Inflammatory airway disease (IAD) is a noninfectious lung disease defined by the presence of respiratory clinical signs at work, exercise intolerance, or both, but without clinical signs of labored breathing at rest (even after a moldy hay challenge), along with lower airway inflammation and airway hyperresponsiveness. One of the most common tests to diagnose IAD is cytological analysis of bronchoalveolar lavage fluid (BALF). Affected horses can have increases in BALF eosinophils, lymphocytes, and macrophages, but the most common diagnostic cytological alterations are increases in BALF mast cells or neutrophils. The prevalence of IAD is high in some populations and can affect horses of any age and work use.[2, 3] The pathophysiology of IAD is unknown, but allergy and environment may be contributing factors. Similarly, recurrent airway obstruction (RAO) is a nonseptic lung inflammatory disease in which environment is important, but where only neutrophils accumulate in the lower airways. In comparison with human asthma, the expression of some inflammatory cytokines has been studied in horses with RAO and variable patterns of expression of cytokines from the Th2 and Th1 family have been reported by means of various methods.[6-9]
The pathophysiology of IAD is unclear. Some IAD horses have mostly mast cell accumulation in their airways whereas others have mostly neutrophils and some have a combination of both. Previous studies showed some clinical differences among horses with BAL accumulations of eosinophils, mast cells, or neutrophils. Bronchoalveolar lavage fluid mastocytosis and eosinophilia is associated with airway hyperreactivity whereas only neutrophilia is associated with cough.[4, 10, 11] In addition, the mast cells and neutrophils usually are activated differently by the immune system. It is thus possible that the chemokines and inflammatory cytokines involved in these IAD phenotypes are different.
The goal of this study thus was to describe the gene expression profiles in the BALF cells of 2 groups of horses with IAD: those with BALF mastocytosis (IAD-Mast), and those with BALF neutrophilia (IAD-Neutro). A comparison between analytical methods of gene expression by quantitative polymerase chain reaction (QPCR) also is provided.
Materials and Methods
This study was approved by the Animal Care Committee of the Health Science Centre at the University of Calgary. The authors used the REFLECT statement guidelines to report this study.
We compared the mRNA expression of cytokines and chemokines in 10 normal horses (control) and 17 horses with IAD (IAD-Total) volunteered for the research project and examined in the field by the authors. Each horse had a history taken using a questionnaire, physical examination, and BALF procedure done. The control horses were from the same barns as the horses with IAD and had (1) no history of respiratory clinical signs either at rest or during exercise (ie, no coughing, nasal discharge, labored breathing, based on the answers to the questionnaire), (2) normal BALF cytology, (3) normal physical examination, including absence of coughing on laryngeal and tracheal palpation, and (4) no history of a recent medical condition (Table 1). The inclusion criteria for the IAD horses were (1) respiratory clinical signs during exercise, but no history of labored breathing at rest, (2) normal physical examination, (3) BALF cytology with increased mast cell (IAD-Mast) or neutrophil (IAD-Neutro) percentages. The IAD-Mast horses (n = 8) had BALF with 5–10% mast cells, <0.2% eosinophils, and <10% neutrophils (Table 1). The IAD-Neutro horses (n = 9) had BALF with >10% neutrophils and <2.5% mast cells, <0.2% eosinophils (Table 1). The horses were of various breeds, both sexes, and 9.8 (±5.4 SD) years of age. The history questionnaire also covered the management of the horses, which were housed in a variety of indoor and outdoor conditions and were used for pleasure, performance, breeding, or racing (Table 1).
