Quantification of mucosal leucocyte endothelial cell interaction by in vivo fluorescence microscopy in experimental colitis in mice

Abstract

Leucocyte recruitment to sites of intestinal inflammation is a crucial, multi-step process that leads ultimately to the accumulation of cells in the inflamed tissue. We established a new in vivo model system of experimental colitis to quantify leucocyte–endothelial cell interaction and leucocyte extravasation in the inflamed mucosa of the colon. Furthermore, we investigated the pathophysiological role of ICAM-1 in the intestinal microcirculation in vivo. Using the model of dextran sodium sulphate (DSS)-induced acute and chronic colitis in mice, in vivo microscopy was performed in the colonic submucosal postcapillary venules and the submucosal collecting venules in normal or inflamed murine colonic segments. ICAM-1 expression was blocked by an anti-ICAM-1 monoclonal antibody or by suppressing NF-κB activation by gliotoxin. Significant increases in leucocyte adhesiveness (51-fold in postcapillary venules, 30-fold in collecting venules, P < 0·01) and extravasation (6·5-fold) could be demonstrated as early as day 2 of DSS-application in acute colitis (P < 0·01). This was paralleled by increases in both the histological damage scores and myeloperoxidase activities. In chronic dextran sodium sulphate-induced colitis significant increases in leucocyte–endothelium interactions and leucocyte extravasation were observed. Blocking ICAM-1 expression with a monoclonal antibody or gliotoxin, leucocyte sticking and extravasation were significantly down-regulated in vivo compared to controls (> 70%; P < 0·01). This new model system offers the possibility to specifically assess the role of adhesion molecules in the colonic mucosa in vivo as well as to investigate and quantify the effectiveness of experimental therapeutic approaches in acute or chronic intestinal inflammation.

Introduction

Leucocyte recruitment into sites of inflammation is a multi-step process that leads to the accumulation of cells in the inflamed tissue. This process is regulated by a whole array of different mediators such as cytokines, chemokines or bacterial products and occurs largely in the region of the microvasculature. Proinflammatory mediators, such as tumour necrosis factor (TNF-α) or endotoxin, up-regulate the expression of adhesion molecules on the endothelial cells lining the sites of inflammation. The newly expressed adhesion molecules are recognized by ligands on blood leucocytes, and leucocyte–endothelium interactions involving members of the selectin-, integrin- and immunoglobulin (Ig) superfamily occur. The initial binding, the so-called ‘rolling’, which takes place between selectins, is reversible, whereas firm adhesion (‘sticking’) is mediated between integrins on leucocytes and members of the Ig-superfamily on endothelial cells. Several studies have shown an increased expression of β2-integrins, ICAM-1 and E-selectin in mucosal specimens from patients with inflammatory bowel diseases (IBD) (reviewed in [1,2]). Furthermore, Binion et al. have demonstrated recently that cultured human intestinal microvascular endothelial cells (HIMEC) from chronically inflamed tissue from patients with IBD exhibit substantially more leucocyte binding than HIMEC from uninvolved tissue [3], underscoring the potential therapeutic benefits of blocking leucocyte attachment in chronic intestinal inflammatory disorders in the gut.

DSS, a sulphated polysaccharide, reproducibly induces acute and chronic colitis in Balb/c mice [4]. Histologically the acute phase is characterized by a destruction of the colonic epithelial layer and a polymorphonuclear cell infiltrate involving the mucosa and submucosa [4–6], whereas the chronic phase is characterized by a mainly mononuclear cell infiltrate with a mainly intact epithelial layer and only a few scattered ulcerations [4]. In acute colitis, an increase in mucosal myeloperoxidase activity and significant decreases in the mucosal antioxidant levels and oxygen radical scavenging capacity [7], as well as increased concentrations of macrophage-derived cytokines such as IL- 1, IL-6, TNF and granulocyte-macrophage colony-stimulating factor (GM-CSF) are found [8,9]. Furthermore, ICAM-1 expression in acute DSS-induced colitis is up-regulated [10] and treatment with an antisense oligonucleotide to ICAM-1 or a monoclonal anti-ICAM-1 antibody reduced the severity of colitis, indicating the importance of this integrin in this animal model [10,11].

