INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE SMALL INTESTINAL epithelium undergoes constant renewal. New epithelial cells are produced in the crypts. The majority of new cells migrate from the crypt to the villus and undergo apoptosis at the villus tip. The life span of rat absorptive cells from birth to death is about 2-3 days (36, 46). During upward migration along the villus column, cells express specific genes to perform digestive and absorptive functions. Studies of epithelial cell differentiation using important functional markers, such as lactose-phlorizin hydrolase (LPH) and sucrase-isomaltase (SI), have shown that the expression of these enzymes takes place in a temporal and spatial manner during the cell life span (36, 40, 49). However, little information is available on the molecular mechanisms by which epithelial cell differentiation is regulated.

An approach to the analysis of regulatory mechanisms for differentiation is to induce damage and exfoliation of differentiated villus cells and then to investigate during recovery the molecular processes leading to the differentiation of incompletely differentiated or newly generated cells. These injury and repair processes can be produced by occlusion of intestinal blood flow for a short period followed by reperfusion. An episode of ischemia-reperfusion (I/R) is known to provoke sequential events beginning with desquamation of villus cells, followed by a transient increase of crypt cell production and finally the migration and differentiation of the replenished villus cells for functional recovery (4, 12, 29, 35). Specific gene expression during recovery from I/R would allow for the analysis of the processes and regulatory mechanisms of cell differentiation. We have used this experimental model to examine absorptive cell expression of brush-border membrane (BBM) integral proteins, such as intestinal alkaline phosphatase (IAP), LPH, and SI. In this study, changes in IAP, LPH, and SI mRNA levels and enzyme activities after I/R have been measured to determine whether each BBM enzyme recovers in coordination with its encoding mRNA and whether the rate of recovery is similar for each enzyme. The results show that the activity of each enzyme recovers in parallel with the encoding mRNA but that the time for recovery differs between enzymes. The coordinated changes in transcript abundance and enzyme activity suggest that the regulation occurs either by increased transcription and/or by mRNA stabilization.

Among the BBM enzymes studied in this report, SI is thought to be regulated primarily at the transcriptional level (17). Recently, an evolutionarily conserved promoter element of the SI gene, SI footprinting 1 (SIF1), has been identified in mice and humans. The binding of SIF1 to a homeodomain protein (Cdx-2) is required for promoter activity (37, 38). The presence of an SIF1 binding protein not related to Cdx-2 has been reported in the intestine expressing SI in adult rats (15). The possible regulatory role of SIF1 binding proteins on SI expression under physiological conditions in vivo has not been examined. To test the possibility that SIF1 protein is related to the upregulation of SI expression, we examined the change of SIF1 binding activity during recovery after I/R. We reasoned that if the SIF1 binding protein is involved in the upregulation of SI expression, an increase of SIF1 activity might occur.

To examine whether the recovery of BBM enzymes is a phenomenon specific to membrane proteins, we determined the expression of the cytosolic protein ferritin (Ft) after I/R. Ft was chosen for the study because 1) similar to SI, it exhibits a pattern of increasing expression as cells migrate upward along the villi (11, 50); 2) it sequesters intracellular iron and plays an important protective role against oxidative stress-induced cellular damage after I/R (14, 16); 3) Ft expression is regulated predominantly at the translational level by iron regulatory proteins (IRPs) (14, 16), in contrast to the transcriptional regulation of SI (16); and 4) both the IRP and SIF1 binding proteins are redox sensitive (25-27, 37).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and surgical induction of segmental intestinal I/R. Twenty-one Sprague-Dawley rats, bred and raised in our animal quarters and weighing 200-250 g, were used. The rats were fed overnight with 5% glucose saline solution, anesthetized by administration of 50 mg/kg ip pentobarbital sodium, and subjected to laparotomy to exteriorize the small intestine from the ligament of Treitz to the ileocecal valve. Three rats were killed without further manipulation, and 10-cm jejunal segments 15 cm below the ligament of Treitz were collected as normal jejunum to serve as the external control. In the other animals, a 10-cm jejunal loop beginning at 20 cm below the ligament of Treitz was marked with two silk-thread ligatures, which also ligated the arterioles and venuoles in the mesenteric margin of the intestine to interrupt collateral circulation. To induce an episode of segmental I/R within the jejunum, the blood flow of mesenteric blood vessels connecting to the marked jejunal loop was interrupted by clamping for 30 min (ischemia) and then released for blood recirculation (reperfusion). After the induction of I/R, the small intestine was returned to the peritoneal cavity and the wound was closed in two layers with 3-0 silk sutures. The rats were provided with 5% glucose solution ad libitum and were killed at 3, 6, 11, 24, and 48 h after reperfusion. The I/R-injured jejunal loop, 0.5 cm away from proximal and distal end markers, was removed and flushed with 10 ml of ice-cold saline. The jejunal mucosa was gently scraped and collected for biochemical assays and for total RNA isolation. For internal controls, the mucosa of a 10-cm jejunal segment just above the ischemic reperfused jejunum (I/R jejunum) was also collected and designated as control jejunum.