Table 1. Description of the population of horses studied: 10 were control horses and 17 horses had inflammatory airway disease (IAD) based on clinical signs and bronchoalveolar lavage analysis. (The IAD-Mast horses [n = 8] had a bronchoalveolar lavage fluid [BALF] with 5–10% mast cells, <0.2% eosinophils, and <10% neutrophils. The IAD-Neutro horses [n = 9] had a BALF with >10% neutrophils and <2.5% mast cells, <0.2% eosinophils.)
|Normal (n = 10)|
F (n = 5)
G (n = 5)
|11.5 ± 5.9|
TB (n = 3)
QH, Paint and QH cross (n = 3)
Other (n = 4)
Sport/Recr./lessons (n = 7)
Racing (n = 2)
Breeding (n = 1)
Outside (n = 5)
Outside/inside (n = 5)
Arena/outdoor (n = 7)
Outdoor (n = 3)
|No (n = 10)||No (n = 10)|
|IAD-Mast (n = 8)|
G (n = 5)
F (n = 3)
|9.9 ± 5.4|
TB (n = 4)
QH, Paint and QH cross (n = 3)
Other (n = 1)
Sport/Recr./lessons (n = 5)
Racing (n = 2)
Breeding (n = 1)
Outside (n = 3)
Outside/inside (n = 1)
Inside (n = 4)
Arena/outdoor (n = 4)
Outdoor (n = 4)
No (n = 4)
Yes (n = 4)
No (n = 1)
Yes (n = 7)
(n = 9)
G (n = 6)
F (n = 3)
|(8.5 ± 5.2)|
TB (n = 2)
QH, Paint and QH cross (n = 3)
Other (n = 4)
Sport/Recr./lessons (n = 7)
Racing (n = 1)
Breeding (n = 1)
Outside (n = 5)
Inside (n = 4)
Arena/outdoor (n = 7)
Outdoor (n = 1)
|Yes (n = 9)||Yes (n = 9)|
Sample Collection and Processing
The BALF procedure was done as previously described. Briefly, BAL was performed with a flexible videoendoscope and 2 sequential boluses of 250 mL 0.9% sterile sodium chloride. Samples were immediately stored at 4°C. Differential counts were performed on slides prepared by means of a cytospin (1,000 rpm for 4 minutes) and stained with a Modified Wright Giemsa. Two 50 mL aliquots of BALF were centrifuged at 700 × g for 10 minutes before removing the supernatant and resuspending the cell pellets in 1.5 mL of RNAlater1 and then immediately freezing them at −80°C for the QPCR experiments.
RNA Extraction and cDNA Synthesis
Total RNA was extracted with the RNeasy Mini Kit.1 The yield and purity of the extracted RNA were measured with the Nanodrop ND-1000 spectrophotometer2 at optical density (OD) A260/A280 nm. Approximately 500 ng total RNA was retro-transcribed with the OmniscriptRT Kit1 in combination with RNaseOUT2 and Oligo(dT) primers.3 After cDNA synthesis, a cDNA cleanup was performed with the QIAquick PCR Purification Kit.1
Primers for glyceraldehyde 3-phosphate dehydrogenase (GAPDH), IL-17, IL-8, IL-1β, IL-5, IL-6, IL-10, IL-12 p35, and eotaxin-2 (Table 1) were designed by the Primer3 software based on horse sequences from the Ensembl database as previously described, so that the predicted amplicon would span exon-exon boundaries. The primers were tested by a BLAST analysis against the Ensembl database, verified by MFold, and tested for optimal annealing temperature.
Primer sequences used for IL-4 and IFN-γ were previously described.
Specificity of the products was verified for each reaction with a high resolution gel electrophoresis and resulted in a single product with the desired length; in addition a melting curve analysis was done, which resulted in single product specific melting temperatures.
Quantitative Real-Time Polymerase Chain Reaction (PCR)
The PCR reactions had a final volume of 25 μL and consisted of 13 μL PerfeCta™ SYBR®Green SuperMix, 40 nM forward and reverse primer, 2 μL of cDNA, and 7 μL nuclease free H2O. The PCR reactions were performed on a MX3005P machine4 with initial denaturation (95°C, 5 minutes), 45 cycles of denaturation (95°C, 1 minute), annealing (30 seconds), extension (70°C, 30 seconds), followed by the melting curve (60–95°C). Reactions were executed in triplicate. Each cytokine was run on a separate plate with negative controls (2 μL of water in place of cDNA).
