SEARCH

SEARCH BY CITATION

Keywords:

  • horse;
  • foal;
  • predictors;
  • maternal;
  • immunoglobulin;
  • IgA

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Reasons for performing the study

Prior to the start of endogenous production of immunoglobulins (Igs), absorption of maternal Igs is important to protect against pathogens in the early neonatal period. It is possible that mare- or foal-associated factors may influence neonatal IgA concentrations.

Objectives

The temporal relationships among serum and milk IgA concentrations in Thoroughbred mare–foal pairs were explored to determine if periparturient mare- and foal-associated factors contribute to the prediction of foal serum IgA concentrations.

Methods

Blood and milk samples as well as complete veterinary records, were collected for 84 Thoroughbred mare–foal pairs from one month before to 2 months after parturition. Samples were tested using enzyme-linked immunosorbent assay (ELISA) for concentrations of IgA. Pairwise correlation coefficients were estimated (P<0.01) and simple linear regression used to investigate unconditional associations between mare IgA levels, mare and foal risk factors and foal serum IgA concentration at 12 h. Backwards, stepwise elimination of nonsignificant factors was used to create a final model.

Results

There were significant temporal relationships among mare serum IgA and among colostrum and milk IgA concentrations within mares (P<0.01). Mare serum IgA concentrations up to one month before parturition were associated with foal serum IgA concentrations at all time points and with colostrum and milk IgA concentrations. Mare serum IgA at -28 days and parity were associated with foal serum IgA concentration at 12 h (P<0.001).

Conclusions

Mare serum IgA concentrations up to 28 days before parturition, together with mare parity, are indicative of neonatal foal serum IgA concentrations.

Potential relevance

Mare serum and colostrum IgA concentrations may be useful peripartum predictors of neonatal mucosal immune status, enabling earlier intervention to prevent the consequences of mucosal infections.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Newborn foals are unable to mount an effective humoral response to infection [1-6] and are therefore dependent upon colostral ingestion and absorption of maternal immunoglobulin (Ig) for protection against pathogenic microorganisms while their own immunity is developing [2, 6-9]. If there is insufficient transfer of passive Ig via the mare's colostrum in the first 12–24 h after birth, the foal is at a considerably higher risk of sepsis, bacteraemia and localised infections [10].

Previously published studies in horses have focused on IgG as it has been considered the most critical Ig for the neonate and comprises the majority of Ig in equine colostrum [11]. The transfer of IgG from dam to offspring remains an active field of research but a better understanding of the role of other equine Igs in the neonate, such as IgA, is required [12, 13]. Emerging research suggests that IgA levels may indeed play a more critical and persistent role than IgG for immunological protection from pathogens that gain access via the mucosa [13]. The transfer of colostral IgA from neonatal serum to the nasal mucosa of the foal has been demonstrated but, to date, little is known of the factors that may predict neonatal serum concentrations [13, 14]. A better understanding of the factors that govern neonatal IgA concentrations may be of value in managing foals, particularly in situations where there are significant risks of exposure to mucosal pathogens [15].

Published studies have suggested that certain mare-associated factors, including parity, mare age and gestational length may be associated with IgG concentration in both the foal and mare [1]. There has been little objective investigation into the utility of mare-associated factors as prepartum predictors of neonatal Ig concentrations, including IgA. The objective of this study was to determine if maternal IgA concentrations and previously postulated mare-associated factors were predictive of transfer of IgA concentrations in the foal. Associations among maternal IgA concentrations in the serum, colostrum and milk and serum IgA concentrations in the foal, and partitioning of foal- and mare-associated factors were investigated.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Study population

The mare–foal pairs were recruited from a Thoroughbred stud located in the Barossa Valley, South Australia, following informed consent and approval from the University of Adelaide Animal Ethics Committee. Veterinary records were obtained for each mare–foal pair and the presence or absence of mare- and foal-associated factors recorded (Tables 1 and 2).

