Systemic inflammation in horses with heaves is poorly characterized.
Systemic inflammation in horses with heaves is poorly characterized.
To assess acute phase proteins (APP) and inflammatory cytokine profiles in serum of healthy horses and horses with heaves.
Six healthy horses and 6 heaves-affected horses belonging to the University of Montreal.
Prospective, observational study. Healthy and heaves-affected control horses were exposed to a 30-day natural challenge with hay and straw to induce clinical exacerbation of heaves. Serum samples were obtained by venipuncture before (T0) as well as after 7 (T7) and 30 days (T30) of stabling. Serum APP (haptoglobin, serum amyloid A protein [SAA] and C-reactive protein [CRP]) and cytokines (IL-2, IL-4, IFN-α, IL-10, IFN-γ, and CCL-2) were measured using singleplex or multiplex ELISA.
Serum haptoglobin concentrations were significantly higher in heaves-affected horses at all time points with no overlap with those of healthy controls. They were also significantly increased by antigen challenge in both controls (T7) and horses with heaves (T7 and T30). Serum SAA was detected more frequently in heaves-affected horses compared with healthy controls at T7. There was no difference in serum concentrations of CRP, IL-10, IFN-γ, and CCL-2 between groups, whereas IL-2, IL-4, and IFN-α remained undetectable in all samples.
In heaves, haptoglobin is a marker of both acute and chronic systemic inflammation, whereas high concentrations of SAA indicate acute inflammation.
acute phase proteins
nuclear factor κB
serum amyloid A protein
lymphocyte T helper
tumor necrosis factor
maximal changes in pleural pressure
Heaves affects mature horses and is characterized by episodes of labored breathing at rest caused by bronchoconstriction, neutrophilic inflammation of the airways, mucus accumulation, and airway remodeling. Neutrophilic airway inflammation in horses with heaves can be reversed by antigen withdrawal. However, evidence suggests that the inflammatory processes are not completely shut down as residual peripheral airway bronchoconstriction, elevated smooth muscle cell turnover surrounding the airways, and higher nuclear factor κB (NF-κB) activity are observable in asymptomatic horses with heaves. Although heaves is a disease of the airways, peripheral blood leukocyte activation[5-7] and increased concentration of circulatory inflammatory mediators[8-10] have been observed in affected horses during disease exacerbation suggesting that the inflammatory process might not be limited to the lungs.
There is an increasing interest for the systemic component of chronic inflammatory diseases in human and veterinary medicine. It is thought that systemic inflammation magnifies the local response and leads to immunological reactions distant from the lungs. Therefore, it is hypothesized that therapies targeting bronchospasm and local inflammation in chronic airway diseases using inhaled bronchodilators and corticosteroids, respectively, might not be optimal for long-term treatment and health improvement. Asthma, a disease that shares many pathophysiological features with equine heaves, is considered to be a systemic disease, as an increase in several inflammatory markers has been observed in the blood of affected patients. These include immunity-related mediators (eg, cytokines, eicosanoids, and cyclooxygenase products, IgEs) and the acute phase markers CRP,[13, 14] haptoglobin, fibrinogen, and SAA. Acute phase proteins (APP) are typically induced by proinflammatory cytokines and are produced by the liver after trauma or infection, but are also detected in a number of noninfectious conditions. Systemic inflammation in patients with chronic airway diseases is thought to contribute to comorbidities.[11, 16-19] In equine medicine, APPs are increasingly used to determine prognosis and monitor response to treatment in infectious and noninfectious diseases or traumatic injuries (reviewed in Ref. ).
The aim of this study was to better characterize the systemic inflammation present in horses affected with heaves. Selected APPs as well as a panel of cytokines were quantified in the serum of horses with heaves in the asymptomatic and symptomatic clinical phases as well as healthy controls kept in a similar environment.
