Materials and Methods

Animals—Two groups of horses (RAO-susceptible and healthy control horses) were evaluated concurrently during this experiment. The RAO-susceptible group consisted of 2 mares and 4 geldings that had a median age of 17 years (range, 13 to 21) and weighed between 410 and 600 kg. This group was comprised of 3 Quarter Horses, 1 Paso Fino, 1 Dutch Warmblood, and 1 Arabian. All horses in this group consistently developed RAO when fed dusty hay while housed in a stable bedded with wood shavings.¹⁵ The RAO phenotype was characterized by the development of pulmonary neutrophilia (≥ 25% of the total nucleated cell count in the BALF) and an accentuated breathing effort, with the ΔPpl_max exceeding 15 cm H₂O.¹⁶ The control group consisted of 4 mares and 1 gelding that had a median age of 14 years (range, 11 to 16 years) and weighed between 500 and 540 kg. This group was comprised of 3 Thoroughbreds, 1 Dutch Warmblood, and 1 Quarter Horse cross. The control horses failed to develop pulmonary neutrophilia or an increase in ΔPpl_max when fed dusty hay in a stable bedded with wood shavings.

Several months before the experiment began, horses were dewormed by administration of ivermectin^a and vaccinated against tetanus, equine influenza, eastern equine encephalomyelitis, western equine encephalomyelitis,^b and rabies.^c Before the start of the experiment, all horses were confirmed healthy on the basis of results of physical, endoscopic, and thoracic radiographic examinations as well as results of hematologic and serum biochemical analyses. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Cornell University and were in accordance with guidelines established by the National Institutes of Health.

Experimental design—Both groups of horses (RAO-susceptible and control horses) were maintained on pasture for 3 months prior to the study to ensure the RAO-susceptible horses were in clinical remission at the start of the experiment. The experiment began when 1 randomly selected control horse and 1 randomly selected RAO-susceptible horse were moved to separate 4 × 4-m box stalls and fed dusty timothy-alfalfa hay. Horses were allowed ad libitum access to water. Twenty-four hours later, pulmonary function tests were performed and BALF cells were collected from each pair of horses. The ΔPpl_max values were measured in each unsedated horse by use of an esophageal balloon catheter. In each horse, the mean ΔPpl_max was obtained for 10 to 15 breaths.

Each horse was then sedated by administration of xylazine hydrochloride^d (0.6 mg/kg, IV), and bronchoalveolar lavage was performed with approximately 300 mL of warm (37°C) sterile saline (0.9% NaCl) solution. The saline solution was administered and subsequently aspirated through the endoscope biopsy channel of a 2.1-m videoendscope^e that was wedged in a sixth- to ninth-generation bronchiole of the right or left lung lobe. Before the lavage, airways were anesthetized by infusion of 25 mL of a solution of 2% lidocaine hydrochloride.^f Samples of BALF were collected in siliconized containers that were placed on ice for transport to the laboratory, where they were processed within 1 hour after collection. The horses were returned to their stalls and fed dusty hay for 13 additional days before being tested and undergoing sample collection again. The process was repeated for the next pair of control and RAO-susceptible horses until all horses had been evaluated. The results of the 2-week dusty hay challenge on pulmonary mononuclear gene expression have been reported.¹⁷

Preparation of BALF cells—Cells in samples of BALF were processed for cytologic analysis and in vitro incubation. For each BALF sample, a cytocentrifugation preparation was made by mixing 900 μL of BALF with 100 μL of 1% DL-dithiothreitol.^g Slides were stained,^h and differential cell counts were performed by examination of 200 cells. The remaining BALF was filtered through sterile gauze and centrifuged (200 × g for 15 minutes at 10°C) to obtain cell pellets for immunomagnetic separation.