ICAM-1 gene expression is regulated by the transcription factor NF-κB, which also controls the transcription of a diverse array of genes such as adhesion molecules (ICAM-1, VCAM-1), chemokines (MCP-1) or cytokines (TNF-α, IL-1, IL-6, IL-8, IL-12) [12]. NF-κB is up-regulated in intestinal inflammation, as shown in experimental animal models and in patients with inflammatory bowel disease [13]. Recently we and others could demonstrate that by application of gliotoxin, which blocks NF-κB activation [14], intestinal inflammation is significantly ameliorated in several animal models of acute and chronic colitis [15,16].

The aim of this study was to design a new model to investigate the leucocyte–endothelium interactions in the intestinal microcirculation during acute and chronic intestinal inflammation. We therefore developed a technique to quantify the rolling and sticking of leucocytes in mucosal colonic postcapillary and collecting venules as well as mucosal leucocyte extravasation in acute and chronic DSS-induced colitis. Furthermore, we performed studies blocking ICAM-1 either by application of anti-ICAM-1 monoclonal antibody (MoAb) or by locally blocking NF-κB activation with gliotoxin.

Materials and methods

Animals

Female Balb/c mice weighing 18–20 g were used for experiments. Animals were housed in a room maintained at 22°C and kept on standard laboratory pellet food (150 mg/kg Vit E, H1003, Alma, Kempten, Germany). All experiments were performed in accordance with the German legislation on the protection of animals.

Induction of colitis

Control mice received normal drinking water. Established protocols were used for induction of acute and chronic colitis [4]. Acute colitis was induced by giving 5% DSS (mol. wt 40 000, ICN, Eschwege, Germany) orally in drinking water for 7 days followed by 1 day of normal drinking water. Chronic colitis was established after four cycles of 5% DSS dissolved in the drinking water for 7 days followed by normal drinking water for 10 days each time.

The number of animals for each experiment is indicated in the legends or in the tables of the Results section.

Microsurgical technique

After premedication with atropine (0·1 mg/kg body weight s.c.) animals were anaesthetized with a constant flow of oxygen (33%), isoflurane (0·4 vol%) and nitrous oxide. Animals were placed in a supine position on a heating pad for maintenance of the body temperature between 36°C and 37°C as measured via a rectal thermistor. The left carotid artery and jugular vein were cannulated by fine polyethylene catheters (inner diameter: 0·28 mm; Portex Lythe, Kent, UK) for continuous recording of mean arterial pressure (MAP) and heart rate, for injection of fluorescent dyes for in vivo microscopy and for substitution of volume loss (40 ml/h/kg Ringer's lactate i.v). After transverse laparotomy the descending colon was mobilized (see Fig. 1).

Fig. 1.

In vivo microscopy set-up. The mouse is placed on a heating pad and an intra-arterial catheter for continuous recording of blood pressure (MAP) and an intravenous catheter (i.v.) for injection of fluorescent dye is inserted. All images are recorded by a videocamera (VC) attached to the microscope (M) and connected to a monitor and a video recorder (VCR) for off-line analysis.

In vivo microscopy

Part of the mobilized left colon (approximately 1 cm length) was exteriorized on a special mechanical stage and covered with a cover slide. The stage was placed on a computer-controlled microscope desk, allowing for a repeated scanning of identical microvessels during the experiment. Throughout the experiment the tissue was superfused with 37°C Ringer's lactate. In vivo microscopy was performed in modification of the technique described by Gonzalez et al. [17] for small bowel isografts. We used a technical set-up as described by Harris et al. [18] including a Zeiss microscope (axiotech vario 100 HD, × 20 objective) with a 100-W HBO mercury lamp, attached to a Ploemo-Pak illuminator with I_2/3 filter block (Leitz, Wetzlar Germany) for epi-illumination. All images were recorded by a video camera (HG16 PCO, Kelheim, Germany) attached to the microscope and connected to a monitor and a S-VHS video recorder (AG7355, Panasonic, Osaka, Japan). Quantitative off-line analysis was blinded using a custom-made computer-assisted analysing system.

