Rapid reversal of human intestinal ischemia-reperfusion induced damage by shedding of injured enterocytes and reepithelialisation

Abstract

Background: Intestinal ischemia-reperfusion (IR) is a phenomenon related to physiological conditions (e.g. exercise, stress) and to pathophysiological events (e.g. acute mesenteric ischemia, aortic surgery). Although intestinal IR has been studied extensively in animals, results remain inconclusive and data on human intestinal IR are scarce. Therefore, an experimental harmless model for human intestinal IR was developed, enabling us to clarify the sequelae of human intestinal IR for the first time.

Methods and findings: In 30 patients undergoing pancreatico-duodenectomy we took advantage of the fact that in this procedure a variable length of jejunum is removed. Isolated jejunum (5 cm) was subjected to 30 minutes ischemia followed by reperfusion. Intestinal Fatty Acid Binding Protein (I-FABP) arteriovenous concentration differences across the bowel segment were measured before and after ischemia to assess epithelial cell damage. Tissue sections were collected after ischemia and at 25, 60 and 120 minutes reperfusion and stained with H&E, and for I-FABP and the apoptosis marker M30. Bonferroni's test was used to compare I-FABP differences. Mean (SEM) arteriovenous concentration gradients of I-FABP across the jejunum revealed rapidly developing epithelial cell damage. I-FABP release significantly increased from 290 (46) pg/ml before ischemia towards 3,997 (554) pg/ml immediately after ischemia (p<0.001) and declined gradually to 1,143 (237) pg/ml within 1 hour reperfusion (p<0.001). Directly after ischemia the intestinal epithelial lining was microscopically normal, while subepithelial spaces appeared at the villus tip. However, after 25 minutes reperfusion, enterocyte M30 immunostaining was observed at the villus tip accompanied by shedding of mature enterocytes into the lumen and loss of I-FABP staining. Interestingly, within 60 minutes reperfusion the epithelial barrier resealed, while debris of apoptotic, shedded epithelial cells was observed in the lumen. At the same time, M30 immunoreactivity was absent in intact epithelial lining.

Conclusions: This is the first human study to clarify intestinal IR induced cell damage and repair and its direct consequences. It reveals a unique, endogenous clearing mechanism for injured enterocytes: rapid detachment of damaged apoptotic enterocytes into the lumen. This process is followed by repair of the epithelial continuity within an hour, resulting in a normal epithelial lining.

Figure 1. Histological analysis of H&E-stained jejunal sections (100×) shows a normal epithelial lining in untreated tissue (A), and upon 30 minutes of ischemia (B).

However, subepithelial spaces (*) appear after 30 minutes ischemia (B). After 25 minutes reperfusion shedding of mature enterocytes into the lumen is found (C). Within 60 minutes of reperfusion the epithelial lining is resealed, while debris of apoptotic, shedded enterocytes (*) are found in the lumen (D).

Figure 2. Immunolocalization of I-FABP in red (3-amino-9-ethylcarbazole, AEC) (100×) in the control jejunum not subjected to ischemia-reperfusion (A) shows an abundant cytosolic presence of I-FABP in the epithelial cells of the upper half of the villus.

Upon 30 minutes ischemia (B), cytosolic I-FABP staining is decreased in mature enterocytes with abundant staining in the subepithelial spaces. A decreased cytosolic staining is still observed after 25 minutes reperfusion (C). Within 60 minutes reperfusion, I-FABP cytosolic positive cells are part of the resealed epithelial barrier (D), while shedded I-FABP containing enterocytes are found in the debris in the lumen.

Figure 3. Characterization of the basement membrane with collagen IV staining in red (AEC) shows collagen IV positive cells directly beneath the epithelial layer in control jejunum (100×).

(A). Upon 30 minutes of ischemia, a clear retraction is found of the collagen IV positive cells from the basal pole of the epithelial cells at the tip of the villus, causing subepithelial spaces (B). After 25 minutes reperfusion, the retracted basement membrane is still observed (C). Within 60 minutes reperfusion, the collagen IV positive basement membrane is again attached to the epithelial lining (D).

Figure 4. Arteriovenous concentration differences of I-FABP (mesenteric venule minus radial artery), a marker for acute enterocyte damage, across the isolated jejunal segment show rapid significant release of I-FABP after ischemia, while concentration differences of I-FABP gradually reduce following reperfusion.

* p<0.001, # p<0.001.

Figure 5. Apoptosis, assessed by M30 staining in red (AEC).

An M30 positive cell (arrow) is observed at the top of the villus in normal jejunum (100×) (A). Upon 30 minutes ischemia, no M30 positive staining is found (B), while after 25 minutes reperfusion intense M30 immunostaining is observed at the villus tip near the shedding of mature epithelial cells into the lumen (arrows) (C). At 60 minutes reperfusion M30 immunoreactivity is no longer detectable in the intact, resealed epithelial barrier, while debris of M30 positive, shedded epithelial cells is observed in the lumen (D).

Figure 6. Histological analysis of H&E stained AgarCyto-fixed debris, appearing from the lumen of the jejunum after 30 minutes ischemia and more than 60 minutes ischemia, shows sheets of enterocytes (100×).

(A), containing many M30 positive cells (arrows) in red (AEC) (100×) (B).