Indirect immunofluorescent staining of SI and LPH. Digestive functional recovery depends on the expression of functional proteins at an appropriate cellular location. Intestinal SI and LPH are expressed in the BBM of villus but not crypt cells (36). To determine the expression and cellular location of these enzymes, we used indirect immunofluorescent staining methods with monospecific rabbit anti-rat SI or LPH antiserum as the probe (48). In this study, a 1-cm segment at the center of the I/R jejunum and a distal segment of control jejunum were cut open, laid on a paraffin block, and fixed with 3% paraformaldehyde-1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) at 4°C for 30 min. The fixed tissues were washed in phosphate buffer, transferred to a 20% sucrose-phosphate buffer solution for 20 min, embedded in O.C.T. compound, frozen, and sectioned at 5-µm thickness. Sections from tissues at different time points were mounted on the same slide to ensure that all tissues were stained under the same conditions. The slides were examined as described previously (48). Briefly, tissue sections were 1) washed with PBS; 2) preincubated with 10% control rabbit serum at room temperature for 10 min, washed once with PBS, and incubated with rabbit antiserum against rat SI or LPH (1:500 dilution) for 1 h; 3) washed three times with PBS, 10 min each; 4) incubated with FITC-conjugated goat-IgG anti-rabbit IgG (1:250 dilution) for 30 min; 5) washed three times with PBS; and 6) mounted in Immunomount and observed under a Nikon microscope with an epifluorescence attachment.

Biochemical assays and Western blots. IAP, SI, and LPH activities were measured using the substrates phenylphosphate, sucrose, and lactose as described previously (48, 49, 52). Protein content was measured by the method of Bradford (5).

Northern blot analysis. Mucosal RNA was isolated according to the method of Chomczynski and Sacchi (10).

Preparation of nuclear extracts. For the determination of SIF1 activity, nuclear extracts were prepared from isolated nuclei. The nuclei were isolated by a modified method described by Lamers et al. (20). Briefly, mucosal scrape (~200 mg) was washed with 20 ml ice-cold PBS, suspended in 10 ml of cell lysis buffer consisting of 10 mM HEPES (pH 7.5), 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol (DTT), 200 mM sucrose, 0.5% Nonidet P-40 (NP-40), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1 µg/ml leupeptin, and 1 µg/ml aprotinin, homogenized by 60-80 strokes in a Dounce homogenizer with pestle B to obtain >80% lysed cells (monitored by phase-contrast microscopy), filtered through four layers of cheesecloth, and centrifuged at 800 g for 5 min to obtain the crude nuclear pellet. The pellet was resuspended in 25 ml of cell lysis buffer without NP-40 but containing 1.65 M sucrose. The crude nuclear suspension was layered on a sucrose cushion solution (2 M sucrose, 2 mM MgCl₂, and 10 mM Tris · HCl, pH 7.5) and centrifuged 25,000 g for 1 h. The nuclear pellet was then resuspended in 300 µl nuclear extract buffer (20 mM HEPES, pH 8.0, 20% glycerol, 0.1 mM KCl, 0.1 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF, and 1 µg/ml each of leupeptin and aprotinin), sonicated at 15% power output for 5 s, and centrifuged at 12,000 g for 1 min. The nuclear extract was transferred to fresh tubes and stored at 72°C in aliquots.

