Endotoxin (bacterial LPS) causes an intense systemic inflammatory reaction with symptoms resembling acute sepsis, including fever, tachycardia, hypotension, coagulation disorders, and lethality. Monocytes and macrophages react on minute amounts of LPS with the massive release of inflammatory mediators such as TNF-α and IL-1β. In typical shock organs such as the lung, liver, and the mesentery, these macrophage products cause a local capillary leak due to endothelial activation, resulting in increased leukocyte infiltration. Thus, endothelium-leukocyte interaction constitutes the initial event in endotoxin-mediated organ failure.
Tolerance to endotoxin can be induced by single or repeated administration of LPS and is characterized by a profound reduction of typical pathophysiological effects of subsequent high-dose endotoxin challenge such as attenuation of lethality. Two types of endotoxin tolerance have been demonstrated: a late type of endotoxin tolerance that develops after weeks (late endotoxin tolerance), depending on the existence of specific O-antigenic antibodies, and an early type of endotoxin tolerance (early phase tolerance) (1). The mechanisms underlying the development of the early phase tolerance are not fully understood, although it was demonstrated that this type of endotoxin tolerance was associated with a reduced inflammatory mediator response, including decreased TNF-α and IL-1β response in endotoxin-tolerant animals.
Because endotoxin tolerance can be induced in nu/nu mice lacking functional T cells, its development has been attributed to innate immune cells such as macrophages and the reprogramming of their actual gene activities by the pretreatment with endotoxin. Although it has been suggested that ET was specific for endotoxins as a substance class, it turned out that at least at higher doses for tolerance induction, it extents to other TNF-α-mediated syndromes either induced by bacterial compounds or for example, I/R syndromes in the kidney (2), heart (3), or liver (4).
Both the LPS-induced pathologic alterations and I/R injury have been shown to be critically mediated by endogenous mediators, in particular, TNF-α (5). As a direct consequence of increased TNF-α production, endothelial cells in the microcirculation get activated, and massive leukocyte adhesion and tissue infiltration occurs. Because the inflammatory mediator response is markedly attenuated by early-phase endotoxin tolerance, we hypothesized that its induction might protect from LPS toxicity by preventing extravasation of leukocytes. To study this issue, the LPS-mediated interaction of leukocytes and endothelium in the mesenteric and hepatic microcirculation in rats was analyzed with and without prior establishment of endotoxin tolerance.
MATERIALS AND METHODS
All experimental procedures were performed under the German legislation on protection of animals. The local animal protection committee of the government reviewed the experimental protocol. After a 2-week period of acclimatization to the laboratory (housing in stainless-steel cages in a temperature- and humidity-controlled room with free access to tap water and standard laboratory animal chow), male CD Rats (350-400 g; Charles River, Sulzbach, Germany) were anesthetized by i.m. injection of 10% ketamine and 2% xylazine (1 mL kg−1 body weight; Sigma Chemical, Deisenhofen, Germany) every 45 min and underwent preparatory surgery. The right internal jugular vein was cannulated for the application of fluorescein isothiocyanate (FITC)-right sacroanterior position (RSA), rhodamine G6, LPS, and volume replacement, with a rate of 0.3 mL h−1. The arterial blood pressure and heart rate (HR) were continuously monitored after connecting to a catheter in the right carotid artery by a pressure transducer (Combitrans monitoring set; Braun, Melsungen, Germany).
Rat mesenteric and hepatic intravital microscopy
For video intravital microscopy (IVM), a left lateral laparotomy and exteriorization of a small segment of ileum mesentery were performed as previously described (6). After exposition, the mesentery was directly coated with 37°C warmed paraffin liquid (Merck, Darmstadt, Germany).
After transverse laparotomy, the left liver lobe was placed carefully on a specially designed concave-shaped aluminum stick. To exhibit the microcirculation, a cover glass fixed to an aluminum stick was replaced on the left liver lobe without pressure.
During the preparatory procedure, the animals were placed on a heating pad for maintenance of body temperature. After preparation, rats were replaced on an adjustable microscope stage. For assessment of body temperature, a temperature probe was placed rectal. After completion of the preparation, the animal was allowed to equilibrate for 45 min. Blood samples were drawn from the arterial catheter for determination of soluble intercellular adhesion molecule (ICAM), l-selectin, and stimulation of whole blood.
