Secretory immunoglobulin A and immunoglobulin G in horse saliva

Veterinary Immunology and Immunopathology 180 (2016) 59–65

Contents lists available at ScienceDirect

Veterinary Immunology and Immunopathology journal homepage: www.elsevier.com/locate/vetimm

Secretory immunoglobulin A and immunoglobulin G in horse saliva Anna-Karin E. Palm a,1 , Ove Wattle b , Torbjörn Lundström b , Eva Wattrang a,c,∗ a Section of Immunology, Department of Biomedical Sciences and Veterinary Public Health, Swedish University of Agricultural Sciences, SE-751 23 Uppsala, Sweden b Department of Clinical Sciences, Swedish University of Agricultural Sciences, SE-750 07 Uppsala, Sweden c Department of Microbiology, National Veterinary Institute, SE-751 89 Uppsala, Sweden

a r t i c l e

i n f o

Article history: Received 4 July 2016 Received in revised form 24 August 2016 Accepted 8 September 2016 Keywords: Equine Secretory IgA IgG Saliva Mucosal immunity

a b s t r a c t This study aimed to increase the knowledge on salivary antibodies in the horse since these constitute an important part of the immune defence of the oral cavity. For that purpose assays to detect horse immunoglobulin A (IgA) including secretory IgA (SIgA) were set up and the molecular weights of different components of the horse IgA system were estimated. Moreover, samples from 51 clinically healthy horses were tested for total SIgA and IgG amounts in saliva and relative IgG3/5 (IgG(T)) and IgG4/7 (IgGb) content were tested in serum and saliva. Results showed a mean concentration of 74 ␮g SIgA/ml horse saliva and that there was a large interindividual variation in salivary SIgA concentration. For total IgG the mean concentration was approx. 5 times lower than that of SIgA, i.e. 20 ␮g IgG/ml saliva and the inter-individual variation was lower than that observed for SIgA. The saliva–serum ratio for IgG isotypes IgG3/5 and IgG4/7 was also assessed in the sampled horses and this analysis showed that the saliva–serum ratio of IgG4/7 was in general approximately 4 times higher than that of IgG3/5. The large inter-individual variation in salivary SIgA levels observed for the normal healthy horses in the present study emphasises the need for a large number of observations when studying this parameter especially in a clinical setting. Moreover, our results also indicated that some of the salivary IgG does not originate from serum but may be produced locally. Thus, these results provide novel insight, and a base for further research, into salivary antibody responses of horses. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The oral cavity poses many challenges for the horse immune system since this is an entry port for numerous infectious agents, has a prominent resident microbiome and is subject to frequent risk of maladies including traumatic injuries due to e.g. malocclusion or maladjusted tack. However, specific knowledge on the immune defence of the equine oral cavity is limited. From other mammals, mainly humans and laboratory rodents, it is known

Abbreviations: CI, confidence intervals; HRP, horseradish peroxidase; NALT, nasopharynx-associated lymphoid tissue; SIgA, secretory IgA. ∗ Corresponding author at: Department of Microbiology, national Veterinary Institute, SE-751 89 Uppsala, Sweden. E-mail addresses: [email protected] (A.-K.E. Palm), [email protected] (O. Wattle), [email protected] (T. Lundström), [email protected] (E. Wattrang). 1 Present address: The Department of Medicine, Section of Rheumatology, The Knapp Center for Lupus and Immunology Research, The University of Chicago, Chicago, IL 60637, USA. http://dx.doi.org/10.1016/j.vetimm.2016.09.001 0165-2427/© 2016 Elsevier B.V. All rights reserved.

that saliva contains many different antimicrobial proteins and peptides that are involved in the protection against infections (for reviewed see e.g. Brandtzaeg, 2007; Teeuw et al., 2004). Many of these belong to the innate immunity, e.g. mucins, while the major proteins of the specific immunity in saliva are secretory IgA (SIgA) and IgG. Early studies with a limited number of horses, n = 6 and n = 4, also showed that SIgA was the dominant immunoglobulin in horse saliva (Pahud and Mach, 1972; Vaerman et al., 1971). By strict definition saliva is equivalent to the secretions from the salivary glands, i.e. the parotid, sublingual, submandibular and minor salivary glands. In for instance humans the composition of secretions with respect to SIgA content, among other components, varies considerably between these different glands (reviewed in Brandtzaeg, 2007; Eliasson and Carlen, 2010). Moreover, “whole/mixed” saliva (also termed oral fluid) additionally often contains contributions from tears, nasal and airway secretions. Nonetheless, for technical reasons the present study is based on “whole” saliva and the term saliva is used in this wider definition throughout. In studied mammals, salivary SIgA originates from polymeric IgA, mostly in dimer form, produced by plasma cells in the acini and

