Modelling the asthma phenotype: impact of cigarette smoke exposure

Asthma is a respiratory disease that is increasing in prevalence globally. Airborne pollutants such as cigarette smoke (CS, direct and passive) and traffic/industrial pollution are reported to increase asthma susceptibility, cause quality of life issues and enhance symptom severity, frequency of attacks and disease exacerbations [1‐23]. Smoking and passive smoking has also been shown to adversely impact on the effectiveness of standard treatment such as inhaled corticosteroid (ICS) in asthmatics [24‐28] and worsen disease outcome [29]. Despite the fact that asthma is a severe and debilitating illness, a significant proportion of asthma patients smoke or are exposed to passive smoke [30]. As many as half of all adult asthma patients may be active, or previous smokers [13, 14]. Thus with the increase in airborne pollution levels, especially in developing countries, and continued exposure to CS (either directly or passively), it is important to try and understand the mechanism by which these pollutants impact on asthma pathogenesis and whether this contributes to treatment-resistance.

Within allergic asthma, exposure to allergen results in a biphasic bronchoconstrictor response. Immediate bronchoconstriction as a result of exposure is termed the Early Asthmatic Response (EAR) and typically occurs within 1 h of contact with aeroallergen. The Late Asthmatic Response (LAR) refers to a more prolonged bronchoconstriction event taking place approximately 3–8 h following contact with allergen. The LAR is often used within clinical studies exploring new therapeutic options with which to treat asthma and as such is considered to be a clinically relevant endpoint [11, 31].

Airway Hyper-Responsivity (AHR) is a cardinal feature of the asthma phenotype. It is defined as an increased sensitivity to inhaled stimuli resulting in narrowing of the airways, which would not usually occur in healthy individuals. This response manifests as excessive bronchoconstriction and airflow limitation, resulting in shortness of breath and chest tightness. Stimuli of AHR include pollution, allergens, cold air and spasmogens such as Methacholine (MCh). The endpoint of AHR in allergic asthmatics exposed to CS has been investigated but results are sparse and conflicting.

Many features of allergic asthma have been successfully modeled in rats and mice. The Brown Norway rat is considered to be one of the most suitable rat strain for use as an allergic asthma model. This particular strain is a high IgE producer, it produces robust responses to allergens (distinct EAR and LAR) and the infiltration of allergic airway inflammation is considered to be similar to that seen in asthmatic patients [31, 32]. The mouse is also considered to be an advantageous model of allergic asthma due to the possibility of the application of genetically modified (GM) strains and the fact that it comprises a highly characterised immune system.

The aim of this study was to determine the effect of CS co-exposure on the phenotype and treatment sensitivity of rodent models of allergic asthma. In order to investigate this, rodent models of allergic asthma were co-exposed to CS and endpoints of the Late Asthmatic Response (LAR), Airway Hyper-Responsivity (AHR) and airway cellular burden were assessed. The effectiveness of gold standard asthma treatments (i.e. ICS, LABA and LAMA) were also investigated within these models. It was hypothesised that the allergic asthma models exposed to CS would exhibit enhanced LAR and AHR responses and the efficacy of standard asthma treatments would be diminished within these groups.

Airborne pollutants such as CS (direct and passive) are known to increase asthma symptoms, severity, frequency of attacks and disease exacerbations and to adversely impact the effectiveness of standard treatment such as inhaled corticosteroid (ICS) in asthmatics. Despite this, the levels of smoking in asthmatic patients are still high; with some estimates suggesting that smoking asthmatics in developed countries represent approximately one quarter of all sufferers. Thus it is important to try and understand the mechanism by which pollution impacts on asthma pathogenesis and treatment. To investigate this effect we determined how CS altered the asthma phenotype in rodent models of allergic asthma. Our studies showed that CS co-exposure increased the magnitude of the LAR, but actually inhibited the AHR signal. CS co-exposure did not appear to impact on cellular burden (above and beyond an additive effect) or treatment effectiveness. This is the first pre-clinical study to comprehensively examine the impact of CS co-exposure on the asthmatic phenotype, and the data demonstrates that these models have many parallels with clinical observations suggesting their usefulness for future investigations.