Table 1. Mare-associated factors identified from the veterinary records of the study population used in data analysis (84 mares)
Mare factors (n = 84)FrequencyAverageRange
Age in years104–20
Parity 
Primiparous8
Multiparous76
Gestation length (days) (n = 75)344316–371
Number foaled/month 
August14
September28
October28
November13
December1
Placental discharge (mins) 695–540
Fetal loss8%
Assisted foaling13%
Premature lactation2%
Placentitis1%
Premature placental separation7%
Retained placenta2%
Uterine haemorrhage1%
Inadequate colostrum1%
Poor udder conformation1%
Rejection of foal1%
Peripartum colic5%
Gastric ulcers1%
Vaginal haematoma1%
Medical treatment   
Strangles and tetanus vaccine92%
Sulphadimidine, trimethoprim, bromohexine4%
Altrenogest1%
Mycobacterium cell wall stimulant1%
Table 2. Foal-associated factors identified from veterinary records of the study population used in data analysis (84 foals)
Foal factorsFrequencyAverageRange
Foal weight (kg)835432–67
Gender   
Males34  
Females50
Standing time (min)8310330–365
Suckling time (min)8210939–272
Colostrum supplementation21
Volume (ml)225
Time post partum (min)344316–371
Rhodococcus equi infection2%
Diarrhoea24%
Gastric ulcers14%
Omphalophlebitis4%
Enteritis1%
Septic arthritis2%
Foot abscess2%
Other infection2%
Strangles-tetanus vaccine @ 60 days96%
Transphyseal periosteal elevation5%

Sample collection

Blood samples were collected aseptically from the jugular vein, allowed to clot and the serum aspirated and stored at -80°C in cryovials for batch analysis. Colostrum and milk samples were collected by hand from a randomly chosen teat and stored at -80°C in cryovials for batch analysis. Blood and milk sampling times corresponded to the 2010 breeding season management schedule of the stud. Prepartum serum sampling times for mares at -28 and -14 days were estimated based on an average gestation length of 339 days. The first colostrum sample was collected at parturition and the first foal serum sample at 12 h post partum. Thereafter mare serum, milk and foal serum samples were collected on Days 19, 24, 36, 47 and 60 post partum.

Immunoglobulin A measurements

Equine enzyme-linked immunosorbent assay (ELISA) IgA test kits1 were used as per manufacturer's instructions to determine IgA concentrations in serum, colostrum and milk. Samples were diluted (50 mmol/l Tris, 0.14 M NaCl, 1% BSA, 0.05% Tween 20) 1:1000 to ensure that measured concentrations were within the standard range of the assays. Samples were tested in duplicate and the results averaged.

Assay evaluation

All samples collected from each mare–foal pair were tested on the same ELISA plate and replicated over a second plate run in parallel. Duplicate raw optical density (OD) measurements of each standard, serum, milk and colostrum specimen were averaged. The coefficient of variation (CV) was calculated using standard deviation (s.d.) and mean (μ) of each replicate pair (CV = s.d./μ). Then CV estimates were plotted against the means to investigate the repeatability across different sample types (i.e. serum, colostrum, milk).

Statistical analysis

Descriptive statistics

The data was evaluated for normality using frequency histograms. Mean ± s.d. were calculated for IgA concentration at each time point, for mare and foal serum, colostrum and milk.

Analyses

All pairwise correlation coefficients among IgA concentrations in mare serum, mare colostrum and milk and in foal serum at any collection time were estimated and tested against null correlation at the P<0.01 level. Simple linear regression was used to investigate unconditional association between recorded factors (Tables 1 and 2) and the IgA concentration in the foal serum at 12 h. As true predictors, mare serum concentrations at -28 and -14 days were included in the unconditional association investigation. Marginally significant unconditional associations were included in multiple factor model building. After exploring co-linearity and first order interaction among unconditional predictors, the models were constructed using backwards stepwise elimination of nonsignificant factors. Only the best predictors were included in the final model and intervening factors were removed. Conventional diagnostics were run to check assumptions of homoscedasticity, normality and linearity. All analyses were performed using the statistical package Stata 11.2.2

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Assay evaluation

Coefficients of variation (CV) estimates of the different sample types (i.e. standard, serum, colostrum, milk) were similar. Overall, 99% control samples, 99% serum samples, 97% milk samples and 99% colostrum samples CVs lay within 20%. There was no evidence of different operating characteristics of the ELISA among the serum, colostrum and milk samples (Supplementary Fig 1).

figure

Figure 1. Temporal relationships of average mare serum, foal serum, colostrum and milk immunoglobulin A (IgA) concentrations with 95% confidence intervals (mg/dl) from 28 days before to 60 days after parturition. Associations between a) mare serum at -28 days with serum from -14 to 47 days, b) mare serum at -14 days with serum from 19 to 60 days, c) associations between colostrum with milk from 24 to 60 days, d) association between milk at 19 days with samples from 24 to 60 days, e) associations between foal serum at 12 h with serum at 19, 24 and 60 days, f) associations between foal serum at 19 days with serum at 24 and 47 days and g) association between foal serum at 47 days with 60 days; level of significance P<0.001.

