igg subclass responses to pigeon intestinal mucin are related to development of pigeon...

IgG subclass responses to pigeon intestinal mucin are relatedto development of pigeon fanciers’ lung

C. I. BALDWIN, A. TODD*, S. J. BOURKE†, A. ALLEN‡ and J. E. CALVERT

Departments of Immunology and‡Physiological Sciences, The Medical School, University of Newcastle upon Tyne,Newcastle upon Tyne,*Public Health Laboratory, Cumberland Infirmary, Carlisle and†Department of RespiratoryMedicine, Royal Victoria Infirmary, Newcastle upon Tyne, UK

Summary

Background Pigeon fanciers’ lung (PFL) is a form of extrinsic allergic alveolitis. Affectedindividuals produce antibodies to various pigeon antigens, and the resulting immunecomplexes are thought to initiate the disease. However, high antibody titres also occur insome asymptomatic individuals. Previously attention has focused on protein antigens, butwe have recently identified pigeon intestinal mucin as a novel antigen in PFL.Objective To determine the relationship between IgG subclass antibodies to pigeonintestinal mucin and the development of pigeon fanciers’ lung.Methods Sera were collected from 250 pigeon fanciers, who also completed a clinicalquestionnaire. Sera were screened for precipitating antibodies to pigeon serum anddroppings. Individuals with symptoms and precipitating antibodies were considered tohave classical PFL. Serum IgG and IgG subclass antibodies to pigeon intestinal mucin andpigeon serum proteins were investigated by quantitative enzyme-linked immunosorbentassay (ELISA).Results Very high titres of IgG antibodies against pigeon mucin were found in allprecipitin-positive individuals. A strong positive correlation was seen between titres ofantibodies to mucin and to serum proteins, but this was not due to crossreactivity. Nosignificant differences in IgG titres to either mucin or pigeon serum proteins were foundbetween individuals with PFL and asymptomatic precipitin positive fanciers. IgG1 andIgG2 were the major subclasses of anti-mucin, with lower titres of IgG3. Patients with PFLhad significantly higher titres of IgG1 to mucin than asymptomatic, precipitin-positiveindividuals. In contrast, no significant differences were seen between PFL and asympto-matic precipitin-positive sera with respect to the subclass titres against pigeon serum proteins.Conclusion The high titres of anti-mucin IgG in sera of all individuals with PFL, togetherwith the finding that high IgG1 titres to mucin are associated with the development ofdisease confirm pigeon intestinal mucin as an important antigen in PFL.

Keywords: pigeon fanciers’ lung, extrinsic allergic alveolitis, mucin, IgG subclass, human,lung, allergy, antibody

Clinical and Experimental Allergy, Vol. 28, pp. 349–357. Submitted 26 May 1997; revised9 October 1997; accepted 16 October 1997.

Introduction

Pigeon fanciers’ lung (PFL) is one of the group of diseasescaused by the repeated inhalation of organic dusts, collec-tively known as extrinsic allergic alveolitis (EAA). The

disease occurs in susceptible individuals who are regularlyexposed to pigeons, and is associated with both respiratoryand systemic symptoms occurring some 4–8 h after expo-sure to antigen. Respiratory symptoms typically include adry cough, breathlessness and chest tightness, whilst sys-temic symptoms include fever, sweating, myalgia, and a’flu’-like sensation.

The pathogenesis of PFL has been shown to be related to

Clinical and Experimental Allergy,1998, Volume 28, pages 349–357

349q 1998 Blackwell Science Ltd

Correspondence: Dr C.I. Baldwin, Department of Immunology, TheMedical School, University of Newcastle upon Tyne, Framlington Place,Newcastle upon Tyne, NE2 4HH, UK.

hypersensitivity reactions to pigeon antigens. The delaybetween exposure and the onset of symptoms, togetherwith the presence of antibodies to pigeon antigens in symp-tomatic pigeon breeders [1–4], suggests a role for immunecomplex (type III) reactions. However, antibodies alsocommonly occur in exposed asymptomatic individuals [2],although symptomatic subjects are reported to have higherantibody levels than their asymptomatic counterparts [5,6].Moreover, the finding of large numbers of infiltrating Tlymphocytes in the lesions and granuloma formation suggestcell-mediated hypersensitivity may also be important [7].