60

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65

proximal ducts of major and minor salivary glands (for reviewed see e.g. Brandtzaeg, 2007; Teeuw et al., 2004). The J-chain of polymeric IgA is subsequently bound to the polymeric-immunoglobulin receptor at the basolateral membrane of a secretory glandular cell, internalised with the receptor and transported to the apical surface of the cell where proteolytic cleavage of the receptor results in the exocytosis of IgA still bound to the N-terminal part of the receptor (secretory component), i.e. SIgA. In the horse, all the genes of the IgA system have been identified (Lewis et al., 2010; Wagner et al., 2003) and the general structure of SIgA, e.g. binding of secretory component to dimeric IgA, seem to conform to that of other mammals (Lewis et al., 2010; McGuire and Crawford, 1972; Pahud and Mach, 1972). Moreover, it has been shown that a SIgA-like immunoglobulin was produced by horse submandibular glands cultured in vitro (Hurlimann and Darling, 1971). The initial activation of mucosal polymeric IgA-expressing Bcells mainly takes place in mucosa-associated lymphoid tissue and it has been suggested that e.g. in humans nasopharynx-associated lymphoid tissue (NALT) may be the most important inductive site for B-cells destined to the salivary glands (Brandtzaeg, 2007). Regarding NALT, the horse has five tonsils in the oro-, laryngo- and nasopharynx that form the Waldeyer’s ring (Casteleyn et al., 2011). IgA-expressing lymphocytes have been screened in four of these (Kumar and Timoney, 2005a,b,c, 2006) and were most frequent in the tonsil of the soft palate. The palatine and lingual tonsils also had prominent IgA-expressing cell populations while the tubal tonsil harboured very few IgA-expressing lymphocytes. Salivary IgG on the other hand, is considered to derive mostly from serum and to reach the saliva via gingival crevices (Brandtzaeg, 2007) although some salivary IgG may be locally produced. For example, human minor gland saliva contained around 30 times more IgG than corresponding parotid gland saliva and a local IgG origin was suggested (Smith et al., 1987). In human whole saliva the IgG content was approx. 14 times lower than that of IgA (Brandtzaeg, 2007). In the early studies of horse saliva, IgG was also detected (Pahud and Mach, 1972; Vaerman et al., 1971) and the IgG concentration was in mean approximately eight times lower than that of IgA (Pahud and Mach, 1972). Values for levels and subtypes of salivary antibodies in healthy individuals and their normal variation have been documented for some animal species but so far not for the horse. This information is important when assessing antibody responses to oral pathogens such as caries bacteria. The present study was therefore undertaken to increase the knowledge on presence of SIgA and IgG in saliva from healthy horses. For that purpose, currently available horse IgA reagents were characterised, and an ELISA for quantification of SIgA was set up. Saliva samples collected from fifty-one healthy horses were subsequently screened for SIgA, total IgG, IgG4/7 (IgGb) and IgG3/5 (IgG(T)) content.

2. Materials and methods 2.1. Antibodies For detection of IgA, murine anti-horse IgA monoclonal antibodies clone K129.2G5 (2G5; #MCA629, Bio-Rad, www.bio-radantibodies.com) and clone K129.3E7 (3E7; (Crouch et al., 2004), kind gift from Dr Karin Haverson, Department of Clinical Veterinary Science University of Bristol, UK), and polyclonal goat anti-horse IgA AAI34 (Bio-Rad) were used. For detection of IgG, polyclonal goat AffiniPure anti-horse IgG heavy and light chain (anti-IgG; #108005-003) and horseradish peroxidase (HRP) conjugated AffiniPure anti-horse IgG Fc fragments (anti-IgG Fc-HRP; #108-035-008) from Jackson ImmunoResearch, https://jireurope.com, were used. For detection of IgG4/7 and IgG 3/5, murine monoclonal antibod-

ies clone CVS39 (CVS39; #MCA1901) and clone CVS38 (CVS38; #MCA1900), respectively (Lewis et al., 2008; Sheoran et al., 1998), from Bio-Rad were used. Secondary HRP conjugated polyclonal rabbit anti-mouse Ig (#P0260) and rabbit anti-goat Ig (#P0449) from Dako, http://www.dako.com, were also used in immunoblotting and ELISA tests.

