INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

SEVERAL MECHANISMS of tissue injury have been proposed, the relevance and interplay of which under in vivo conditions are incompletely understood. The injury mechanisms in many pathological conditions, such as hemorrhagic shock (4, 19), diabetes (18), and ischemia-reperfusion injury (11, 14), have been proposed to be associated with activated leukocytes. Ischemia can result from myocardial infarction, organ transplantation, cerebral stroke, and local trauma. Impaired reflow after ischemia may be caused by leukocyte capillary plugging (7), swelling of endothelial cells (16), and leukocyte venular adhesion (6, 9, 10). Once the leukocytes adhere, they can also emigrate out of the vessels and release cytotoxic products into the tissue (11). Activated neutrophils have the ability to release a complex assortment of agents, including oxidants and proteases. Endothelial cells may also produce significant levels of oxygen free radicals (19). The combined production of oxygen free radicals and proteolytic enzymes could contribute to the cell injury that follows tissue activation, but few studies have directly documented the time course of parenchymal cell injury under in vivo conditions.

N-formylmethionyl-leucyl-phenylalanine (FMLP) and platelet-activating factor (PAF) are two substances commonly used to induce leukocyte activation and experimental inflammation. FMLP is a tripeptide that induces chemotaxis of human neutrophils (15). FMLP also increases neutrophil stiffness, actin assembly, and retention in the pulmonary vasculature (23).

PAF is a lipid mediator of endothelium-granulocyte interaction that results in a granulocyte-dependent increase in microvascular permeability (12). In the rat mesentery, the location of PAF activity corresponds to the localization of sticking granulocytes in postcapillary venules (20), and PAF-induced adherent leukocytes mediate the filtration of fluid and protein across intestinal capillaries (13). In addition to promoting an adhesive interaction between circulating leukocytes and endothelium after ischemia-reperfusion, PAF also promotes leukocyte extravasation (13). Both FMLP and PAF have been used effectively to induce leukocyte adherence and emigration as well as a reduced leukocyte rolling velocity by superfusion over the rat mesentery (1).

In addition, the application of FMLP and PAF can result in an "activation" of platelets, endothelial cells, macrophages, mast cells, or any other cell located in a tissue. The effects of these compounds on leukocytes are quite well characterized, but their effects on the tissue in the absence of leukocytes have not been studied. Thus the phrase "tissue stimulation" was chosen to express the changes in the tissue that occur in response to suffusion with FMLP and PAF.

Propidium iodide (PI) has been demonstrated to be an effective indicator of tissue injury (8, 19). PI is a positively charged moiety and therefore is unable to cross the cell membrane. When a cell is damaged and unable to maintain its membrane potential, PI is able to enter the cell, where it binds to the nuclear DNA and becomes fluorescent. Thus the number of PI-positive cells can serve as a measure of the number of nonviable cells present in a tissue. Dichlorofluorescin diacetate (DCFH) is a hydroperoxide-sensitive fluorescent probe that can be used to detect hydroperoxide production (21). It diffuses into viable cells in a nonfluorescent form and is intracellularly oxidized by hydroperoxide to yield fluorescent dichlorofluorescein (DCF). A 1:1 stoichiometry has been observed between the hydroperoxide present and the DCF produced.