Antibody activity against a range of pigeon antigens hasbeen described. Antigenic material, defined by sera frompigeon fanciers, can be detected by immunofluorescence onpigeon mucosal stroma, epithelial cells lining the villi andcrypts and intestinal secretions within and emanating fromthe crypts [8–10]. The nature of these antigens was assumedto be proteins secreted through the gastrointestinal wall. Inrecent years attention has focused on IgA as the majorantigenic component in pigeon secretory materials. This hasbeen demonstrated in feather bloom and pigeon droppings[11,12], both present in abundance in pigeon lofts, and canalso be detected in the apex of villous epithelial cells [13]and pigeon intestinal mucus secretions [8].

More recently we have identified antibodies to another,unrelated antigen, pigeon intestinal mucin [8]. Such anti-bodies were associated specifically with sera from pigeonfanciers but, like antibodies to IgA, were found in bothsymptomatic and asymptomatic individuals. When the IgGsubclass composition of these antibodies was qualitativelyexamined, IgG1 antibodies to pigeon IgA and IgG1 andIgG2 antibodies to mucin were demonstrated in all seradesignated precipitin positive against pigeon faeces andbloom, however, IgG3 antibodies to mucin were preferen-tially found in symptomatic individuals [14]. This qualita-tive difference in IgG subclass reactivity to mucin insymptomatic as compared to asymptomatic pigeon fancierssuggested that mucin may be of particular importance in thepathogenesis of PFL.

In this study we have used enzyme-linked immuno-sorbent assays to assess quantitatively IgG and IgG subclassresponses to pigeon serum proteins and pigeon intestinalmucin in a large group of pigeon fanciers, to determine therole of these responses in PFL.

Materials and methods

Subjects and clinical status

These studies were undertaken with ethical committeeapproval and informed consent was obtained from all donorsprior to answering the questionnaire or donating blood.

During two conventions of pigeon fanciers, in Peterlee

and Blackpool in 1995/1996, and two visits to markingstations in 1995, a questionnaire was completed by 250individuals under the supervision of attending clinicians.Information relating to the nature, frequency and severity oflate (4–8 h) respiratory and systemic symptoms related topigeon contact was obtained as well as further relevantbackground details such as degree of pigeon contact, smok-ing history and other respiratory symptoms. A clinicaldiagnosis of symptoms related to pigeon fanciers’ lungwas based on the presence of at least one classic respiratorysymptom, such as shortness of breath or persistent drycough, and at least one classic systemic symptom, such asfever with shivering or aching muscles, occurring on at leastthree separate occasions 4–8 h after pigeon contact [15].

Blood samples were collected into preservative freeheparin. Plasma was removed following centrifugationand stored as aliquots at – 808C. Plasma was also collectedfrom 40 laboratory workers who had no previous contactwith pigeons or other species of bird.

Precipitating antibodies to pigeon faeces and pigeonserum were detected by counter current immunoelectro-phoresis (CIE) [16] using antigen and positive control seraobtained from Microgen Ltd. Only those sera that demon-strated precipitins to both pigeon faeces and pigeon serumwere regarded as positive. Use of the two antigens elim-inates false positives arising from antibodies to commonbacterial constituents in the faeces (e.g. teichoic acid), orreactions to avian sera such as have been described incoeliac disease. Individuals were classified into four groups:

Group A symptomatic with precipitating antibodiesGroup B asymptomatic with precipitating antibodiesGroup C symptomatic without precipitating antibodiesGroup D asymptomatic without precipitating antibodies.Only those individuals in Group A were considered to

have classic Pigeon Fanciers’ Lung.

Antigens

Mucin was prepared from the intestines of freshly killedpigeons as described [8] with minor modifications. Theintestines were gently washed through with water, opened,and the mucus gel carefully scraped from the mucosalsurface into 100 mL of pH 6.5 phosphate buffer containingproteinase inhibitors (1 mm phenyl methyl sulphonyl fluor-ide, 5 mm EDTA and 5 mmN- ethylmaleimide) [17]. Theresulting mixture was homogenized in a low speed blenderfor 1 min, centrifuged at 10 000g for 1 h at 48C and thesupernatant collected as soluble mucus. This was immedi-ately fractionated in a CsCl gradient (starting density 1.42 g/mL) and centrifuged towards equilibrium for 24 h at 40 000g.The resultant gradients were divided into nine equal fractionsand the density determined by weighing a known volume[18]. Following extensive dialysis small aliquots were

