Cytokines and chemokines are key regulators of the inflammatory response, with both proinflammatory (e.g., TNF-α, granulocyte-macrophage colony-stimulating factor [GM-CSF], interleukin-1 [IL-1], IL-6, IL-8, and IFN-γ) and anti-inflammatory (e.g., IL-10) effects. Macrolides appear to decrease the production of proinflammatory cytokines that are detrimental to the host (48
In an earlier clinical study, erythromycin (600 mg daily for 3 months) reduced IL-8 and neutrophil elastase in bronchoalveolar lavage (BAL) fluid from subjects with chronic airway diseases, including those infected with P. aeruginosa
). In subjects with DPB, treatment with erythromycin or roxithromycin for up to 24 months decreased IL-1β, IL-8, and neutrophils in BAL fluid (244
). Other studies have shown that macrolides suppress proinflammatory cytokines in human airway washings and in serum (17
IL-8 is one of the cysteine-X-cysteine (CXC) chemokines and is a potent neutrophil chemoattractant. Erythromycin inhibits release of IL-8, epithelial cell-derived neutrophil attractant (ENA-78), and macrophage inflammatory protein 1 (MIP-1) from macrophages and leukocytes (248
). Clarithromycin can suppress LPS-induced IL-8 production in human peripheral monocytes and the human monocytic leukemia cell line THP-1 (140
). Oishi and associates (213
) reported that erythromycin inhibited IL-8 production from human neutrophils stimulated with formalin-killed P. aeruginosa
. Erythromycin and clarithromycin also showed a concentration-dependent suppression of IL-8 release by eosinophils isolated from atopic subjects (149
Physiological concentrations of erythromycin and clarithromycin inhibit IL-8 mRNA and protein in bronchial epithelial cells from healthy subjects and those with chronic inflammatory airway diseases (283
). Erythromycin at a concentration of 10 μg/ml inhibited IL-6, IL-8, and soluble intercellular adhesion molecule 1 (sICAM-1) secretion from human bronchial epithelial cells (BEC) stimulated by endotoxin (135
). Erythromycin decreased production of TNF-α and IL-6 in human whole blood stimulated with heat-killed Streptococcus pneumoniae
). Similar effects have been observed using roxithromycin, which suppressed the production of IL-6, IL-8, and GM-CSF in the human epithelial cell line Bet-1A stimulated with IL-1α (133
). Media conditioned with TNF-α-stimulated A549 human airway epithelial cells prolonged neutrophil survival in culture, an effect that was abolished by pretreatment of A549 cells with macrolides, which inhibited TNF-α-induced GM-CSF expression at both the protein and mRNA levels (333
). While azithromycin stimulated TNF-α secretion from murine CF airway epithelial cells (75
), azithromycin at 10 ng/ml in primary BEC from human lung transplant recipients inhibited release of IL-8 and GM-CSF after stimulation with LPS (201
). This result may be due to the time of measurement, as it is characteristic of immunomodulation to both stimulate and suppress cytokine suppression, depending on the length of exposure.
As immunomodulatory drugs rather than immunosuppressive drugs, macrolides do not show simple time-dependent suppression of proinflammatory cytokines. Shinkai and associates (255
) confirmed the multiphase response of IL-8 release from normal human BEC stimulated with LPS in the presence of macrolides. Addition of 10 μg/ml clarithromycin immediately decreased IL-8 production, then potentiated the LPS-evoked response (2.5-fold), and subsequently normalized IL-8 to the LPS-untreated control level. This effect was preceded by phosphorylation of ERK1/2, a p44/42 MAPK (255
). These findings are similar to those reported by Culić et al. for human neutrophils (49
). The reason for this nonlinear response may be related to interaction with cross talk of intracellular signal transduction or oscillation-damping feedback (3
). ERK mediates normal human bronchial epithelial (NHBE) cell proliferation, and clarithromycin affects the transition from G1
phase to S phase of the cell cycle and delays growth of NHBE cells (256
). This effect is consistent with a previous in vivo
study in which the oral administration of roxithromycin to healthy BALB/c mice increased levels of IL-1 and IL-2 activity for the initial 14 days, followed by decreases to the control level after 42 days of therapy (153
). It has been proposed that macrolides first activate leukocytes and then suppress cytokine production in the presence of inflammatory priming (49
). Acute neutrophil activation by a macrolide can facilitate the killing of microorganisms, while the suppression of chronic inflammation may limit airway damage (221
). These studies suggest that macrolides initially increase the production of proinflammatory cytokines but that this then rapidly normalizes via immunomodulation.