EMSA. The DNA binding activity of SIF1 in the nuclear extract was determined by EMSA. Oligonucleotides of the rat SIF1 element (5'-TGTGAAAGTGCAATAAAACTTTATGAGTA-3' and 5'-TGACTACTCATAAAGTTTTATTGCAC-3') were synthesized according to the rat nucleotide sequence reported by Hecht et al. (15) and labeled with [³²P]dATP with the use of a filling reaction in the presence of a Klenow fragment of DNA polymerase 1 to form a double-stranded SIF1 motif corresponding to base pairs 62 to 29 from the transcriptional start site. The labeled probes were separated from free [³²P]dATP on an NAP-5 column (Pharmacia Biotech, Piscataway, NJ). EMSA was performed in 15 µl containing 10-15 µg nuclear extract protein, 20 mM HEPES, pH 7.5, 60 mM KCl, 10 mM MgCl₂, 0.5 mM DTT, 1 mM EDTA, 12.5% glycerol, 1 µg poly(dI-dC), and 1 × 10⁵ counts per minute ³²P-labeled DNA. The mixture was incubated at room temperature for 30 min and analyzed by electrophoresis on a 5% nondenaturing polyacrylamide gel, using high-iron conditions described by Ausubel et al. (1). To examine whether the protein-DNA complexes contained the SIF1 binding factor Cdx-2, supershift assays were performed using affinity-purified antibody against mouse Cdx-2 (kindly provided by Dr. Peter R. Traber) (38).

Statistical analysis. ANOVA was computed from the experimental data. When the F value obtained from ANOVA was significant, Bonferroni's test was applied to test for differences among groups. When comparing values of I/R jejunum and control jejunum in the same animal, paired Student's t-test was used. P < 0.05 was considered significant.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Changes in the intestinal mucosal abundance of IAP, LPH, SI, and Ft H transcripts after I/R. Figure 1 shows the changes in the abundance of IAP, LPH, SI, and Ft H transcripts in normal jejunum, internal control jejunum, and I/R jejunum after reperfusion by Northern blot analysis. All transcripts were significantly decreased in I/R jejunum after 3 and 6 h of reperfusion (Figs. 1 and 2). Local segmental I/R also produced a modest reduction in all transcripts in the control jejunum; the decrease was significant for IAP and LPH mRNAs, but not SI and Ft H mRNAs, compared with the mRNAs of the normal jejunum (Figs. 1 and 2). The recovery of the transcripts in I/R jejunum occurred to different extents and at different times: LPH, SI, and Ft H mRNA levels had recovered and increased to levels ~50% higher than normal levels at 12 h of reperfusion (P < 0.05 for all transcripts), whereas IAP mRNA did not recover until 48 h (Figs. 1 and 2). By 24 h of reperfusion, the LPH and Ft H mRNAs had subsided, whereas SI mRNA increased further and then decreased to a level still higher than that of normal jejunum at 48 h (Fig. 2). The overshoot phenomenon of mRNA during recovery also occurred in control jejunum, in which LPH and SI mRNA increased to a level higher than that in normal jejunum at 12 and 24 h, respectively (Fig. 2).