To visualize leukocyte-endothelial cell interaction in sinusoids and postsinusoidal venules of the liver and postcapillary venules in mesentery, rhodamine 6G (0.3 mg kg−1; 0.05%; mol wt, 479; Sigma) was administered intravenously, and an I2.2 filter block (excitation, 530 - 560 nm; emission, >580 nm; Leica, Wetzlar, Germany) was used. The visualization of the microvascular leakage in the microcirculation of the gut and the contrast of plasma was enhanced by i.v. application of FITC-labeled albumin (FITC-RSA, 0.6 mL kg−1; 7.5%; Sigma). For optimal visualization, an I3 filter block was used (Leica). Intravital microscopy was performed using a camera microscope (Leica DMLM) with long-distance objective HCPL Fluotar 10×/0.30, 20×/0.40 BD (Leica). It was equipped for transillumination and epi-illumination. We used 10× oculars and a 5× objective for fluorescence microscopy and 0.60/40× long working distance objectives for leukocyte counting. The fluorescence microscopy was provided by a mercury lamp (HBO, 50 W; Osram, Munich, Germany). The video images were transferred to a monitor (MT-H 1480; Panasonic, Osaka, Japan) by a video camera (Kappa CF 8/4 NIR) and a digital recorder (DSR 20 P, both Sony Corp., Tokyo, Japan) and analyzed off-line.
Four groups (n = 6 per group) were used for the measurement of the microcirculation of mesentery, and four groups (n = 6 per group) were used for the visualization of the microcirculation of the liver. For the induction of endotoxin tolerance, rats (groups tolerant/LPS (TOL/LPS) and tolerant/sodium (TOL) were injected intraperitoneally with 0.5 mg kg−1 LPS at day 1 and 1 mg kg−1 at days 2 and 3 (Escherichia coli, serotype 055:B5; Sigma). Normal control rats were injected intraperitoneally with the same volume of pyrogen-free 0.9% sodium chloride (sodium/LPS [N/LPS], sodium/sodium [N]; for definition of groups, see Table 1). On the fifth day, rats were injected intravenously with 5 mg kg−1 of LPS (Salmonella abortus equi; Sigma) to induce endotoxemia or a comparable volume of sodium chloride. The IVM started together with the application of LPS and was referred to as t = 0. All animals received 1 mL−1 kg−1 h−1 physiological saline after cannulation of the jugular vein until the end of the experiment.
Measurement of leukocyte-endothelial interaction
One-minute recordings of four to six single unbranched postcapillary venules with a diameter of 20 to 40 μm and 70 μm in length were taken from each mesentery every hour over a total time of 3 h. Micro-vessel diameter and length were measured off-line using a digital image processing system. For assessment of leukocyte-endothelial interaction in postsinusoidal venules, four nonoverlapping randomly chosen acini of the left liver lobe were scanned over a period of 60 s every 1 h over a total time of 3 h.
A leukocyte was considered to be adherent (sticker) when it was stationary for more than 30 s. When it was stationary for less than 30 s but in contact to the endothelium, it was defined as roller. A new leukocyte appearing in tissue was defined as leukocyte after diapedesis. The number of leukocytes was counted off-line using frame-by-frame analysis and expressed as mean values per square millimeters of inner surface, assuming a cylindrical cross section of the vessel. The number of perfused sinusoids divided by the total number of sinusoids observed was defined as sinusoidal perfusion and given as percentage.
Measurement of permeability
One-minute recordings of a single area of mesenteric vasculature were performed every hour throughout the observation period of 3 h. The recorded digital video was evaluated off-line using the system Cap Image (Dr. Zeintl, Heidelberg, Germany). It consists of an IBM-compatible PC with an IP 8/AT image-processing card (Matrox; Dorval, Quebec, Canada) with real-time videotape digitalization. The video recorder was connected through an RS-232 interface. Video images produced by fluorescence microscopy were evaluated with regard to the light intensity of a selected area of the video picture. The increase in light intensity after application of FITC-RSA can be assumed to be an increase in number of gray values, presupposing a linear relationship between fluorescent dye concentration and camera output value as gray value. Because of different total light intensities of gray values at the start of each experiment, gray value was assumed to be 100%. Extravasation of FITC-RSA was determined as the difference in percent of the light intensity over time and expressed as integrated optical density.
Measurement of TNF-α, soluble ICAM, and soluble E-selectin
The blood samples for preparation of the whole blood culture were collected 10 min prior each experiment. Whole blood (250 μL) was diluted 1:1 with 1 mL RPMI 1640 medium containing 100 U mL−1 penicillin, 100 μg mL−1 streptomycin, and glutamine, and incubated in 24-well flat bottom tissue culture plates at 37°C in an atmosphere of 5% CO2 in air. Blood cultures were set up as duplicates and incubated with 500 ng mL−1 LPS (from Salmonella friedenau, kindly provided by H. Brade, phenol extracted, free of Toll-like receptor-Ligands, Borstel, Germany) for 18 h before supernatants were collected after centrifugation for TNF-α detection by means of enzyme-linked immunoabsorbent assay (ELISA) (Hölzel Diagnostica, Cologne, Germany). The lower detection limit was 15 pg mL−1 of recombinant rat TNF-α.