Control horses (6 mares, mixed breeds) had no record of lung disease, weighed 490 ± 5 kg (mean ± SD, range 470–533 kg), and were 15.7 ± 1.4 years old (range 11–20). Horses with heaves (2 geldings and 4 mares, mixed breeds) had been diagnosed on the basis of their history of respiratory disease characterized by the development of airway obstruction and neutrophilic inflammation upon exposure to moldy hay. They weighed 501 ± 10 kg (range 442–550) and were 21 ± 1.2 years old (range 18–26). All horses were kept in the same environment at least 1 week before, and during the study, they were dewormed regularly and vaccinated annually (8 months before the study). All experimental procedures were performed in accordance with the Canadian Council of Animal Care and approved by the University of Montréal Animal Care Committee.
The study was performed during the months of January and February. Heaves-susceptible horses were kept in a paddock and fed alfalfa pellets and sweet feed twice a day for 2 months to ensure clinical remission of the disease. Control horses were then added to the herd for an additional 7 days, after which all horse were stabled and exposed to hay and straw for 30 days. Blood was collected before (day 0: T0), as well as 7 days (T7) and 30 days (T30) after the beginning of antigenic challenge. Pulmonary function tests and bronchoalveolar lavages (BAL) were performed after blood sampling procedures at T0 and T30, to avoid stress-related effects on APP or cytokine concentrations.
Blood was obtained by venipuncture into dry sterile tubes and allowed to clot at room temperature before being centrifuged. Serum was stored in aliquots at −80°C until the time of APP analysis. Samples were transferred to −20°C before cytokine quantification.
Bronchoalveolar lavages was performed on standing, sedated horses (xylazine 0.2–0.5 mg/kg IV, butorphanol 10–20 μg/kg IV, and detomidine 5–10 μg/kg IV) using a flexible videoendoscope (length 1.6 m, 13.3 mm) and two 250 mL boluses of prewarmed sterile isotonic saline instilled into the bronchus, as previously described. Differential cell counts were obtained from 400 cells stained with a modified Wright's solution.1
Maximal changes in pleural pressure (ΔPpl), lung resistance (RL), and elastance (EL) were obtained using a heated pneumotacograph and esophageal balloon, and a dedicated computer software, as previously described.
Clinical scoring was performed by blind observers during stabling to monitor the development of clinical signs in both groups. Briefly, a score from 0 to 4 was attributed to abdominal movement (0: no abdominal effort; 4: severe and marked abdominal movement) and nasal flaring (0: no flaring; 4: severe, continuous flaring). Abdominal and nostril scores were added for a maximal score of 8. Scores ≥5 indicated the presence of airway dysfunction.
Acute phase markers were quantified in serum samples using commercial ELISA kits according to the manufacturer's instructions. C-reactive protein (CRP) was quantified using Kamiya Biomedical horse CRP ELISA and haptoglobin was quantified using Kamiya Biomedical Horse haptoglobin ELISA. Serum amyloid A (SAA) was quantified using an ELISA test (Tridelta Development LTD “PHASE” TM Serum Amyloid A Assay) previously validated in horses. Plates were washed using an automatic plate washer (ELX50 Autostrip washer2) and absorbance was obtained using Power Wave X340 plate reader.3 The assay standard curves ranged from 3.13 to 100 ng/mL for haptoglobin (serum dilution 1 : 32,000), 0.3125–20 ng/mL for SAA (serum dilution 1 : 2,000), and 5–320 ng/mL for CRP (serum dilution 1 : 1,000).