Antigen challenge triggered cellular recruitment in sensitised animals as previously reported [35, 39]. Similarly exposure to CS caused the expected increase in airway neutrophilia [40]. Co-exposure of the allergic asthma models with CS appeared not to alter the cellular profile above and beyond an additive effect (i.e. neutrophil number). Similar increases in neutrophil numbers have been reported in asthmatics that smoke [41, 42] and it is believed that this cell type plays an important role in the pathophysiology of asthma and is linked to the “asthma COPD overlap syndrome”. Furthermore, Meghji et al. have recently shown similar eosinophilia data in human asthmatics demonstrating that smoking status does not alter the levels following antigen challenge [43]. Interestingly there are some reports that eosinophil numbers are reduced in asthmatics that smoke [14, 44]. This observation could depend on a number of factors including the level of smoke exposure/pack years, asthmatic status and time of sampling. The published preclinical data from studies examining the effect of CS co-exposure is varied, with some reporting reductions and others augmentation in cellular inflammation (the main focus is often eosinophil numbers) [45‐59]. These disparate findings appear to be largely due to variations in CS co-exposure protocols.

Treatment with a clinically relevant corticosteroid, budesonide, inhibited the allergen induced cellular inflammation in the model systems as expected [35, 36], whilst it failed to impact on the CS induced neutrophilia as previously shown [38, 60, 61]. In our studies, co-exposure with CS did not appear to impact the effectiveness of budesonide treatment, a similar result was published by Song et al. [59]. Surprisingly few clinical studies have described the effects of steroid treatment on airway inflammation in smoking asthmatics; the studies tend to report lung function or asthma control as the primary endpoint. In addition, if pulmonary cellular inflammation is described, it is typically only eosinophilia that is reported, therefore there is little direct evidence on the effects of steroids on other inflammatory cells in smoking asthmatics. ICS have been shown to reduce sputum eosinophils in asthmatics, but not in smoking asthmatics in short term and long term studies [62], but others have shown that ICS do improve sputum eosinophils and ECP in smokers and non-smokers alike [29]. Therefore, the effect of smoking on the anti-inflammatory effects of steroids in asthmatics is currently controversial.

A striking observation is the apparent blockade of AHR in the model systems, whether it was driven by an allergic response to OVA or HDM. A similar finding was recently reported in asthmatic smokers that were exposed to a range of antigens and challenged with inhaled MCh [43]. As stated by the authors, it is not clear what the clinical significance is of this observation. One could speculate that as it is well known that smoking does increase clinical symptoms, the measurement of airway reactivity could be clinically irrelevant. Another group has reported that smoke challenge increases AHR in asthmatics but these experiments were performed using a sub-population of asthmatics that have previously reported to be sensitive to CS [22, 23]. Furthermore, the change was observed in only 30% of this sub-population and a similar number were affected in non-asthmatics. Other pre-clinical studies have reported similar findings with CS co-exposure inhibiting the AHR [54, 55]. Currently the mechanism by which CS causes this effect is not known. Melgert et al. (2004) suggested it was through the reduction of cellular inflammation in their model, but this seems unlikely as in our model systems since cellular inflammation was not decreased. There has been some speculation as to whether CS could be directly or indirectly evoking bronchodilation. Indeed CS is known to contain carbon monoxide which has been reported to reduce mouse AHR [63]; furthermore, CS can induce the release of bronchodilation substances such as PGE₂ and nitric oxide [64]. In addition, CS contains nicotine, which conceivably could alter AHR. We believe, however, that these mechanisms are unlikely as normal airway reactivity to inhaled spasmogen was not altered by CS exposure, and the model systems presented with a strong LAR signal. Both these end points should be altered if CS was causing bronchodilation. Other possible mechanisms by which CS co-exposure reduces the AHR signal could be through the reduction of the mediators driving the AHR and the many cytokines suggested to be involved such as IL-5, IL-13 and IL-17 [65‐71] or the production of mediators reported to inhibit AHR like TGFb [72, 73]. Indeed it has been reported that CS co-exposure increases levels of TGFb [74]. Unfortunately measurement of these end points is not possible in our studies as they were designed to focus on cellular inflammation and AHR, and not cytokine levels (the optimum time for cytokine measurements is much earlier) [75]. Another possible mechanism by which CS alters AHR could be due to an impact on airway smooth muscle (ASM), either the increased ability to contract [76] or the remodelling changes reported such as increased ASM thickness via antigen induced increase in proliferation/migration associated with the AHR phenotype [77]. Of the published studies, some have suggested CS increases proliferation, some have suggested inhibition and others to modulate the contractile response, thus this mechanism is still a possibility but needs to be further investigated [78‐86]. Finally, CS could be causing remodelling in the airway which subsequently impacts on AHR. Indeed it has been reported that CS increases airway remodelling in pre-clinical asthma models [48, 52, 87].