Azithromycin slows the rate of lung function decline in CF patients. Cigana and colleagues (41
) used three CF (IB3-1, 16HBE14o-AS3, and 2CFSMEo−) and two isogenic non-CF (C38 and 16HBE14o-S1) airway epithelial cell lines to investigate whether azithromycin reduced TNF-α mRNA and protein levels. Azithromycin reduced TNF-α mRNA levels and decreased TNF-α secretion, to approximately the levels in the isogenic non-CF cells. NF-κB and specificity protein 1 (Sp1) DNA-binding activities were also significantly decreased after azithromycin treatment. Similar to these findings, azithromycin significantly reduced lung TNF-α levels, and this was associated with inhibition of neutrophil recruitment to the lung in a murine model stimulated with Pseudomonas
In addition to effects on neutrophils, macrolides probably inhibit eosinophil recruitment. Macrolides inhibit secretion of the eosinophil-chemotactic cytokines RANTES and eotaxin (245
). Erythromycin significantly decreased TNF-α-induced eotaxin mRNA as well as eotaxin release from human lung fibroblasts. Roxithromycin at 5 mg/kg of body weight inhibited formation of IL-5 by mouse spleen cells (152
). These doses are comparable to those in humans, and it was speculated that oral administration of roxithromycin can inhibit the function of Th2-type lymphocytes.
Human studies demonstrated that macrolides can reverse the serum IFN-γ/IL-4 (Th1/Th2) ratio (229
). Roxithromycin dose dependently and significantly inhibited both IL-4 and IL-5 secretion by T cells prepared from both healthy and allergic rhinitis donors without affecting IL-2 and IFN-γ levels (15
). However, there are also contradictory data suggesting a shift from Th1 to Th2 cytokine production after the use of macrolides. Park et al. (219
) quantified changes in Th1 and Th2 cytokines in BAL fluid from patients with DPB after long-term treatment with erythromycin. After the treatment, IL-2 and IFN-γ levels in the BAL fluid were significantly decreased and the IL-4, IL-5, and IL-13 levels were significantly increased. In the ovalbumin-sensitized rat, a model of Th2 stimulation, BAL fluid and lung tissues were examined for mRNAs for cytokines (171
). In these experiments, erythromycin did not decrease Th2-related cytokines or mRNAs. Clarithromycin and other macrolides were reported to inhibit Th1 cytokines, such as IL-2 and TNF-α, more markedly than Th2-type cytokines produced by mitogen-stimulated human T lymphocytes in vitro
More recent studies have evaluated the effect of macrolides on T-cell regulation by dendritic cells (DCs). Azithromycin and clarithromycin significantly upregulated the expression of CD80, a costimulatory molecule for T-cell activation, from murine bone marrow-derived DCs (270
). Azithromycin, but not clarithromycin, increased the production of IL-10, and clarithromycin, but not azithromycin, inhibited the production of IL-6 by DCs. Moreover, azithromycin increased IL-10 and clarithromycin significantly decreased IL-2 when naive splenic T cells were cocultured with DCs stimulated with LPS and exposed to macrolides. These data suggest that macrolides modulate the functions of DCs and that these effects may be different among macrolides.
