Conceived and designed the experiments: RM JD DK RvD CD WB. Performed the experiments: RM JD DK RvD WB. Analyzed the data: RM JD DK RvD CD WB. Contributed reagents/materials/analysis tools: RM JD DK CD WB. Wrote the paper: RM JD RvD CD.
Current address: Spaarne Hospital, Hoofddorp, the Netherlands
W.A. Buurman is shareholder of the company Hycult Biotechnology (Hbt, Uden, the Netherlands) which provided the human MBL and I-FABP ELISAs. These ELISAs, as used, are commercially available worldwide.
Recently, we observed that small-intestinal ischemia and reperfusion was found to entail a rapid loss of apoptotic and necrotic cells. This study was conducted to investigate whether the observed shedding of ischemically damaged epithelial cells affects IR induced inflammation in the human small gut.
Using a newly developed IR model of the human small intestine, the inflammatory response was studied on cellular, protein and mRNA level. Thirty patients were consecutively included. Part of the jejunum was subjected to 30 minutes of ischemia and variable reperfusion periods (mean reperfusion time 120 (±11) minutes). Ethical approval and informed consent were obtained. Increased plasma intestinal fatty acid binding protein (I-FABP) levels indicated loss in epithelial cell integrity in response to ischemia and reperfusion (p<0.001 vs healthy). HIF-1α gene expression doubled (p = 0.02) and C3 gene expression increased 4-fold (p = 0.01) over the course of IR. Gut barrier failure, assessed as LPS concentration in small bowel venous effluent blood, was not observed (p = 0.18). Additionally, mRNA expression of HO-1, IL-6, IL-8 did not alter. No increased expression of endothelial adhesion molecules, TNFα release, increased numbers of inflammatory cells (p = 0.71) or complement activation, assessed as activated C3 (p = 0.14), were detected in the reperfused tissue.
In the human small intestine, thirty minutes of ischemia followed by up to 4 hours of reperfusion, does not seem to lead to an explicit inflammatory response. This may be explained by a unique mechanism of shedding of damaged enterocytes, reported for the first time by our group.
Recently, we demonstrated the ability of the human small intestine to rapidly restore its epithelial architecture after massive epithelial sloughing and gut barrier damage in response to a thirty minutes ischemic episode and variable reperfusion periods.
The presence of apoptosis or cellular damage in reperfused ischemic organs has cardinal implications in the pathogenesis of ischemia reperfusion (IR).
The inflammatory response induced by IR is essentially characterized by different contributors. Early after cellular injury various cytokines, such as tumor necrosis factor-α, interleukin-1, -6, -8 or -10 and others, are expressed.
The presence of apoptotic and necrotic cells following ischemia reperfusion of the small intestine appears to be of critical importance in the development of an inflammatory reaction that contributes to tissue injury. From this we hypothesized that the remarkable ability of the human small intestine to effectively shed apoptotic cells, is a way of losing an important endogenous inflammatory trigger. This may serve as a unique intestinal rescue mechanism, which potentially attenuates the development of an intense IR induced inflammatory reaction. Therefore, we studied time related characteristics of IR induced inflammation of the human small intestine in a newly developed human ischemia reperfusion model of the jejunum. The results indicate that reperfused ischemic jejunum effectively discards apoptotic and damaged epithelium, restores the epithelial gut barrier and thereby prevents the development of a local inflammatory response.
The study was approved by the Maastricht University Medical Center ethical committee. Written consent of all patients was obtained before inclusion. Consecutively included patients were scheduled to undergo pancreatic surgery; twenty eight underwent a pylorus preserving pancreatico-duodenectomy (modified Whipple's procedure) for benign or malignant pancreatic disease and two underwent a Frey's procedure for chronic pancreatitis. During pancreatico-duodenectomy a variable length of jejunum is usually resected in continuity with the specimen. Similarly, in creating a Roux-limb for the pancreatico-jejunostomy in Frey's procedure, it is often necessary to resect a small segment of jejunum. We took advantage of this, which enabled us to study IR induced cell damage in a harmless human jejunal IR model. To that purpose, the most distal part of the jejunum to be resected with the Whipple specimen, or the most proximal part of the Roux-limb to be used as a pancreatico-jejunostomy in the Frey's procedure, was used to isolate a 6 cm segment of jejunum. The jejunum of all patients was carefully checked for any signs of underlying pathology and inflammation during operation and before the start of the IR procedure. Due to an incomplete reperfusion 2 patients were later excluded from further analysis, all having had a modified Whipple's procedure.