The two activators FMLP and PAF were used in combination with the two fluorescent probes PI and DCFH to examine details of the mechanisms of cell injury after tissue activation in the rat mesentery and cremaster muscle. After tissue activation, first with a single stimulus and then with a combination of stimuli, the number of adherent leukocytes and nonviable cells and hydroperoxide production were examined in these two tissues. A combination of stimuli was used to more closely mimic an actual inflammation in which multiple cytokines and mediators may be present. These measurements were also made after infusing an anti-rat antibody to explore the effects of leukocytes on tissue injury mechanisms. The level of preactivation in the animals was measured using the nitro blue tetrazolium (NBT) test to determine its effect on the extent of tissue injury. In two studies by Barroso-Aranda et al. (3, 5), it was shown that the level of preactivation of the leukocytes inversely correlated with the survival rate after hemorrhagic and endotoxic shock. It is believed that there is some circulating plasma factor that may prime the leukocytes, thereby making them more sensible to stimuli. Thus preactivation is a measure to which leukocytes are primed because of plasma factors. The aim of this study was therefore to examine the mechanisms of leukocyte adherence and extravasation, endothelial cell hydroperoxide production, and preactivation on tissue injury.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Male Wistar rats (260-320 g, Charles River Laboratories, Wilmington, MA) were anesthetized with an injection of pentobarbital sodium (Nembutal; 30 mg/kg im). All animal protocols were reviewed and approved by the Animal Subjects Committee of the University of California, San Diego. The femoral artery and vein were cannulated with PE-50 tubing for measurement of blood pressure and to administer additional anesthesia as necessary. Body temperature was maintained at 37°C, and arterial blood pressure and heart rate were monitored. The rats were then randomized into one of two groups, normal or neutropenic. After this division, both the mesentery and the cremaster were studied in the same animal in a random order. The neutropenic group received an intravenous infusion of an anti-leukocyte antibody directed against the rat leukocyte common antigen (LC-A, CD45; Serotec, London, UK). This antibody is specific for leukocytes, and its binding causes the leukocytes to become trapped in the lung and spleen (22). Normal animals received a vehicle infusion. White blood cell counts were made before and after the antibody was administered as well as after the first tissue was studied. The blood pressure initially dropped 40-50 mmHg after the antibody was infused. The experiment was continued after the blood pressure had returned to normal, usually after 30-45 min. The rats were placed on a heated stage maintained at 37°C and covered with a thermal blanket at the completion of the surgery. The stomach and scrotum of the animal were shaved for surgery.

Surgical procedure. The mesentery and cremaster muscle were viewed subsequently in random order by intravital microscopy in the same animals. For the mesentery preparation, the abdomen was opened via a midline incision and the ileocecal portion of the mesentery was carefully withdrawn from the animal with saline-saturated, cotton-tipped swabs and draped over the microscope stage. Once a sector was selected, the remainder of the exposed tissue was covered with plastic wrap to prevent loss of moisture. In cases in which the mesentery was the first tissue studied, the mesentery was placed back in the abdomen at the conclusion of the experiment and the skin incision was closed by a suture.

Intravital microscopy. Throughout the exteriorization procedures and during the experiments, both preparations were maintained at 37°C and continuously superfused with Krebs-Henseleit bicarbonate-buffered solution saturated with a 95% N₂-5% O₂ gas mixture. After the exteriorization, five to six vascularized areas were selected for observation. The observations were made with a ×40 water-immersion lens (numerical aperture 0.75, Zeiss). Brightfield and fluorescent images were recorded with a color charge-coupled device camera (model VI-470, Optronics, Goleta, CA, frame rate = 1/60th s for brightfield and 1/8th s for fluorescent light) and stored on a video recorder (model AG3600, Panasonic Matsushita Electric). The selected areas were videotaped under baseline conditions, and then PI and DCFH were added to the superfusate solution at final concentrations of 1 and 5 µM, respectively. The concentrations were chosen on the basis of previous studies by Suematsu et al. (19, 21) in which these concentrations were shown to be effective. At each measurement time point, the selected area was videotaped for ~30 s using transillumination. PI and DCFH fluorescence were measured under epi-illumination using two different filter sets and a fluorescent microscope attachment (Ploempak, Leitz, Wetzlar, Germany). After a 15-min equilibration period, the areas of interest were again viewed and images were videotaped. The tissue was activated using FMLP at a concentration of 1 × 10⁸ M in the superfusate solution. The areas of interest were videotaped 15, 30, and 60 min after tissue stimulation. Next, PAF was added to the superfusate at a final concentration of 1 × 10⁸ M, and the same measurements were made after 15 and 30 min. Ethyl alcohol was added to the tissue preparation at the conclusion of the measurements to determine the total cell numbers in the observation fields and to compute the fraction of cells with PI-positive nuclei. A similar protocol was then performed in the other tissue.