The microcirculation of the submucosa was visualized for determination of leucocyte endothelium interaction in 10 randomly selected postcapillary (PV) and collecting venules (CV), respectively (magnification: 600×). Leucocytes were stained in vivo with acridine orange 0·02% (Sigma Chemical, St Louis, MO, USA), dissolved in normal saline solution and injected intravenously at a concentration of 0·1 mg/kg/min, and were classified as rolling, adherent or non-adherent cells with regard to their interaction with the endothelial lining. In each vessel segment adherent leucocytes were determined as cells that did not move or detach from the endothelium within an observation period of 30 s; data are given as number of cells per mm² endothelial surface [19]. Rolling leucocytes are defined as cells moving at a velocity less than two-thirds of the central bloodstream [20] and are expressed as a percentage of all leucocytes passing the observed vessel segment within an observation period of 30 s.

For visualization of the microcirculation, fluorescein isothiocyanate (FITC)-labelled dextran (0·1 ml, 5% FITC dextran 150, Sigma Chemical, St Louis, MO, USA) was injected intravenously. In order to analyse mucosal microcirculation and lymphocyte extravasation, a longitudinal incision (approximately 20 mm) along the antimesenteric border was performed by microcautery to obtain access to the intestinal mucosa. Functional capillary density (FCD), defined as the length of red blood cell-perfused mucosal capillaries per observation area, was assessed using the computer-assisted analysing system. The amount of extravasated leucocytes in the mucosa was obtained by counting the acridin-orange labelled leucocytes, lying close to the mucosal vessels, in the same observation area as of the assessment of the FCD and calculated as leukoctes/mm² mucosal surface.

In vivo microscopy was performed between 20 and 70 min following laparotomy. At the end of the experiment animals were sacrificed by exsanguination for collection of tissue.

Experimental design

Experiment 1

In acute colitis in vivo microscopy was performed on days 0, 2, 4, 7, 14 and 21 after the start of DSS application. In chronic colitis, in vivo microscopy was performed 4 weeks after the last cycle of DSS administration.

Histology

Mice were killed on days 0, 2, 4, 7, 14 and 21 of the experiment. A histological inflammatory score was used for each animal by a blinded observer as described previously [15]. In our hands the most prominent DSS-induced intestinal inflammation is located in the distal colon of Balb/c mice. A score of 0–4 (4 being the most severe) was assigned for each of the following [1]: epithelial loss and shortening of the crypts [2], inflammatory infiltration and [3] area of extension. Each of the serially coded cross-sections of the distal third of the colon were evaluated by a blinded observer and an average score was calculated.

Electron microscopic analysis

For electron microscopic analysis the abdominal aorta was cannulated at the end of in vivo microscopy experiments. After clamping the thoracic aorta and opening of the inferior vena cava for free flow, the colon was perfused via the aorta first with 37°C Ringer's lactate, then with a modified Karnovski's solution (2% glutaraldehyde and 2% paraformaldehyde in 0·1 m cacodylate buffer, pH 7·1) at 37°C for approximately 20 minutes. Small tissue samples were immersed in this solution for further fixation at room temperature for the next 2 h. Thereafter about 1 mm³ tissue samples were routinely double-fixed in 0·1 m cacodylate buffered 1% OsO₄ for 2 h, dehydrated in graded ethanol and embedded in the EmBed-812 epoxy resin. After 48 h heat polymerization at 60°C, semithin (0·8 µm) sections were cut, stained with toluidine blue, and after selection of an appropriate area of interest the epon block was trimmed for ultrathin sectioning. The ultrathin (80 nm) sections were cut with a diamond knife on a Reichert Ultracut-S ultramicrotome and double-stained with 1% aqueous uranyl acetate and lead citrate solutions for 10 min each. The sections were examined in a Zeiss EM902 electron microscope operating at 80 kV.

Myeloperoxidase activity

Colonic myeloperoxidase (MPO) activity was determined as described previously [15]. Briefly, colonic tissue was homogenized in 1 ml of 50 mmol/l potassium phosphate buffer (pH 6·0) containing 0·5% (wt/vol) hexadecyltrimethylammonium hydroxide and centrifuged at 11 000 r.p.m. at 4°C for 20 minutes; 10 µl of the supernatant was transferred into phosphate buffer (pH 6·0) containing 0·17 mg/ml 3,3′-dimethoxybenzidine and 0·0005% H₂O₂. MPO activity of the supernatant was determined by measuring the H₂O₂-dependent oxidation of 3,3′-dimethoxybenzidine and expressed as units per gram of total protein. Total protein content of the samples was analysed using a bicinchoninic acid protein assay kit (BCAI; Sigma).

