The horses in our study showed changes in pulmonary function and clinical scores that confirmed a period of respiratory crisis with neutrophilic inflammation and airway obstruction and also a period of remission. Surprisingly, we were unable to detect significant differences in either mRNA or protein expression in investigated cytokines in endobronchial tissue between provocation and remission. The only statistically significant finding was a small (1.1-fold down-regulation of TNF-α mRNA in endobronchial biopsies in RAO horses after provocation). However, given the extreme sensitivity of real time-PCR methodology this difference was unlikely of clinical relevance.
Earlier studies report conflicting results regarding IL-6 and IL-10 bioactivity or mRNA levels in BAL cells from RAO horses after challenge. However, these results may be due to differences in sampling time or the type of inhalation provocation [20, 21]. Regarding IL-10, others have shown that inhalation challenge with lipopolysaccharide (LPS) elevates IL-10 mRNA expression in both BAL alveolar macrophages  and peripheral blood cells in RAO-susceptible horses . However inhalation challenge with other agents failed to induce this up-regulation in BAL macrophages . Thus it is possible that not only the sampling time, but also the mode and individual components of challenge agent may influence the dynamics of cytokine regulation. In our study, the negative correlation of bronchial TNF-α levels to IL-6 mRNA levels in controls examined on pasture may indicate the time-dependent changes in the expression of these two cytokines  or simply be coincidental. The positive correlation of IL-10 mRNA expression with TNF-α measured by IHC (rs = 0.967, p = 0.007) may have reflected the immunoregulatory properties of IL-10 as a counterbalance to TNF-α .
There are also several other possible reasons that this study did not detect alteration of these cytokines under exacerbation. Firstly, the time point for sample collection in correlation to provocation might have failed to capture a temporally related enhancement in examined cytokine expression. Others workers have shown time-dependent expression of inflammatory mediators in horse airways that occur within hours or days after provocation [13–16, 18, 20]. In contrast, our samples were collected only once RAO-horses had developed clinical signs of impaired pulmonary function, (ΔPplmax more than 1.5 kPa), rather than at a predetermined time after provocation. Alternatively, it can be questioned whether the bronchial epithelial tissues obtained were representative for the study. As shown in earlier studies in human COPD and asthma patients, there is variability of distribution of inflammatory cells in bronchial tissue which can require examination of biopsies from more than one airway generation to increase the statistical power to detect differences between individuals [25, 26]. Of the six biopsies we obtained from each horse, the first two were from the first generation bronchus and examined by PCR. The remaining four biopsies were obtained progressively more distally in the lung and from these morphologically representative biopsies were then selected for further analysis by IHC. As well, the small biopsy size, the sampling of the epithelial layer without underlying deeper mucosa, and the fact that biopsies are generally sampled from more proximal airways and from subcarinae, are additional limitations to the sampling procedure. To secure antigen preservation and optimal morphology for IHC studies, we chose plastic resin embedding, GMA, used in several human studies together with acetone fixation containing protease inhibitors. This embedding technique ensures both excellent antigen preservation for IHC and morphologic overview . When planning the study, the choice of equine antibodies was limited. At present there are several equine specific antibodies available for the investigated cytokines.
The primers used in PCR have been evaluated and validated in other studies. Moreover, companion studies in these RAO subjects have shown increased expression during crisis of other inflammatory mediators in BAL and epithelial cells collected by bronchial brushings and in biopsies [27–29], indicating ongoing inflammatory process. Thus, our findings suggest that, within the complex interrelation of regulatory cytokines, IL-6, IL-10 and TNF-α may not play a key role in the inflammation in bronchial tissues in horses with RAO.
Apart from IL-10 for the RAO horses at pasture, the lack of correlation of the cytokine mRNA to protein levels of IL-6, IL-10 and TNF-α was not unexpected as regulation of actual protein level is likely more complex than a direct relationship to amount of mRNA and the rate of translation and thus the amount of mRNA does not necessarily directly correlate with the level of protein expressed .