2.2. Immunoblotting For dot blots, 5 ␮l sample was applied on nitrocellulose membranes (Trans-Blot transefermedium nitrocellulose membrane 0.45 ␮m; Bio-Rad Laboratories Ltd., www.bio-rad.com) pre-soaked in phosphate buffered saline (pH 7.0; PBS). When reduced samples were analysed these were boiled for 5 min in SDS-PAGE sample buffer with 15.4 mg dithiothreitol/ml prior application to the membranes. Dot blot membranes were air dried before immunoblotting. For Western blots, proteins were transferred from separation gels onto the nitrocellulose membranes using a fully submerged Mini Trans blot cell (Bio-Rad Laboratories) at 150 mA for 60–75 min. Both types of membranes were blocked in PBS with 5% dried milk for 45 min, incubated with the primary antibody for 60 min, washed and subsequently incubated with HRP conjugated secondary antibody for 60 min (when applicable). Antibodies were diluted in PBS with 5% dried milk and 0.1% Tween 20 (SigmaAldrich, www.sigmaaldrich.com) and washing was performed using PBS with 0.1% Tween 20 with 3 changes of PBS for 5 min each. Blocking, antibody incubation and washing was performed at room temperature on a shaker. After the final antibody incubation membranes were washed 3 times in PBS with 0.1% Tween 20 and once in PBS and then developed using Sigma FASTTM DAB tablets (SigmaAldrich) or Amersham ECLTM (GE Healthcare Life Sciences, www. gelifesciences.com) according to the manufacturers’ instructions. In immunoblots primary antibodies 2G5, 3E7 and AAI34 were used at 1 ␮g/ml, anti-IgG Fc-HRP was used at 1.6-0.8 ␮g/ml. Secondary antibodies anti-mouse Ig-HRP and anti-goat Ig-HRP were used at dilutions 1:1000 and 1:2000, respectively.

2.3. Poly-acrylamide gel electrophoresis (PAGE) Samples were separated according to molecular weight using Tris/glycine 4–10% poly-acrylamide gels in a vertical mini-gel system (Bio-Rad Laboratories). Electrophoresis was carried out at constant voltage, 200 V, for approx. 60 min. Three different PAGE types were used: native PAGE, sodium dodecyl sulphate (SDS)PAGE with non-reduced samples and SDS-PAGE with reduced samples. In native PAGE the Amersham High Molecular Weight Calibration Kit (GE Healthcare Life Sciences) was used according to the manufacturer’s instructions. Molecular weight standards and samples were analysed on 4%, 6%, 7%, and 8% poly-acrylamide gels, respectively. A molecular weight standard curve was constructed by determining the retardation coefficient KR for each of the proteins in the molecular weight standard using Ferguson plots (Gallagher, 2001). Subsequently, KR for proteins in the samples was determined and the molecular weight calculated by linear regression of the standard curve. In SDS-PAGE, Spectra Multicolor High Range Protein Ladder (#26625, Thermo Fisher Scientific, www. thermofisher.com) was used as molecular weight standard. For reduction, samples were boiled for 5 min in sample buffer with 15.4 mg dithiothreitol/ml. A molecular weight standard curve was constructed by determining the relative mobility, Rf , of standard proteins and plotting these against the log10 of their molecular weights. The Rf of sample proteins was determined and their molecular weight was calculated by regression analysis on the linear interval of this standard curve. After electrophoresis, gels were

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65

either stained with Coomassie brilliant blue or proteins were transferred on to nitrocellulose membranes for immunoblotting. 2.4. Purification of SIgA from horse saliva Approximately 50 ml of pooled saliva collected from 15 clinically healthy horses was centrifuged (470 × g for 5 min) to remove large debris and filtered through a 5 ␮m filter. The saliva pool was then precipitated by drop wise addition of an equal volume of 4 ◦ C saturated ammonium sulphate solution under constant stirring. The mixture was stirred for a further 30 min at room temperature and the precipitate collected by centrifugation at 2000 × g for 15 min. The precipitate was washed (2000 × g for 15 min) twice in the original sample volume of 50% ammonium sulphate solution and subsequently dissolved in 1 ml PBS. This solution was then dialysed against PBS for ≥ 24 h with 5 buffer changes. The sample solution was diluted with an equal volume of 20 mM sodium phosphate buffer, pH 7.0 and applied to a HiTrap Protein G HP column (GE Healthcare Life Sciemces) according to the manufacturer’s instructions. Effluents from sample application as well as eluates of proteins bound to the column were collected and tested for SIgA and IgG content by dot blot analysis. SIgA containing effluents were pooled and concentrated using Vivaspin 20 (10 000 MWCO, Sartorius Stedim Biotech, www.sartorius.com) according to the manufacturer’s instructions and the protein content was determined by Bradford analysis. 2.5. ELISAs for SIgA, total IgG, IgG4/7 and IgG3/5 Saliva and serum samples were tested for SIgA, total IgG, IgG4/7 and IgG3/5 content, respectively, using four different ELISAs. The SIgA ELISA was set up using antibody AAI34 at 1 ␮g/ml for coating and antibody 2G5 at 1 ␮g/ml for detection followed by anti-mouse Ig-HRP at 1 ␮g/ml as tracer. The SIgA concentration in the samples was calculated by linear regression from serial dilutions of a SIgA standard (see above) and the ELISA’s linear range of detection was between 8.5 and 0.25 ng SIgA/ml. The total IgG ELISA was set up using antibody anti-IgG at 1 ␮g/ml for coating and anti-IgG Fc-HRP at 16 ng/ml as tracer. The total IgG concentration in the samples was calculated by linear regression from serial dilutions of ChromPure horse IgG whole molecule standard (IgG standard; #008-000-003, Jackson ImmunoResearch) and the ELISA’s linear range of detection was between 14 and 0.88 ng IgG/ml. The IgG4/7 and IgG3/5 ELISAs were set up using antibodies CVS39 and CVS38, respectively, both at dilution 1:100 for coating and anti-IgG Fc-HRP at 16 ng/ml as tracer. The total IgG standard was used to determine relative values for IgG4/7 and IgG3/5 concentrations. Seven 2-fold dilutions of the standard starting at 112 ␮g total IgG/ml were included to create a standard curve and the concentration of the highest optical density value within the linear range of detection was set to 100%. Relative IgG4/7 and IgG3/5 concentrations were then calculated by linear regression from the serial dilutions of the IgG standard. In all ELISAs 0.05 M Na2 CO3 /NaHCO3 buffer, pH 9.6, was used for coating and 4% bovine serum albumin (BSA; Sigma-Aldrich) in PBS was used for blocking and as diluent. All ELISAs were performed in flat-bottomed 96-well plates (MaxiSorp, NuncTM , ThermoFisher Scientific, www.thermofisher.com) and throughout PBS with 0.05% Tween 20 (Sigma-Aldrich) was used as wash buffer and an in house substrate buffer (1 mM 3,5,3 ,5 -tetrametylbenzidine in 0.1 M potassium citrate, pH 4.2, with 0.007% H2 O2 ) was used for visualisation of antibody binding. This reaction was stopped at a time point standardised for each ELISA with 2 M H2 SO4 and the A450 was measured in an ELISA reader. All samples were tested at different dilutions in order to ascertain values within the linear range

