Differential regulation of CCL-11/eotaxin-1 and CXCL-8/IL-8 by Gram-positive and Gram-negative bacteria in human airway smooth muscle cells

Background Bacterial infections are a cause of exacerbation of airway disease. Airway smooth muscle cells (ASMC) are a source of inflammatory cytokines/chemokines that may propagate local airway inflammatory responses. We hypothesize that bacteria and bacterial products could induce cytokine/chemokine release from ASMC. Methods Human ASMC were grown in culture and treated with whole bacteria or pathogen associated molecular patterns (PAMPs) for 24 or 48 h. The release of eotaxin-1, CXCL-8 or GMCSF was measured by ELISA. Results Gram-negative E. coli or Gram-positive S. aureus increased the release of CXCL-8, as did IL-1β, LPS, FSL-1 and Pam3CSK4, whereas FK565, MODLys18 or Poly I:C did not. E. coli inhibited eotaxin-1 release under control conditions and after stimulation with IL-1β. S. aureus tended to inhibit eotaxin-1 release stimulated with IL-1β. E. coli or LPS, but not S. aureus, induced the release of GMCSF. Conclusion Gram-positive or Gram-negative bacteria activate human ASMC to release CXCL-8. By contrast Gram-negative bacteria inhibited the release of eotaxin-1 from human ASMCs. E. coli, but not S. aureus induced GMCSF release from cells. Our findings that ASMC can respond directly to Gram-negative and Gram-positive bacteria by releasing the neutrophil selective chemokine, CXCL-8, is consistent with what we know about the role of neutrophil recruitment in bacterial infections in the lung. Our findings that bacteria inhibit the release of the eosinophil selective chemokine, eotaxin-1 may help to explain the mechanisms by which bacterial immunotherapy reduces allergic inflammation in the lung.


Background
In diseases such as asthma, the airway smooth muscle fulfills a contractile role. However, airway smooth muscle cells (ASMC) can also acquire a secretory phenotype with the ability to release inflammatory mediators including cytokines, chemokines [1][2][3][4], and lipid hormones [5] when stimulated with cytokines or oxidants [6,7].
The role of bacteria in airway inflammation is currently the subject of investigation. For example, bacterial infections may exacerbate asthmatic symptoms and eradication therapy appears to provide some therapeutic benefits [8,9]. On the other hand, low level inoculation immunotherapy with bacterial components may provide therapeutic benefit in the treatment of allergic asthma, presumably by redressing the balance between type1 (Th1; T helper 1) and type 2 inflammatory pathways [10]. In addition to asthma, it has been recognized for some time that bacterial infections are the underlying cause of exacerbations in some patients with COPD [11]. It is therefore important that we understand how bacteria and other pathogens are sensed by airway cells. Human ASMCs express Toll like receptors (TLRs), including TLR2 and TLR4 [12,13]. Stimulation of ASMCs with pathogen associated molecular patterns (PAMPs) including LPS induce either low or undetectable levels of cytokines [12][13][14], although the sensing of LPS, for example, by ASMCs can be enhanced dramatically by co-culture with monocytes [12]. However the effects of whole bacteria on release of cytokines from human ASMCs has not previously been studied. Moreover, the effect of bacteria or associated PAMPs on the eosinophil-selective chemokines released by airway smooth muscle is not completely understood.
In the current study we have used whole Gram-negative Escherichia coli and Gram-positive Staphylococcus aureus as model organisms, as well as a range of well characterized PAMPs, to compare directly their effects on the release of CXCL-8, which recruits neutrophils, and eotaxin-1, which recruits eosinophils. The level of GMCSF, which is active on both eosinophils or neutrophils was also measured.

