased after ischemic challenge
compared to sham controls (Figure. 10B). Enteral instillation of glucose diminished the
increase of IκBα phosphorylation caused by ischemia (Figure. 10B). Lastly, the
phosphorylation of Akt correlated with the phosphorylation of mTOR, Bad, and
FoxO1/3a in ischemic tissues with glucose instillation (Figure. 10C).