Concentration of soluble ICAM (s-ICAM) in serum was determined with an ELISA test kit (Hölzel Diagnostica). The lower detection limit was 15 pg mL−1 of recombinant s-ICAM (rat). Concentration of soluble leukocyte L-selectin in serum was determined with the rat L-selectin ELISA (Duo set; R&D Systems, Minneapolis, Minn). The lower detection limit was 15 pg mL−1.
Immunohistochemistry for E-selectin
Elliptical biopsies were taken from four rats of the right liver lobe of the previously described animal groups. For immunohistochemical analysis, excised tissue was oriented for cross-sectional analysis and snap-frozen in Tissue-Tek cryomolds (Miles Corp, Elkhart, Ind) and stored at −80°C. Frozen tissue was sectioned (5 μm), collected on glass slides, fixed in acetone (5 min), and stored at −20°C for immunohistochemical analysis.
Immunohistochemical staining of liver tissue using chromogen was performed as follows. All steps were performed at 4°C using an antirat cell and tissue staining kit (HRP-AEC; R&D). Serial sections were warmed 5 min at room temperature, re-hydrated in phosphate-buffered saline, blocked with 3% hydrogen peroxide, rat serum, avidin, and biotin, then incubated overnight with 10 μg mL−1 primary antibody goat antirat E-selectin (R&D). Treated slides were washed and incubated with 5 μg mL−1 antigoat biotinylated immunoglobulin (Ig)G as secondary antibody (R&D) for 1 h. Washed slides were then incubated for 30 min with high-sensitivity streptavidin-horseradish peroxidase conjugate (R&D), washed, and incubated with the substrate 3-amino-9-ethylcarbazole chromogen (R&D) for 10 min. Mayer hematoxylin (Fisher Diagnostics, Fair Lawn, NJ) was used to counterstain. Slides were preserved by using an Aqueous mounting medium solution (R&D). We incubate control sample with nonimmune Ig of the same isotype for isotype control, with serum for null control, and with antibodies that were preincubated with the corresponding immunogen for absorption control.
Data are expressed as mean ± SD. Values of one group are compared by Wilcoxon test, whereas differences between different groups are evaluated by Mann-Whitney U test. P values less than 0.05 were considered as significant. In the case of multiple testing, the Bonferroni correction was used. Data analysis and statistics were performed using the Statistical Package for Social Sciences (SPSS) version 12.0 for Windows (SPSS Inc., Chicago, Ill). Statistical procedure is indicated in the figure legend.
Induction of endotoxin tolerance
Four groups of animals were studied: N (NaCl on days -4, -3, and -2, and NaCl on day 0), N/LPS (NaCl on days -4, -3, and -2, and LPS on day 0), TOL (LPS on days -4, -3, and -2,and NaCl on day 0), and TOL/LPS (LPS on days -4, -3, and -2, and LPS on day 0) (Table 1). In all groups, MAP, HR, and rectal temperature (T) were similar at the beginning of experiments (Table 2). There was a significant increase in T and HR at 2 h and in T, HR, and MAP at 3 h after LPS administration (N/LPS versus N). The endotoxin-mediated rise in T and HR and the changes in MAP were completely prevented in LPS-tolerant animals (Table 2).
Because it has been shown that the development of ET is associated with a strongly reduced cytokine response, it was tested whether whole blood isolated of the diverse groups differed in their response to LPS ex vivo with regard to TNF-α synthesis. Intraperitoneal pretreatment with 0.5 mg LPS (E.coli) on day 1 and 1.0 mg kg−1 bodyweight on days 2 and 3 led to a significant reduction of TNF-α synthesis after in vitro LPS incubation (0.5 μg mL−1; LPS, S. friedenau for 4 h) (Fig. 1).
Systemic LPS challenge causes a general disturbance of the hepatic microcirculation. As a result of this effect, the total number of perfused sinusoids decreased significantly in LPS treated in comparison to control animals. The number of perfused sinusoids in ET animals, however, after an additional LPS challenge was on the level of control animals. The systemic administration of LPS in ET animals showed a slight, although not statistically significant increase of sinusoids without perfusion (Fig. 2).