Cell pellets were resuspended in 400 μL of PBSS containing 0.5% bovine serum albumin and 2mM EDTA.^g Approximately 5 × 10⁷ cells were incubated (15 minutes at 4°C) with 100 μL of murine monoclonal IgMⁱ (reactive against equine granulocytes), washed with buffer, and then incubated (15 minutes at 4°C) with 100 μL of rat anti-mouse IgM that contained magnetic beads.^j Cells were applied to columns,^k and unlabeled (negatively selected) cells were eluted and evaluated cytologically, as described previously. Cell separation was performed at 4°C.

Eluted cells were resuspended in RPMI medium^l with 10% bovine growth serum,^m and 5 × 10⁶ cells/mL were seeded in a 24-well culture plate.ⁿ Each well received 1 of the following 4 treatments: 100 μL of PBSS, 100 μL of LPS^g solution (final concentration, 10 μg/mL), 100 μL of a β-glucan^g (from brewer's yeast) solution (final concentration, 12.5 pg/mL), and 100 μL of a hay dust solution prepared by diluting the stock solution 1:100.

The hay dust stock solution was prepared as described elsewhere^17,18 by use of the same dusty hay that had been fed to induce RAO in the susceptible horses. Briefly, flakes of hay were agitated onto a clean surface and sieved through a grid (2 × 3 mm), and fine particles were separated from coarse particles by use of a dual-vortex vacuum cleaner.^o Ten milliliters of sterile saline solution was added to each gram of hay dust, and the solution was filtered through a 100-μm mesh screen. The hay dust solution was then exposed to γ-radiation (1.5 Mrad for 12 hours, with samples maintained on dry ice) to sterilize the solution^p and to prevent bacterial and fungal overgrowth of the macrophage-lymphocyte cultures. Preliminary data had revealed that irradiated hay dust increased gene expression of IL-8 in a manner qualitatively similar to that induced with nonirradiated hay dust solution. Fungal particulates in the stock solution that were ≤ 4 μm in diameter were counted with a hemocytometer. Endotoxin and β-glucan concentrations were measured by a commercial laboratory.^q Each milliliter of the stock hay dust solution contained 10⁸ fungal particulates, 2 μg of endotoxin, and 12.5 ng of β-glucan. Thus, each well that was treated with the hay dust solution (100 μL of a 1:100 dilution of the stock solution) contained 10⁵ fungal particulates, 2 ng of endotoxin, and 12.5 pg of β-glucan.

Depending on the number of pulmonary mononuclear cells available after immunomagnetic separation, replicates of the 4 treatments were performed whenever possible and used as backup samples. Cells were incubated for 24 hours at 37°C in 5% CO₂. Harvested cells were frozen at 80°C until gene expression was measured.

Measurement of gene expression—Gene expression in the treated mononuclear cells was measured by use of real-time reverse transcription PCR assays.^r Genes that were evaluated included those for IL-23 (the p19 and p40 subunits), IL-17, IL-8, CXCL2, and IL-1β. Expression of each target gene was standardized to the housekeeping gene for β-actin. For each sample of BALF cells for each treatment, the total RNA was extracted,^s genomic DNA was destroyed,^t and cDNA was synthesized.^u During cDNA synthesis of each sample, a negative control sample (ie, one that was lacking reverse transcriptase) was also prepared and reserved for subsequent PCR assays to confirm that genomic DNA had been destroyed. The PCR reaction mixtures contained 2.5 μL of cDNA and 25 μL of master mix, for a final volume of 27.5 μL. For each cDNA sample, triplicate reactions were carried out on each plate for detection of target genes. Negative control samples of nuclease-free water were included on each plate. Primer and probe sequences used were validated as described elsewhere.¹⁷

In a real-time PCR assay, the C_T represents the number of PCR cycles that take place before amplicon is initially detected, and this value ranges from 0 to 40. Gene expression was reported as the ΔC_T or as the fold-change in gene expression.^v The ΔC_T reflects the difference between the C_T of the target gene and the C_T of the housekeeping gene (β-actin). In general, the smaller the ΔC_T, the greater the amount of mRNA (cDNA) in the sample. The fold-change method is calculated as 2^−(ΔΔC_T). Within a specific group of horses, the ΔΔC_T of a target gene represented the difference between the ΔC_T for a specific treatment (eg, LPS-treated cells) relative to the ΔC_T of the PBSS (control) treatment.