Experiment 2

Treatment with intraperitoneal administration of anti-ICAM-1 (40 µg/mouse) (Pharmingen) or an isotype control antibody was performed on day 3 after the start of DSS administration (n = 5/group). Four hours after the antibody application in vivo microscopy was performed.

Gliotoxin (Sigma, St Louis, MO, USA) was dissolved in sterile dimethylsulphoxide (DMSO) at a concentration of 25 mg/ml. Treatment with intracolonic administration of gliotoxin (50 µg/mouse) dissolved in 0·25 ml sterile PBS (final DMSO concentration 0·32% in 0·25 ml PBS) or PBS only containing the same amount of DMSO (0·32%), or saline was performed on days 1 and 2 after the start of DSS administration (n = 4/group). On day 3 in vivo microscopy was performed.

Immunohistochemistry

For immunohistochemistry experiments animals received 5% DSS for 7 days and were treated with phosphate buffered saline (PBS) or gliotoxin (40 µg/day by intraperitoneal administration) from days 4–7. ICAM-1 expression in the colonic mucosa of control, PBS- and DSS/gliotoxin-treated animals was examined by the immunoperoxidase method. Frozen colonic tissues were cut into 5-µm serial sections and fixed in acetone at 4°C for 10 minutes. Endogenous peroxidase activity was inhibited by treatment with methanol containing 0·3% H₂O₂ for 30 minutes, followed by a washing step with PBS for 5 minutes. The sections were incubated with 1% bovine serum albumin (Sigma Chemical, St Louis, MO, USA) for 30 minutes followed by a biotinylated monoclonal antibody against murine ICAM-1 (1 : 500 dilution; Pharmingen) for 2 h at room temperature. The sections were washed three times for 5 minutes and then incubated with the avidin–biotin peroxidase complex for 30 minutes at room temperature (Vectastain Elite ABC reagent; Vector Laboratories). The peroxidase reaction was developed in 0·03% 3-3′-diaminobenzidine tetrahydrochloride (Sigma) until a brown reaction product could be observed. All sections were counterstained with haematoxylin.

Statistical analysis

All data are presented as median and range. Statistical analyses were performed using the Kruskall–Wallis statistic and the Mann–Whitney U-test. Differences were considered significant at P < 0·05. Box plots represent median in solid lines, mean with a broken line.

Results

Experiment 1: leucocyte rolling, sticking and extravasation in acute and chronic DSS-induced colitis

Adhesion to the endothelium and subsequent migration into the subendothelial tissue can be seen in the mucosa of mice after DSS administration (Figs 2 and 3). As shown in Fig. 4a,b, compared to normal mice we observed a 33-fold and 12-fold increase of leucocyte sticking in the mucosal PV and submucosal CV in the colon on day 2 of DSS application, reaching a maximum at day 7 with a 51-fold and 30-fold increase, respectively (P < 0·01). This was paralleled by augmented rolling in the PV (P < 0·05) but not in the CV. In chronic DSS-induced colitis leucocyte rolling nearly doubled in PV and CV (P < 0·01) and leucocyte sticking was augmented 20-fold in the PV and 19-fold in the CV (P < 0·01; P < 0·05, respectively) (Fig. 5a,b).

Fig. 2.

Micrograph of mononuclear cells adhering (a) and penetrating (b) through endothelial cells in a colonic postcapillary venule 4 days after starting DSS application.

Fig. 3.

In vivo microscopy of mucosal microcirculation in acute DSS-induced colitis. (a) Sticking leucocytes in the submucosal CV of a control, non-inflamed colonic segment. (b) Sticking leucocytes in the submucosal CV 7 days after first DSS application. (c) Colonic mucosa with visible crypts and extravasated leucocytes 7 days after starting DSS application.

Fig. 4.

