ABSTRACT
It has been known for many years that in addition to their antimicrobial actions, macrolide antibiotics have a number of secondary effects on various host cells and their functions (1). For example, more than 40 years ago it was suggested that troleandomycin (TAO) may be of benefit in the treatment of inflammatory conditions of the airways (1), and in 1970 erythromycin and TAO were shown to improve the clinical status of steroid-dependent asthma (2). There followed a 145
We have been investigating mechanisms of injury to human respiratory epithelium and the ability of the macrolide/ketolide group of antimicrobial agents to attenuate effects on ciliary function and on structural integrity of the ciliated epithelium. We have been particularly interested in determining possible mechanisms of cytoprotection. In these studies the source of human ciliated epithelium
has been nasal brushings obtained from the inferior nasal turbinate of healthy volunteers, which were sampled using a bronchoscopy cytology brush as described previously (35). Ciliary beat frequency was measured using a photo-transistor technique (36), and effects on structural integrity of the ciliated epithelium were assessed by means of a visual scoring system (27). Superoxide production by human polymorphonuclear leukocytes (PMNL) was measured using a luci-genin (bis-A-methylacridinium nitrate)-enhanced chemiluminescence (LECL) method (37), and membrane stabilization was determined using a hemolytic assay(38) .In this model of airway disease we used the bioactive phospholipids, platelet-activating factor (PAF), lyso-PAF, and lysophosphatidylcholine (LPC) as mediators of inflammation, in order to mimic pathophysiological conditions in vivo in various clinical conditions. For example, the bioactive phospholipids are important putative mediators of asthma, and PAF is the only potential mediator that can mimic all the important pathophysiological manifestations of this disease(39) . The injurious effects of the bioactive phospholipids were studied in the absence and presence of PMNL. RESULTS AND DISCUSSION
All three bioactive phospholipids-PAF, lyso-PAF, and LPC-caused significant, irreversible, dose-dependent slowing of ciliary beating and damage to the structural integrity of ciliated epithelium at concentrations > 1 pg/mL, and these effects were progressive over 4 hours (40). There were no striking differences in the injurious effects of the individual PL. The effects were most likely mediated via nonspecific cytotoxicity as a result of the detergent-like, membrane disruptive activity of the PL. The injurious effects in this experimental system were not due to reactive oxidant production by contaminating leukocytes in the epithelial strips since neither catalase nor superoxide dismutase, either individually or in combination, was able to reduce the injury induced by PAF (40). The macrolide agents clarithromycin and roxithromycin, as well as the azalide azithromycin (final concentration 20 |Hg/mL) were unable to attenuate the toxic effects of the PL (40). However the ketolide agent, HMR 3004 (final concentration 20 pg/mL) was able to attenuate almost completely the ciliary slowing and damage to structural integrity induced by all three PL (final concentration 2.5 pg/mL) (Table 1) (41).The amount of ciliary slowing and damage to structural integrity of ciliated epithelium induced by the PL was significantly increased in the presence of PMNL, such that the effects on ciliary function and structural integrity of the epithelium at concentrations of 1 pg/mL of the PL in the presence of PMNL (final concentration 1 X 106/mL) were equivalent to those of the PL alone at 5 pg/mL (40). In this experimental system direct epithelial injury may have been present, but injurious effects appeared to be largely due to reactive oxidant release
from PL-primed PMNL. Confirmatory evidence for this was that PAF and LPC were shown to cause dose-related activation of superoxide production of PMNL, while catalase alone or in combination with superoxide dismutase was able to attenuate both the slowing of CBF and epithelial damage induced by all three PL significantly (38,41). In addition, preincubation of the PMNL with any of the macrolide, azalide, or ketolide agents prior to their exposure to the PL, followed by co-incubation with the epithelial cells was associated with almost complete attenuation of these injurious effects (Table 2) (40,41). Further studies have documented that the ketolide agents, and in particular HMR 3004, cause dose-related inhibition of superoxide production by PMNL activated by four different stimuli of membrane-associated oxidative metabolism, namely, A-formyl-Met-Leu-Phe (FMLP), the calcium ionophore, A23187, phorbol 12-myristate-13 acetate (PMA), and opsonized zymosan, all of which use different transductional mechanisms to activate the superoxide-generating system of phagocytes, NAPDH-oxidase (42).The cytoprotective effects of these agents appear to be related to their membrane-stabilizing ability. These effects were demonstrated in a hemolytic assay whereby pretreatment of erythrocytes with each of these agents in similar concentrations to those that inhibited superoxide production antagonized the lytic effects of all three PL (38,41,42). In terms of potency, HMR 3004 was much more effective than the other ketolide, HMR 3647, with both ketolides being considerably superior to roxithromycin (42). The latter has been shown to have membrane-stabilizing properties equal to those of clarithromycin but superior to
those of azithromycin or erythromycin (42). These differences in potency would also account for differences in cytoprotection afforded by the various agents in the different experimental systems. The most potent membrane-stabilizing agents, the ketolides, are the only ones that appear to be able to stabilize the ciliated epithelial membrane against direct cytolytic effects of the PL, whereas all three classes of agents are able to decrease reactive oxidant-induced epithelial injury by stabilizing the PMNL membrane against sensitization by lower concentrations of PL.We also attempted to establish the relative cytoprotective potencies of clarithromycin, roxithromycin, and azithromycin for ciliated epithelium, but no significant differences were noted in experiments with PMNL (40). However, the ketolides were more effective cytoprotective agents than the macrolides/azalides, with comparative results similar to those in the experiments described above. There was also a slight difference in the relative potencies of the two ketolide agents tested, in that the cytoprotection afforded by 3004 was almost complete at the concentrations used (20 pg/mL), whereas there were still mild residual injurious effects with HMR 3647 (41). The differences in membrane-stabilizing ability and cytoprotection of the different agents in experiments with and without PMNL are almost certainly related to differences in their levels of intracellular accumulation, with the accumulation of HMR 3004 being the most impressive (41-43).Our most recent study appears to confirm the distinct relationship between the anti-inflammatory activities of the macrolides/azalides/ketolides and their ability to inhibit superoxide production via membrane-stabilization (44). In this study, it was confirmed that clarithromycin, a 14-membered macrolide, with anti
inflammatory activity, caused dose-related inhibition of superoxide production by activated PMNL and also protected erythrocytes against PL-induced hemolysis, whereas spiramycin, a 16-membered macrolide of the group, known not to have anti-inflammatory activity, did not inhibit superoxide production and possessed only weak membrane-stabilizing properties (44). These experiments have been complemented by additional data showing that spiramycin may in fact increase superoxide production by stimulated PMNL (45), as well as IL-6 production by human monocytes in vitro (46).