Statistical analysis—Differences between control and RAO-susceptible horses were detected by use of a Wilcoxon rank sum test. A value of P ≤ 0.05 was considered significant. Because multiple comparisons were made, a Bonferroni correction was applied. This was calculated by dividing α (0.05) by the number of genes that were compared in each objective. All computations were performed by use of a statistical software program.^w Results are reported as mean ± SD.

Results

Pulmonary function tests and cellular composition of BALF—After 24 hours of stall confinement and exposure to dusty hay, the RAO-susceptible horses had a mean ± SD ΔPpl_max of 15.7 ± 17.3 cm H₂O, compared with 5.2 ± 1.3 cm H₂O for control horses (P = 0.20). The mean total nucleated cell count in BALF samples from RAO-susceptible horses (313 ± 177 cells/μL) was not significantly (P = 0.78) different from that in BALF samples from the control horses (316 ± 81 cells/μL). The RAO-susceptible horses developed an increase in mean BALF neutrophil percentage (23 ± 15%), but this value was not significantly (P = 0.17) different from that of the control horses (13 ± 8%). That the RAO-susceptible horses were in the developmental phase of RAO (and representative of the RAO phenotype) was subsequently confirmed by detection of additional increases in the median BALF neutrophil percentage (41%) and the mean ΔPpl_max (44 cm H₂O) after 14 days of continuous dusty hay exposure.¹⁷

Immunomagnetic removal of neutrophils—The mean percentage of macrophages and lymphocytes in BALF samples was 54 ± 16% and 33 ± 12%, respectively, for the control horses and 42 ± 15% and 32 ± 19%, respectively, for the RAO-susceptible horses. After immunomagnetic removal of neutrophils, the mean percentages of macrophages and lymphocytes were 57 ± 18% and 43 ± 18%, respectively, for the control horses and 55 ± 11% and 44 ± 12%, respectively, for the RAO-susceptible horses. The mean percentages of macrophages and lymphocytes used for the in vitro studies were not different between the 2 groups of horses (P = 0.77 and P = 0.91, respectively).

Within-group gene expression of pulmonary mononuclear cells—Within-group increases in target gene expression after treatment with solutions of hay dust, LPS, or β-glucan relative to expression of the respective gene in PBSS were summarized (Table 1). In pulmonary mononuclear cells isolated from both groups of horses, hay dust treatment resulted in a > 2-fold upregulation in the expression of all target genes. The LPS treatment increased gene expression of IL-23p40, IL-8, CXCL2, and IL-1β > 3-fold in RAO-susceptible horses. With the exception of IL-23p40, the gene expression of which increased nearly 10-fold, LPS treatment failed to upregulate target gene expression in cells from control horses more than 2-fold. Relative to cell treatment with PBSS, cell treatment with β-glucan was not associated with increased target gene expression in either group of horses, with the exception of the p40 subunit of IL-23, the gene expression of which increased nearly 3-fold in control horses.

Between-group differences in gene expression of pulmonary mononuclear cells—Box-and-whisker plots were constructed to depict the results of standardized gene expression of IL-23p19, IL-23p40, and IL-17 in pulmonary mononuclear cells after each of the 4 treatments (solutions of PBS, LPS, β-glucan, and hay dust; Figure 1). There were no significant (P > 0.20) differences in gene expression in cytokines of pulmonary mononuclear cells isolated from the 2 groups of horses after 24 hours of treatment with PBSS. Furthermore, there were also no significant (P > 0.20) differences in gene expression of IL-23 or IL-17 between the 2 groups after 24-hour exposure to hay dust, LPS, or β-glucan.