The leukocyte-endothelial interaction was analyzed in two distinct vessel areas, the sinusoids itself and the postsinusoidal venules. In the sinusoids, LPS challenge of normal rats caused a significant increase of rolling and of sticking leukocytes, reaching maximal levels 180 min after LPS application. In contrast, in endotoxin tolerant rats, the increased adherence of leukocytes in the sinusoids was significantly lower at all time points after endotoxin challenge (Fig. 3, A and B). The examination of postsinusoidal venules revealed opposite effects. LPS challenge in nontolerant rats caused an increase of rollers and stickers in the postsinusoidal venules. In ET rats, even higher numbers of rollers and stickers were determined (Fig. 3, C and D).
To determine if the peculiar adhesive behavior of leukocytes in the liver was specific for this organ, mesenteric microcirculation was investigated after LPS challenge and compared between tolerant and normal rats. Leukocyte rolling and sticking were significantly increased at time points 1, 2, and 3 h in the N/LPS group in comparison to the beginning of the experiment. Interestingly, in tolerant animals, leukocyte rolling in postcapillary venules was already enhanced without LPS stimulation and further increased over the experiment at 2 and 3 h. After LPS stimulation, the endotoxin-tolerant group revealed no further increase of rolling leukocytes in the mesentery (Fig. 4A). The pattern of leukocyte adherence (stickers) paralleled the effects observed in the postsinusoidal venules. Examination of the ET group with LPS challenge revealed an increase of sticking leukocytes that was significantly higher than the increase in nontolerant rats (Fig. 4B). When leukocyte diapedesis was investigated, it was found that after LPS challenge, an increasing number of leukocytes were present in the interstitium as a consequence of enhanced sticking and rolling (Fig. 4C). In endotoxin-tolerant animals, the number of transmigrated leukocytes in the interstitium was higher than in the nontolerant group. Therefore, increased adhesion of leukocytes in the postcapillary venules of the mesentery is followed by transmigration of the adhering cells (Fig. 4D). The number of perfused postcapillary venules in the mesenterial microcirculation is very similar to the amount of perfused sinusoids in the liver. We also observed a decrease in perfused postcapillary venules of the mesentery after LPS shock in nontolerant animals. This endotoxin effect was prevented by ET induction (Fig. 4E). Endotoxin tolerance prevented the reduction of functional venular and capillary density after LPS challenge significantly, in contrast to animals without LPS pretreatment and LPS application on day 0.
To evaluate the function of the endothelial barrier, which can be affected by direct interaction with endotoxin or indirectly by oxidative burst activity of adhering leukocytes, we evaluated the integrity of the capillary vessel by means of FITC-labeled albumin leakage detection. The number of gray levels did not differ between the groups at baseline. In control animals and in both groups with LPS pretreatment (TOL, TOL/LPS), no significant increase or decrease in macromolecular leakage was observed. In contrast, in animals without LPS-pretreatment and with LPS treatment at the beginning of IVM, the number of gray levels increased from 100% at time point 0 h to 304% at 1 h, 388% at 2 h, and 384% at 3 h (Fig. 4D). These findings suggest that ET completely prevented the LPS-induced endothelial damage.
Level of s-ICAM
Leukocyte adherence is commonly regulated by expression of ICAM and E-selectin on the endothelial cell and L-selectin as counterpart on leukocytes. It is known that LPS causes an up-regulation of ICAM on endothelial cells (29). Because ICAM and L-selectin can also be shedded from the surface and be detected in soluble forms in the serum, we analyzed serum samples for soluble ICAM and L-selectin. The concentrations of s-ICAM at the beginning of the IVM already showed significantly higher levels in both ET groups compared with the N and N/LPS group, indicating a long-lasting activation of the endothelium during tolerance induction (Fig. 5A). After 1.5 and 3 h, there was no further increase of s-ICAM levels in the TOL/LPS group. LPS-treated animals without induction of ET showed an increase of s-ICAM at 3 h. Probably due to surgical stress, s-ICAM increased after 3 h without additional LPS stimulation, so that there were no significant differences between the four groups at 3 h. In contrast, L-selectin increased significantly in both tolerant and nontolerant rats 3 h after LPS injection. The increase in circulating L-selectin was slightly (but significantly; P = 0.048, Wilcoxon rank-sum test) higher in tolerant rats (Fig. 5B). This fact may give evidence for increased adhesion molecule expression on leukocytes of tolerant animals.