Rolling, sticking and leucocyte extravasation in acute DSS-induced colitis. Colitis was induced as described in Table 1. Leucocyte rolling (a) as well as leucocyte sticking (b) is significantly increased on day 2 of DSS-induced colitis in submucosal PV and CV compared to control animals. Stickers are defined as cells adhering to the endothelium longer than 30 s. (c) Leucocyte extravasation is significantly increased as early as day 2 after start of DSS-application and persisted until day 14. *P < 0·01, #P < 0·05 (n = 5/group or n = 5/time-point after start of DSS application). Box plots represent median in solid lines, mean with a broken line. (a and b) □, Collecting venules; ░, postcapillary venules.

Fig. 5.

Rolling, sticking and leucocyte extravasation in chronic DSS-induced colitis 4 weeks after the last application of DSS. Leucocyte rolling (a) as well as leucocyte sticking (b) is significantly increased in chronic DSS-induced colitis in submucosal PV and CV compared to control animals. (c) Leucocyte extravasation is still significantly increased 4 weeks after the last cycle of DSS in chronic colitis compared to control animals. *P < 0·01, #P < 0·05 (n = 5/group). Box plots represent median in solid lines, mean with a broken line. (a and b) □, Collecting venules; ░, postcapillary venules; (c) □, control; ░, chronic colitis.

We then assessed the extent of leucocyte extravasation into the colonic tissue. On day 2 the number of freshly extravasated cells increased 6·5-fold (P < 0·05) and remained at this high level during the entire observation period (P < 0·01) (Fig. 4c). Most interestingly, the high number of the extravasated leucocytes in the colon persisted until day 14, 7 days after the end of DSS application, and then decreased on day 21. In chronic colitis 4 weeks after the last cycle of DSS, the number of extravasating leucocytes was increased 11-fold compared to normal mice (P < 0·01) (Fig. 5c). Furthermore, due most probably to chronic inflammation, we found an increase in the functional capillary density in chronic colitis (256 [243–289] mm/mm² in chronic colitis vs. 187 [165–192] mm/mm² in normal mice, P < 0·01).

The in vivo findings paralleled histological findings and the colonic myeloperoxidase activity (MPO) in the acute DSS model. The histological score increased steadily with a delayed concomitant increase in MPO activity, which is a marker of neutrophil infiltration (Table 1) [21]. In chronic DSS colitis no significant increase of MPO could be demonstrated, probably since most of the infiltrate consists of mononuclear cells (data not shown).

Table 1.

Histology and MPO activity in acute DSS-induced colitis; 5% DSS was given in drinking water from days 1–7. The mice (n = 5/day) were killed on the days indicated and the extent of colonic inflammation was estimated histologically (max. grade 4, median and range) in a blinded fashion

	Day 0	Day 2	Day 4	Day 7	Day 14	Day 21
Histology
Median (range)	0·0	0·5	1·7	3·5	2·5*	0·5
(0·0–0·2)	(0·0–1·0)	(1·2–3·0)	(2·7–4·0)*	(2·0–3·7)*,**		(0·0–1·0)
MPO
Units/g protein	0·2	1·4	1·3	4·4	7·9	2·2
Median	(0·0–1·2)	(0·0–3·8)	(0·0–1·9)	(2·0–17·3)*	(5·4–12·8)**	(1·4–4·2)
(range)

MPO activity was assessed as described in the Methods section.

^*P < 0·05 vs. day 0,

^**P < 0·01 vs. day 0.

Experiment 2: the role of ICAM-1 in leucocyte–endothelial cell interactions and extravasation in acute DSS-induced colitis

To demonstrate the mechanisms of leucocyte adhesion and extravasation in this model we chose to perform in vivo studies blocking ICAM-1, which has been proven to be effective in DSS-induced colitis [10,11]. Using immunohistochemistry, control colon tissue demonstrated only minimal ICAM-1 immunoreactivity, whereas on days 2 and on 4 the colon had increased anti-ICAM immunoreactivity, scattered throughout the lamina propria, but associated predominantly with capillary endothelial cells (data not shown). On day 7 of DSS application ICAM-1 immunoreactivity was clearly increased, but more diffuse in areas of cellular infiltration (Fig. 6). To delineate further the implications of ICAM-1 in leucocyte trafficking into the colon, we blocked endothelial ICAM-1 expression with a monoclonal antibody against ICAM-1 at submaximal colonic inflammation (day 3). The sticking of leucocytes in the PV and CV was significantly suppressed by 78% (5-fold, P < 0·01) and 85% (7-fold, P < 0·05), respectively, compared to animals which were treated with the isotype control antibody. Furthermore, extravasation was significantly down-regulated by 70% (3-fold, P < 0·01) (Fig. 7).