View larger version (94K): [in this window] [in a new window]   Fig. 1.   Northern blot analysis of intestinal alkaline phosphatase (IAP), lactose-phlorizin hydrolase (LPH), sucrase-isomaltase (SI), and ferritin heavy-chain (Ft H) mRNA and 28S rRNA in intact normal jejunum (N-J), internal control jejunum (C-J), and ischemic reperfused jejunum (I/R-J). Rats were fed a glucose solution overnight and were either killed (control) or subjected to surgical induction of jejunal segmental ischemia for 30 min, followed by reperfusion for recovery. Animals were killed at indicated times after reperfusion, and the ischemia-reperfusion (I/R) segment and the control segment (immediately above the I/R segment) were collected for analysis (see MATERIALS AND METHODS for details). About 10 µg total RNA per sample was size separated by electrophoresis in a 1% formaldehyde agarose gel and transferred to a Nylon membrane. The same membrane was sequentially blotted with 32P-labeled probes, and the probe was removed by stripping after each blot. All mRNAs were detected by a storage Phosphorimaging system. Reduction of all mRNAs in I/R-J was noted at 3 and 6 h. Restoration of mRNA levels occurred at different times after reperfusion for each mRNA. 28S rRNA showed that equivalent amounts of RNA samples were applied to each lane.

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Effect of I/R on the abundance of mucosal IAP, LPH, SI, and Ft H transcripts. Radioactive signals from Northern blots as shown in Fig. 1 were quantitated by volume integration using a Phosphorimaging system with ImageQuant application software (see MATERIALS AND METHODS for details) and were normalized to 28S rRNA in each sample. Data are means ± SE, expressed as percentage of normal levels (n = 3-4). + P < 0.05 and ++ P < 0.01 vs. C-J; * P < 0.5 and ** P < 0.01 vs. N-J.

Changes in villus height and mucosal IAP, SI, and LPH activities and Ft content after I/R. Figure 3 shows morphological changes in the intestine after I/R, resulting in a significant decrease in villus height (Fig. 4). Intestinal damage that occurred at 1 h after reperfusion was readily observed at the boundary between control jejunum and I/R jejunum (Fig. 3A). The control jejunum (Fig. 3A, right) showed intact villi, whereas the I/R jejunum (Fig. 3A, left) showed upper villus cells extruding into the lumen (Fig. 3A). By 3 h, the I/R jejunum had lost most upper villus cells and showed significantly decreased villus height (Figs. 3B and 4). The villus height did not increase until 24 h after reperfusion (Figs. 3C and 4) and had recovered to normal levels at 48 h (Figs. 3D and 4). In contrast to the villus height, the crypt depth showed a statistically insignificant trend of increasing rather than decreasing depth after reperfusion (data not shown).

View larger version (134K): [in this window] [in a new window]   Fig. 3.   Micrographs of intestinal mucosa after I/R (bar = 25 µm). A: boundary between I/R-J (left) and C-J (right) 1 h after reperfusion. Upper villus cells were extruded in the area of I/R-J, whereas villus remained intact in C-J. B, C, and D: I/R-J at 3 (B), 24 (C), and 48 h (D) after reperfusion, respectively. Villus height decreased ~60% at 3 h, partially recovered at 24 h, and fully recovered by 48 h.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Effect of I/R on villus height. Villus height was measured from villus tip to crypt-villus junction under ×200 magnification. Data are means ± SE (n = 3-4).  P < 0.01 and * P < 0.05 vs. control jejunum.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Effect of segmental jejunal I/R on IAP, lactase (L), and sucrase (S) enzyme activity at 3, 6, 12, 24, and 48 h after reperfusion. Enzyme units (U) are expressed as micromoles of substrate hydrolyzed per milligram protein per minute. Data are means ± SE (n = 3-4). Dashed line across each panel represents mean enzyme activity of N-J. + P < 0.01 and ++ P < 0.05 vs. C-J; * P < 0.05 and ** P < 0.01 vs. I/R-J at immediate earlier time point.

View larger version (27K): [in this window] [in a new window]   View larger version (23K): [in this window] [in a new window]   Fig. 6.   Changes in Ft protein after I/R determined by Western blots. A: representative Western blot analysis of Ft in mucosa of N-J, C-J, and I/R-J. Aliquots of mucosal homogenates containing 10 µg protein were subjected to SDS-PAGE under denaturing conditions, the protein was transferred to a nitrocellulose membrane, and Ft was detected by rabbit antiserum to rat Ft, using the enhanced chemiluminescence blot kit (see MATERIALS AND METHODS for details). B: Ft protein was quantitated by transmittance densitometry, using volume integration with ImageQuant application software. Data are expressed as percentage of Ft in the N-J (means ± SE, n = 3). * P < 0.05 vs. C-J; § P < 0.05 vs. immediate earlier time point.