Download figure to PowerPoint

Veterinary record data

A summary of the relevant data collected from records is given in Tables 1 and 2. Sufficient samples were considered to have been collected from the majority of time points in 84 of 124 mare–foal pairs recruited for the study; only data for these 84 pairs were included in statistical modelling. Insufficient data were available for the remaining pairs due to early departure from the stud or farm management activities that prevented sampling at some time points.

Immunoglobulin A patterns

Mean mare serum IgA concentration decreased from 16% over the sampling period. All mare serum time points were correlated with each other (P<0.01) except the serum concentrations at -28 and 60 days post parturition (P<0.05). In other words, mares with higher IgA concentrations in their serum at -28 days tended to have a higher concentration in their serum during the rest of the study period (Fig 1). The mean milk IgA concentration was highest at parturition (i.e. colostrum) and declined 44% by Day 19, after which concentrations remained relatively constant for the rest of the study period (Fig 1). Milk concentrations of IgA at time points were correlated with each other (P<0.01), except between colostrum and 19 days (P<0.05). Mean foal serum IgA concentration was highest at 12 h and declined 64% by Day 19, after which there was a progressive increase for the remainder of the study period (Fig 1). Foals serum IgA concentrations that were higher 12 h after parturition declined more rapidly during the first 19 days than those with lower initial serum concentration. Nevertheless, foals with higher serum concentrations at 12 h tended to have higher concentrations at 19 days. Foal serum IgA concentrations were correlated with each other from 12 h to 24 days (P<0.01), discontinuity was observed between Days 24 and 47 post parturition and then concentrations between Days 47 and 60 were correlated (Fig 1).

Predictive model of IgA concentrations in foal serum at 12 h

The foal serum concentration at 12 h was significantly and unconditionally associated with mare serum concentration at 28 and at 14 days prior to parturition, colostrum concentration, mare age, parity and with foal birth weight. Representative scatter plots are shown in Figure 2. Due to the collinear relationship between mare serum concentration at -28 days and -14 days, only the foremost predictor was kept for model development. Similarly, mare age and parity were highly colinear but only parity was included as it best explained foal serum concentration at 12 h. A quadratic term for parity was added to the model to better fit its nonlinear association with foal serum concentration at 12 h. Colostrum concentration and foal birth weight were excluded from the final model due to the intervening effect of mare serum concentration at -28 days and parity. The final model included mare serum concentration at -28 days and mare parity as 2 independent and highly significant predictors (r = 0.71, P<0.01). Statistical diagnostic procedures did not reveal any significant evidence of violation of the model assumptions. For an increase of 1 mg/dl of IgA in the mare serum 28 days before parturition, the model predicted an increase of 0.52 mg/dl of IgA in the foal serum at 12 h (P<0.001). The association between parity and foal serum concentration at 12 h was not linear (Fig 3). The foal serum concentration at 12 h was predicted to increase up to 6 parities (increase of 52.2 mg/ml of IgA at 6 parities compared with 0 parities) and progressively decreased beyond this. There was no significant association between mare serum concentration at -28 days and parity (P = 0.68).

figure

Figure 2. Scatter plots of associations for a) mare serum immunoglobulin A (IgA) concentration -28 days with foal serum IgA 12 h, b) mare serum IgA -14 days with foal serum IgA 12 h, c) colostrum IgA at parturition with foal serum IgA 12 h, d) mare serum IgA -28 days with colostrum at parturition (mg/dl), significance level P<0.01.