61

of detection for each ELISA and all sample dilutions were tested in duplicate. Standard curves and no-sample controls were included on every plate. For immunoglobulin concentrations in saliva, group mean values were calculated with 95% confidence intervals (CI) and values with non-overlapping CIs indicate statistically significantly differences.

2.6. Horses, clinical oral examination and sample collection The collection of samples for this experiment was approved by the Ethical Committee for Animal Experiments, Uppsala, Sweden (decisions C 90/7 and C 128/10). Animals included in the study were selected from horses that attended the equine clinic at the Swedish University of Agricultural Sciences (Uppsala, Sweden) or the animal dental clinic in Söderköping (Sweden) for dental health check-ups. Included horses were clinically healthy, had not been treated with either antibiotics or anti-inflammatory drugs at least one month prior to sampling and upon oral examination had no active inflammatory or infectious lesions such as large mucosal wounds, periapical infection, periodontal disease, grade I or more of infundibular caries or more than isolated areas with mild peripheral cemental caries. Horses with minor mucosal abrasions due to e.g. malocclusion or maladjusted tack, which were considered “normal findings” in the horse population were included. A total of 51 horses were included in the study, 23 mares, 26 geldings and 2 stallions of the following breeds/breed groups: 13 standardbred trotters, 19 Swedish warmbloods, 7 of other warmblood breeds, 8 of pony breeds, 2 Icelandic horses, 1 thoroughbred and 1 of unknown breed. The age of the horses ranged from 2 to 18 years with a mean of 10 years (± 1 year, 95% confidence interval (CI)). Blood and nasal secretion samples were collected prior to sedation of the horse. Saliva samples were collected either prior to sedation or at initiation of sedation when the mouth was opened for application of the oral speculum. Saliva samples were collected neat directly from the oral cavity with a suction device designed for clearing airway mucus in neonatal puppies and kittens (#902500, SweVet Piab, http://shop.swevet.se). Typically 0.5–1 ml of saliva was recovered from each individual by this method. Saliva samples were centrifuged at 470 × g for 5 min to remove debris and subsequently stored at −20 ◦ C until analysis. Due to shortage of saliva from several horses we were not able perform all the antibody analyses, i.e. SIgA, total IgG, IgG4/7 and IgG3/5, on some saliva samples. Blood samples were collected from the jugular vein into evacuated glass tubes without additive (BD Vacutainer® , http://www.bd.com/ vacutainer/). Blood samples were stored over night at 4 ◦ C, centrifuged at 2000 × g for 10 min and the serum collected and stored at −20◦ C until analysis. Samples of nasal secretions were collected on cotton swabs that were subsequently transferred to test tubes with 0.5 ml PBS and stored at −20 ◦ C until analysis.