View Full Size

Box-and-whisker plots of degrees of gene expression of IL-23 subunit p19 (A), IL-23 subunit p40 (B), and IL-17 (C) in pulmonary mononuclear cells isolated from BALF from 5 healthy horses (shaded boxes) and 6 RAO-susceptible horses (white boxes) after incubation for 24 hours with PBSS (control solution), hay dust solution (HDS), LPS solution, or β-glucan solution. Expression of each target gene (represented by ΔC_T) was standardized to the gene expression of β-actin. Boxes represent the 25th to 75th percentiles, the horizontal bar within each box represents the median value, and the whiskers represent the 95th percentiles. The smaller the ΔC_T value, the greater the amount of cDNA (ie, mRNA) in the sample. Notice that scales differ among the 3 cytokines. Gene expression of IL-17, IL-23p19, and IL-23p40 did not differ between the 2 groups of horses for any cell treatment. Outlier observations are represented by asterisks and circles.

Citation: American Journal of Veterinary Research 70, 10; 10.2460/ajvr.70.10.1277

• Download Figure
• Download figure as PowerPoint slide

Box-and-whisker plots were created to depict the ΔC_T values for gene expression of IL-8, CXCL2, and IL-1β in pulmonary mononuclear cells after each of the 4 treatments (Figure 2). Treatment with PBSS was not associated with a difference between the 2 groups with respect to expression of IL-8 (P = 0.02), CXCL2 (P = 0.06), and IL-1β (P = 0.08). After cells were exposed to hay dust or LPS, there were no differences between the 2 groups in the magnitude of upregulation of expression of these 3 chemokines (all values of P > 0.32). After treatment with β-glucan, gene expression of IL-8 (P = 0.008) and IL-1β (P = 0.01) was greater in BALF cells from control horses than in BALF cells from RAO-susceptible horses. However, for either group, this degree of expression did not differ from that associated with the PBSS treatment.

View Full Size

Box-and-whisker plots of degrees of gene expression of IL-8 (A), CXCL2 (B), and IL-1β (C) in pulmonary mononuclear cells isolated from BALF from healthy horses (shaded boxes) and RAO-susceptible horses (white boxes) after incubation for 24 hours with PBSS (control solution), hay dust solution, LPS solution, or β-glucan solution. Expression of IL-8, CXCL2, and IL-1β did not differ between the 2 groups of horses for PBSS, hay dust, or LPS treatment. †Gene expression for control horses is significantly (P < 0.017) greater than that for RAO-susceptible horses, but did not differ between β-glucan and PBSS treatment for either group of horses. See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 70, 10; 10.2460/ajvr.70.10.1277

• Download Figure
• Download figure as PowerPoint slide

Discussion

The present study was undertaken to investigate the potential role of pulmonary mononuclear cells in inciting airway neutrophilia during the early stages of RAO. The experimental design enabled examination of the inherent (cell treatment with PBSS) and ligand-induced (cell treatment with solutions of hay dust, LPS, and β-glucan) cytokine gene expression in cultured pulmonary macrophages and lymphocytes isolated from RAO-susceptible and control horses. The data suggested that neither stimulated nor nonstimulated pulmonary mononuclear cells isolated from RAO-susceptible horses during the developmental phase of the disease had a greater degree of gene expression of selected cytokines and chemokines than did cells isolated from control horses. The in vitro responses of equivalent numbers of pulmonary mononuclear cells isolated from the 2 groups of horses were reported. Because these data and those of others⁵ have suggested that the total numbers of nucleated cells, pulmonary lymphocytes, and alveolar macrophages in the BALF of RAO-susceptible horses were not greater than those in control horses during the first 24 hours of dusty hay exposure, the in vitro results most likely reflected the in situ responses (devoid of the neutrophilic contributions).