The first part of surgery (opening of the abdominal wall, installation of a large self-retaining retractor and exploration for metastases and localization of the major anatomical structures and tumor) was according to standard procedures. From this moment the 6 cm part of jejunum, which was going to be studied, was identified and care was taken that the vasculature of the studied jejunum consisted of 1 central mesenteric arteriole and venule. This was achieved as described before.
Paraffin-embedded sections (4 µm) were prepared from healthy and the final IR jejunum sample. These sections were stained using haematoxylin and eosin or immunostained for endothelial adhesion molecules E-selectin (CD-62-E, monoclonal antibody ENA1, kindly provided by Hycult Biotechnology (Hbt, Uden, the Netherlands)), intercellular adhesion molecule-1 (ICAM-1, monoclonal antibody HM1, Hbt) and mucosal vascular addressin cell adhesion molecule 1 (MADCAM-1, monoclonal antibody 314G8, Hbt), neutrophils by myeloperoxidase staining (polyclonal antibody, A0398, DakoCytomation, Glostrup, Denmark) or human neutrophil defensin 1–3 (monoclonal antibody, D21, Hbt) and complement component C3c (A0062, DakoCytomation). Sections were deparaffinized, rehydrated and blocked for 1 hour at room temperature using 5% BSA in phosphate buffered saline (PBS) or tris buffered saline (TBS) for HM1 and MPO. Properly diluted primary antibodies were incubated in PBS or TBS supplemented with 0.1% BSA for 1 hour at room temperature. After washing, specific antibody binding was detected using specific peroxidase conjugated secondary antibodies (Jackson Immunoresearch, West Grove, PA) or a biotinylated swine anti rabbit secondary antibody (E0432, DakoCytomation) for detecting MPO specific binding. Next, signal enhancement was achieved by peroxidase conjugated avidin biotin complexes. Staining was visualized by 3-amino-9-ethylcarbazole (AEC) followed by a hematoxylin (Sigma, St. Louis, MO) nuclear counterstaining. Microscopic analysis of several high power fields (n = 5) from different paraffin embedded tissue samples was carried out by a blinded observer.
Arteriovenous concentration differences of plasma Intestinal-Fatty Acid Binding Protein (I-FABP) levels were determined by Enzyme-Linked ImmunoSorbent Assay (ELISA) before and after ischemia to measure intestinal mucosal cell damage.
The I-FABP ELISA (HK406, Hbt) was performed according to the manufacturer's guidelines. Myeloperoxidase (MPO) was measured in tissue samples (healthy vs the final IR jejunum sample) that were homogenized according to the later described Western blot protocol. MPO was measured using a commercially available solid-phase ELISA kit based on the sandwich principle (Hbt) in aliquots containing equal amounts (10 µg) of total protein. The assay was performed in accordance with the manufacturer's guidelines.
Arteriovenous concentration differences in plasma tumor necrosis factor-α (TNFα) levels were determined by ELISA before and after ischemia as well as during reperfusion. The TNFα ELISA was performed as described before.
Western blot analyses for human C3 deposition on reperfused ischemic tissue samples and healthy controls was performed as described before, with minor modifications.
Total RNA was extracted from healthy and all IR damaged jejunum samples using a SV Total RNA Isolation System according to manufacturer's guidelines (Promega, Madison, WI). Total cDNA was synthesized with oligo (dT) primers and Molony murine leukemia virus reverse transcriptase. Specific primers (Sigma) for amplification of hemoxygenase-1, hypoxia inducible factor 1-α (HIF-1α), interleukin-6 and -8, TNFα and complement component C3 specific cDNA were designed (
Proteins | Primers | |
Forward | Reversed | |
HIF-1α | ||
HO-1 | ||
TNFα | ||
IL-6 | ||
IL-8 | ||
C3 | ||
RPLP-0 | ||
Cycophilin A |
Data are expressed as means ± SEM and were analyzed by either paired or unpaired two-tailed Student's t-test depending on sample collection. Bonferroni's multiple comparison test was used (after significant repeated measures ANOVA) to compare I-FABP or LPS arteriovenous concentration differences in time. Data were analyzed using Prism 4.01 for Windows (Graphpad Software, San Diego, CA). A p-value≤0.05 was considered statistically significant.
The mean (± SEM) age of included patients (n = 28) was 65 (±2) years and 64% were male. Isolated jejunum samples were subjected to 30 minutes of ischemia followed by a mean total reperfusion time of 120 (±11) minutes (
Distribution of maximal reperfusion times of the isolated jejunum, following 30 minutes of ischemia (n = 28).