In summary, there are clear and compelling data demonstrating that the 14- and 15-membered (but not the 16-membered) macrolide antibiotics can decrease the hypersecretion of proinflammatory cytokines and chemokines in cell culture, in animal models of disease, and in persons with chronic inflammatory pulmonary diseases. This effect appears to be nonlinear, and thus truly immunomodulatory, with an initial and short-lived increase in inflammation followed by a sustained decrease of cytokine production and secretion to normal, noninflamed levels, conceptualized as a “resetting” of the circuits. As discussed later, many of these effects appear to be medicated through inhibition of ERK1/2 or downstream transcription factors.
Adhesion molecule expression.
Adhesion molecules are necessary for neutrophils and other inflammatory cells to migrate into the airway in response to inflammatory signals. Mac-1 (CD11b/CD18) expression on peripheral neutrophils from patients with DPB is higher than that on cells from healthy volunteers (159
). Roxithromycin treatment caused a significant decrease in the expression of Mac-1, which was associated with decreased neutrophil numbers in BAL fluid. There were no changes in lymphocyte function-associated antigen 1 (LFA-1) expression (159
). Downregulation of the integrin CD11b/CD18 has also been reported with erythromycin (173
). Erythromycin inhibited LPS-induced neutrophil recruitment to the middle ear of rats, in part by downregulating L-selectin and Mac-1 expression on peripheral blood neutrophils (61
Roxithromycin at therapeutic concentrations inhibited neutrophil adhesion to epithelial cells and decreased the expression of ICAM-1 on IFN-γ-treated epithelial cells (133
). It was reported that soluble ICAM-1 derived from cultured epithelial cells was also reduced by erythromycin treatment (135
). A subsequent study has shown that 14-membered macrolides directly inhibit vascular adhesion molecule 1 (VCAM-1) mRNA induction and leukocyte migration into the lung in a mouse model of bleomycin-induced lung injury (169
In summary, macrolides have been shown to decrease stimulated expression of adhesion molecules, which may contribute to resolution of airway neutrophilic inflammation.
(ii) Chemical mediators.
Neutrophil elastase stimulates degranulation of mucin granules and release of mucin glycoprotein by goblet cells and submucosal glands. This protease also regulates the expression of IL-8, ICAM-1, and MUC5AC mRNAs (252
). Macrolides have been thought to have anti-elastolytic properties in the inflamed airway (98
). Erythromycin can act as an alternate substrate inhibitor of neutrophil elastase, and flurythromycin can inactivate this enzyme in vitro
). In asthmatics treated with clarithromycin, there was a decrease in sputum neutrophil elastase and matrix metalloproteinase 9 (MMP-9) (260
Macrolides may modulate the lipoxygenase pathway of arachidonic acid metabolism. Leukotriene B4
), a metabolite of arachidonic acid, is an important chemotactic factor for neutrophils and is elevated in patients with chronic airway disease. Erythromycin and roxithromycin reduced LTB4
in BAL fluid or epithelial lining fluid (ELF) in subjects with DPB. This was associated with decreased neutrophil numbers and neutrophil chemotactic activity (203
Endothelin-1, a potent bronchoconstrictor and vasoconstrictor, is also a mediator of airway inflammation (68
). Bronchial smooth muscle cells have specific binding sites for endothelin-1, and BEC of asthmatic patients release large amounts of active endothelin-1 (314
). Erythromycin and clarithromycin suppress endothelin-1 expression and release by human BEC, but this was not seen with josamysin, a 16-membered macrolide, or with tacrolimus, a very large macrolide (284
Erythromycin dose dependently attenuates the contractile response of isolated human bronchial strips to electrical field stimulation, suggesting that macrolides may block bronchoconstriction by inhibiting neurotransmitter release or the cholinergic response in airway smooth muscle (286
Neutrophil recruitment and activation are usually followed by neutrophil clearance and resolution of inflammation. In chronic inflammatory lung disease, neutrophils continue to be recruited to the airway, die by necrosis, and release their granule contents, increasing lung damage. Macrolides may induce apoptosis of activated neutrophils (12
). Erythromycin increased cyclic AMP (cAMP) in neutrophils in vitro
, and it accelerated apoptosis at 24 h (12
). Azithromycin has also been reported to promote apoptosis of neutrophils, but this effect was not seen in the presence of Streptococcus pneumoniae
). However, in other studies, no direct effects of macrolides on neutrophil apoptosis and survival were observed, although shortening of neutrophil survival was mediated indirectly through inhibition of GM-CSF release from epithelial cells or IL-8 production in activated neutrophils (307
Roxithromycin induces apoptosis of anti-CD3-activated Jurkat T cells in vitro
by enhancing Fas-Fas ligand and caspase-3 but not caspase-8 (118
). Similarly, apoptosis of unstimulated peripheral lymphocytes from healthy subjects is induced by macrolides through the Fas-Fas ligand pathway (107
). Clarithromycin and azithromycin can also increase the phagocytosis of apoptotic epithelial cells and neutrophils by alveolar macrophages (92
Although macrolides may promote neutrophil apoptosis as a mechanism to promote the resolution of inflammation, these data have been difficult to confirm in vivo and may be of limited clinical relevance.