Microvascular measurements. The videotapes were played back, and the vessel lengths and diameters were measured from the transillumination images using the software package NIH Image. The number of adherent cells in the vessels was also counted. An adherent cell was defined as a leukocyte on the vessel wall that did not move for the entire 30-s observation period. The number of PI-positive nuclei was counted by placing a grid over the video monitor and counting the total number of fluorescent nuclei within the tissue. Hydroperoxide production was quantified from the DCFH images. The average pixel intensity in a 7 × 7 pixel area was measured in two separate areas in the tissue (background) and in three different areas in the endothelial cells of the vessel walls, using NIH Image. The average values for each area (background and endothelium) were found, and their difference was used to estimate the level of hydroperoxide production.

Neutrophil activation. The commercially available NBT test (Sigma, St. Louis, MO) was used to evaluate the level of preactivation of each animal. The plasma of the experimental animals was incubated with the blood of a single donor animal to determine the relative level of leukocyte activation present in each of the experiments. Eighty microliters of blood from the donor animal and twenty microliters of plasma from the experimental animals were incubated for 15 min at 37°C, and the NBT test was carried out on this sample. One hundred microliters of the NBT solution from the kit were added to the one hundred-microliter blood solution and incubated for 10 min at 37°C and 10 min at room temperature. Four blood smears per sample were prepared and stained with Wright stain for leukocyte identification. The slides were dried in a slide warmer overnight. With a Leitz microscope and a ×100 oil immersion objective, 100 neutrophils from each sample were located on the slides and the number of NBT positive neutrophils was counted. NBT-positive cells were identified by the presence of NBT crystals. The level of preactivation is reported as a percentage of NBT-positive cells.

Statistical analysis. Measurements were analyzed using a Friedman repeated-measures analysis of variance on ranks (SigmaStat). Where significant differences were found, Dunn's method was used to compare the data with the baseline within the groups, and a pairwise multiple comparison was also used to examine differences between the normal and neutropenic animals. All values are reported as means ± SE, and statistical significance was set at P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

There were no significant differences in blood pressure and heart rate during the experimental period (data not shown). The systemic leukocyte counts were reduced significantly from an average of 5,969 to 2,518 cells/mm³ and remained at this reduced level until the examination of the second tissue, when the average was 2,393 cells/mm³. Differential counts were not performed as a part of this study. However, in a later study with the same antibody and same experimental protocol a differential count was performed and there was no change in the ratio of the subpopulations.

The efficacy of the anti-neutrophil antibody is further illustrated by the measurement of leukocyte adherence in postcapillary venules (Fig. 1). In normal animals, the addition of FMLP resulted in a significant increase in adherence that was sustained over the entire observation period, involving both tissues. This indicates that FMLP was effective in stimulating both tissues equally. In the neutropenic animals, the adherence was significantly reduced to almost zero in both tissues. Thus the administration of the antibody effectively prevented leukocyte adherence and thereby extravasation.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Mean ± SE values (n = no. of animals) for leukocyte adherence are shown for mesentery and cremaster in both normal and leukopenic animals. In normal animals, addition of N-formylmethionyl-leucyl-phenylalanine (FMLP) resulted in a significant increase in adherence that was sustained over entire observation period in both tissues, indicating that FMLP effectively activated both tissues. In neutropenic animals, adherence was reduced to almost zero in both tissues, indicating that administration of antibody effectively prevented leukocyte adherence and thereby extravasation. PAF, platelet-activating factor. + P < 0.05 vs. baseline; * P < 0.05 vs. normal.