Epithelial cell integrity was lost in response to 30 minutes of ischemia followed by various periods of reperfusion. This was illustrated by increased plasma I-FABP levels in the venous outflow of the reperfused jejunum, when compared to arterial circulating levels and I-FABP levels prior to the IR period (* p<0.001.
A) Arteriovenous concentration differences of I-FABP (mesenteric venule minus radial artery) across the isolated ischemic jejunal segment show rapid release of I-FABP into the circulation at reperfusion (early R vs healthy * p<0.001). B) Shedding of damaged epithelial cells into the intestinal lumen during the first thirty minutes of reperfusion. C) Rapid repair of the intestinal epithelial lining, observed following 1 hour of reperfusion.
Disruption of the gut barrier may lead to translocation of microbiota and pro-inflammatory microbial products, such as endotoxin associated with gram negative bacteria (lipopolysacharide (LPS)). Therefore we assessed the translocation of LPS following the rapid and temporary disruption of the epithelial lining in response to IR observed in our model.
Illustrative of rapid gut barrier repair, no translocation of LPS could be detected during the IR procedure. Translocation of LPS was calculated from arteriovenous differences in LPS plasma levels across the isolated bowel segment during reperfusion (Δ (V–A). p = 0.18).
In order to demonstrate a physiological response to anoxia and to reperfusion stress, the mRNA expression profiles of hypoxia inducible factor-1α (HIF-1α) and heme oxygenase-1 (HO-1) were determined in healthy and IR damaged whole jejunum samples (n = 12). The HIF-1α expression, which regulates both physiologic and pathophysiologic responses to ischemia, increased during reperfusion (
Gene | Healthy | IR | Fold increase (95%-CI) | P - value |
HIF-1α | 1.00 | 1.91 | 1.91 (1.15–2.67) | 0.02 |
HO-1 | 1.00 | 1.58 | 1.58 (0.22–2.94) | 0.34 |
TNFα | 1.00 | 1.18 | 1.18 (0.56–1.80) | 0.54 |
IL-6 | 1.00 | 12.31 | 12.31 (−5.29−29.91) | 0.18 |
IL-8 | 1.00 | 15.86 | 15.86 (0.59–31.13) | 0.06 |
C3 | 1.00 | 4.09 | 4.09 (1.74–6.43) | 0.01 |
p-value≤0.05 was considered statistically significant (n = 12).
In order to assess the inflammatory reaction over the course of IR, serial venous blood samples from the reperfused jejunum were analyzed for different inflammatory markers. Additionally, the cytokine response in reperfused whole jejunum tissue samples was studied using Q-PCR. Endothelial cell activation was assessed by immunohistochemistry on paraffin embedded whole jejunum tissue samples before and after IR. No increase in TNFα protein levels was detected in venous plasma samples, collected from the isolated jejunum during the procedure; e.g. prior to ischemia, directly after ischemia and every 30 minutes during reperfusion. Compared to the TNFα levels in arterial circulation, AV-differences in TNFα levels revealed no release of TNFα from the gut (data not shown). In line, no increase in TNFα mRNA was detected in IR damaged jejunum samples, when compared to healthy jejunum controls from the same patient (
mRNA expression of cytokines IL-6 (
Classic in the development of extensive IR damage is the early expression of various adhesion molecules by activated endothelial cells. Presence of E-selectin, ICAM-1 and MADCAM-1 were analyzed using immunohistochemistry. Only 3 patient samples demonstrated sporadic and minimal expression of E-selectin (data not shown), ICAM-1 and MADCAM-1 (
Compared to healthy control jejunum tissue, reperfused jejunum samples in response to 30 minutes of ischemia show an only moderate increase in ICAM-1 (upper panel) and MADCAM-1 (lower panel) expression demonstrated by terracotta red staining of the AEC (indicated by arrows). Original magnification 200x.