Localization of immunoreactive SI and LPH. The normal jejunum showed a pattern of gradual increase in SI expression along the villus axis (Fig. 7A). SI was undetectable in the crypts, appeared at the villus base, and increased to plateau in the upper fourth of the villus (Fig. 7A). The SI was predominantly located at the BBM (Fig. 7A). In samples stained with control antiserum, the BBM was devoid of immunoreactivity, whereas some of the undefined nonepithelial cells showed nonspecific fluorescent staining (not shown). I/R did not alter subcellular localization of SI but markedly decreased the villus height and reduced BBM immunoreactivity at 3 or 6 h after I/R (Fig. 7C). The exfoliated cells remaining in the lumen of I/R jejunum showed positive fluorescent staining (Fig. 7C). By 12 h, immunoreactivity was detected not only in BBM of villus cells but also at the apical surface of cells in the crypt bottom (Fig. 7D). By 24 h, SI immunoreactivity in the crypt cells had subsided (Fig. 7E). By 48 h, I/R jejunum had completely recovered villus height, with the villi covered by epithelial cells showing increased SI immunoreactivity throughout the BBM of villus cells (Fig. 7F). It is worth noting that upregulation of SI expression might occur in cells at the villus base, as these cells had levels of SI immunoreactivity comparable to those at the upper villus (Fig. 4F). LPH was also localized at the BBM of the villus cells. Changes in LPH immunoreactivity after reperfusion were the same as those of SI, except there was no fluorescent stainability in the BBM of the crypt cells at 12 h after reperfusion (data not shown).

View larger version (81K): [in this window] [in a new window]   Fig. 7.   Localization of SI expression by indirect immunofluorescent staining (bar shown in A = 35 µm for A, B, C, E, and F and 100 µm for D). A: normal jejunum. Immunoreactive SI is localized in brush-border membrane (BBM; arrows) of absorptive cells with increasing immunofluorescence from villus base to a plateau at the upper fourth of the villi. Continuity of the immunoreactive BBM is interrupted by goblet cells (arrowheads). Unidentified leukocytes in the villus core show nonspecific fluorescent staining. B: control jejunum 6 h after distal segmental I/R. Villus height and vertical gradient of immunoreactive SI at the BBM are similar to those of normal jejunum. C: I/R-J at 6 h after I/R. Compared with control jejunum, I/R-J shows smaller and shorter villi, which are lined with epithelial cells showing weak SI staining at BBM. Some exfoliated upper villus cells in the lumen retain immunoreactivity. D: I/R-J at 12 h after I/R. Immunoreactive SI is detected not only in the BBM of the villus absorptive cells (arrows) but also in cells located at the crypt base (arrowheads). No fluorescent activity is detected in goblet cells. E: I/R-J at 24 h after I/R. Villi are still short but are covered by the epithelium with strongly stained fluorescent BBM. Cells at base of crypts show no detectable immunoreactive SI. F: I/R-J at 48 h after I/R. Villi have recovered to normal height and are lined by epithelium expressing higher levels of SI at the BBM throughout the villus. High fluorescent intensity is consistent with high specific and total activity measured biochemically.