3. Results 3.1. Characterisation of equine IgA reagents Assessment of recognition of different forms of IgA by antiequine IgA antibodies 2G5, 3E7 and AAI34 was undertaken using dot blot analysis of native or reduced aliquots of the same serum, saliva or nasal secretion samples (Table 1). This analysis showed that antibodies 2G5 and 3E7 only detected IgA in native samples. Moreover, 2G5 detected IgA in saliva or nasal secretions but not in serum. Hence, it was concluded that 2G5 recognises an epitope on the native SIgA molecule while 3E7 recognises native IgA, not necessarily in the SIgA form. Polyclonal antibody AAI34 detected

62

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65

Table 1 Summarised results for reactivity of anti-equine IgA antibodies 2G5, 3E7 and AAI34. Antibodies were tested in dot blot analysis of native or reduced aliquots of the same samples of pooled saliva, nasal secretions and serum, respectively. Samples Saliva

Nasal secretions

Serum

Name

Type

Native

Reduced

Native

Reduced

Native

Reduced

2G5 3E7 AAI34

Murine, monoclonal Murine, monoclonal Goat, polyclonal

+ + +

− − +

+ + nt

− − nt

– + +

− − +

+: positive, −: negative, nt: not tested.

both native and reduced IgA in both saliva and serum and some of its epitopes are thus likely on the heavy ␣-chain. 3.2. Molecular weight of equine IgA The molecular weight of IgA was determined in PAGE and Western blot using three different systems: native PAGE, SDS-PAGE with non-reduced samples and SDS-PAGE with reduced samples (Fig. 1). In native PAGE, a high molecular mass band was detected in saliva samples (an example with a 6% poly-acrylamide gel is shown in Fig. 1A) and was confirmed as SIgA in Western blot with antibody 2G5 (Fig. 1B). By calculation of KR using Ferguson plotting and linear regression using the molecular weight standard, the molecular mass of SIgA was determined to 395 kDa in the native PAGE system. In SDS-PAGE with non-reduced samples, saliva, ammonium sulphate precipitated saliva and serum was analysed on 6% polyacrylamide gels and IgA was detected in Western blot with antibody 3E7 (Fig. 1C). In saliva samples a band determined to 402 kDa was detected which was confirmed as SIgA using antibody 2G5 (not shown). In serum a band determined to 324 kDa was detected with antibody 3E7 that was designated as dimeric IgA. In addition, a more faint and diffuse band estimated to 149 kD was detected in serum which was designated as monomeric IgA. The relative intensity of the dimeric and monomeric IgA bands (Fig. 1C) suggests that the dimeric form of IgA was dominant in horse serum. Moreover, based on the difference in molecular weight between SIgA in saliva and dimeric IgA in serum this corresponds to an apparent molecular weight of 78 kDa for the secretory component. In SDS-PAGE with reduced samples, saliva, ammonium sulphate precipitated saliva and serum was analysed on 10% poly-acrylamide gels and IgA was detected in Western blot with antibody AAI34 (Fig. 1D). In saliva and serum samples a distinct band determined to 60 kDa was designated as the heavy chain of IgA. In the unpurified saliva sample a band determined to 137 kDa was visible (Fig. 1D), which probably corresponds to incompletely reduced monomeric IgA. In serum a low molecular weight band of approx 28 kDa was visible, which could correspond to the J-chain. However, this gel system was not calibrated for the low molecular range. Taken together, we have determined the molecular weights of horse SIgA, dimeric and monomeric IgA and the heavy ␣-chain. 3.3. Purification of SIgA from horse saliva A SIgA preparation of known concentration was needed for ELISA determination of SIgA content in horse saliva. When analysing horse saliva for protein content in PAGE we found that SIgA seemed to be the dominant high molecular weight protein (see e.g. Fig. 1A). We therefore performed ammonium sulphate precipitation to purify and concentrate SIgA. Dot blot analysis showed that the precipitated preparation indeed contained SIgA but also some IgG (Fig. 2). Protein G affinity chromatography was therefore performed to remove IgG. We found that when the sample preparation had passed through the column, SIgA but not IgG could be

