C3a and C5a enhance granulocyte adhesion to endothelial and epithelial cell monolayers: epithelial and endothelial priming is required for C3a-induced eosinophil adhesion

Abstract

Effects of the anaphylatoxins C3a and C5a on eosinophil and neutrophil adhesion to HUVEC and to primary culture human bronchial epithelial cells (HBEC) were investigated. Activities on both leukocytes and on structural cells were examined. C3a upregulated β₂ integrin expression and caused shedding of L-selectin on eosinophils, but had no effect on neutrophil adhesion molecule expression. C5a upregulated β₂ integrins and caused shedding of L-selectin on both eosinophils and neutrophils. The potency of C5a was equivalent on both cell types; however, the magnitude of the changes in each of these adhesion molecules was significantly greater in neutrophils than eosinophils. Neither C3a nor C5a altered expression of ICAM-1, VCAM-1, E-selectin or P-selectin on either HUVEC or HBEC. C5a induced adhesion of both neutrophils and eosinophils to unstimulated HUVEC or HBEC, and adhesion was further enhanced when HUVEC and HBEC were “primed” with TNF-α and IFN-γ, respectively. C3a failed to enhance adhesion of either eosinophils or neutrophils to unprimed HUVEC or HBEC, and enhanced only eosinophil adhesion to cytokine-primed HUVEC or HBEC. Similar to C3a, C3a_desArg and a C3a-analog peptide E7 also enhanced eosinophil adhesion only to cytokine-primed HUVEC and HBEC. These results support the traditional view of anaphylatoxins as leukocyte-specific mediators. The specificity of C3a for eosinophils implicates this molecule as a potential participant in allergic inflammation. The pro-adhesive effects of C3a_desArg suggest that this molecule, previously characterized as a spasmogenically inactive derivative of C3a, may also alter leukocyte dynamics and migration. Finally, activation of endothelium may represent an important control mechanism for C3a-mediated adhesion preventing unchecked eosinophil adhesion to uninflamed systemic vasculature.

Introduction

The anaphylatoxins C3a and C5a are bioactive peptides generated through cleavage of C3 and C5, respectively, as a consequence of complement activation. C3a and C5a generation results from a diverse set of pathological conditions including autoimmune disease, trauma and sepsis (Heideman and Hugli, 1984; Moxley and Ruddy, 1987; Hack et al., 1989; Buyon et al., 1992; Belmont et al., 1994; Meade et al., 1994). The importance of anaphylatoxin generation is evident in diverse models of pulmonary inflammation, including immune complex-induced lung injury, infusion of cobra venom factor and infection with Pseudomonas aeruginosa (Bozic et al., 1996; Hopken et al., 1996; Mulligan et al., 1996).

The physiological effects of the anaphylatoxins are mediated principally through their actions on leukocytes. The role of C5a as an activator of neutrophils has been exhaustively investigated, and C5a is virtually a gold standard for leukocyte activating factors. C5a activates neutrophils by most parameters, including stimulation of granule release, phagocytosis, oxygen radical formation and production and release of active arachidonic acid products [reviewed in (Ember et al., 1998)]. Perhaps the most thoroughly characterized property of C5a is its ability to induce neutrophil deformation and chemotaxis. C5a has been characterized as a neutrophil chemoattractant in a number of in vitro and in vivo models, and recent studies have demonstrated C5a to be a dominant endogenous chemoattractant; i.e., it is capable of desensitizing neutrophils to other chemotactic factors (Kitayama et al., 1997). Conversely, the chemotactic responses to C5a are not inhibited by pre-exposure to other host-derived neutrophil chemoattractants. Neutrophil infiltration into the lung is considered to be a major deleterious component of several acute and chronic inflammatory conditions, most notably acute asthma, ARDS and cystic fibrosis (Shoemaker et al., 1980; Bedard et al., 1993; Sur et al., 1993). In these and other syndromes, neutrophil accumulation has been noted at intravascular sites, in the interstitum, and in the airway itself (Beasley et al., 1989; Gibson et al., 1989; Sur et al., 1993). These findings suggest that interactions with both endothelial cells and airway epithelial cells may be important components of neutrophil-mediated injury.

While the effects of C5a as a neutrophil chemoattractant are well recognized, its effects as a chemotactic signal for eosinophils have been less thoroughly studied and are frequently not acknowledged. Nevertheless, C5a has been shown to induce chemotaxis of eosinophils with an efficacy and potency equivalent to that observed for neutrophils (Morita et al., 1989; Daffern et al., 1995). Eosinophil accumulation is of particular interest in allergic airway disease. In allergic rhinitis, as well as in allergic and non-allergic asthma, eosinophils are the predominant invasive leukocyte and are implicated as principal mediators of the histopathology of these diseases (Beasley et al., 1989; Masuyama et al., 1993). Eosinophils are also frequently found within the airways, particulary during exacerbations of allergic and asthmatic responses, suggesting that they traverse both endothelial and epithelial barriers (Beasley et al., 1989; Gibson et al., 1989). The identification of potential eosinophil chemoattractants is therefore of particular interest in the study of allergic disease.

In contrast to the extensive characterization of the cellular effects of C5a, the role of C3a as an inducer of leukocyte activation is still ill-defined. While neutrophils appear to express functional receptors for C3a, as evidenced by an intracellular calcium flux in response to C3a, neutrophils appear to be essentially devoid of physiologically relevant responses to C3a (Norgauer et al., 1993; Takafuji et al., 1994; Daffern et al., 1995). In contrast, C3a has been shown to be an effective agonist for eosinophil activation, and induces granule release, generation of reactive oxygen intermediates and chemotaxis of eosinophils in in vitro assays (Takafuji et al., 1994; Daffern et al., 1995). C3a has been investigated as an eosinophil chemoattractant in vivo, but its role has not yet been firmly established (Discipio et al., 1999).

Recent evidence suggests that the effects of C3a and C5a may not be limited to leukocytes, but may extend to “structural cells” (i.e., endothelium and epithelium) as well. C5a has been reported to upregulate P-selectin on microvascular endothelial cells (Foreman et al., 1994), and receptors to C5a have been identified by immunohistochemical methods on airway epithelium (Haviland et al., 1995). P-selectin contributes to the initial rolling of various leukocyte subtypes along the vasculature, which precedes firm adhesion and transendothelial migration, and is rapidly but transiently upregulated by inflammatory stimuli including histamine, IL-1 and TNF-α (Patel and McEver, 1997). Increased expression of P-selectin by C5a therefore could enhance acute leukocyte accumulation in the lung. Northern blot analysis of C3a receptor expression suggested a relatively high level of C3aR mRNA in lung tissue, again implicating potential binding sites on pulmonary epithelial or endothelial cells (Ames et al., 1996). Direct effects of C3a on endothelial or epithelial cells have not been reported. Bronchial epithelial cells have been reported to produce IL-8 in response to C5a if previously primed with cigarette smoke extract (Floreani et al., 1999). However, we have been unable to duplicate these results.

Due to the importance of neutrophils and eosinophils in inflammatory pulmonary diseases, and the potential role of the anyphylatoxins in the recruitment and activation of these cells, we investigated the effects of these mediators on leukocyte adhesion to endothelial and epithelial cells of the lung. Our results suggest that while C5a alone is a potent inducer of eosinophil and neutrophil adhesion, C3a promotes adhesion only of eosinophils, and only under conditions where endothelium or epithelium has been activated. The effects of both anaphylatoxins appear to be leukocyte-specific, because treatment of either epithelial or endothelial cells with C3a or C5a affected neither adhesion of leukocytes nor expression of known adhesion molecules.