Changes in SIF1 activity. Figure 8 shows that SIF1 binding proteins in the nuclear extract from the mucosa of normal jejunum and control jejunum produced two major protein-DNA complexes, A and B, confirming an earlier report (37). In addition, there were faint complexes, C and D, with greater mobility. SIF1 binding activity was present at low levels in both normal and control jejunum (Fig. 8A) and was significantly increased in the I/R jejunum throughout the first 12 h of reperfusion (Fig. 8). Interestingly, complexes A-D showed specific temporal patterns after reperfusion: the increase of complex A did not occur until 6 and 12 h; the increase of B was detected at 1 h and continued during the 12-h period (Fig. 8A); complex C increased from 1 to 6 h and decreased to a low level by 12 h; and complex D increased at 1 h and decreased quickly to the background level. Complexes C and D were reproducibly detected in our laboratory. It is possible that the proteins in C and D complexes might be the degraded products of activated proteins in A and B complexes after tissue injury in I/R jejunum, as complexes C and D were not present in the nuclear extract of I/R jejunum at 12 h. The SIF1 protein binding was specific, as all complexes were not detected when a 100-fold concentration of competitor DNA was included in the assay (Fig. 8A). Incubation of the nuclear extract with rabbit anti-mouse Cdx-2 produced a supershift band with a concomitant decrease and/or disappearance of complex A in control jejunum and I/R jejunum at 1 h and in I/R jejunum at 12 h after reperfusion (Fig. 8A, last 3 lanes). The supershift indicates that Cdx-2 or a related protein participated in the formation of complex A. SIF1 binding activity was present in the nuclear extract of epithelial cells but not submucosal cells (data not shown).

View larger version (79K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Fig. 8.   Electromobility shift assay (EMSA) of SI footprinting 1 (SIF1) protein binding activity in jejunal nuclear extracts. A: rats were fed 5% glucose saline overnight and subjected to surgical induction of segmental I/R. Intact rats were killed after anesthesia to serve as normal controls. The mucosal scrape was collected from normal jejunum (N-J), internal control jejunum (C-J), and ischemic reperfused jejunum (I/R-J) for isolation of the nuclear extract. An aliquot of nuclear extract containing 15 µg protein was used for each assay. Low levels of SIF1 binding activity were observed in A, B, C, and D bands in N-J and C-J. These four bands became conspicuous in I/R-J 1 h after reperfusion, and each complex showed specific temporal changes during the first 12 h after reperfusion. The DNA-protein interaction was not detected when 100-fold competitor was included in the assay. Incubation of nuclear extract from 1 h C-J and 1 and 12 h I/R-J with polyclonal antibodies against Cdx-2 for 30 min produced a supershift band in association with the decrease of complex A (last 3 lanes). B: changes in SIF1 binding activity in C-J and I/R-J. Radioactive signals of A and B complexes, as shown in Fig. 8A, were detected by a storage Phosphorimaging system and quantitated with ImageQuant software. Data are means ± SE (n = 4). Significant increases in the activity of SIF1 binding protein occurred in I/R-J at 1 h, and this increase was retained during the 12 h of the experiment. * P < 0.01 vs. C-J.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The mechanisms by which temporal and/or spatial regulation of cell differentiation occurs along the crypt-villus axis during the short 2- to 3-day life span of enterocytes are poorly understood. The present studies were undertaken to examine molecular events occurring as the regenerated and/or immature cells differentiate after the loss of upper villus cells induced by I/R. A short 30-min period of ischemia with subsequent reperfusion has been reported to cause upper villus cell exfoliation (4). A prolonged 90-min period of ischemia followed by reperfusion destroys the villus with a 90% decrease in the activity of disaccharidases (18). We observed that 30 min of ischemia followed by reperfusion caused a 60% decrease in the specific activity of IAP and 40-50% reduction in both lactase and sucrase activities at 3 and 6 h. The decrease was not limited to the BBM enzymes, as cytosolic Ft showed a 40-65% decrease at 1-6 h after I/R. The decrease of the BBM proteins and Ft appears to be primarily the result of premature exfoliation of upper villus cells, as the shortened villi were covered by the epithelium expressing low levels of SI immunoreactivity. The decrease in mRNA levels would not be predicted from the usual mRNA distribution along the villus axis; as IAP, LPH, and SI transcripts are higher in the lower than in the upper villus (38, 39, 45), the loss of upper villus cells should cause a greater decrease in enzyme activity than in mRNA levels. The present observation that after I/R BBM mRNAs were reduced, in some instances to a greater degree than the reduction of enzyme activities (Figs. 2 and 3), suggests increased mRNA degradation after I/R.