Cytokines were quantified in undiluted serum samples from healthy and heaves-affected horses at T0, T7, and T30 using species-specific singleplex ELISA (CCL2 and IL-2) or multiplex ELISA (IL-4, IL-10, IFN-α, and IFN-γ), as described previously. In brief, monoclonal antibodies against recombinant equine cytokines were produced either in a mammalian expression system using the equine IgG1 heavy chain (anti-equine IL-4 coupled to bead 33; anti-equine IL-10 clone 492-2 coupled to bead 34; anti-IFN-α clone 29B coupled to bead 35) or anti-CCL2 (coupled to bead 37), anti-IFN-γ3 (coupled to bead 36), and anti-IL-2 (Wagner et al, unpublished data; polyclonal, coupled to bead 37) were coupled to fluorescent beads. Mixed coupled antibodies (final concentration of 105 beads/mL each) were added to each well of the microtiter plate before adding samples or standard curves. The latter were prepared using serial 3-fold dilutions of supernatants containing recombinant cytokine/IgG-fusion proteins. Biotinylated antibodies specific for the cytokines were then added (anti-equine IL-10 clone 165-2; anti-equine IFN-α clone 240–2; anti-equine IL-4 clone 25 [Wagner et al, unpublished data]; anti-CCL2 [Wagner et al, unpublished data]; anti-IL-24 and anti-INFγ5) followed by streptavidin–phycoerythrin.6 The assay was analyzed in a Luminex IS 100 instrument7 and the data were reported as median fluorescence intensities. Calculation of the cytokine concentrations in samples was performed according to the logistic 5p formula (y = a + b/(1 + (x/c)^d)^f).8 The multiplex intra- and interassay variability has been described elsewhere. The detection threshold was >15 pg/mL for IL-2, INFα, and IL-10, >40 pg/mL for IL-4, 1 pg/mL for CCL2, and >10 U/mL for INFγ.
Data were analyzed using GraphPad Prism 5 software.9 An arbitrary value corresponding to one tenth of the lowest standard concentration was attributed to samples with undetectable APPs or cytokines. Samples with OD values exceeding the range of quantification were attributed a value of twice the highest standard concentration. Serum cytokine and APP data were analyzed using repeated-measure two-way variance analysis (ANOVA) on log10-transformed values with groups (controls, heaves) as a between-subject factor and time (T0, T7, and T30) as a within-subject factor. A one-way ANOVA was performed to determine which time point accounted for the significant effect of time on the average serum CCL2 concentrations in the 2 groups combined. For SAA and IL-10, Fisher's exact test was used to evaluate different frequencies of detection in serum samples from healthy and heaves-affected horses. In addition, values were analyzed for association between markers of systemic inflammation and age, lung function, and BALF neutrophilia in the 2 groups, separately. Spearman correlation P values were corrected for multiple testing using the Bonferroni method. Otherwise, differences were considered significant when P values were less than .05.
Horses with heaves were asymptomatic and had normal BALF cytology (BALF neutrophils <5%) while living outside in a paddock (T0). However, the mean neutrophil percentage at baseline was significantly higher in heaves-affected horses horses (mean ± SD, 3.0 ± 1.66%, Fig 1A) compared with control horses (0.61 ± 0.52%). Two of the 6 heaves-affected horses had evidence of mild airway obstruction at T0 (RL < 1.5 cmH2O/L/s, EL < 1.1 cmH2O/L, ΔPplmax, or both <12 cmH2O) and there were significant differences between groups in mean baseline maximal changes in pleural pressure (ΔPPLmax: 5.0 ± 1.46 versus 9.3 ± 2.37, for control and heaves, respectively, Fig 1C) and lung resistance (RL: 0.41 ± 0.19 versus 0.81 ± 0.39 cmH2O/L/s, for control and heaves, respectively, Fig 1D).
Exposure to hay for 30 days resulted in airway obstruction in horse with heaves only, as indicated by their increase in values of lung function parameters (Fig 1). Clinical scores in horses with heaves were also significantly higher than control horses at T7 (5.0 ± 1.7 versus 2.25 ± 0.4) and T30 (5.6 ± 1.1 versus 2.2 ± 0.4) (not shown). Control horses showed a significant increase in BALF neutrophils at T30 when compared with baseline (12.3 ± 6.1% versus 0.61 ± 0.52%), although this increase was significantly less than that observed in heaves-affected horses (3.0 ± 1.7% to 37 ± 3.7% from T0 to T30, respectively).
