Peritoneal cavity has immunological functions as an anti-bacterial defense (Mizgerd et al., 1998; Sasaki, 1999). Once bacterial infection occurs, leukocytes are rapidly recruited from the circulation into the peritoneal surface. Mesothelial cells are actively involved in the inflammatory processes as the final barrier to leukocyte emigration, and as the ground on which leukocytes function in the inflamed area. Intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), belonging to the immunoglobulin superfamily, have been shown to be the main cell adhesion molecules expressed by activated mesothelial cells (Jonjić et al., 1992; Cannistra et al., 1994; Klein et al., 1995; Liang and Sasaki, 2000). Both adhesion molecules are ligands for several integrins expressed on the leukocyte surface, i.e., ICAM-1 for both LFA-1 and Mac-1 (Diamond et al., 1990; Springer, 1990; Dustin and Springer, 1991), and VCAM-1, together with fibronectin, for VLA-4 (Elices et al., 1990).
LFA-1 (αLβ2) is distributed in all types of leukocyte (Dustin and Springer, 1991), and Mac-1 (αMβ2) is mainly in the neutrophil and monocyte/macrophage (Springer 1994). VLA-4 (α4β1) is distributed over almost all types of leukocyte except neutrophil (Hemler, 1990). Interaction between those integrins and their ligands is involved in the leukocyte extravasation and tissue migration (Osborn, 1990; Springer, 1994; Mizgerd et al., 1998), and also in the leukocyte adhesion to the mesothelial cells (Jonjić et al., 1992; Andreoli et al., 1994; Cannistra et al., 1994; Liberek et al., 1996). The spatial microlocalization of those cell adhesion molecules has been suggested to play primary roles in such interaction (Erlandsen et al., 1993; Tohya and Kimura, 1998; Sasaki et al., 1998), but little is known about integrins on peritoneal leukocytes. Recently, we showed the spatial distribution of ICAM-1 and VCAM-1 on the liver mesothelial cells of the LPS-stimulated mice, in which non-stimulated mesothelial cells consistently expressed ICAM-1 on their microvilli, whereas the LPS-stimulated mesothelial cells expressed VCAM-1 and an increasing amount of ICAM-1 on their microvilli (Liang and Sasaki, 2000). To know what this localization of cell adhesion molecules means, the spatial distribution of their ligand integrins on leukocytes gives essential information.
By immuno-SEM based on hypothermic preservation of cell morphology with UW solution (Wahlberg et al., 1986; Sasaki et al., 1993, 1996), we attempted to localize integrins LFA-1, Mac-1, and VLA-4 on activated leukocytes in the cecal perforation-induced peritonitis. The ligands for the above integrins, ICAM-1, VCAM-1, and fibronectin, were also localized on the peritoneal surface structures.
MATERIALS AND METHODS
The investigation was approved by Animal Care Committee of our university. ICR male mice weighing 40–45 g (SLC, Hamamatsu, Japan) were used throughout this study. They received a cecal perforation using 20-gage needles under anesthesia with pentobarbital sodium solution (0.04 mg/gbw). After survival times of 3, 12, and 24 hr, they were sacrificed under deep anesthesia, and their ceca were removed.
The cecum was immersed immediately in hypothermic University of Wisconsin (UW) solution at 4°C for morphological preservation (Wahlberg et al., 1986). The UW solution contains the following per 1,000 ml: lactobionic acid (Sigma, St. Louis, MO) 35.83 g, raffinose (Sigma) 17.84 g, allopurinol (Sigma) 0.14 g, glutathione (Sigma) 0.92 g, adenosine (Sigma) 1.34 g, KH2PO4 3.40 g, MgSO4 0.60 g, NaOH 0.08 g. Its pH was adjusted to 7.4 with 1N KOH solution. The cecal peritoneum was first reacted with primary antibodies diluted with the hypothermic UW solution at 4°C for 1 hr. The rat monoclonal antibodies against mouse integrins LFA-1, Mac-1, and VLA-4 were as follows: anti- αL (M17/4), αM (M1/70), β2 (C71/16) and α4 (9C10). Those antibodies were purchased from Pharmingen (San Diego, CA). The rat monoclonal antibodies KAT (Antigenix America, NY) and 429 (Pharmingen) were used to detect mouse ICAM-1 and VCAM-1 respectively. The rabbit polyclonal antibody to mouse fibronectin was purchased from Biogenesis (Poole, UK). After rinsing in UW solution at 4°C, the peritoneum was immersed in the goat anti-rat or rabbit IgG conjugated with 15-nm gold particles (British BioCell International, Cardiff, UK) diluted with the hypothermic UW solution at 4°C for 1 hr. After rinsing in UW solution at 4°C, they were fixed in 2.5% glutaraldehyde/ 45 mM cacodylate HCl, pH 7.2 for 4 hr. Then they were washed in 180 mM sucrose/80 mM cacodylate HCl, pH 7.2 overnight at 4°C. Negative controls were performed omitting the primary antibody step, or using rat IgG instead of the primary antibody. In the preliminary study, the blocking treatment with normal goat serum had not altered significantly the staining pattern and background intensity of the non-treated specimens.