E-selectin expression in the liver
To further evaluate the expression of endothelial adhesion molecules, we performed immunostaining for E-selectin in sections of the liver before and 3 h after endotoxin injection. In liver sections before LPS injection, only minimal staining for E-selectin was observed in the parenchyma and around central veins. The background staining around the central veins was a little bit more intense in liver sections from ET rat. After LPS injection, we found an obvious increase of E-selectin staining in both ET and control animals. Interestingly, the distribution of E-selectin staining differed between tolerant and nontolerant rats. Although E-selectin staining in nontolerant rats was homogeneously distributed over the parenchyma of the liver, the staining in tolerant rats was predominantly found around the central veins. This distribution of E-selectin expression parallels the differential behavior of leukocyte adhesion (Fig. 6, A-D).
Tolerance to endotoxin is a well-established phenomenon that entails altered host reactions, including mitigated lethality, immune responses, and metabolic changes after repeated LPS administration. Tolerance to endotoxin prevents tissue and organ injury due to endotoxin shock, I/R injury, and thermal damages (7). The missing inflammatory response under these conditions has been attributed to reduced cytokine secretion. In addition, ET also induced nonspecific protection against infections with a number of different bacterial and fungal organisms (8). Although all of the syndromes that are attenuated by ET include endothelium leukocyte interactions as a pathogenic component, it has not been studied so far how ET act upon these elements of microcirculation. We therefore investigated the microcirculatory alterations of the liver and mesentery after induction of ET.
Our data demonstrate that ET has distinct effects on microcirculatory reactions. First, whereas ET decreased LPS-induced rolling and firm adhesion in the sinusoidal vasculature, its induction enhanced adhesion of leukocytes to postsinusoidal venules of the liver and postcapillary venules of mesentery. The reduction of leukocyte adhesion in the sinusoids was accompanied by improvement of the sinusoidal perfusion. In addition, macromolecular efflux in mesentery and the reduction of perfused vessels were prevented by ET despite the increased leukocyte diapedesis in the mesentery. It is reasonable to assume that the differential pattern of leukocyte adhesion in the sinusoids in comparison to the mesenteric postcapillary venules and the postsinusoidal venules of the liver is the result of differential expression of adhesion molecules in sinusoidal or venular endothelium.
Adhesion and diapedesis in most organs takes place in postcapillary venules (9). Chosay et al. (10) investigated in a model of endotoxin-induced liver damage both types of vessels. They found that despite adherence of neutrophils in sinusoids and venules, only the adherence in sinusoids is responsible for extravasation and cellular injury. Comparable results were found in a model of hepatic I/R injury demonstrating the sinusoidal adherence of neutrophils as relevant for the observed tissue injury (11). Our IVM study clearly demonstrated a significantly reduced adhesion of leukocytes to sinusoidal endothelium despite an increase in leukocyte adherence in postsinusoidal venules in the liver. At least the first aspect can be the explanation for liver protection that was found in various studies after induction of ET (12, 13). There is only little evidence for mechanisms of leukocyte recruitment in sinusoidal microvasculature after ET. The adherence of leukocytes to the vascular wall depends on several mechanisms. The first step of leukocyte-endothelial interaction is assumed to be dependent on leukocyte rolling (14). Leukocyte firm adhesion and diapedesis is mediated by molecules such as P-, L- and E-selectins, the integrins (CD11/CD18), and members of the Ig superfamily (ICAM-1/ICAM-2).Expression of adhesion molecules is differentially regulated through a wide number of stimuli such as endotoxin, cytokines, and radicals (15).
However, beside the activation of leukocytes or the endothelial cells, a reduction of the sinusoidal perfusion is also an important feature of the induction of hepatic injury in inflammation. Some authors are convinced of the fact that sinusoidal adhesion is not generated at all through adhesion molecules (16).
In our experiments, the reduced leukocyte adhesion in the sinusoids after ET induction was accompanied by an improved perfusion of the sinusoids. Therefore, local hemodynamic mechanisms can also explain the changes in leukocyte adhesion in the sinusoids.
However, others state the expression of ICAM-1 as crucial for leukocyte adherence in the sinusoids. Antibodies against ICAM-1 improved sinusoidal perfusion by reduction of sinusoidal leukocyte adhesion in a hepatic I/R injury (17). Uhrig et al. (18) investigated the microcirculation of the liver after induction of ET and challenge after 24 h. The authors demonstrated that the induction of ET in rats resulted in a reduced leukocyte adhesion and improved microcirculation in the sinusoidal microvasculature after LPS challenge as a consequence of decreased CD 54 (ICAM-1) surface expression. The authors concluded that a pretreatment with LPS leads to a reduced up-regulation of ICAM-1 after LPS injection.