In the normal mesentery, the cell death fraction as measured by the fraction of PI-positive nuclei was not significantly changed compared with the baseline until 60 min, at which time it reached 42% (Fig. 2). In the neutropenic mesentery, the fraction of PI-positive cells did not reach significant levels until 75 min, when it was 47%. There was no significant difference in the fraction of PI-positive cells in the mesentery between the normal and neutropenic cases. The PI-positive cells were located in the tissue itself, and a positively stained endothelial cell was almost never seen. Cell injury was not seen in the normal cremaster until PAF was applied, and it only affected 5% of the total cells present in the tissue. However, this increase was not significant, although it was a definite trend. In the neutropenic cremaster muscle, cell injury was abolished even after the combined stimulation with FMLP and PAF.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Mean ± SE values (n = no. of animals) for fraction of nuclei that are propidium iodide (PI) positive. PI-positive cells were distributed randomly throughout tissue in observed field, were not localized near venules, and were almost never endothelial cells. There were no significant differences in PI staining in mesentery between normal and leukopenic animals, indicating that leukocytes are not a major mechanism in resultant tissue injury in mesentery. In cremaster muscle, leukocytes are a major contributor to tissue injury, because tissue damage was abolished during leukopenia. + P < 0.05 vs. baseline.

There was a much greater difference in DCFH fluorescence intensity because of oxidative stress between the background and the endothelium in the mesentery than in the cremaster (Fig. 3). The neutropenic mesentery showed the largest change in DCFH fluorescent intensity, indicating the highest hydroperoxide production. This difference was increased significantly at each time point compared with the baseline. The intensity difference in the normal mesentery was not significantly changed from the baseline. There were only low levels of DCFH fluorescence observed in the cremaster despite extensive leukocyte adhesion.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Mean ± SE values (n = no. of animals) for difference in dichlorofluorescin acetate fluorescence intensity (I DCFH) between background and endothelial cells, which is a measure of hydroperoxide production in endothelial cells. Levels of hydroperoxide production are significantly increased in leukopenic mesentery, indicating that it could be a possible mechanism for tissue injury, whereas in cremaster very little hydroperoxide was produced. + P < 0.05 vs. baseline; * P < 0.05 vs. normal.

To examine the possible relationship between cell injury and hydroperoxide production, a plot of the difference in DCFH fluorescence intensity versus the fraction of PI-positive cells in the mesentery was made (Fig. 4). From the linear regression, it is evident that there is no correlation between the two parameters (r² = 4.4 × 10⁴). This suggests that hydroperoxide production may not be a primary factor in the tissue injury reaction in the mesentery after FMLP or PAF stimulation.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Plot in mesentery of difference in I DCFH vs. PI-measured cell death fraction. A linear regression to data is also shown, and there appears to be no correlation between the 2 parameters (r2= 4.4 × 104). Hydroperoxide production is evidently not a major mechanism for tissue injury in mesentery.

To examine a possible relationship between preactivation and tissue cell injury, the percentage of NBT-positive neutrophils and the PI-measured cell death fraction in the mesentery after 90 min were plotted against each other (Fig. 5). No correlation was evident between the two parameters (r² = 2.9 × 10³). Similar results were obtained when the NBT counts were plotted against cell death at other observation times (data not shown). This would suggest that preactivation is not a major mechanism of tissue injury in the mesentery after FMLP and PAF stimulation.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Correlation between percentage of nitro blue tetrazolium (NBT)-positive neutrophils and PI-measured cell death fraction after 90 min in mesentery. A linear regression fitted to data shows no correlation between the 2 parameters (r2 = 2.9 × 103). Here again, preactivation does not appear to be a major mechanism for tissue injury in mesentery.