350 C. I. Baldwinet al.

q 1998 Blackwell Science Ltd,Clinical and Experimental Allergy, 28, 349–357

analysed for carbohydrate using a PAS method [19] andprotein. The fractions containing the majority of mucin(density 1.487–1.517 g/mL) were pooled and fractionatedon a second CsCl gradient (starting density 1.47 g/mL) andcentrifuged towards equilibrium as previously. The resultantgradients were fractionated into nine equal fractions and thedensity, carbohydrate and protein concentrations estimatedas previously. The fractions containing the majority of themucin (density 1.454–1.519 g/mL) were pooled, stored asaliquots at – 808C and used in the subsequent tests.

Three racing pigeons were ex-sanguinated. Sera werecollected, pooled, and stored as aliquots at – 808C.

Assays for total immunoglobulins

Total plasma IgG, IgA and IgM (mg/mL) were quantified byautomated rate nephelometry using a Beckman array proteinsystem. Total IgE (IU/mL) was measured by sandwichELISA, using a polyclonal goat anti-human IgE (e-chainspecific; Sigma Chemical Co) to coat the plates and amonoclonal anti-human IgE (clone HB121, American TypeCulture Collection) as the second antibody. Sample IgEconcentrations were obtained by reading off a standardcurve generated using National Institute of Biological Stan-dards second international reference preparation 75/502.

Assays for antibodies to pigeon serum and pigeon intestinalmucin

IgG and IgG subclass antibodies were quantified by ELISA.All washes were carried out in PBS/Tween, and the diluentwas PBS/Tween/3% BSA unless otherwise stated. Nunc-immuno Maxisorp 96 well plates were coated overnight atroom temperature with pigeon serum (1/20 000) or pigeonintestinal mucin (0.3mg/mL) in 0.1 m sodium phosphatepH 7.0. After washing, 100mL of PBS/Tween/3% BSA wasadded to each well for 30 min. Plates were washed andserum samples, appropriately diluted in PBS/Tween/BSA,were added to the top wells, and doubling dilutions weremade down the plate. After incubation at room temperaturefor 2 h, plates were washed and then incubated with 100mLof a 1/5000 of horseradish peroxidase (HRP)-conjugatedrabbit anti-human IgG (Dako Ltd, Denmark) for 90 min atroom temperature. Plates were washed, developed with OPDand read as above. Absorbance was plotted against serumdilution, and the titre was calculated for each serum as thereciprocal of the dilution giving an optical density of 0.2.

IgG subclass antibodies were quantified essentially asabove with the following modifications. The test sera wereincubated on the plates for 4 h at room temperature. Afterwashing, optimal concentrations of subclass specific mono-clonal antibodies were applied overnight at 48C: HP6012for IgG1, HP6010 for IgG3, HP6011 for IgG4 (Bionostics

Ltd, UK) and HP6002 for IgG2 (American Type CultureCollection). The following day plates were washed threetimes in PBS/Tween and then incubated with a 1/2000dilution of HRP-conjugated rabbit anti-mouse IgG (Dako,Denmark).

Determination of functional affinity

Functional affinity of antibodies was estimated by twoELISA methods:

The concentration of free antigen required to inhibit bind-ing of antibody to antigen-coated plates was determined (highfunctional affinity antibodies requiring lower free antigenconcentrations than antibodies of low functional affinity)[20]. Sera were diluted to give an OD of 1 for each subclass,and were added to antigen-coated plates containing a rangeof concentrations of the same antigen in 50mL of diluent.After mixing the contents of each well, plates were incu-bated for 4 h at room temperature, and the subclass ELISAwas performed as above. The concentration of free antigenwas plotted against percentage inhibition, and from this theconcentration needed to give 50% inhibition was calculated.

The second method involved elution of antibodies fromantigen-coated ELISA plates by chaotropic ions [21]. Serawere added to antigen-coated plates at a dilution giving anOD of 1 for each subclass, incubated for 4 h at roomtemperature and then washed. Ammonium thiocyanate in0.1 m phosphate buffer pH 6.0 was added to the wells at arange of concentrations and incubated for 15 min at roomtemperature. After washing, subclass-specific monoclonalantibodies were added and the assays completed as describedabove. The concentration of thiocyanate giving 50% inhibi-tion was determined (antibodies of high functional affinityrequiring higher concentrations to inhibit).