In contrast, in an in vitro system of ET induction in human endothelial cells, a down-regulation of ICAM-1 could not be shown (19). The systemic levels of s-ICAM were significantly higher in patients who developed organ dysfunction than patients with uneventful surgery (20, 21). Miwa et al. (22) previously found an uptake of s-ICAM in cardiac tissue, and this uptake was associated with decreased cardiac function. It was hypothesized that this was due to entrapment of s-ICAM in the ischemic phase of coronary spasm as a result of the binding of s-ICAM to leukocytes, with subsequent activation and adhesion of the leukocytes to the endothelium. In our own results, we found an increase in s-ICAM in the circulation. There is strong evidence that there is a positive correlation of s-ICAM levels in serum and cell-bound ICAM after different stimuli (23, 24). Soluble ICAM suppresses cell-mediated cytotoxicity of lymphocytes by binding to lymphocyte function associated antigen 1 and acts as an immunosuppressive factor by inhibiting the ICAM-1/lymphocyte function associated antigen 1 system (25). Interestingly, Maruo et al. pointed out that s-ICAM is involved as a local immunosuppressive factor and adhesion molecule in metastasis. In line with this, we found an initial up-regulation of s-ICAM, an increase in cell adhesion and diapedesis, and a decrease in cell damage. We therefore assume that s-ICAM can be involved in this phenomenon after induction of ET. However, 3 h after LPS administration, no differences in s-ICAM levels between tolerant and nontolerant rats were observed, which argues against a differential regulation of ICAM during ET. In the same manner, the soluble form of L-selectin was found to be increased after LPS injection in tolerant and nontolerant rats in a comparable degree. Selectins are considered to mediate leukocyte rolling, which is enhanced in tolerant rats in the mesentery and the postsinusoidal venules of the liver. Because serum levels of soluble L-selectin were similar in tolerant and nontolerant rats, the data do not support a protective effect of ET on the basis of L-selectin regulation. E-selectin expression was also evaluated by immunostaining. We found an increase in E-selectin immunostaining after LPS challenge in both tolerant and nontolerant rats. However, the distribution of this expression differed between tolerant and nontolerant animals. Although the expression in naive rats was predominantly found in the liver parenchyma assumably representing expression sinusoidal microvasculature, in tolerant rats, the expressions was most obvious around the central veins. The expression of E-selectin seems to parallel the differential effects of leukocyte adhesion in these different vessel areas. Therefore, regulation differential regulation of E-selectin during ET can serve as an explanation for the effects on leukocyte adhesion observed after ET induction. However, ET does not necessarily have to influence different vasculature endothelial cells in E-selectin expression. Cell-cell interaction can also be responsible for the observed differences. There is strong evidence that Kupffer cells and hepatocytes affect the regulation of adhesion molecules in sinusoidal endothelial cells. Edwards et al. (26) demonstrated that the recruitment of lymphocytes depends on the presence of hepatocytes, and that these effects are mediated through different adhesion molecules, including E-selectin. In addition, depletion of Kupffer cells leads to a reduced diapedesis of polymorphonuclear neutrophils and expression of P-selectin after administration of LPS (27). Therefore, hepatocytes may affect leukocyte migration, and this interaction can be an explanation for the observed differences concerning differential patterns of leukocyte adhesion in venules of the mesentery and the liver, in contrast to sinusoids.
Astiz et al. (28) studied the therapeutic value of administration of a detoxified LPS derivate (monophosphoryl lipid A [MLP]) to mice before induction of a peritonitis by cecal ligation and puncture. Intraperitoneal injection of MLP was more effective in reducing mortality rate than i.v. MLP administration. Lehner et al. (7) demonstrated that i.p. accumulation of leukocytes and enhanced inactivation of intraperitoneally injected Salmonella typhimurium were dependent upon the route of LPS pretreatment. The author suggested that the accumulation of a large number of granulocytes at the site of previous LPS treatment was responsible for the enhanced host defense. However, it is still not documented if these leukocytes, mainly consisting of neutrophilic granulocytes, originated strictly from the i.p. compartment or rather transmigrated from the mesentery after challenge with LPS. An important finding in our own results was the demonstration of a significantly increased binding and diapedesis of leukocytes in postcapillary venules of the mesentery after secondary LPS stimulation. This can be an explanation for the enhanced host defense observed after ET induction.