A significant decrease of Ft, but not BBM enzymes, was found in the control jejunum at 3 and 6 h after I/R of distal jejunum. Since the precocious exfoliation of upper villus cells did not occur in the control jejunum, the decrease of Ft must have resulted either from a decrease in Ft synthesis, an increase in Ft degradation, or both. The decrease of Ft content without a parallel decrease in the Ft mRNA levels suggests either translational repression of Ft mRNA or accelerated Ft protein degradation. Direct identification of the actual mechanism will depend on measurements of Ft synthetic and degradative rates. It is well established that Ft synthesis is negatively regulated at the translational level through the binding of IRP to the IRE on the 5' untranslated region of the Ft mRNA (14, 16). The increase of IRP activity has been reported in cultured cells under oxidative stress, resulting in the suppression of Ft mRNA translation (16, 25-27). The decrease of Ft mRNA translation in the control jejunum may well be related to the oxidative stress induced by I/R in the more distal jejunal segment. The phenomenon of remote tissue damage after intestinal I/R has been reported in the lung and liver (8, 9, 13) and is thought to be mediated by activated leukocytes and inflammatory cytokines in the circulation (8, 9, 13, 28, 34). The precise mechanism by which a circulatory mediator causes translational repression of Ft mRNA in the control jejunum is unknown. Physiologically significant levels of tumor necrosis factor- (TNF-) and interleukin-1 (IL-1) have been reported to be released from human intestinal segments after 30 min of ischemia followed by reperfusion (44). Moreover, intestinal I/R has been reported to elevate plasma concentrations of TNF-, IL-1, IL-6, and endotoxin (9, 28). It is likely that these circulatory cytokines released from the I/R jejunum enhance expression of adhesive molecules in the endothelium of remote organs (8, 13, 23). Subse- quently, these adhesive molecules would recruit and activate circulating leukocytes to peripheral tissues (23, 28). In our experimental model, activated macrophages and/or neutrophils that lodged in the mucosa of control jejunum may release hydrogen peroxide and nitric oxide, which subsequently activate IRP and repress Ft synthesis (16, 25-27). This distant tissue effect takes place later than in the local tissues, as the decrease of Ft protein was detected at 3 h in control jejunum compared with 1 h in I/R jejunum. The indirect tissue effects might also have caused the reduction the IAP, LPH, and SI mRNAs in the control jejunum. In Caco-2 cells, cytokines such as IL-6 and interferon- have been reported to decrease SI but not LPH synthesis, and the magnitude of the downregulation is greater than that of upregulation by TNF- (53). Thus segmental intestinal I/R elicited effects on the expression of diverse genes in the local and distant enterocytes; the effects were both gene specific and occurred at different levels of regulation.

The LPH and SI activity and immunoreactive SI recovered quickly, by 12 h after I/R. The increase of LPH and SI expression occurred in the absence of an appreciable increase in the mucosal weight, a phenomenon consistent with earlier reports that the recovery of villus cellularity is independent of BBM hydrolase or functional recovery (12, 18, 31). Our data showed that LPH and SI enzyme recovery occurred before, rather than after, the restoration of villus cellularity (18, 31). Differences in the depth of mucosal damage might account for this variance from earlier reports, as the duration of ischemia was 30 min in the present study and 90-180 min in other studies (18, 31).

Although changes in SI immunoreactivity observed with indirect immunofluorescent staining methods might not reflect tissue SI content, the observation of parallel changes in SI immunoreactivity and sucrase activity measured by biochemical assay after I/R implies a quantitative increase in SI. These parallel changes are particularly conspicuous in 48 h I/R jejunum, in which the BBM throughout the villus showed high SI immunoreactivity (Fig. 4F) and intestinal SI activity was significantly higher than normal (Fig. 2 and Table 1). The presence of immunoreactivity in the BBM of the crypt bottom cells at 12 h after I/R, but not later, suggests a transient activation of the SI gene that is normally not expressed in the crypts. Whether the activation of the SI gene in the crypts is related to a local increase in SIF1 binding activity remains to be determined.