Human airway smooth muscle cell isolation and culture
Main lobar bronchi obtained from patients undergoing lung resection for carcinoma of the bronchus were used for dissecting human ASMC, as previously described [4]. Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% foetal calf serum FCS supplemented with sodium pyruvate, L-glutamine (2 mM), non-essential amino acids (1:100), penicillin (100 U ml -1 )/streptomycin (100 μg ml -1 ) and amphotericin B (1.5 μg ml -1 ) in a humidified atmosphere at 37°C in air/ CO 2 (95:5 % vol/vol). Confluent cells were passaged with 0.25% trypsin and 1 mM EDTA. Cells at passages 3-7 from 6 different donors were used in the studies described below. In pilot studies ASMC were initially characterized by positive immunostaining for calponin, smooth muscle α-actin and myosin heavy chain. There after they were characterized by their typical morphology and phenotype.
Preparation of E. coli and S. aureus E. coli, reference strain 0111.B4, and S. aureus H380 were used. The S. aureus H380 was isolated from clinical blood culture and stored frozen in 15% glycerol. This was streaked onto agar plates prior to inoculation of single colonies into RPMI-1640 medium with 10% FCS and glutamine at 37°C overnight. The bacteria were pelleted by centrifugation at 800 × g then washed in sterile saline twice. The pellets were re-suspended in sterile saline. In order to quantify the cell density, aliquots of the bacterial suspension were serially diluted and plated onto agar. The bacteria in the bacterial suspensions were killed by heat treatment for 45 min at 80°C. The resultant suspensions were plated to confirm the sterility. Suspensions were adjusted to 10 10 -10 12 colony-forming units (CFUml -1 ) and then frozen with 20% glycerol in aliquots before use in cell culture experiments.

Flow cytometric analysis of TLR2 and TLR4 surface expression human SMC
To determine cell-surface TLR expression, 4 × 10 4 cells fixed in 3.2% formaldehyde were preincubated in 50 μL of staining buffer (0.5% BSA and 0.1% sodium azide in PBS) containing 10% human serum and 30 μg of human IgG (Sigma) for 30 minutes on ice to block Ig receptors as described previously, [13]. Briefly phycoerythrin-conjugated anti-human TLR2 (clone TL2.1) or TLR4 (clone HTA125), or the relevant phycoerythrin-conjugated isotype control antibodies (IgG2a,κ or IgG1,κ; 0.75 μg each; eBioscience, San Diego, Calif) were then added, and the cells were incubated for a further 60 minutes on ice. A FACScan flow cytometer (Becton Dickinson Immunocytometry Systems [BIDS], Oxford, United Kingdom) was used, and typically 10,000 events were acquired in the viable cell region of the forward light scatter/side light scatter plots. Analysis was performed with CELLQuest software (BIDS). The fluorescence signal was calculated as the geometric mean fluorescence intensity of the gated ASMC population.

Effect of selective TLR3 ligand Poly I:C on chemokine release
Cells were treated in triplicate with either medium only or with Poly I:C (0.01 -10 μg/ml) alone or Poly I:C (1 or μg/ ml) in the presence of E. coli (10 6 CFU ml -1 ), S. aureus (10 6 CFU ml -1 ) or LPS at 0.01 μg ml -1 for 24 h before chemokine release was measured as described earlier.

Cell number and cell viability
The cell viability was assessed by the mitochondrialdependent reduction of MTT to formazan on the remaining cells following all treatments mentioned above. Briefly, following the removal of supernatant, the remaining cells were incubated with 1 mg ml -1 MTT in serum free medium at 37°C for 15 min. The MTT was removed and DMSO was added. The OD was read at 550 nm.

Data and Statistical Analysis
Data are presented as mean ± SEM. Data were compared using one-way analysis of variance (ANOVA) followed by Newman-Keuls post-hoc t-test to determine statistical differences after multiple comparisons using Prism software. A probability value of less than 0.05 was considered significant.