Despite the loss of the AHR phenotype in the models following CS exposure, the LAR remains a clear feature; a similar observation was made in smoking human asthmatics [43]. Indeed, our data suggests that CS co-exposure actually enhances this cardinal feature of asthma. It is therefore tempting to speculate that it is this symptom of asthma that is central to the detrimental impact CS has on asthmatics.

As far as we know, we are the first to examine the effect of CS co-exposure on the LAR in a preclinical model. It is currently not clear how CS is exacerbating the LAR signal. One could speculate that as previous data has strongly implicated the TRPA1 - sensory nerve – parasympathetic axis in the LAR [31] that CS is somehow modulating elements of this pathway. Indeed it is well known that CS contains elements like acrolein which can activate TRPA1 [88, 89]. Further, TRPA1 is the molecular target for by-products of oxidative stress including Reactive Oxygen Species (ROS) and other electrophilic compounds, including hypochlorite and hydrogen peroxide which are linked to CS exposure [90‐94]. As CS alone did not cause a “LAR” like response, it would seem that CS induced exacerbation of the response is not simply due to an increase of TRPA1 activator(s). One possible reason for the synergy between CS and antigen challenge could be that CS is increasing the sensitivity of airway sensory nerves to TRPA1 activators. Indeed we, and others, have observed that CS exposure can increase sensory nerve responses to TRPV1 ligands [95], furthermore we have unpublished data that suggests that TRPA1 responses are also increased. It is interesting to note that whilst we do not yet know the mechanism by which CS exacerbates LAR, current therapies such as ICS and LABA can combat this symptom of asthma. Furthermore, the inhibition of LAR in this model with glycopyrrolate confirms previous finding using another LAMA, tiotropium [31, 96].

The aim our investigation was to determine the effect of CS co-exposure on the phenotype and treatment sensitivity of rodent models of allergic asthma. In order to investigate this, rodent models of allergic asthma were co-exposed to CS and endpoints of the Late Asthmatic Response (LAR), Airway Hyper-Responsivity (AHR) and airway cellular burden were assessed. The impact of ICS, LAMA and LABA were also observed within these models.

In summary, we found that the magnitude of LAR within the allergen sensitised models increased with co-exposure to CS and is concordant with our initial hypothesis. Divergent with our hypothesis; ICS, LAMA and LABA attenuated the LAR across both CS exposed and non-exposed groups. Interestingly the AHR was attenuated with exposure to CS. This was accompanied by an increase in neutrophilic inflammation, and although ICS was successful in attenuating overall cellular inflammation, the enhanced neutrophil populations observed remained undiminished.

We suggest that the data from these studies have parallels with clinical findings and that these model systems may be useful tools in helping to understand how exposure to airborne pollutants such as CS can alter the asthmatic phenotype. We propose that these model systems will be extremely useful in future research and will provide the opportunity to identify novel targets for asthma.