detected in the effluent (Fig. 2). IgG but not SIgA was detected in eluates from the column (not shown). The SIgA containing effluent samples were subsequently pooled and concentrated. The protein content was determined and the preparation was used as “gold standard” for SIgA determination by ELISA. 3.4. Quantification of SIgA and total IgG in horse saliva Total levels of the SIgA and IgG were determined in saliva collected from healthy horses (Fig. 3). The levels of salivary SIgA ranged from <0.5 ng/ml to 351 ␮g/ml saliva and showed a large individual variation (74.3 ± 21.2 ␮g/ml, n = 45; mean ± 95% CI). Mares had lower average SIgA values than geldings (mares: 63.6 ± 22.7 ␮g/ml, n = 21; geldings: 89.3 ± 38.0 ␮g/ml, n = 22; mean ± 95% CI) but this difference was not statistically significant. The salivary content of total IgG varied between 0.6 and 128 ␮g/ml (19.8 ± 6.5 ␮g/ml, n = 44; mean± 95% CI) and the inter-individual variation was lower than that of salivary SIgA (Fig. 3). There was no significant difference in total IgG content between mares and geldings (mares: 18.6 ± 7.7 ␮g/ml, n = 20; geldings: 19.2 ± 11.4 ␮g/ml, n = 22; mean ± 95% CI). In individual horses the level of salivary SIgA was in general approximately 5 times higher than that of IgG (SIgA–IgG ratio 5.3 ± 1.6, n = 37; mean ± 95% CI) although in some samples IgG was equal to or dominated over SIgA. 3.5. Relative concentrations of IgG4/7 and IgG3/5 in horse saliva and serum The saliva–serum relationship of the two IgG isotypes IgG4/7 and IgG3/5 was determined to gain insight in the origin of salivary IgG in healthy horses. For IgG4/7 the saliva–serum ratio was in general approximately 4 times higher than the saliva–serum ratio of IgG3/5 (Fig. 4). When comparing IgG4/7 and IgG3/5 saliva/ IgG3/5; median: serum ratios within individual horses, IgG4/7 D 2.5, mean: 9.4 ± 8.8 (± 95% CI), n = 27, this relationship also indicated that the IgG isotype composition in saliva differed from that in serum. 4. Discussion We performed a screening of SIgA and total IgG in saliva from healthy horses. The amount of salivary SIgA recorded for these fiftyone horses was in mean lower but in the same range as that earlier reported for six horses (mean 120 ␮g/ml, range 40–300 ␮g/ml) by (Pahud and Mach, 1972). Compared to similar values for other species, the amount of SIgA in horse saliva was lower compared to that reported for whole saliva from humans (Brandtzaeg, 2007; Sari-Sarraf et al., 2008) and cats (Harley et al., 1998). The levels of SIgA in saliva from the horses we sampled showed a large individual variation. This is in analogy with what has been observed for e.g. human and rodent salivary SIgA where the variation has been attributed to a number of physiological, e.g. stress, hormonal status and age, and technical factors, e.g. fluid composition, sampling technique and method of analysis, influencing SIgA content

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65

63

Fig. 1. Poly-acrylamide gel electrophoresis and Western blot of horse IgA samples. (A & B) Native PAGE (6%); (A) stained with Coomassie brilliant blue, (B) Western blot for SIgA using antibody 2G5. (C) SDS-PAGE (6%) with non-reduced samples, Western blot for whole molecule IgA using antibody 3E7. D) SDS-PAGE (10%) with reduced samples, Western blot for IgA using antibody AAI34. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Samples: (1 and 2) horse saliva pools; (3) ammonium sulphate precipitated horse saliva; (4) horse serum (diluted 1:10); (MW) molecular weight marker. Arrowheads indicate detected bands. For details see Materials and methods.

Fig. 2. Dot blot analysis of horse saliva preparations subjected to protein G affinity chromatography for (A) IgG using antibody anti-IgG Fc-HRP and (B) SIgA using antibody 2G5. Samples: (PS) ammonium sulphate precipitated horse saliva pool prior to chromatography; (1) one column volume effluate during sample application; (2) first effluate with sample (approx. 1 sample volume); (3) second; (4) third; (5) fourth wash effluates (3× column volume); (neg) negative control (10% BSA in PBS).

(reviewed in Brandtzaeg, 2007). In the present study care was taken to standardise factors such as sample treatment and SIgA detection methodology but some factors such as the influence of stress on the horses were impossible to control in this clinical setting. This emphasises the importance of including a large number of observations in studies of salivary SIgA. Nevertheless, in the present study geldings tended to have higher salivary SIgA than mares, which is in agreement with what has been observed for humans where males had higher concentrations, and higher secretion rates, of SIgA in saliva than females (Booth et al., 2009; Eliasson and Carlen, 2010). In the present study, the amount of salivary total IgG was in mean similar to that reported for six horses (mean 15 ␮g/ml, range 5–20 ␮g/ml) by (Pahud and Mach, 1972). The amount of IgG in horse saliva was also similar to that in whole saliva from healthy humans (Brandtzaeg, 2007) but lower than that reported for cats (Harley et al., 1998). The SIgA–IgG ratio in saliva in the present study was similar to that reported for horses by (Pahud and Mach, 1972) (SIgA–IgG ≈ 8) and to that reported for cats (SIgA–IgG ≈ 5; Harley et al., 1998) and lower than that reported for healthy humans (SIgA–IgG ≈ 14; Brandtzaeg, 2007). In our material the total IgG levels in saliva showed a lower inter-individual variation compared to that of SIgA which could indicate that the major source of salivary IgG was passive diffusion from serum, e.g. via gingival crevices, rather than active secretion. In other mammals, local production/secretion of IgG in saliva have also been proposed (Brandtzaeg, 2007; Smith et al., 1987). For instance in human patients with

64

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65

Antibody concentration (µg/ml saliva)

351 350 200

150

100

50

0

SIgA

IgG

Fig. 3. Individual values for total SIgA and IgG in saliva from healthy horses. Horizontal lines depict mean values. SIgA was measured in samples from 50 horses and values for 45 horses are shown in the figure. In 5 samples SIgA levels were below the levels of detection for the ELISA and was <0.5 ng SIgA/ml in 2 samples and <2.5 ng SIgA in 3 samples. Total IgG was measured in samples from 44 horses, all included in the figure.