Fig. 6.

Expression of ICAM-1 in colonic mucosa of control animals, animals with DSS-induced colitis treated with vehicle or gliotoxin as described in the Methods section (magnification × 40).

Fig. 7.

Leucocyte sticking and extravasation but not leucocyte rolling is significantly down-regulated after application of MoAb against ICAM-1 compared to animals which were treated with an isotype-matched antibody. Treatment with intraperitoneal administration of anti-ICAM-1 (40 µg/mouse) (Pharmingen) or an isotype-control antibody was performed on day 3 after the start of DSS administration (n = 5/group). Four hours after the antibody application in vivo microscopy was performed. *P < 0·01, #P < 0·05 (n = 5/group). Box plots represent median in solid lines, mean with a broken line. □, Collecting venules; ░, postcapillary venules; (c) isotype control antibody, ░, anti-ICAM antibody.

ICAM-1 gene expression is partially controlled by a NF-κB binding site in its upstream promotor [22]. We have demonstrated recently that by blocking NF-κB activation with the proteasome inhibitor gliotoxin, DSS-induced acute colitis could be down-regulated [15]. Furthermore, ICAM-1 immunoreactivity is clearly diminished in the colonic tissue of mice treated with gliotoxin with DSS-induced acute colitis (Fig. 6). Therefore we investigated whether, by locally applying gliotoxin, we could affect endothelial–leucocyte interaction. By giving 50 µg gliotoxin to mice during the induction of DSS-colitis on days 1 and 2, we observed a significant down-regulation of leucocyte–endothelial interaction (75% in PV, P < 0·05; 7% in CV), as well as extravasation in vivo (64%, P < 0·01) (Table 2).

Table 2.

Intrarectal gliotoxin administration significantly reduces leucocyte sticking and extravasation in submucosal PV and intramural CV compared to DSS/vehicle-treated animals. Acute colitis was induced by oral administration of 5% DSS dissolved in the drinking water given for 3 days to Balb/c mice. Treatment with intracolonic administration of gliotoxin (50 µg/mouse, days 1 and 2) or the vehicle (DMSO) was performed (n = 4/group)

	Leucocytes/mm2 venular surface
	Sticker in PV	Sticker in CV	Leucocytes/mm2 mucosa Extravasation1
Control	10 (0–16)	5 (0–8)	17 (9–22)
DSS/vehicle	196 (140–287)	34 (17–50)	221 (134–309)
DSS/gliotoxin	48 (21–65)†	31 (15–46)	81 (92–63)*
(50 µg days 1 + 2)

Data are given as median and (range).

^*P < 0·01,

^†P < 0·05 vs. DSS/vehicle.

Discussion

Using intravital microscopy it is possible to study mechanisms of inflammation in a specific organ at the microcirculatory level. The intestinal microcirculation plays a pivotal role in the regulation of the immunological homeostasis of the gut as well as the leucocyte recruitment associated with inflammation. Here we demonstrate a new model in which it is possible to quantify leucocyte–endothelial interaction and leucocyte extravasation in the mucosa and submucosa of the colon during acute and chronic intestinal inflammation. Increased leucocyte rolling and adherence in the mucosal microcirculation is seen as early as 2 days after the start of DSS administration. At the same time an augmentation of extravasated leucocytes in the mucosa and submucosa of the colon could be found in vivo. This correlated with the histological and biochemical assessments of inflammation. The slight increase of the histological score on days 2 and 4 is corroborated by Steveca et al., who found a broad variation of the severity of histological infiltration after 3 days of exposure to DSS [23]. However, the delayed increase of MPO activity on day 7 of DSS administration, which is characterized histologically by a marked infiltration by inflammatory cells, indicates a low sensitivity of this marker of neutrophil infiltration for slighter increases of the inflammatory infiltrate on days 2 and 4. In the chronic phase of DSS-induced colitis significantly enhanced adhesion and extravasation of leucocytes could still be observed 4 weeks after the last DSS administration, compared to the decrease of extravasation 3 weeks after the induction of acute colitis. Together with the significant increase in the functional capillary density, this lasting activation of leucocyte–endothelial interaction suggests the involvement of different or additional mechanisms in the adhesion and extravasation processes in chronic compared to acute DSS-induced colitis, which needs further elucidation in the future.