To examine the possible relationship between preactivation and hydroperoxide production, a plot of the percentage of NBT-positive neutrophils versus the difference in DCFH fluorescence in the mesentery was made (Fig. 6). Again, no correlation was found (r² = 0.01). This would indicate that preactivation is not a mechanism favoring hydroperoxide production after FMLP and PAF stimulation of the rat mesentery.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Correlation between percentage of NBT-positive neutrophils and I DCFH in mesentery after 90 min. A linear regression to data indicates that there is no correlation between the 2 parameters (r2 = 0.01). This suggests that preactivation is not a mechanism that determines hydroperoxide production.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The results of the present study indicate that cell injury in the mesentery and the cremaster muscle develops by different mechanisms. Stimulation of the tissue with FMLP resulted in similar levels of leukocyte adherence in both tissues yet different amounts of cell injury. The fact that the elimination of leukocyte with an anti-leukocyte antibody had no effect on the amount of cell injury in the mesentery, whereas the injury was abolished in the cremaster muscle, is further evidence supporting the hypothesis of different mechanisms of injury.

It would appear that leukocyte adherence and extravasation are the major mechanisms involved in injury in the cremaster, although the low level of injury overall makes interpretation of these data difficult. The damage in the normal cremaster first appeared after 75 min of tissue stimulation, which also coincided with the addition of the second stimulus, PAF. This increase in tissue injury was not significant, but a definite trend was seen. The addition of the anti-neutrophil antibody abolished the leukocyte adherence and resulted in elimination of the slight damage in the cremaster. Thus the observation that tissue injury in the cremaster is possibly mediated by leukocytes is in agreement with the results of Korthuis et al. (11), who showed that leukocytes play an important role in vascular injury associated with ischemia-reperfusion in the dog gracilis muscle. The low levels of hydroperoxide production observed in this study are an indication that this pathway is not a major contributor to the injury in the cremaster. This does not agree with the study by Suematsu et al. (19), who showed that oxygen radicals promote granulocyte accumulation and adhesion to endothelium. Thus their proposal of a two-tiered mechanism, whereby oxygen radical production leads to leukocyte recruitment that is then ultimately responsible for the tissue injury, is only partially supported by our results in skeletal muscle. However, our use of FMLP to directly stimulate the leukocytes may have simply bypassed the oxygen radical mechanism that is critical in early ischemia-reperfusion injury.

In the mesentery, leukocyte adherence and extravasation appear to play no observable role in tissue injury. There was no reduction in the amount of tissue injury in the leukopenic animals in this tissue, indicating that neither leukocyte adherence or extravasation is a mechanism of tissue injury. This is in agreement with Suematsu et al. (19), although they were looking at skeletal muscle. They observed an attenuation of the injury using a xanthine oxidase inhibitor to cause a significant reduction in the oxygen radical production. In the current study in the mesentery, there was evidence for the production of hydroperoxide, so its presence could explain the differences between the mesentery and the cremaster.

Oxygen-derived free radicals have been shown to play a role in ischemia-reperfusion injury, and a major source of these damaging compounds is thought to be xanthine oxidase (17, 19). Thus ischemia and the resulting oxidative stress are believed to lead to an increased production of oxygen radicals in the endothelial cells via the xanthine oxidase pathway. This hypothesis is based on the observation that attenuation of the production of oxygen radicals by either xanthine oxidase inhibition or oxygen radical scavengers reduces tissue injury (11). Although an increased hydroperoxide production was observed in the current study with experimental stimulants, the lack of a correlation between the total tissue injury and the production of hydroperoxide would seem to rule out this mechanism. However, it is possible that the hydroperoxide does play a role in tissue injury at some later time and was thus not observed in this study. The damage caused by hydroperoxide could have also been reduced as a result of washout of potentially damaging molecules through the continuous superfusion. However, if such a large washout of the tissue occurred, one would also expect leukocyte adherence and extravasion to be greatly reduced. The removal of inflammatory mediators would result in only minimal activation of the leukocytes, which is not the case in this model. Thus, although hydroperoxide production was detected, it did not appear to have a causal relationship to tissue injury.