Nothing is known about rolling and sticking after induction of ET in the microcirculation of the mesentery in vivo. However, several studies were performed in isolated endothelial cells after repetitive treatment with LPS in vitro. Lush et al. demonstrated in an in vitro investigation with LPS-pretreated human umbilical vein endothelial cell that a second LPS challenge resulted in a reduction of nuclear factor-κB and E-selectin expression, but unchanged ICAM-1 expression. They observed a reduced adhesion of polymorphonuclear neutrophils on human umbilical vein endothelial cell after induction of tolerance in vitro. They concluded that this is the reason for the reduced leukocyte infiltration in different organs after ET induction (29). In a study with human intestinal microvascular endothelial cells (HIMECs), Ogawa et al. (30) demonstrated that HIMEC rendered endotoxin-tolerant in vitro showed a reduced leukocyte adhesion and decreased superoxide production, whereas ICAM-1 surface expression remained unchanged. Only blocking monoclonal antibodies against E-selectin and vascular cell adhesion molecule 1 reduced leukocyte adhesion to HIMEC (30).
These in vitro findings are somehow in contrast to our findings to our observation of an increase of rolling, adhesion, and diapedesis of leukocytes in the mesentery after ET induction. However, induction of ET induction in vivo and repetitive LPS treatment of isolated cells in vitro, also referred to as in vitro tolerance, probably have completely different underlying mechanisms so that comparisons have to be evaluated cautiously.
Interestingly, we observed a significant reduction of macromolecular leakage in the mesentery despite the enhanced leukocyte adhesion after induction of ET. In line with this observation, Walther et al. (31) demonstrated that a blocking of leukocyte adherence to the endothelium by fucoidin, an L- and P-selectin inhibitor, was not accompanied by a reduction of macromolecular extravasation, indicating that leukocyte-endothelial interaction does not play a pivotal role in the development of capillary leakage and endothelial damage in mesentery after LPS administration (32). In our own results, the adhesion and diapedesis of leukocytes after LPS challenge increased significantly despite a decrease in macromolecular extravasation of FITC-labeled albumin.
In the case of endotoxin shock or sepsis, a clinical relevance of ET can be barely imagined. However, leukocyte adhesion is also a central pathophysiological event in all I/R injuries. A better understanding of the regulation of leukocyte adhesion during ET can open prophylactic approaches with induction of ET, for example, in the field of solid organ transplantation.
In summary, ET is associated with an inhomogeneous pattern of leukocyte adherence in liver and intestine. A reduced leukocyte adherence was noticed in liver tissue, and this may explain the observed organ protection after endotoxin shock and I/R injury in tolerant animals. In contrast, the elevated leukocyte number in the mesentery of endotoxin-tolerant animals may be responsible for the better resistance to bacterial infection. These phenomenological findings needs further investigations to characterized the microcirculatory alterations after induction of endotoxin tolerance.
1. Greisman SE, Hornick RB: The nature of endotoxin tolerance
. Trans Am Clin Climatol Assoc
2. Godet C, Goujon JM, Petit I, Lecron JC, Hauet T, Mauco G, Carretier M, Robert R: Endotoxin tolerance
enhances interleukin-10 renal expression and decrease ischemia-reperfusion renal injury in rats. Shock
3. Schulz R, Post H, Sakka S, Wallbridge DR, Heusch G: Intraischemic preconditioning. Increased tolerance to sustained low-flow ischemia by a brief episode of no-flow ischemia without intermittent reperfusion. Circ Res
4. Dominguez FE, Siemers F, Flohe S, Nau M, Schade FU: Effects of endotoxin tolerance
on liver function after hepatic ischemia/reperfusion injury in the rat. Crit Care Med
5. West MA, Heagy W: Endotoxin tolerance
: a review. Crit Care Med
6. Leithauser B, Schumacher J, Lendemans S, Tillmanns H, Matthias FR: Antithrombin attenuates microvascular leakage and leukocyte-endothelial interaction in response to endotoxin. Semin Thromb Hemost
7. Lehner MD, Ittner J, Bundschuh DS, van Rooijen N, Wendel A, Hartung T: Improved innate immunity of endotoxin-tolerant mice increases resistance to Salmonella enterica