Haptoglobin was detected in all serum samples. Mean concentrations were significantly higher in serum from horses affected with heaves compared with controls at all time points, with no overlap in data points between groups (Fig 2A). The mean ± SD serum haptoglobin concentrations in controls versus heaves-affected horses were, respectively, 410 ± 211 versus 1090 ± 212 μg/mL at T0, 795 ± 379 versus 1663 ± 280 μg/mL at T7, and 448 ± 188 versus 1446 ±528 μg/mL at T30. Horses with heaves had significantly higher haptoglobin serum concentrations at T7 and T30 compared to T0, whereas control horses had significantly higher serum haptoglobin concentrations at T7 compared with other time points.
C-reactive protein remained undetected in serum samples from 1 control horse and 1 heaves-affected horse at all time points. There was no significant difference in mean serum CRP concentrations between healthy controls and heaves-affected horses at any time point (Fig 2B).
Serum amyloid A remained undetectable in all serum samples from control horses except in 3 samples obtained from 2 horses. Six serum samples from 4 heaves-affected horses also had SAA concentrations under the detection threshold. Serum samples from 3 heaves-affected horses had SAA concentrations exceeding the assay quantification range at T7 (n = 1), and at T30 (n = 2) and were attributed a value of 80 μg/mL. There was a significant difference in the frequency of SAA detection (Fisher's exact test, P = .015) and in the mean serum concentrations between heaves-affected and control horses at T7 (mean ± SD: 15.75 ± 31.51 versus 3.22 ± 7.73 μg/mL; two-way ANOVA, P < .05) (Fig 2C).
Chemokine ligand-2 remained undetectable in serum samples from 1 healthy and 1 heaves-affected horse at all time points. There was an overall significant decrease in CCL-2 concentrations between T0 and T7 in serum samples from both groups of horses (8.34 ± 6.32 versus 5.31 ± 5.74 pg/mL, respectively, one-way ANOVA). There was no significant difference between groups or time points in IL-10 and IFN-γ serum concentrations (Fig 3). IFN-α, IL-4, and IL-2 remained undetectable in all serum samples.
Neither lung function parameters nor BALF neutrophil percentages were significantly correlated with serum APPs or cytokines concentrations in heaves-affected horses (Table 1). However, CCL2 concentrations were positively associated with lung resistance (rs = 0.89, P = .03) and ΔPPLmax (rs = 0.84, P = .058) at T30, but did not reached the Bonferroni-corrected statistical significance threshold (P < .0125). Similarly, INFγ was positively associated with neutrophil percentages at baseline (rs = 0.93, P = .0167). None of the APP concentrations or lung function and inflammation parameters significantly correlated with age.
|% BAL Neutrophils||E L||R L||ΔPPLmax|
|Hapto-globin||−0.20 (.71)||0.54 (.30)||−0.058 (.92)||−0.14 (.80)||0.37 (.50)||−0.77 (.10)||0.086 (.92)||−0.26 (.66)|
|SAA||0.34 (.50)||−0.6789 (.1361)||0.28 (.56)||−0.093 (.92)||0.091 (.92)||−0.37 (.50)||−0.03 (1.00)||0.031 (1.00)|
|CRP||−0.35 (.50)||−0.66 (.18)||0.029 (1.00)||0.029 (1.00)||−0.20 (.71)||−0.029 (1.00)||−0.31 (.56)||0.086 (.92)|
|IFN-γ||0.93 (.0167)||−0.1429 (.8028)||0.23 (.66)||−0.20 (.71)||0.26 (.66)||0.26 (.66)||0.43 (.42)||0.086 (.92)|
|CCL2||−0.64 (.18)||−0.41 (.42)||0.41 (.42)||0.890 (.033)||−0.1429 (.8028)||0.64 (.18)||0.26 (.66)||0.84 (.058)|
This study shows that proteins of the acute phase are increased in the peripheral blood of horses with heaves. More specifically, serum haptoglobin concentrations were found to be above those of healthy horses in all clinical phases of the disease, and were further increased by antigenic challenge. SAA was also transiently increased by antigen challenge only in heaves-affected horses; however, these markers did not correlate with lung function parameters or BALF neutrophil percentages. In addition, this study reveals that CRP is not a good marker of chronic inflammatory airway disease in horses and that none of the cytokines detected in serum samples in this study were upregulated in heaves. In addition, these results suggest that the presence of subclinical airway inflammation in heaves-susceptible horses even with strict environmental dust control may lead to sustained systemic inflammation (ie, haptoglobin synthesis) or inversely, residual airway inflammation may derive from persistent systemic inflammation if it reveals to be an intrinsic feature of heaves pathology.