The extent of preactivation has been shown to correlate with survival after hemorrhagic and endotoxic shock (3, 5). In both these studies, higher levels of leukocyte activation, as measured by the number of NBT-positive neutrophils, were found to be correlated with the survival rate. We therefore tested to what degree there may be a similar correlation between tissue injury and the level of preactivation in the current protocol. In the current study, however, no correlation was found between preactivation and either hydroperoxide production or tissue injury. Although this result may seem surprising, it is not entirely out of line, because there was no difference in the amount of tissue injury in the mesentery between normal and leukopenic animals. If leukocytes do not play a role in tissue injury in the mesentery, then one would not expect to find a correlation between the level of preactivation and the tissue injury. The apparent discrepancy in the results of the two studies could be because of the different stimuli used or could lie in tissue differences, because survival depends on the function of multiple organs and in the current study only two organs were examined. Such a lack of a correlation also suggests that in the mesentery the mechanism of tissue injury is not related to leukocytes.

It is interesting to note that one major difference between the mesentery and the cremaster is that the mesentery has a large number of mast cells, whereas the cremaster has relatively fewer. One possible mechanism to account for the tissue injury observed in the mesentery is that it is associated with mast cell degranulation. In a pilot study using compound 48/80, an agent that causes mast cell degranulation, an analogous pattern and amount of tissue damage and hydroperoxide production were observed. Such an observation further strengthens the probability that mast cells may be a mechanism of injury in the mesentery.

Although the addition of an anti-leukocyte antibody did not completely eliminate leukocytes from the circulation, leukocyte adhesion and the subsequent extravasation were effectively blocked. The actual leukocyte counts were reduced by only ~50% with the addition of the antibody, so that a significant number of leukocytes were still present in the circulation. However, because the number of adherent cells was reduced to almost zero, it seems that those circulating leukocytes were functionally not able to adhere. This event actually might be expected. It is conceivable that the addition of the antibody did result in the removal of most of the leukocytes from the circulation, but in response to the ensuing leukopenia, a large number of immature leukocytes are then released from the bone marrow. As a consequence, the number of leukocytes in the circulation is still significant, but they are not fully functional.

The blood pressure drop associated with the injection of the antibody could also have resulted in some sort of preconditioning in the cremaster but not the mesentery. In the mesentery, the tissue injury observed was the same in both groups, which would indicate that no protective effect was conferred. In the cremaster, there was virtually no injury observed in the leukopenic group, so some sort of preconditioning could have occurred. However, although we cannot rule out this mechanism, it seems more likely that the injury improvement is the result of the leukopenia, given the large amount of evidence in the literature indicating that leukocytes play a role in skeletal muscle injury.

In conclusion, there are a variety of mechanisms by which an inflammatory stimulus can cause cell injury in the cremaster muscle and the mesentery. After FMLP and PAF stimulation, leukocyte adherence and extravasation may be the cause of tissue injury in the cremaster, whereas another mechanism seems to be responsible for the injury in the mesentery. In both tissues similar levels of leukocyte adherence occurred, yet the elimination of leukocyte adherence had no effect on tissue injury in the mesentery. Despite the fact that there was a significant amount of hydroperoxide production in the mesentery, the lack of correlation between tissue injury and hydroperoxide production would seem to indicate that the latter is also not by itself a major mechanism of injury in the mesentery. No correlation was seen between preactivation and the indexes for either cell injury or hydroperoxide production, indicating that this pathway is not a major contributor to tissue injury in the mesentery after topical stimulation with FMLP and PAF. A striking difference between the cremaster and the mesentery is the number of mast cells present. In light of this difference and both the temporal and spatial pattern of tissue injury observed in the mesentery as well as the data from two pilot studies with compound 48/80, it appears likely that mast cell degranulation is a major mechanism of tissue injury in the mesentery, and this possibility needs to be studied further.