Download figure to PowerPoint

figure

Figure 3. Associations between parity number, mare serum immunoglobulin (IgA) concentration -28 days and foal serum IgA concentrations (mg/dl) at 12 h; level of significance P<0.001.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

The concentrations of IgA measured in Australian Thoroughbreds are comparable to those found in in a study of 52 European Thoroughbred foals [11]. The rapid reduction of IgA concentration in the foal serum from 12 h to 19 days after parturition is likely due to gut closure and utilisation of passively transferred IgA and is consistent with the response observed for IgG [16, 17]. Differences in the decline of serum IgA concentrations observed among foals may be due to inherent differences in the timing of gut closure but more data is required to determine the significance of the different rates of decline in IgA in these neonates. The higher IgA concentrations in some of foals at 19 days may confer greater mucosal protection, but may also be of concern when planning vaccination programmes for foals due to interference by maternal antibodies [2]. In agreement with others, the increase in foal serum IgA commencing after day 24 suggests endogenous production of IgA by these foals [14]. In agreement with the current study, onset of IgA synthesis in foals has been reported to occur at approximately 1–2 months of age [18]. Mean periparturient mare serum IgA concentrations were comparable with values reported in a small group of Thorougbred mares [11] and another of mixed-breed horses [16] but lower than a study of 27 mixed-breed adults [14]. In contrast, colostrum and milk concentrations of IgA were found to be lower than those described in previous studies [11, 16]. The significance of persistent IgA levels in mares' milk after closure of the window for absorption of maternal IgA appears to have had limited exploration in the horse [8]. However, investigations in human neonates have demonstrated that IgA and other factors within breast milk may make an ongoing contribution to local passive protection and immunological development [19].

Of the published studies on the influence of periparturient factors on failure of passive transfer (FPT) or neonatal sepsis, few have recruited sufficient horse numbers to utilise multivariate techniques [11]. The significant association found between mare parity and foal serum demonstrates another relationship that may influence foal serum IgA concentration. This association was, in part, explained by the fact that parity influenced foal weight at birth, which by itself was associated with foal serum concentration at 12 h (i.e. foal weight at birth was an intervening factor between parity and foal serum concentration). Previous studies on IgG have found that parity and mare age are associated with IgG concentrations in both the foal and mare [1, 21]. The association between parity and foal birth weight is also consistent with a previous report [21].

To the authors' knowledge, the current study represents the first attempt to statistically evaluate and model such a large range of periparturient factors with respect to neonatal foal IgA concentration. Nevertheless, the low incidence of some of the mare- and foal-associated factors in the modest study population restricted the statistical power of this investigation to effectively model many of them. A larger field study may be required to address this limitation but would perhaps require the selection of less well managed studs with higher incidences of these factors.

The associations found among mare serum, colostrum and foal serum in the current study and the possibility of predicting foal IgA concentrations based on mare parity and serum IgA, highlight the complexity of the mechanisms by which foals obtain protection from maternal antibodies. The partitioning of the mare's contribution to transfer of passive antibody to the foal is useful but, once born, the ability of the neonate to nurse and effectively absorb Ig must also be considered. The prepartum forecast of IgA concentrations in the foal may enable planning for intervention with vaccination or other strategies to optimise IgA concentrations in the dam, with the goal of significantly improving foal survival and reducing the risk of sepsis via mucosal infection [16, 19].

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

The authors acknowledge Mr Sam Hayes and the staff of Cornerstone Stud for facilitating and assisting with sample collection for this study, and thank Sasha Lanyon for her invaluable technical assistance. Professor Howarth is supported by the South Australian Cancer Council Collaborative Research Fellowship.

Authorship

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Christopher Riley conceived the study, and he and Caitlin Jenvey developed the study design with the cooperation of the manager of the stud at which the samples were collected. Data collection and study execution was performed by Caitlin Jenvey and Christopher Riley. Data analysis and interpretation were performed by Charles Caraguel, Caitlin Jenvey and Christopher Riley. Further data interpretation and critical review of study design was provided by Gordon Howarth. All authors contributed to the preparation and editing of the manuscript.

Manufacturers' addresses
  1. 1

    Bethyl Laboratories Inc., Montgomery, Texas, USA.

  2. 2

    Stata, version 11.2 Stata Corporation, College Station, Texas, USA.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors' declaration of interests
  8. Source of funding
  9. Acknowledgements
  10. Authorship
  11. References
  12. Supporting Information
FilenameFormatSizeDescription
evj648-sup-0001-si.zip734K

Fig S1: Scatter plots of mean optical density (OD) and coefficients of variation (CV %) for (a) control standards, (b) mare and foal serum samples, (c) milk samples, and (d) colostrum samples.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.