The results of this study support and further extend previous reports suggesting that systemic inflammation is a component of clinical exacerbation of heaves (reviewed by Leclere et al). Herein, both SAA and haptoglobin, but not CRP, were induced with antigenic challenge in heaves-affected horses. The results also indicate that systemic inflammation remains present during clinical remission of the disease as high haptoglobin concentrations were found in asymptomatic heaves-susceptible horses. This is in agreement with a recent study that observed high serum TNF concentrations in heaves-susceptible horses kept in a low dust environment and free of clinical signs. Further support for a persistent systemic inflammation in asymptomatic horses is the increased adhesion of peripheral blood neutrophils to fibronectin-coated plastic and priming for proinflammatory cytokine expression in response to bacterial-derived products.
Haptoglobin has been shown to be increased 5–10 fold in horse serum as a result of inflammation and infection. Its primary recognized role is to scavenge hemoglobin released into the circulation because of hemolysis or normal red blood cell turnover. Equine haptoglobin is a glycoprotein with an alpha and beta chain subunit structure, resembling human haptoglobin type 1–1. Normal concentrations in adult horses are ~1,000 μg/mL, whereas higher concentrations are reported in foals,[24, 35] suggesting that haptoglobin concentrations decrease with age. In this study, haptoglobin concentrations were lower than reported previously, which may be attributable to the fact that the horses studied were aged horses. Haptoglobin concentrations in asymptomatic heaves-susceptible horses were ~2.5 fold higher than control horses at T0, and rose up to ~3 fold after exposure to natural antigenic challenge for 30 days. This is similar to findings in human patients with asthma or chronic obstructive pulmonary disease.[15, 36] Importantly, serum haptoglobin concentrations in heaves-affected horses did not overlap those of controls kept under identical environmental conditions, suggesting that haptoglobin may be used as a biological marker for heaves. Interestingly, control horses also showed an increase in haptoglobin serum concentrations after hay exposure. This suggests that the magnitude, but not the uniqueness of the systemic inflammatory response distinguishes heaves-affected from healthy horses. Of note, however, haptoglobin concentrations returned to baseline in controls at T30, reinforcing the hypothesis that heaves underlies a defect in negative feedback mechanisms of inflammation. Monitoring of haptoglobin in horses with lower inflammatory airway diseases may possibly be of help in selecting therapies, as haptoglobin concentrations are higher in serum of asthmatic patients refractory to β2-agonist treatment and requiring systemic corticosteroids.
The main source of haptoglobin in acute phase responses is the liver. Interestingly, a heavily glycosylated form of haptoglobin was found to be synthesized and stored by human neutrophils in specific granules during their maturation and released after their activation. It may be hypothesized that peripheral blood, airway neutrophil degranulation, or both after activation by inflammatory factors present in the blood (eg, TNF[10, 31]) may constitute a source of haptoglobin in heaves-affected horses. Studies have shown that haptoglobin is largely induced in the lungs of mice after LPS challenge and produced by human alveolar macrophages, alveolar epithelial cells type II, and bronchiolar cells. As plasmatic haptoglobin decreases neutrophil functions through binding to specific membrane receptors and counteracting calcium signalization and respiratory burst activation, it may be hypothesized that this APP plays a protective role and prevents overwhelming inflammation.