We validated GAPDH as the most stable reference gene in BAL cells of horses with IAD treated with corticosteroids in a previous study and used it as a reference.
We corrected for differences in efficiency between PCR reactions by the LinRegPCR software for each individual sample.
We used the relative expression software tool (REST) with PCR efficiencies correction and normalization by a reference gene as previously validated,[18, 19] and compared them with the 2−ΔΔCt method. The REST analysis uses the following calculations with efficiency correction:
The “Delta-delta” method used:
where E is efficiency of the PCR reaction and Ct is the crossing point above background fluorescence.
Traditional approaches of QPCR analysis can provide an average expression value, but the statistical analysis of results is challenging because ratio distributions do not have a standard deviation. In addition to reporting the magnitude in the ratio of expression between groups, REST integrates a statistical analysis with a statistical randomization algorithm to calculate if the variation in the mean cycle threshold values between 2 groups is significant (see below for details). Another benefit from using REST is that it calculates a 95% confidence interval for expression ratios by a bootstrapping technique without normality or symmetrical assumption, thus providing critical information for data interpretation.
Normality of the distribution of the BAL fluid cell counts and age of the horses were tested by a Shapiro-Wilk normality tests. Parametric and nonparametric analysis of variance (ANOVA) were used to assess differences between the groups in macrophage and lymphocyte cell counts (Kruskal-Wallis one-way ANOVA) and neutrophils, mast cells, and eosinophils (one-way ANOVA followed by a Tukey Least Significant Difference method if indicated) cell counts. The REST software uses a P(H1) test for the statistical analysis that represents the probability of the alternate hypothesis that the difference between the sample and the control group is because of chance only. The hypothesis test performs 10,000 random reallocations of samples and controls between the 2 groups and counts the number of times the relative expression on the randomly assigned group is greater than the sample data. Correlation between cytokines expression and BALF cell percentages was assessed by a Pearson test. P-values < .05 were considered significant throughout the study.
The aim of this study was to compare the mRNA expression of cytokines and chemokines in BALF cells from 17 IAD horses versus control horses. We report for the first time the ratios of expression of inflammatory cytokines from the Th1 and Th2 family in the BALF of horses with IAD. The expression of cytokines from the Th2 family was increased in the cells from the BALF of horses with IAD. In addition, the expression of IL-17, IL-8, IL-4, and IL-5 cytokines in the BALF differed between subpopulations of IAD horses with increased BALF mast cells versus neutrophils.
We used the QPCR method with relative quantification and chose a reference gene that had been specifically validated for IAD horses. The stability of reference genes must be validated because variations in expression level between experimental conditions can alter results. In addition, as another important methodological precaution, we used efficiency correction of the PCR reactions as previously described. A more traditional QPCR analysis method, called the “2 delta-delta Ct” method, assumes that the efficiency of all the PCR reactions is the same and optimal. Reaction efficiency varies among samples and small differences in PCR efficiency can affect the conclusions of a study. Therefore, a newer method (used in REST, see methods) has been proposed that includes a correction for the differences in efficiency among PCR reactions. However, because in the present study we optimized the PCR reaction such that efficiencies were high for all primers, the efficiency correction did not affect our conclusions. Indeed, results by the −2ΔΔCt method without efficiency correction were comparable with those reported with REST (Table 3). This underlines the importance of having optimized PCR reactions for robustness of the analysis.
Because the efficiency correction did not present a net advantage in our study, the main benefits from the REST analysis were the provision of a value for the changes in ratios of gene expression, a statistical analysis to assess the significance of these changes, and a standard error and 95% confidence interval of the ratios. Validated statistical randomization algorithms and bootstrapping of the data are necessary and used in REST to obtain this information because ratios of expression obtained with other methods do not have standard deviation values. This also explains why previous studies[23, 24] did not report ratios of expression between IAD and control groups, but rather the presence or absence of a statistical difference in expression of cytokines among groups without providing an order of magnitude. Despite these precautions, our results are reported at the mRNA level, which can be subjected to post-transcriptional regulation and thus not reflect true protein concentrations. Additional studies with quantitative protein detection methods in BALF will be necessary to verify the results reported here at the protein level.