0,06

Saliva - serum ratio

0,05

0,04

0,03

sa

0,02

0,01

0

IgG4/7

IgG3/5

Fig. 4. The saliva–serum ratios of IgG isotypes IgG4/7 (n = 32) and IgG3/5 (n = 31), respectively, for healthy horses. The box encloses 50% of the data with the median value displayed as a horizontal line, the limits of the box represents the upper and lower quartile. Whiskers mark the maximum and minimum values, excluding outliers. Open circles represent outliers defined as values greater than the upper quartile, or smaller than the lower quartile, +1.5× the interquartile distance.

study by (Pahud and Mach, 1972) the free equine secretory component was determined to approx. 80 kDa. Furthermore, recombinant equine secretory component expressed in mammalian cells had a molecular mass of approx. 75 kDa (Lewis et al., 2010). Thus, our extrapolation of 78 kDa for the equine secretory component was in agreement with these earlier results. In addition to SIgA (Pahud and Mach, 1972) also found free secretory component in horse saliva. This was however not directly evident in our analyses which could be due to technical reasons since we were focussed on SIgA, e.g. antibody 2G5 may not recognise free secretory component. For IgA in horse serum, our results indicate that the dimer form dominates over monomer IgA, which is in analogy with earlier reports (Lewis et al., 2010; McGuire and Crawford, 1972; Pahud and Mach, 1972; Vaerman et al., 1971). In that respect horses conform to the norm of many mammals and are distinct from humans that have a strong domination of monomer IgA in serum (Woof and Kerr, 2004, 2006). To enable quantification of SIgA by ELISA a preparation of purified equine SIgA was needed. Initially we attempted purifying SIgA from horse saliva with ion-exchange chromatography using a protocol based on that described by (Sugiura et al., 2000). Although technically this worked well, the amount of SIgA recovered was too small to be of practical use. In the PAGE analysis of horse saliva we found that SIgA seemed to be the dominating high molecular mass protein and that ammonium sulphate precipitation significantly reduced other proteins. The precipitated saliva did however still contain some IgG. We therefore evaluated protein G affinity chromatography for removal of the contaminating IgG. We found that equine SIgA did not bind protein G which was in agreement with what has been reported for equine serum IgA (Sugiura et al., 2000). Regarding horse IgG, (Lewis et al., 2008) showed that all IgG subclasses except IgG5 bound to protein G. In accordance, earlier studies showed that all IgG subclasses bound protein G except a fraction of IgG(T), i.e. IgG3 and IgG5, that did not bind (Sugiura et al., 2000) or bound weakly (Sheoran and Holmes, 1996). After subjecting the present ammonium sulphate precipitated saliva to protein G we could not detect any remaining IgG using a sensitive ECL dot-blot assay. We thereby concluded that the SIgA was sufficiently purified for the intended use. To conclude, this study provides useful basic knowledge on the horse IgA system including some of the available horse IgA reagents. In addition a comprehensive screening of horse salivary antibodies was performed, which will provide a basis for further understanding of the immune defence of the equine oral cavity and its infectious and inflammatory diseases. Conflict of interest The authors declare no financial or other conflicts of interests. Acknowledgements

chronic hepatitis C infections, IgG responses to various hepatitis C virus epitopes differed between whole saliva and serum (Maticic et al., 2003), indicating a local source of IgG. This is in analogy with our analysis of IgG isotypes IgG3/5 and IgG4/7, which showed that the IgG isotype composition in horse saliva was not a complete mirror of that in serum. Taken together, this indicates that although serum derived IgG may be a major part of total IgG in horse saliva locally derived IgG also contributes. The molecular mass of equine SIgA, dimer and monomer IgA and the ␣-heavy chain determined in the present study were in accordance with earlier reported results. Using size exclusion chromatography (Pahud and Mach, 1972) determined SIgA to 400 ± 50 kDa, dimer IgA to 340 ± 40 kDa and monomer IgA to 175 ± 12 kDa. Using SDS-PAGE (Sugiura et al., 2000) determined monomer IgA to 148 kDa and the ␣-heavy chain to 56 kDa. In the

This study received financial support from the Swedish Foundation for Equine Research. The authors wish to thank Dr Karin Haverson and professor Christopher Stokes for help with the IgA reagents, Kerstin Nordling for help with purifying IgA, Dr David Morrison for statistical support, Kajsa Dalunde for technical assistance and all the horse owners for invaluable contributions. References Booth, C.K., Dwyer, D.B., Pacque, P.F., Ball, M.J., 2009. Measurement of immunoglobulin A in saliva by particle-enhanced nephelometric immunoassay: sample collection, limits of quantitation, precision, stability and reference range. Ann. Clin. Biochem. 46, 401–406. Brandtzaeg, P., 2007. Do salivary antibodies reliably reflect both mucosal and systemic immunity? Ann. N. Y. Acad. Sci. 1098, 288–311.