Statistical analysis

Statistical analysis was performed using Student’st-testwhere the data were normally distributed and the Mann–Whitney U-test where the data were non-parametric. Thesignificance of contingency tables was assessed by chi-squared analysis. All statistical calculations were carried outusing a computer based, commercially available, statisticspackage (Minitab data analysis software). Values ofP< 0.05were considered significant.

Results

Subjects

Details of subjects with respect to clinical status, age, degreeof pigeon exposure, smoking history, and mean total serumimmunoglobulin levels are shown in Table 1.

IgG subclass responses to pigeon351


One hundred and twenty-six of the 250 individuals hadprecipitating antibodies to pigeon antigens, and of these 48were in Group A, fulfilling our criteria for the diagnosis ofPFL, and 78 had precipitating antibody with no apparentsymptoms of PFL (Group B). Of the 124 exposed indivi-duals who did not have precipitating antibodies, 43 reportedavian-related symptoms indistinguishable from group A(Group C) and 81 were asymptomatic (Group D).

The mean age of all four clinical groups was similar.Degree of exposure to pigeons was comparable for all fourgroups, in terms of numbers of birds, how long they had keptpigeons, and the time spent in the loft.

Current smokers were significantly less likely to have

disease than ex-smokers or non-smokers (x2¼ 39.3,P<0.001) (Table 1).

Groups A and B had significantly higher serum IgGconcentrations than Groups C and D (P< 0.0001). Themean IgA concentration of Group A was above the normalrange for the North-East of England Health Region (0.64–2.97 mg/mL); all other values fell within local normalranges. Group A had significantly higher total IgA thanGroups C and D (P<0.013) but not significantly higher thanGroup B. There were no significant differences in total IgMor total IgE in any of the clinical groups.

IgG responses to pigeon serum and pigeon mucin

The median titre of IgG to pigeon serum and pigeon mucinin each clinical group and 40 control (unexposed) indivi-duals is shown in Tables 2 and 3. Groups A and B had veryhigh titres of antibodies to both antigens, significantlyhigher than Groups C and D (P< 0.0001). This accordswith the fact that groups A and B had precipitatingantibodies to both pigeon serum and pigeon faeces detect-able by CIE. There were no significant differences in IgGtitre to pigeon serum or pigeon mucin between Groups Aand B or Groups C and D. The four groups comprisingthe pigeon fanciers (Groups A–D) had significantly highertitres of IgG to both pigeon serum and pigeon mucinthan the unexposed control individuals in Group E(P<0.0001).

There was a strong positive correlation between the IgGtitre to pigeon mucin and the IgG titre to pigeon serum(Spearman’s Rank correlationr¼ 0.892,P<0.0001) (Fig.1). This did not appear to be due to crossreacting antibodies:when the anti-mucin ELISA was inhibited by titrating infree antigen 100% inhibition was achieved by adding pigeonmucin but a maximum of 20% inhibition was achieved by



Table 1. Background details of 250 pigeon fanciers in study

Group A Group B Group C Group D

Number of subjects 48 78 43 81Mean age in years 50.8 48.2 47.3 44.3Smoking history

Current smoker 1 10 17 33Ex-Smoker 21 25 14 17Non-Smoker 26 43 12 31

Exposure to pigeonsNumber of pigeons kept 73.6 89.2 84.9 74.0Years kept pigeons 29.4 28.2 24.1 26.5Hours contact per week 24.1 26 23.1 24.3

Serum Immunoglobulins1

IgA (mg/ml) 3.12 2.85 2.59 2.65IgM (mg/ml) 1.33 1.23 1.24 1.22IgG (mg/ml) 13.88 12.72 10.46 10.26IgE (IU/ml) 111 134 141 391

1 Mean immunoglobulin concentration.

Table 2. Median IgG and IgG subclass titres to pigeon serum (interquartile range in brackets)

A B C D E

IgG 51 700 39 000 2700 1900 125(25 100–148 700) (19 700–108 500) (20–6200) (220–6000) (50–327)

IgG1 56 100 30 500 1600 900 30(22 600–165 000) (15 000–94 000) (10–7500) (90–5900) (20–70)

IgG2 5150 4050 320 150 8(1900–10 750) (1100–11 470) (20–1300) (30–1100) (5–43)

IgG3 80 45 10 5 5(32–183) (20–153) (5–20) (5–20) (5–5)

IgG4 35 30 20 5 5(5–120) (5–93) (5–80) (5–40) (5–5)

adding pigeon serum (data not shown). Similarly no greaterthan 20% inhibition of the anti-pigeon serum ELISA wasobtained by titrating in soluble mucin, suggesting there isonly minor crossreactivity between the two antigens.