Preparation for TEM
Tissues were post-fixed in 1% OsO4/90mM sucrose/40 mM cacodylate HCl, pH 7.2 for 1 hr. They were dehydrated with a graded series of ethanol and acetone, and embedded in Epok 812 (Oken, Tokyo, Japan). Ultra-thin sections were cut and collected on copper grids. After staining with uranyl acetate and lead citrate, they were observed with a JEOL JEM-1200 transmission electron microscope (TEM) at an accelerating voltage of 80kV.
Preparation for SEM
Small pieces of tissues (5×5×0.5 mm) were post-fixed in 1% OsO4/90mM sucrose/40 mM cacodylate HCl, pH 7.2 overnight. After dehydration in a graded series of ethanol and substitution with isoamyl acetate, they were dried with a critical point drying method. The peritoneal surface of the specimens was coated by OsO4 to a thickness of 3-nm with an osmium plasma coater (Nippon Laser & Electronics Lab., Nagoya, Japan), and observed with a JEOL JSM-6000F field emission scanning electron microscope (FESEM). The surface ultrastructure of the specimens was observed with secondary electron imaging, while the gold labeling was detected with backscatter electron imaging (BSE).
Peritoneum of Normal Cecum
By secondary electron imaging with FESEM, the cecal peritoneum was covered by monolayer of squamous and polygonal mesothelial cells (Fig. 1A). The center of the cell body bulged due to the nucleus, while the marginal cytoplasm was thin. The free surface of the cells had microvilli, showing heterogeneity in number and length in different areas, i.e., few short in the central bulge, while many high in the margin. Leukocytes were rare on the mesothelial layer. By TEM observation, the mesothelial cells lay on the basement membrane, under which thin layer of loose connective tissue existed containing collagen fibers (Fig. 1B). This loose connective tissue layer had few vasculatures and almost no leukocytes.
Peritoneum in Peritonitis
The cecal peritoneal surface appeared diverse morphologic changes 3, 12, and 24 hr after perforation by the secondary electron imaging with FESEM (Figs. 2, 3A–D). The deformation and detachment of the mesothelial cells were observed 3 hr after perforation. Several types of enterobacteria dispersed on the peritoneal surface, but leukocytes were few in such inflamed area (Fig. 2A). In contrast, many leukocytes had appeared on the inflamed peritoneal surface 12 and 24 hr after perforation (Figs. 2B, 3A–D), forming large aggregations occasionally. The observation was focused on the period at 24 hr after perforation, when a diverse relationship between the mesothelial cells and leukocytes could be seen (Fig. 3). Mesothelial cells remaining in the inflamed area had microvilli with bulbous tips, on which leukocytes were found (Fig. 3A). Retraction and detachment of the mesothelial cells occurred in some areas, causing the exposure of the underlying tissues (Fig. 3B). In other areas, the peritoneum was almost completely denuded of mesothelial cells and basement membranes, in which many leukocytes adhered to the exposed smooth muscle layer covered by thin connective tissue (Fig. 3C). A filamentous network, presumably fibrin, was commonly observed extending between leukocyte aggregation (Fig. 3D). A diversity existed on the surface structures of those leukocytes, e.g., some leukocytes had a round shape and flat surface with a few microvilli, whereas others had rich ruffles/microvilli (Fig. 3A–D). TEM observation showed most of the emigrated leukocytes to be monocytes/macrophages, and others to be neutrophils, lymphocytes, and eosinophils. Some leukocytes were in contact with the cell body or microvilli of the mesothelial cells (Fig. 3E), but others directly adhered to the connective tissue in the denuded area of the mesothelial cells. Many leukocytes existed in the connective tissues, but less in the underlying smooth muscle layer (Fig. 3F).
Immuno-SEM Localization of Cell Adhesion Molecules
Three distinctive membrane domains of leukocytes were set up to describe the immuno-SEM findings, i.e., cell body, ruffles, and microvilli (Fig. 4A). By BSE imaging with FESEM, the gold particles corresponding to the four integrins αL, αM, β2 and α4 were localized on leukocytes, and almost negative on the mesothelial cells or the thin connective tissue overlying the exposed smooth muscle layer. The density of immunolabeling highly varied from cell to cell in each staining. The gold particles for αL and α4 were preferential on the ruffles/microvilli of some leukocytes (Figs. 4B and 5B), whereas those for αMand β2 were distributed randomly on the membrane of the cell body as well as on the ruffles/microvilli (Figs. 4C and 5A). Besides the single gold particles, clustering of them was occasionally observed in each immunolabeling.