The expression of LPH and SI, and perhaps IAP, is known to be regulated essentially at the level of transcription (17). The present study found that the recovery of IAP, LPH, and SI mRNAs and their protein products occurred in parallel, agreeing with transcriptional regulation. It is worth noting that BBM mRNAs did not recover in parallel, as lactase and sucrase returned to normal levels at 12 h, whereas IAP did not return to normal levels until 48 h after I/R, indicating that each gene is differentially regulated. The concomitant recovery of LPH and SI appears to be more than coincidental. A cis-element conferring positive regulation for LPH expression contains the nucleotide sequence TTTAT (NF-LPH1), which is very similar to the cis-element with a palindromic SIF1 sequence TTTAC (37, 38, 42). Both sequences have been demonstrated to compete for binding to Cdx-2 or nuclear factors closely related to Cdx-2 (41). The significant increase in the activity of SIF1 binding protein detected in the I/R jejunum might be related to the upregulation of both LPH and SI expression during recovery after I/R. A recent report suggests that the rat intestinal SIF1 binding protein might not be related to Cdx-2 (15). However, the present EMSA demonstrated that the antibody against mouse Cdx-2 produced a supershift band with the disappearance of complex A (Fig. 8A). These data provide evidence indicating that the SIF1 binding protein in complex A of the rat intestinal nuclear extract is either Cdx-2 or a Cdx-2-related protein, consistent with earlier findings in mouse, pig, and human intestine (37-42). Complexes B and C shown in Fig. 8A might be equivalent to complexes IV and III reported in adult rats (15), as the anti-Cdx-2 antibodies altered neither mobility nor activity of these complexes. The identity of complexes B and C, as well as D, remains to be defined.

In addition to SIF1 binding and/or Cdx-2 proteins, other transacting factors might also be involved in the upregulation of SI expression, since a significant increase in the SI mRNA, but not LPH mRNA, occurred in the control jejunum at 24 h without an increase of SIF1 binding activity (Figs. 2 and 8). Recently, a consensus binding site for GATA zinc finger transcription factors, in addition to the SIF1 site, has been identified as necessary for transcriptional activation of SI (35). Thus it is possible that the increase of an undefined GATA protein occurred in control jejunum, resulting in the increase of SI expression.

Similar to BBM proteins, mucosal Ft mRNA increased to higher than normal levels at 12 h after I/R (Fig. 2). This recovery may reflect an increase of Ft H gene transcription, probably mediated by oxidative stress activation of nuclear transcription factor-B (NF-B), which has been reported to enhance Ft gene expression in response to TNF-1 in cultured mouse fibroblasts (19). Our recent preliminary data showed that NF-B is activated in I/R jejunum prior to the recovery of Ft, although it is not known whether the rat Ft H gene contains B enhancer elements (51). Alternatively, increased Ft H gene transcription may be directly related to an increased intracellular free iron pool after intracellular Ft degradation. It has been reported that expansion of the intracellular free iron pool and increased Ft expression occurred after oxidative stress in the liver (7). The present data showing the recovery of Ft H mRNA, but not Ft protein, at 12 h after I/R indicate that the free iron pool in the I/R jejunum was not high enough to totally derepress mRNA translation. Further study of rat Ft H gene structure is needed to confirm the role of NF-B in the upregulation of Ft expression.

The rapid recovery of intestinal BBM enzymes and Ft mRNAs and mucosal cellularity after I/R injury underscores the high adaptability of the intestinal epithelial cells for regeneration and differentiation. The recovery apparently depends on the genetic programs of surviving crypt and lower villus cells. The present observation that in surviving cells SIF1 binding protein and/or Cdx-2 activity was increased in association with SI and perhaps LPH recovery suggests a mechanism of transcriptional activation of SI and LPH genes. Recently, overexpression of Cdx-2 in Caco-2 cells has been shown to stimulate SI and LPH expression (22). Other genes related to cell survival and cell proliferation must also be activated after I/R. In preliminary studies, we observed that I/R increased NF-B and activation protein-1 (AP-1) activity (Ref. 51 and unpublished data); NF-B has been reported to play a role for cell survival (2, 43), and AP-1 activity is known as an early response to growth stimulation (32, 33).