We found that cytokines typically classified from the Th2 family, namely IL-4, IL-5, IL-6, and IL-10, were overall upregulated in the horses with IAD whereas IFN-γ, which is traditionally classified as a Th1 cytokine, was not. These results are in agreement with a previous study where no difference was found in the expression of IFN-γ in 7 horses with IAD, but the mRNA expression of IL-13, classified as a Th2-cytokine, was upregulated more than 8 times. A recent study on race horses with exercise intolerance and IAD however, found an increase in both IL-4 and IFN-γ in the BALF cells. Another study recently found no difference in the expression of cytokines from the Th1 and Th2 family in horses with IAD versus controls. However, the magnitude of the difference in the ratios of cytokines expression was not provided in these studies, making the physiologic interpretation and comparison with our data difficult. Visual inspection of these previous results indicates that the magnitude of the differences in cytokines expression between groups often was small. One possible explanation for the differences between studies may be the QPCR analytical methods. In the 2nd study, β-actin was used as a reference gene without prior validation, but it has been recently shown that β-actin is unstable in IAD horses treated with corticosteroids. It is thus possible that β-actin expression also is altered in IAD and affected the results. Similar to our study, Lavoie et al used GAPDH for reference, but the analysis was performed by calculating absolute concentrations of amplicons obtained for each gene, which has not been validated, and uses a standard curve dilution method that does not take into account the variation in efficiency along all the dilutions of the curve. It also does not correct for variation in efficiencies of individual PCR reactions. Another possible explanation for the discrepancies between studies may be that the phenotype of the horses used was different. For example, a majority of the horses studied by 1 group had exercise-induced pulmonary hemorrhage and most of the horses studied by another group were sampled after strenuous exercise on a treadmill and were making noise during exercise. In spite of these differences, we found similarities with the results from these studies, as discussed below.
The definition of IAD encompasses both horses having mostly moderately increased BALF neutrophilia and those having mainly increased BALF mast cells. Because the immunologic pathways involved in the regulation of these 2 cell types are likely to be different, we also analyzed the mRNA expression of cytokines in the IAD-Mast group (increased BALF mast cell percentage) and in the IAD-Neutro group (increased BALF neutrophil percentage) separately. Such analysis revealed that, compared with control horses, IL-5 and IL-6 were significantly upregulated in both groups, but IL-4 was upregulated in the IAD-Mast group only (Fig 3). IL-4 was also the only cytokine to be less expressed in the IAD-Neutro group than in the IAD-Mast group (Fig 3C). Similarly, a recent study found that the amount of IL-4 mRNA was increased in the BAL of horses with increased metachromatic cell percentage, but not in those with increased neutrophil percentage. IL-4 is primarily produced by Th2 cells and mast cells, which probably explains why we found it to be upregulated in the IAD-Mast group, but not in the IAD-Neutro group. IL-4 is a cytokine that, among others properties, induces B-cells to undergo immunoglobulin isotype switching to IgE, suggesting that IgE may play a role in the pathophysiology of IAD in horses with increased BALF mast cells.
The increase in IL-5 expression that we found in all IAD groups is similar to what others reported in RAO and in an ovalbumin sensitization model in ponies.[9, 28] However, the mechanisms to explain this increase are not clear because IL-5 is primarily involved in tissue eosinophils migration. Similar to horses with RAO used in the other reports mentioned above, the population of IAD horses did not have BALF eosinophilia. It is thus possible that IL-5 plays a role in mechanisms other than eosinophil migration in horses.
Contrary to other studies on RAO,[7, 30] we found that the IL-6 mRNA expression is systematically increased in the BALF of the IAD horses. IL-6 is a Th2 cytokine mainly produced by lung epithelial cells that promotes the Th2 differentiation of CD4+ T cells.