Effect of Gram-negative E. coli and associated PAMPs on release of eotaxin-1 from human ASMCs
Under control culture conditions, ASMCs released relatively low or undetectable levels of eotaxin-1. When E. coli (10 6 to 10 8 CFU ml -1 ) was added to human ASMCs, in experiments from cells of donors where 'basal' levels were detectable, a reduction in 'basal' release was noted (data not shown). In subsequent experiments, cells were stimulated with IL-1β in order to induce eotaxin-1 release and the effects of co-treatment with bacteria or PAMPs were studied. E. coli induced a concentration dependent reduction in the release of eotaxin-1 stimulated with IL-1β. This was significant at 10 8 CFU ml -1 (Figure 2). Neither LPS, nor FK565 given alone or in combination affected the release of eotaxin-1 induced by IL-1β (data not shown).

Effect of Gram-positive S. aureus and associated PAMPs on release of eotaxin-1 from human ASMCs
Similar to observations with E. coli, S. aureus (10 7 and 10 8 CFU ml -1 ) did not induce the release of eotaxin-1 from human ASMCs (Figure 2). Furthermore, when cells were co-stimulated with IL-1β, S. aureus tended to reduce eotaxin-1 release, although this did not reach statistical significance ( Figure 2). In line with observations from    whole bacteria, neither Pam 3 CSK4 nor FSL-1 nor MDPLys18 affected eotaxin-1 release by human ASMCs (data not shown).

Effect of Gram-negative E. coli or Gram-positive S. aureus bacteria on GMCSF release by human ASMCs
Under control culture conditions, human ASMCs released low or undetectable levels of GMCSF. Stimulation of cells with E. coli (10 7 and 10 8 CFU ml -1 ) induced increased release of GMCSF ( Figure 3). However, S. aureus (10 7 or 10 8 CFU ml -1 ) had no effect on the release of GMCSF (Fig-ure 3). In the presence of IL-1β, GMCSF release was increased. Neither E. coli nor S. aureus had any additional effect on GMCSF release in cells stimulated with IL-1β ( Figure 3).

Effect of Poly I:C on chemokine release by human ASMCs
Viral PAMPs activate intracellular PRRs. Poly I:C is a synthetic form of viral dsRNA and a selective ligand for TLR3.
Poly I:C (0.01 -10 μg/ml) had no effect on the release of CXCL-8 or eotaxin-1 from human ASMCs stimulated for 24 h (data not shown). When cells were co-treated with Effect of bacteria on GMCSF release from human ASMCs over 24 h  Poly I:C and LPS, E. coli or S. aureus there was a tendency, which did not reach statistical significance, for eotaxin-1 and CXCL-8 to be increased ( Figure 4).

Expression of TLR2 and TLR4 by human ASMCs
As we have shown previously [13], human ASMCs express detectable levels of TLR2 or TLR4 ( Table 3). Incubation of cells with whole Gram-positive or Gram-negative bacteria did not significantly alter levels of TLR expression (Table  3).