The gold particles for ICAM-1 were detected on the leukocytes at random (Fig. 5C), and on the mesothelial cells, confined to their microvilli (Fig. 6A). Those for VCAM-1 were observed on the mesothelial cell microvilli (Fig. 6B), but were few on the leukocytes. The gold particles for fibronectin were detected on the fibrous structures, covering the exposed smooth muscle layer in the denuded area of the mesothelial cells (Fig. 6C). The immunolabeling for fibronectin was also found on the presumable fibrin surrounding the leukocyte aggregation (Fig. 6D). In the control experiments, fewer gold particles dispersed on the leukocytes and peritoneal surface, showing the nonspecific background to be sufficiently low (Fig. 7A–C).
The spatial distribution of cell adhesion molecules on the leukocytes has been studied with immuno-SEM (Erlandsen et al., 1993; Berlin et al., 1995; Hasslen et al., 1995, 1996; von Andrian et al., 1995; Fernández-Segura et al., 1996) and immuno-TEM (Picker et al., 1991; Borregaard et al., 1994; Burns and Doerschuk 1994; Moore et al., 1995; Bruehl et al., 1996; Tohya and Kimura 1998) focusing mainly on leukocyte-endothelium interactions, e.g., diapedesis of the lymphocytes into the organ through high endothelial venules, and extravasation of neutrophils through the activated endothelium in inflamed tissue. Sequential adhesion steps, rolling, arrest, strong adhesion, and transendothelial migration, have been postulated in those phenomena, in which interactions of cell adhesion molecules play a central role (Osborn, 1990; Springer, 1994). L-selectin, a cell adhesion molecule mediating the initial rolling contact of leukocytes with vascular endothelial cells, has been localized preferentially on the ruffles/microvilli of neutrophils, lymphocytes, and monocytes in the blood flow or unactivated condition in vitro (Picker et al., 1991; Erlandsen et al., 1993; Borregaard et al., 1994; Hasslen et al., 1995, 1996; von Andrian et al., 1995; Bruehl et al., 1996; Tohya and Kimura, 1998). β2 integrins mediate the subsequent irreversible adhesion with ICAMs expressed by endothelium. Mac-1 (αMβ2) was distributed on the cell body of unactivated human neutrophils, whereas it was localized on both the membrane of the cell body and also of ruffles/microvilli on activate spreading neutrophils (Erlandsen et al., 1993; Hasslen et al., 1996). LFA-1 (αLβ2) was reported to exist on the cell body of lymphocytes in high endothelial venules (Tohya and Kimura, 1998) and of unactivated TK1 lymphoma cells (Berlin et al., 1995), but its distribution in activated leukocytes remains obscure. The binding between VLA-4 on the leukocytes and VCAM-1 on the endothelium also mediates the strong adhesion (Elices et al., 1990), but little is known about the spatial distribution of VLA-4. Although the mechanism of leukocyte transendothelial migration has not yet been fully clarified, PECAM-1 has been suggested to play a central role (Liao et al., 1997; Nakada et al., 2000).
In the present study, we focused on the leukocyte-peritoneum interactions. Morphological changes of the peritoneum presented here were consistent with the previous study on infectious peritonitis (Verger et al., 1983). The αMimmunolabeling showed the Mac-1 distribution on all membrane domains of activated leukocytes in the inflamed peritoneum, consistent with the previous study in vitro by Erlandsen et al. (1993). The αLstaining showed the preferential distribution of LFA-1 on the ruffles/microvilli of activated leukocytes, different from the known distribution of LFA-1 on leukocytes in the blood flow (Tohya and Kimura, 1998). The total expression of β2 subfamily including LFA-1 and Mac-1 on activated leukocytes was reflected by the ubiquitous distribution of β2 integrin molecule. ICAM-1 was localized on the mesothelial cells restricted to their microvilli as well as on the surface of the leukocytes. Thus, the spatial distribution of LFA-1 and Mac-1, and of their ligand ICAM-1 may be involved in leukocyte interaction such as antigen presentation and T cell-B cell collaboration (Dustin and Springer, 1991; Tohma et al., 1992; Sasaki et al., 1998), and also in leukocyte adhesion to mesothelial cell microvilli. The α4 distribution suggested the predominance of VLA-4 on ruffles/microvilli of activated leukocytes, but the possibility could not be excluded that it included α4β7 (LPAM-1) expression by lymphocytes (Berlin et al., 1995). VCAM-1 was associated with the mesothelial cell microvilli, supported by our previous study (Liang and Sasaki, 2000). Fibronectin was localized on the fibrous structures on the exposed smooth muscle cells, presumably collagen fiber or fibrin (Leak, 1983), either are ligands for fibronectin. VLA-4 capable of binding VCAM-1 and fibronectin may mediate leukocyte adhesion to the mesothelial cells (Cannistra et al., 1994) as well as to the exposed surface covered by collagen and/or fibrin.
The present study showed the spatial distribution of integrins on peritoneal leukocytes to be, at least in part, different from that on circulating leukocytes, supporting our state that leukocyte migration in the peritoneal cavity is via a mechanism distinct from leukocyte extravasation (Sasaki and Liang 2000).
We thank Mr. Kiyokazu Kametani and Ms. Kayo Suzuki, Research Center for Instrumental Analysis, Shinshu University, for their skillful technical assistance.