We found that IL-17 is upregulated, but only in horses with IAD that have a higher BALF neutrophil content. This is similar to the increased expression of IL-17 shown by others in RAO models[32, 33] and suggests that IL-17 contributes to the difference between the 2 phenotypes seen in IAD (ie, IAD-mast versus IAD-neutro). Our finding that IL-17 was upregulated only in the IAD-Neutro group but not in the IAD-Mast group matches well with the role of IL-17 in the chemotaxis and maturation of neutrophils. Interestingly, the expression of IL-23 (which promotes the Th17 response by the IL-23-Th17 axis) was recently found to be increased in the BALF of horses with IAD. In addition, the fact that IL-6 promotes the generation of murine Th17 cells combined with our findings of IL-17 and IL-6 upregulation suggests that IL-6 may also be involved in mechanisms leading to increased IL-17 expression in IAD horses.
Furthermore, we found that IL-8, a potent chemo-attractant and activator for neutrophils, was significantly upregulated in the IAD-Neutro group compared with the IAD-Mast group. This difference in expression suggests a positive correlation between the percentage of neutrophils and the concentration of IL-8 in airway secretions, as previously described. Similarly, Giguere et al also reported a correlation between the expression of IL-8 and the BALF neutrophilia in horses with heaves. IL-17 also plays a role in the chemo-attraction of neutrophils, by inducing the production of IL-8 in airway epithelial cells and of IL-1β in pulmonary macrophages. In addition, we showed that IL-1β, a proinflammatory cytokine that also stimulates the production of IL-8, was upregulated in both the IAD-Mast and the IAD-Neutro groups. The expression of IL-1β was, however, greater in the IAD-Neutro than in the IAD-Mast group. A previous study also found increased expression of IL-1β in the horses with IAD that have a BALF neutrophilia, but did not observe any increase in IL-8 expression. There is also evidence for the involvement of lipopolysaccharide (LPS) in the upregulation of IL-1β in RAO-susceptible horses exposed to Aspergillus fumigatus allergen by inhalation, but the role of LPS and molds in the pathogenesis of IAD is unknown.
Another cytokine we investigated was IL-10, which was upregulated in the cells from the BALF of the total IAD group. However, the increase in IL-10 expression observed in each of the 2 IAD groups separately was not significant. IL-10 promotes the development of a Th2 cytokine profile by inhibiting the IFN-γ production of T-lymphocytes particularly by the suppression of IL-12 production in accessory cells. Although IL-10 was only upregulated significantly in the IAD-total group, this finding relates well with our results on IFN-γ and IL-12 p35 expression in IAD.
Indeed, we found no upregulation of IFN-γ in both the IAD-Mast group and the IAD-Neutro group. This is in agreement with results from a recent study,but is dissimilar to what others found in Standardbred horses with IAD and EIPH. IL-12 p35 was, however, slightly upregulated in the IAD-Mast group, but this finding barely reached significance, suggesting that IL-12 is unlikely to play an important role in the pathophysiology of IAD.
Lastly, because the different cell types from the BALF were not separated and their cytokine production analyzed separately, we cannot conclude that the increased expression of cytokine from the Th2 family in IAD represents a Th2 immunologic response driven by T helper type 2 lymphocytes.
In the present study, we found upregulation of IL-5, IL-6, and IL-10 in horses with IAD concurrently with limited or no upregulation of IFN-γ and IL-12. These findings together strongly suggest an important role for cytokines from the Th2 family in IAD. We also suggest that a distinction be made to understand the pathophysiology of IAD between 2 different IAD phenotypes, namely IAD horses with mainly neutrophil versus mast cell accumulation in their BALF. Our hypothesis is supported by the present results showing a differential mRNA expression of IL-17, IL-4, IL-5, and IL-8 between these 2 groups. IL-17 was only up-regulated in the IAD-Neutro group, which suggests a specific role for IL-17 in the inflammatory reaction observed in IAD horses with neutrophil accumulation in their BALF.