Serum amyloid A is the most commonly APP evaluated in the horses, as it is induced 10–1,000 fold during inflammation. It has previously been used as a blood marker of lower airway inflammation caused by viral or bacterial infections. In humans, both serum and sputum SAA is increased in asthma. The production of local SAA by the airway tissues was also suggested. Herein, all heaves-affected horses had detectable serum SAA after 7 days of challenge with increases in concentrations up to at least 50 times those of baseline, whereas it was detected in only 1 control horse serum at this time point. These results are in agreement with previous reports indicating that blood concentrations of SAA are very low or undetectable in healthy individuals. A limitation of this study is the assignment of arbitrary values to samples with very high or low concentrations, decreasing the power of statistical analysis. Despite uncertainty about the absolute protein concentrations, differences in biological significance between groups at T7 were clearly highlighted. Interestingly, it has recently been shown that SAA can promote airway neutrophilia, activate neutrophils, and retroactively, be degraded by their granule content. SAA is also rapidly degraded by the liver.
The evaluation of the cytokine profile associated with heaves has been the focus of intense research for the last decade and has generated conflicting results.[49-51] This is the first study to our knowledge that characterizes multiple serum cytokines at the protein level in heaves-affected horses. We found that only IL-10, CCL2, and IFN-γ were detectable in a majority of horses, whereas IL-4, IFN-α, and IL-2 remained under detection levels. The latter cytokines are unlikely to contribute to the disease through systemic circulation, as the ELISA assay detects cytokines in the physiological range (low pg/mL). Correlation analysis suggested that CCL2 (also named MCP-1), a chemokine expressed by stimulated human and murine neutrophils promoting the development of Th17-type or Th2-type adaptive immunity, respectively, was positively associated with lung resistance (rs = 0.89) at T30 in heaves-affected horses. Similarly, INFγ showed a positive association with BALF neutrophil percentages in heaves-affected horses in clinical remission (rs = 0.93). Indeed, this Th1-associated cytokine is known for its role in promoting neutrophilic inflammation and the activation of neutrophils. As multiple statistical tests were performed in a relatively small cohort of animals and because cytokine concentration did not vary according to lung obstruction or inflammation, one must be careful in interpreting these results. Further exploration is required to identify the source of these cytokines and their significance in the disease.
Whether systemic inflammation represents an “overspill” of lung inflammation into the bloodstream or an intrinsic feature of heaves remains to be elucidated. In this study, as neither the lung function parameters nor the BALF cytology were correlated with the concentration of APPs, the causal relationship between those events could not be ascertained. Nevertheless, it has been shown in human patients that systemic inflammation in young adults is associated with decreased lung function independently of the presence of asthma, smoking or obesity. Hence, systemic inflammation has been proposed to predate the development of chronic lung diseases, and this hypothesis was recently verified in a 15-year prospective study. From these, it may be hypothesized that the presence of systemic inflammation earlier in life may have contributed to the development of heaves in these horses. Long-term prospective studies investigating systemic inflammatory markers in young horses from heaves-prone families and the development of the disease would also benefit in resolving this issue. In addition, studies where serum APPs are quantified in symptomatic horses treated with inhaled or systemic corticosteroids may give valuable information.
In conclusion, results of this study indicate that heaves is associated with increased serum concentrations of the acute phase proteins, SAA, and haptoglobin. These findings confirm previous assumption that heaves involves systemic inflammatory changes, in both the remission and exacerbation phases. Further investigation is required to assess the usefulness of these markers for the evaluation of response to treatments and the consequences of systemic inflammation on the development of collateral pathologies in affected horses.
Funding: This study was supported by the Natural Sciences and Engineering Research Council of Canada (grant # MOP-94848). This study was presented in an abstract form at the International Conference of the American Thoracic Society 2011.
The authors thank Johanne Vanderstock and Josiane Lefebvre-Lavoie for clinical and technical support, respectively, as well as Guy Beauchamp for statistical analysis.
Conflict of interest: Authors disclose no conflict of interest.
Hematek, Bayer Diagnostics, Elkhart, IN
Biotek instruments, Inc, Winooski, VT
Serotech, Oxford, UK
Kingfisher, St. Paul, CA
Mabtech, Mariemont, OH
Invitrogen, Carlsbad, CA
Luminex Corp, http://www.luminexcorp.com
Luminex 100 Integrated System 2.3, Austin, TX
GraphPad Software Inc, La Jolla, CA