The pathophysiology of the intestinal microcirculation during inflammation by means of intravital microscopy has already been assessed in other animal models. Arndt et al. used the indomethacin model in the rat to demonstrate endothelial–leucocyte interactions during enterocolitis, which is normally located in the mid-jejunum. The shortcoming in this model is that the intestinal microcirculation can not be directly evaluated, since intravital microscopy was performed in the adjacent mesenteric vascular beds and not in the intestinal mucosa itself [24]. Recently, the leucocyte–endothelial cell interactions of colonic submucosal postcapillary venules were studied in the trinitrobenzene sulphonic acid (TNBS) model in the rat [25]. A significant increase of rolling and adherent leucocytes was reported on day 1 after TNBS administration, reaching maximum levels on day 7 and decreasing to normal levels on day 21. However, neither the mucosal microcirculation itself (CV and PV, functional capillary density) can be analysed, nor can the amount of extravasated leucocytes be quantified in this model.

Firm adhesion and emigration of leucocytes involves interactions with a group of cell surface glycoprotein receptors, which include ICAM-1, VCAM-1 or MAdCAM-1. Expression of soluble ICAM-1 (sICAM-1), which is considered to correlate with the level of expression of membrane bound ICAM-1, as well as mucosal ICAM-1 expression are elevated in patients with ulcerative colitis and Crohn's disease [26–32]. ICAM-1 expression is also increased in inflamed tissues in animal models of colitis [11,25,33], and the application of antisense oligonucleotides to ICAM-1 in mice [11] or monoclonal antibodies to ICAM-1 in rats and mice [10,33] attenuates DSS-induced inflammation histologically. In vitro migration of leucocytes across monolayers of endothelial cells could be inhibited significantly by an anti ICAM-1 monoclonal antibody [34], indicating an important role of this adhesion molecule in inflammatory processes. We have now shown that in vivo blocking ICAM-1 significantly reduced the sticking as well as the extravasation of the leucocytes into inflamed intestine, corroborating the above-described studies in vivo.

The expression of ICAM-1 is regulated by the transcription factor NF-κB and can be blocked by inhibiting NF-κB activation. Under steady state conditions NF-κB is present in the cytosol bound to its naturally occurring inhibitor IκB. Translocation of NF-κB into the nucleus, where it binds to κB-specific promotor/enhancer regions of genes such as ICAM-1, only happens upon stimulation by inducers such as IL-1 and TNF, followed by the phosphorylation of IκB which is then degraded in a proteasome dependent pathway [12]. In vitro, using pharmacological inhibitors of IκBα degradation, TNF-induced ICAM-1 expression on intestinal epithelial cells was strongly down-regulated [35]. To block NF-κB activation in our animal model we used the fungal metabolite gliotoxin, which is produced by Aspergillus fumigatus and other pathogenic fungi, including some candida and penicillium species [36]. Gliotoxin inhibits the activation of NF-κB in T and B cell lines in vitro without affecting other transcription factors such as NF-AT [14]. Gliotoxin down-regulates the promotor-driven reporter gene activity [14] as well as ICAM gene expression of ICAM-1 in vitro (H. Herfarth, unpublished results). We were able to demonstrate recently that gliotoxin down-regulates intestinal NF-κB activation and significantly ameliorates DSS-induced histological damage [15]. Intrarectal administration of gliotoxin down-regulated significantly the leucocyte sticking in the PV as well as leucocyte extravasation comparable to the blocking of ICAM-1 by a MoAb. However, further studies are needed to evaluate whether these effects of NF-κB inhibition are due solely to ICAM-1 down-regulation or if other integrins such as MAdCAM-1 are equally affected.

Taken together, our data demonstrate a new possibility for assessing and quantifying leucocyte–endothelium interactions in the colonic mucosa itself during experimental mouse colitis in vivo. Using this model it is possible to investigate further adhesion molecules which may be involved in intestinal inflammation in vivo, and to analyse novel therapeutic approaches which aim to block cell–endothelium adhesive interactions in the inflamed intestine.