In humans, IL-23 and IL-17 play a prominent role in the pathogenesis of chronic inflammatory disorders such as rheumatoid arthritis, Crohn's disease, multiple sclerosis, psoriasis, and asthma.10,14,19,20 With regard to respiratory disease, humans with severe asthma or with nonatopic, neutrophilic-dominated asthma have increased concentrations of IL-17 mRNA or IL-17 protein in the lungs, sputum, BALF, and serum, compared with concentrations in healthy humans. Such increases in IL-17 concentration correlate with the severity of airway hypersensitivity.²¹ In horses chronically affected with RAO (ie, horses that have been exposed to dusty hay for ≥ 10 days), mRNA expression of IL-17 in BALF cells is increased, compared with expression in BALF cells of control horses.^15,22 However, because equine neutrophils also express IL-17 mRNA, the augmented gene expression in chronically affected horses probably reflects that attributable to the high granulocyte numbers in the BALF of diseased horses.¹⁵ Indeed, when equal numbers of pulmonary mononuclear cells from horses chronically affected with RAO are examined in vitro,¹⁷ endogenous or ligand-induced gene expression of IL-23p19, IL-23p40, or IL-17 is no different from that of cells from control horses. Although the IL-23IL-17 axis does not appear to be dysregulated during the chronic stages of RAO, prior to the present investigation, the kinetics of IL-23 and IL-17 expression during the acute phase of RAO had not been examined.

In murine models of pulmonary inflammation, there is an immediate increase in gene expression of pulmonary IL-23 and IL-17 in response to exposure to microbial products. For example, within 2 hours after intranasal instillation of LPS (10 μg) or peptidoglycan (50 μg) in mice, copy numbers of IL-23 mRNA in lung tissue and protein concentrations of IL-23 in BALF increase by > 80%.²³ These effects are presumably mediated through peptidoglycan-induced activation of TLR2 and LPS-induced signaling through TLR4.⁹ Furthermore, within 24 hours of intranasal administration of recombinant IL-23 (1 to 3 μg), increases are evident in mRNA copy numbers and protein concentrations of IL-17 in murine lung cells and BALF, respectively.²³ Such cytokine changes appear to contribute to the ensuing airway neutrophilia because inflammatory cell infiltration is prevented when mice are pretreated IP with anti–IL-17 antibody.²³

Acute activation of the pulmonary IL-23IL-17 axis also occurs after pulmonary exposure to fungal organisms or fungal cell wall components, and this response is also mediated via TLR2 and TLR4 signaling.²⁴ In mice, intranasal instillation of Aspergillus or Candida organisms reportedly increases the IL-23 protein concentration in dendritic cells and neutrophils and increases IL-17 protein synthesis in CD4+ Th-17 cells, γδ+ T cells, and neutrophils.^25,26 Furthermore, 24-hour in vitro exposure to β-glucan of mononuclear cells isolated from the peripherally obtained blood of healthy humans reportedly increases the protein concentration of IL-23 in a dose-dependent manner.²⁷ However, compared with the effect of peptidoglycan treatment at an equivalent concentration (500 μg), β-glucan treatment induces a much smaller increase in the protein concentration of IL-23.²⁷

Exposure of healthy humans to organic dust (generated in swine barns) for as little as 4 hours can result in an increase in copy numbers of IL-17 mRNA and protein concentration of IL-17 in BALF lymphocytes and an increase in the number of BALF neutrophils 24 hours after exposure.²⁸ Alterations in IL-23 protein concentrations or mRNA copy numbers were not measured in that study. In the study reported here, hay dust, which is a composite of fungal particulates, endotoxin, β-glucan, and other matter,¹⁸ considerably upregulated gene expression of IL-23p19, IL-23p40, and IL-17 in pulmonary mononuclear cells. However, the increases detected in cells from the RAO-susceptible horses were no different than those detected in cells from the control horses. Given this result and the finding that the inherent gene expression of these cytokines after BALF cells were treated with PBSS was no different between the 2 groups, we concluded that excessive activation of the IL-23IL-17 axis in lungs of RAO-susceptible horses was not an inciting cause of their airway neutrophilia. Interestingly, our data also revealed that BALF cells from both groups failed to upregulate gene expression of IL-23p19 or IL-17 when exposed to LPS (10 μg) or β-glucan (13 pg). This lack of upregulation suggested that LPS or β-glucan alone was not responsible for changes in gene expression associated with hay dust exposure. Dose-dependent alterations remain to be defined for gene expression of IL-23 and IL-17 in equine pulmonary mononuclear cells treated with various concentrations of LPS or β-glucan.