Airway epithelial cells. (i) Epithelial barrier.
Data suggest that macrolides may help to stabilize the epithelial cell membrane. In the inflamed airway, phospholipases A2
cleave arachidonic acid from the cell wall, and subsequent metabolism of arachidonic acid produces proinflammatory mediators. Feldman et al. have shown that macrolides can stabilize cell walls, protecting them against activated phospholipases (67
). Furthermore, neutrophils potentiated the ciliary dysmotility and epithelial damage caused by proinflammatory phospholipids; this was ameliorated by pretreating the leukocytes with macrolides (67
Junctional complexes between epithelial cells are a physical and chemical barrier. Inhalation of PAF causes ciliary dysmotility and disrupts the tight junction barrier in the rabbit trachea (205
). Oral administration of roxithromycin significantly ameliorated PAF-induced ciliary dysfunction and the increase in epithelial permeability. Azithromycin, but not penicillin or erythromycin, increased the transepithelial electrical resistance (barrier) of human airway epithelial cells cultured on filter supports (16
). In this model, azithromycin induced processing of the tight junction proteins claudin-1, claudin-4, occludin, and junctional adhesion molecule A.
Epithelial β-defensins are components of innate immunity in the airway. Erythromycin increased bactericidal activity of the airway surface liquid taken from cultured human primary tracheal cells, bronchial BEAS-2B cells, and pneumocyte II-like A549 cells (109
). Erythromycin significantly increased the production of human β-defensin-1 and β-defensin-2 mRNA and protein. On the other hand, macrolides appeared to decrease β-defensin-2 levels, but not β-defensin-1 levels, in BAL fluid from DPB patients (90
Thus, although macrolides appear to beneficially affect the epithelial barrier and ciliary function, this may be secondary to the decrease in airway inflammation.
(ii) CFTR expression.
Cystic fibrosis transmembrane ion conductance regulator (CFTR) is a member of the ATP-binding cassette (ABC) transporter superfamily whose function is the transport of a wide variety of substrates. Multidrug-resistant (MDR) protein and a P-glycoprotein also belong to the ABC transporter family and share sequence homology. It was reported that a CF patient who was treated with chemotherapy for fibrosarcoma had a dramatic improvement in lung function and that P. aeruginosa
was no longer isolated from this patient (162
). An increase in MDR mRNA in the patient's nasal epithelial cells was found, whereas no MDR mRNA was detected in a CF subject who had not received chemotherapy. It was speculated that the upregulated MDR protein complemented deficient CFTR function, leading to clinical improvement. Erythromycin is known to upregulate P-glycoprotein expression in epithelial cells (73
), suggesting that this may be a mechanism by which macrolides improve pulmonary function in CF (4
However, Equi and collaborators (63
) demonstrated that neither nasal epithelial MDR nor CFTR mRNA levels in adult subjects with CF were changed after 2 weeks of azithromycin treatment, during which time there was a significant improvement in forced exhaled volume in the first second of exhalation (FEV1
). In addition, data show that neither clarithromycin nor azithromycin changes the nasal potential difference, sodium absorption, or chloride secretion (23
), and azithromycin has no effect on MRP promoter transcriptional activity (42
Although macrolides are clearly beneficial in persons with CF, as discussed below, the data do not support the hypothesis that this is due to either expression or function of the CFTR protein.