Discussion
In the current study we demonstrate that human ASMCs respond directly to Gram-negative or Gram-positive bacteria. We also show that there is a differential relationship between the release of CXCL-8 and eotaxin-1 by these cells when stimulated with bacteria.
Gram-negative E. coli or Gram-positive S. aureus activated human ASMCs to release CXCL-8. CXCL-8 is a potent che-moatractant for neutrophils. Recruitment of neutrophils is an early event at the site of bacterial infection and these observations identify an innate immune function of ASMCs which may be relevant in conditions such as asthma or COPD.
In most cells, the effects of Gram-negative bacteria can be mimicked by LPS, which activates TLR4 [15]. Gram-negative bacteria also contain peptidoglycan which can activate NOD receptors. The effect of E. coli on CXCL-8 by human airway smooth muscle cells was mimicked by LPS, but not by FK565, which activates NOD1 receptors. In some cell types LPS synergizes with FK565 to activate cells [16,17]. However, in the current study using ASMCs, we found no evidence for interaction between these two pathways. These observations suggest that the ability of Gram-negative E. coli to stimulate airway smooth muscle cells to release CXCL-8 is mediated by TLR4 and not NOD1 receptor activation.
Gram-positive bacteria contain PAMPs for TLR2, which can heterodimerize with either TLR1 or TLR6 [18]. In our hands, activation of either the TLR2/TLR1, with Pam 3 CSK4 or TLR2/TLR6, with FSL-1 resulted in a stimulation of ASMCs and release of CXCL-8. Peptidoglycan from Gram-positive bacteria activates mainly NOD2 [16,19]. MDPLys18 which is a selective ligand for NOD2, had no effect on CXCL-8 release by airway smooth muscle cell; these observations suggest that the effects of S. aureus on airway smooth muscle are mediated by TLR2 and not by NOD receptor activation. Effect of selective Poly I:C on chemokine release from human ASMC Interestingly, by contrast to results with CXCL-8, we found that E. coli, but not S. aureus, increased the release of GMCSF by human airway smooth muscle cells. In other cells, the effects of E. coli are mediated principally by TLR4 and those of S. aureus by TLR2 [20]. TLR4 recruits both TRIF/TRAM and MyD88/MAL adapter protein pathways. TLR2, on the other hand, recruits MyD88/MAL only. It may well be, therefore, that CXCL-8 is a MyD88/MALdependent gene whereas GMCSF, like M-CSF [21], or NOSII [22,23], is independent of the MyD88 pathway and is mediated by TRIF/TRAM. By contrast to results obtained with CXCL-8, we found that bacteria did not stimulate the release of eotaxin-1. In fact, in cells co-treated with IL-1β, Gram-negative E. coli inhibited eotaxin-1 release. A similar trend was seen for cells stimulated with IL-1β and S. aureus. If the observations we make here with cells in vitro translate to the in vivo situation, we may expect bacteria to reduce eotaxin-1 in the airways, limit the numbers of eosinophils recruited and reduce the associated respiratory symptoms in allergic airway disease. Interestingly, recent studies have shown that bacterial DNA immunotherapy can successfully inhibit eosinophil recruitment into the airways in models of allergic lung disease [24]. In the current study, we found that the inhibitory effects of E. coli, on eotaxin-1 release were not mimicked by LPS or by FK565 in our experimental conditions. Whilst not tested directly, our observations are consistent with the effects of E. coli on eotaxin-1 release being mediated by DNA and not by TLR4 or NOD receptors. Our data shows that treatment of cells with Poly:IC did not induce CXCL-8 or eotaxin release. We have shown previously that Poly I:C at a concentration of 25 μg/ml induced the release of eotaxin and IL-8 from airway smooth muscle cells [13], but in the current study, we did not find an effect of Poly I:C at 10 μg/ ml although eotaxin release has been reported at this concentration [25]. This may be due to the fact that we treated the cells under different conditions, that is, in the presence of serum, while the other studies used serum-free conditions.
It is important to note that the observations we report here may well have important implications for other respiratory conditions in which bacterial infections are prevalent such as cystic fibrosis. In cystic fibrosis, colonisation of the lungs with bacteria is known to contribute to exacerbations. Moreover, remodelling of airway smooth muscle in this condition [26][27][28] may lead to an alteration in the balance of chemokine release in response to bacteria.

Conclusion
We show that whole Gram-positive or Gram-negative bacteria activate human airway smooth muscle cells to release CXCL-8. Using selective PAMPs, we establish that the effects of whole bacteria on CXCL-8 are likely to be mediated by TLRs and not by NOD receptors. By contrast to CXCL-8, we found that bacteria inhibited the release of eotaxin-1 from human ASMCs. In the case of eotaxin-1 release, the effects of Gram-negative bacteria did not appear to be mediated by TLR4 or NOD pathways. These observations establish a clear but complex innate immune function of ASMCs with regard to bacterial infections. http://respiratory-research.com/content/9/1/30