A.-K.E. Palm et al. / Veterinary Immunology and Immunopathology 180 (2016) 59–65 Casteleyn, C., Breugelmans, S., Simoens, P., Van den Broeck, W., 2011. The tonsils revisited: review of the anatomical localization and histological characteristics of the tonsils of domestic and laboratory animals. Clin. Dev. Immunol. 2011, 472460. Crouch, C.F., Daly, J., Hannant, D., Wilkins, J., Francis, M.J., 2004. Immune responses and protective efficacy in ponies immunised with an equine influenza ISCOM vaccine containing an ‘American lineage’ H3N8 virus. Vaccine 23, 418–425. Eliasson, L., Carlen, A., 2010. An update on minor salivary gland secretions. Eur. J. Oral Sci. 118, 435–442. Gallagher, S.R., 2001. One-dimensional electrophoresis using nondenaturing conditions. Curr. Protoc. Mol. Biol., Chapter 10, Unit 10.12B. Harley, R., Gruffydd-Jones, T.J., Day, M.J., 1998. Determination of salivary and serum immunoglobulin concentrations in the cat. Vet. Immunol. Immunopathol. 65, 99–112. Hurlimann, J., Darling, H., 1971. In vitro synthesis of immunoglobulin-A by salivary glands from animals of different species. Immunology 21, 101–111. Kumar, P., Timoney, J.F., 2005a. Histology and ultrastructure of the equine lingual tonsil I. Crypt epithelium and associated structures. Anat. Histol. Embryol. 34, 27–33. Kumar, P., Timoney, J.F., 2005b. Histology, immunohistochemistry and ultrastructure of the equine palatine tonsil. Anat. Histol. Embryol. 34, 192–198. Kumar, P., Timoney, J.F., 2005c. Histology, immunohistochemistry and ultrastructure of the equine tubal tonsil. Anat. Histol. Embryol. 34, 141–148. Kumar, P., Timoney, J.F., 2006. Histology, immunohistochemistry and ultrastructure of the tonsil of the soft palate of the horse. Anat. Histol. Embryol. 35, 1–6. Lewis, M.J., Wagner, B., Woof, J.M., 2008. The different effector function capabilities of the seven equine IgG subclasses have implications for vaccine strategies. Mol. Immunol. 45, 818–827. Lewis, M.J., Wagner, B., Irvine, R.M., Woof, J.M., 2010. IgA in the horse: cloning of equine polymeric Ig receptor and J chain and characterization of recombinant forms of equine IgA. Mucosal Immunol. 3, 610–621.

65

Maticic, M., Poljak, M., Seme, K., Skaleric, U., 2003. The IgG antibody profile to various antigen regions of hepatitis C virus differs in oral fluid and serum of patients with chronic hepatitis C. Oral Microbiol. Immunol. 18, 176–182. McGuire, T.C., Crawford, T.B., 1972. Identification and quantitation of equine serum and secretory immunoglobulin A. Infect. Immun. 6, 610–615. Pahud, J.J., Mach, J.P., 1972. Equine secretory IgA and secretory component. Int. Arch. Allergy Appl. Immunol. 42, 175–186. Sari-Sarraf, V., Reilly, T., Doran, D., Atkinson, G., 2008. Effects of repeated bouts of soccer-specific intermittent exercise on salivary IgA. Int. J. Sports Med. 29, 366–371. Sheoran, A.S., Holmes, M.A., 1996. Separation of equine IgG subclasses (IgGa, IgGb and IgG(T)) using their differential binding characteristics for staphylococcal protein A and streptococcal protein G. Vet. Immunol. Immunopathol. 55, 33–43. Sheoran, A.S., Lunn, D.P., Holmes, M.A., 1998. Monoclonal antibodies to subclass-specific antigenic determinants on equine immunoglobulin gamma chains and their characterization. Vet. Immunol. Immunopathol. 62, 153–165. Smith, D.J., Taubman, M.A., King, W.F., 1987. Immunological features of minor salivary gland saliva. J. Clin. Immunol. 7, 449–455. Sugiura, T., Imagawa, H., Kondo, T., 2000. Purification of horse immunoglobulin isotypes based on differential elution properties of isotypes from protein A and protein G columns. J. Chromatogr. B Biomed. Sci. Appl. 742, 327–334. Teeuw, W., Bosch, J.A., Veerman, E.C., Amerongen, A.V., 2004. Neuroendocrine regulation of salivary IgA synthesis and secretion: implications for oral health. Biol. Chem. 385, 1137–1146. Vaerman, J.P., Querinjean, P., Heremans, J.F., 1971. Studies on the IgA system of the horse. Immunology 21, 443–454. Wagner, B., Greiser-Wilke, I., Antczak, D.F., 2003. Characterization of the horse (Equus caballus) IGHA gene. Immunogenetics 55, 552–560. Woof, J.M., Kerr, M.A., 2004. IgA function—variations on a theme. Immunology 113, 175–177. Woof, J.M., Kerr, M.A., 2006. The function of immunoglobulin A in immunity. J. Pathol. 208, 270–282.