When healthy humans inhale organic dusts generated in animal-rearing or grain-processing facilities, the ensuing neutrophilic inflammation is accompanied by increased protein concentrations of IL-6, IL-8, IL-1, and TNF-α in BALF.^29,30 These proinflammatory cytokines are derived from immune-system cells (eg, alveolar macrophages, dendritic cells, and neutrophils) and non–immune-system cells (eg, airway epithelial cells)³¹ that respond primarily to the inhaled organic dust components or secondarily, as with a paracrine response, to cytokines initially released (eg, TNF-α, IL-1β, or IL-17).

In an ex vivo study,²⁹ human alveolar macrophages exposed to swine barn dust, LPS, or β-glucan yielded dose-dependent increases in IL-8 protein concentrations. However, the 3 ligands were not equipotent with respect to the degree of concentration increase. Approximately 100 μg of LPS was required to obtain the same IL-8 protein concentration as that induced with 100 μg of swine dust that contained 2 ng of endotoxin. This finding suggests that organic dust components other than LPS increase chemokine production.²⁹ Furthermore, the IL-8 protein concentrations yielded after cell treatment with β-glucan (10 or 100 μg) were modest, compared with those yielded after treatment with LPS (100 μg/mL) or swine dust (100 μg).²⁹

In the present study, the results we obtained were qualitatively similar to results of the aforementioned human study.²⁹ Regardless of whether BALF cells were isolated from RAO-susceptible or healthy horses, cell treatment with hay dust solution (which contained 2 ng of endotoxin) resulted in upregulation of gene expression of IL-8, CXCL2, or IL-1β to a degree approximately twice that induced by cell treatment with 10 μg of LPS. In addition, β-glucan applied to cells at the same concentration as that in the hay dust solution (13 pg) failed to upregulate chemokine gene expression, compared with expression after treatment with PBSS. Our experimental design did not permit us to identify other hay dust components that might have contributed to the augmented hay dust responses, but potential candidates included peptidoglycan, which commonly exists in bioaerosols,^32,33 and reactive oxygen species, which are generated during phagocytosis by macrophages.³⁴ It is also possible that the TNF-α, IL-1β, or both that were initially generated after TLR activation and nuclear factor-κB signaling further enhanced chemokine gene transcription via activation of their own cell surface–associated receptors.^35–37 Nevertheless, whatever the mechanisms involved in proinflammatory gene transcription in our study, those mechanisms that were operational in pulmonary mononuclear cells isolated from RAO-susceptible horses and exposed to hay dust were not different from those in cells of control horses.

Our data failed to support the hypothesis that following 24 hours of dusty hay exposure, the airway neutrophilia that developed in RAO-susceptible horses was a consequence of an imbalance in the IL-23IL-17 axis or an exaggerated chemokine (IL-8, CXCL2, or IL-1β) response of pulmonary mononuclear cells to hay dust. However, when interpreting these results, it is important to consider that changes in gene expression rather than in protein concentrations were detected. Although we found a good correlation between IL-8 gene and protein expression in BALF cells of RAO-affected horses in another study,¹⁵ similar congruence between IL-23, IL-17, CXCL2, and IL-1β mRNA and protein concentrations remains to be reported.