IgG subclass responses to pigeon antigens

Pigeon serumThe median titres for the IgG subclass responses to pigeonserum in each clinical group are shown in Table 2. The titresof all four subclasses in Groups A–D were always signifi-cantly higher than those titres seen in Group E (for IgG1,IgG2 and IgG3P< 0.0008, and IgG4P<0.0163). Groups A

and B had significantly higher titres of all four IgGsubclasses than Groups C and D (P<0.0015), althoughtitres of IgG3 and IgG4 were very low.

There were no significant differences between antibodytitres in Groups A and B or Groups C and D except Group Chad significantly higher titres of IgG4 than Group D(P¼ 0.03).

When the relative amounts of the two major subclasseswere analysed, the ratio of IgG1:IgG2 titres was signifi-cantly higher (P<0.016) for Groups A and B (medianvalues 11.4 and 7.8) than for Groups C and D (4.3 and4.0). However, there was no significant difference betweenGroups A and B.



Table 3. Median IgG and IgG subclass titres to pigeon mucin (interquartile range in brackets)

A B C D E

IgG 165 000 153 000 6100 5000 145(79 400–318 700) (56 100–290 000) (400–26 100) (420–28 500 (78–408)

IgG1 168 000 69 500 2800 1500 55(80 300–300 000) (17 600–176 500) (1200–15 400) (250–6000) (20–190)

IgG2 245 000 230 000 12 000 6700 60(154 500–384 800) (49 800–324 000) (700–63 900) (610–35 850) (20–130)

IgG3 3100 2350 40 10 5(350–17 220) (90–7280) (5–340) (5–115) (5–9)

IgG4 145 40 5 5 5(10–430) (5–462) (5–40) (5–20) (5–5)

Fig. 1. Relationship between antipigeon serum and antimucin antibody levels. Box in bottom left hand corner of graph represents valuesobtained for unexposed controls (meanþ 3xSD).

Pigeon intestinal mucin

The median titres for the IgG subclass responses to pigeonintestinal mucin in each clinical group are shown in Table3. The titres of all four subclasses were always significantlyhigher in Groups A–D as compared with the normalcontrols in Group E (P<0.0019). Groups A and B alwayshad significantly higher titres of all four subclasses thanGroups C and D (P< 0.0015), although titres of IgG4 werevery low.

Group A had significantly higher IgG1 titres than GroupB (P¼ 0.0019). There were no other significant differencesin antibody titres between the four clinical groups.

The ratio of IgG1:IgG2 titres was much lower for anti-mucin than for anti-pigeon serum, with median values of0.77, 0.33, 0.43 and 0.32 for Groups A, B, C and D,respectively. The IgG1:IgG2 ratio was significantly higherin Group A than Group B (P<0.03).

Functional affinity of anti-mucin antibodies

The composition of immune complexes will depend notonly on the amount of each subclass, but also on relativeaffinities. The functional affinities of the anti-mucin anti-bodies of different subclasses were determined in 15 serafrom Groups A and B, and results are shown in Fig. 2. Tomaximize the chance of detecting differences, sera wereselected such that the Group A samples had high IgG1 titresand the group B samples had low IgG1 titres. Two differentELISA methods were used, one relying on inhibition by freeantigen, and the other using thiocyanate elution (data not

shown). Only a proportion of sera had sufficiently high IgG3levels to enable analysis. The results obtained were broadlysimilar for the two methods. No significant difference wasobserved between the functional affinities of the differentsubclasses, and no differences were apparent between serafrom Group A and Group B.