The Toll-like receptors (TLRs) are an evolutionarily conserved family of receptors that function in innate immunity by recognizing some invariant regions in bacterial molecules. TLR2 recognizes peptidoglycan and bacterial lipoproteins, and TLR4 and TLR5 recognize LPS and flagellin, respectively. LPS can promote mucus secretion through TLR4, even in neutropenic animals (287
). Epithelial TLR signaling stimulated by LPS or flagellin is decreased by macrolides, and this appears to be mediated through the MAPK-NF-κB pathway (255
Macrolides have been shown to affect TLR expression. Peripheral blood mononuclear cells stimulated with LPS have increased TLR4 mRNA levels. This was suppressed by clarithromycin and was accompanied by decreased LPS-induced IL-8 production (218
). In contrast, Yasutomi et al. (337
) demonstrated that erythromycin did not affect mRNA levels of TLR4 in monocyte-derived dendritic cells but that TLR2 mRNA was upregulated. In TLR2-transfected HeLa cells infected with S. pneumoniae
, erythromycin did not influence the activation of TLR2 (196
Because the macrolides are antimicrobial drugs and many of the innate immunity TLRs recognize specific bacterial products as an initial inflammatory response, it is tempting to speculate that macrolides may act in part by decreasing TLR expression or function. However, the limited data available do not support this as a primary mode of action for the macrolides' immunomodulatory properties.
Other cells. (i) Fibroblasts.
Roxithromycin was shown to inhibit fibroblast proliferation in nasal polyps from patients treated for 1 month before polypectomy (210
). Similar effects were reported in another study, along with inhibition of IL-8 levels in nasal lavage fluid, which was thought to drive polyp growth (330
). Clarithromycin dose dependently reduced the expression of IL-8 and NF-κB in human adenoidal fibroblasts (279
), and erythromycin significantly decreased TNF-α-induced eotaxin mRNA as well as eotaxin release from human lung fibroblasts (245
). EM703, a novel derivative of erythromycin A (discussed below), inhibited fibroblast proliferation and collagen production in the murine lung fibroblast cell line MLg2908 treated with TGF-β (170
Vascular endothelial growth factor (VEGF) plays a role in tumor growth as well as in inflammation of the respiratory tract by enhancing neovascularization and vasopermeability. Clarithromycin and roxithromycin have been reported to inhibit VEGF production from fibroblasts stimulated by hypoxia or TNF-α (182
MMPs are secreted by a wide variety of cells, including fibroblasts, and are important in the migration of inflammatory cells through the basement membrane (60
). Roxithromycin, but not josamycin, suppressed the production of MMP-2 and -9 in nasal polyp fibroblasts induced by TNF-α in vitro
, accompanied by suppression of MMP mRNA expression through the inhibition of NF-κB and activator protein 1 (AP-1) activation (124
). The inhibition of MMPs by macrolides may reduce the extracellular spread of inflammation.
Kohyama and coworkers (150
) documented that clarithromycin, but not ampicillin, minocycline, or azithromycin, had a concentration-dependent inhibitory effect on migration of human fetal lung fibroblasts (HFL-1 cells) stimulated with thromboxane A2
, whereas there was no effect on HFL-1 cell-mediated gel contraction, another function of fibroblasts at the wound area.
Modulation of fibroblast function by macrolides appears to occur in vivo, perhaps as a result of generalized immunomodulation. Inhibition of fibroblast migration by macrolides may regulate the wound healing response following tissue injury.