serovar typhimurium infection despite attenuated cytokine response. Infect Immun
8. Lehner M, Rayhane N, Fitting C, Lortholary O, Dromer F, Cavaillon JM: Administration of endotoxin associated with lipopolysaccharide tolerance mice against fungal infection. Infect Immun
9. Granger DN, Kubes P: The microcirculation
and inflammation: modulation of leukocyte-endothelial cell adhesion. J Leuk Biol
10. Chosay JG, Essani NA, Dunn CJ, Jaeschke H: Neutrophil margination and extravasation in sinusoids and venules of the liver during endotoxin-induced injury. Am J Physiol
11. Jaeschke H, Farhood AI, Smith CW: Neutrophils contribute to ischemia/reperfusion injury in rat liver in vivo. FASB J
12. Margenthaler JA, Landeros K, Kataoka M, Eilers M, Ku G, Flye W: Effects of endotoxin tolerance
on propionibacterium acnes-primed lipopolysaccharide hepatic injury. J Surg Res
13. Dominguez Fernandez E, Flohé S, Siemers F, Nau M, Schade FU: Endotoxin tolerance
in rats: influence on LPS-induced changes in excretory liver function. Inflamm Res
14. Lindbom L, Xie X, Raud J: Chemoattractant-induced firm adhesion of leukocytes
to vascular endothelium in vivo is critically dependent on initial leukocyte rolling. Acta Physiol Scand
15. Jaschke H, Smith CW: Cell adhesion and migration. III. Leukocyte adhesion and transmigration in the liver vasculature. Am J Physiol Gastrointest Liver Physiol
16. Menger MD, Richter S, Yamauchi J, Vollmar B: Role of microcirculation
in hepatic ischemia/reperfusion injury. Hepatogastroenterology
17. Vollmar B, Glasz J, Menger MD, Messmer K: Leukocytes
contribute to hepatic ischemia/reperfusion via intercellular adhesion molecule-1-mediated venular adherence
18. Uhrig A, Banafsche R, Kremer M, Hegenbarth S, Hamann A, Neurath M, Gerken G, Limmer A, Knolle PA: Development and functional consequences of LPS tolerance in sinusoidal endothelial cells of the liver. J Leuk Biol
19. Lush CW, Cepinskas G, Kvietys PR: LPS tolerance in human endothelial cells: reduced PMN adhesion, E-selectin expression, and NF-kappaB mobilization. Am J Physiol Heart Circ Physiol
20. Law MM, Cryer HG, Abraham G: Elevated levels of s-ICAM correlate with the development of multiple organ failure in severely injured trauma patients. J Trauma
21. van Deventer SJH, Pajkrt D: Neutrophil adhesion molecules and MOF. Intensive Care Med
22. Miwa K, Igawa A, Inoue H: Soluble E-selectin, ICAM-1 and VCAM-1 levels n systemic and coronary circulation in patients with variant angina. Cardiovasc Res
23. Marguet C, Dean TP, Warner JO: Soluble intercellular adhesion molecule-1 (sICAM-1) and interferon-gamma in bronchoalveolar lavage fluid from children with airway disease. Am J Respir Crit Care Med
24. Urbaschek R, McCuskey RS, Rudi V, Becker KP, Stickel F, Urbaschek B, Seitz HK. Endotoxin, endotoxin-neutralizing-capacity, sCD14, sICAM-1, and cytokines in patients with various degrees of alcoholic liver disease. Alcohol Clin Exp Res
25. Koyama S: Immunosuppressive effect of shedding intercellular adhesion molecule 1 antigen on cell-mediated cytotoxicity against tumor cells. JPN J Cancer Res
26. Edwards S, Lalor PF, Nash GB, Rainger GE, Adams DH: Lymphocyte traffic through sinusoidal endothelial cells is regulated by hepatocytes. Hepatology
27. Shi J, Kokubo Y, Wake K: Expression of P-selectin on hepatic endothelial and platelets promoting neutrophil removal by liver macrophages. Blood
28. Astiz ME, Saha DC, Carpati CM, Rackow EC: Induction of endotoxin tolerance
with monophosphoryl lipid A in peritonitis: importance of localized therapy. J Lab Clin Med
29. Lush CW, Gediminas C, Kvietys R: LPS tolerance in human endothelial cells: reduced PMN adhesion, E-selectin expression, and NF-kB mobilization. Am J Physiol Heart Circ Physiol
30. Ogawa H, Rafiee P, Heidemann J, Fisher PJ, Johnson NA, Otterson MF, Kalyanaraman B, Pritchard KA, Binion DG: Mechanisms of endotoxin tolerance
in human intestinal microvascular endothelial cells. J Immunol
31. Walther A, Weihrauch M, Schmidt W, Gebhard MM, Martin E, Schmidt H: Leukocyte-independent plasma extravasation during endotoxaemia. Crit Care Med
32. Croner RS, Lehmann TG, Fallsehr C, Herfarth C, Klar E, Kirschfink M: C1-inhibitor reduces hepatic leukocyte-endothelial interaction and the expression of VCAM-1 in LPS-induced sepsis in the rat. Microvasc Res