Critical in the induction of reperfusion injury is the influx of neutrophils as well as the release of their reactive constituents. The presence of neutrophils was assessed using two different and widely used neutrophil markers, human neutrophil defensin 1–3 (HNP1-3) and myeloperoxidase (MPO), both stored in abundance in azurophilic granules of neutrophils. Analysis of immunostained tissue sections showed that the number of PMN in the reperfused jejunum was similar to the number of PMN observed in healthy control jejunum samples, as detected by immunohistochemical analysis of HNP1-3 (
A) No increase of PMN was observed over the course of reperfusion in response to 30 minutes of ischemia. Detected by specific HNP1-3 staining (AEC, indicated by arrows) the number of PMN in reperfused jejunum did not increase in comparison to healthy tissue (p = 0.90). Original magnification 200x. Clearly the PMN concentrated around the cellular debris collecting in the safe intestinal lumen (right insert in
As a central mediator of IR induced inflammation, complement is involved in the development of organ damage. Complement component C3 is an important constituent of the complement system, of which activation and deposition indicate complement activity. Immunohistochemical analysis of healthy and IR damaged jejunum sections revealed no increase in C3 deposition in response to IR (
A) C3 was not be detected in the reperfused jejunum using immunohistochemistry. Original magnification 200x. B) Western blot analysis, under reducing conditions, revealed no increase in activated human C3 (±75 kD) in whole reperfused jejunum samples, when compared to healthy control tissue (p = 0.14). C) C3 mRNA expression increased over 4-fold during IR of the human jejunum (p = 0.01).
Interestingly, subsequent analysis of the response to IR by the jejunum involving C3 mRNA, using Q-PCR analysis, clearly detected C3 mRNA levels in both healthy and IR damaged whole jejunum samples. During reperfusion, the amount of C3 specific mRNA increased over 4 times in response to 30 minutes of ischemia (p = 0.01,
The present data demonstrate that rapid shedding of IR damaged intestinal epithelial cells attenuates the development of a classically observed vigorous inflammatory response in the reperfused jejunum. When analyzed over a prolonged reperfusion period, no increase in TNF-α, HO-1, IL-6, IL-8 gene expression, expression of adhesion molecules, PMN influx or activation and deposition of complement were observed. However, HIF1- α and C3 gene expression were up-regulated during reperfusion. This possibly illustrated an IR induced regenerative response, aimed at intestinal barrier restoration and prevention of bacterial translocation.
Development of IR induced organ damage is generally described to be characterized by an excessive and vigorous inflammatory response.
It has been demonstrated in the kidney as well as other organs that IR induced inflammation and subsequent organ damage are dependent on the development of widespread apoptosis.
Conclusive data on intestinal ischemia and reperfusion pathophysiology have been obtained from experimental animal studies. Cytokine activity of TNFα, IL-6 and IL-8 have been associated with the inflammatory reaction following IR.
E-Selectin (CD-62-E and formerly known as endothelial leukocyte adhesion molecule-1 (ELAM-1)) has an important role in the recruitment of leukocytes to sites of tissue injury. Its expression by vascular endothelial cells in response to injury is readily induced by cytokines IL-1 and TNFα. Interestingly, our data suggest that despite initial presence of damaged epithelial cells in response to an IR period, the rapid shedding of damaged epithelial debris largely averts endothelial cell activation. Consistent with this observation, PMN infiltration, assessed by total MPO tissue levels or the presence of HNP1-3 positive cells, in IR was not detected in our experiments. MPO, most abundantly released upon PMN activation, has been demonstrated to have a clear effect on the development of IR induced organ damage. Renal IR studies have demonstrated that MPO itself is able to contribute to the development of organ damage.
As a central innate immune regulator, the complement system is activated during intestinal IR and contributes substantially to IR induced inflammation, organ damage and failure.
The recently discovered molecular mechanism that recognizes and responds to excessive cell death consisting of SAP130 and the macrophage inducible C-type lectin (Mincle) fits in our observations.
A mechanism of protection was suggested by the enhanced expression of HIF-1α in the surviving epithelium. Expression of HIF-1α in response to low oxygen tension during ischemia and early reperfusion triggers physiologic responses characterized by activation of functional proteins mucin, P-glycoprotein, intestinal trefoil factor and adenosine A2B receptor, aimed at preventing mucosal inflammation.
Taken together, ischemia, sensed by the small intestine, induces normal physiological and IR induced responses. Our data provide new and compelling evidence of increased HIF-1α and also C3 gene expression during the reperfusion period. However, these responses are not paralleled by an IR induced inflammatory response, since important conditions to an inflammatory reaction have not been met by the absence of dead cells in the reperfused tissue. It is important to realize that these data indicate that the human intestine is more resistant to IR than initially thought. Its ability to shed damaged epithelial cells and repair its ever important barrier function without triggering massive inflammation can be seen as key features that prevent the gut from inflammation following splanchnic ischemia and reperfusion.
The authors wish to thank Jan H. Stoot, MD for his excellent technical assistance during pancreatico-duodenectomy procedures and sample collection.