Discussion

Although the mechanisms involved in the pathogenesis ofEAA are unknown it is generally accepted that the initialacute stages of the disease are caused by immune complexesformed at the alveolar epithelial surface of the lung [22–25].There is a large influx of neutrophils into the lung immedi-ately after exposure to the causative antigen [26,27]. Com-plement derived mediators are thought to account for theaccumulation of neutrophils [23,28] which are primarilyresponsible for the development of the immune complexmediated lung injury [28]. The later development of fibrosiscould be dependent on these sequestered neutrophils releas-ing chemotactic factors for mononuclear cells andmacrophages which lead to the development of the alveolitis.

The starting point of the disease process is therefore theproduction of immune complexes. However, this does notexplain why antibodies to pigeon antigens are also presentin a large number of asymptomatic pigeon fanciers. In thisstudy we have looked for differences in the antibodyresponse between symptomatic and asymptomatic pigeonfanciers that could explain this apparent anomaly. The IgGsubclass composition of the antibody response is likely to beimportant as this will determine the composition of immunecomplexes formed in the lung and influence the induction ofdisease.

Few studies have been carried out on the IgG subclasscomposition of the immune responses that occur in PFL. DeRidder and Berrens [29] suggested that non-precipitatingantibodies, capable of complement consumption in sympto-matic fanciers, were most likely to be IgG3 whilst precipi-tating antibodies, found in patients with PFL andasymptomatic individuals, were most likely to be IgG1and IgG2. Toddet al. [14]. showed that although bothsymptomatic and asymptomatic pigeon fanciers producedIgG1 and IgG2 to pigeon intestinal mucin all symptomaticpatients and only 12% of asymptomatic individuals pro-duced IgG3 to this antigen. All exposed individuals pro-duced IgG1 against pigeon secretory IgA whilst only 10%and 4% of these individuals produced IgG2 or IgG3,respectively, to this antigen. In this study we have quantita-tively assessed the IgG and IgG subclass responses to bothpigeon serum and pigeon intestinal mucin in a large groupof pigeon fanciers in order to see whether there are differ-ences in responses between symptomatic and asymptomaticindividuals. It should be noted that comparisons are only



Fig. 2. Functional affinity of antibodies to pigeon mucin asdetermined by free antigen competition of ELISA: concentrationof antigen (mg/mL carbohydrate) giving 50% inhibition.

made within, and not between subclasses, since the assaysare not necessarily of identical sensitivity [30].

Both symptomatic and asymptomatic precipitin positivepigeon fanciers had similarly high levels of IgG antibodiesto both pigeon serum and pigeon mucin, indicating thatdevelopment of disease is unlikely to be determinedsimply by the magnitude of these antibody responses.Moreover, no differences in subclass composition of theantibodies to pigeon serum were seen between the groups.In contrast, antibodies to mucin did show a significantsubclass difference between symptomatic and asympto-matic individuals, in that the former produced significantlyhigher titres of IgG1. This observation, together with thevery high anti-mucin titres, is consistent with the proposalthat antimucin reactivity is likely to be implicated in thedisease [14]. The reason this antigen was not detected insome earlier studies [31–33] is probably explained by thefact that, due to its large molecular size, it will not diffuseinto gels [8]. Mucin can be isolated from both pigeondroppings and bloom (unpublished observations) and islikely to be present in abundance in pigeon lofts. More-over, we have demonstrated antibodies to avian mucin insera from patients with bird breeders’ lung associated withother species (budgerigars, falcons and cockatiels) (unpub-lished observations).

The strikingly higher titres of IgG2 against mucin ascompared to pigeon serum antigens is of interest. Since theELISA protocol for the two antigens was identical, this islikely to reflect a genuine difference. IgG2 antibody isgenerally associated with responses to polysaccharideantigens, which are T-independent (in contrast to T-dependent protein antigens) [34]. Since mucin is a highlyglycosylated molecule it may well be able to elicit T-independent antibody production, accounting for the hightitres of IgG2.

Since the IgG subclasses differ with respect to theirbiological functions, the subclass composition of an anti-body response is likely to affect the properties of anyresulting immune complexes. The IgG1 and IgG3 sub-classes are generally more effective than IgG2 at activat-ing complement through the classical pathway and bindingFcg receptors [35], and so immune complexes containingthese isotypes are likely to be of greater pathologicalsignificance.

Although IgG3 anti-mucin titres were higher in group Athan group B, this difference was not significant. This differsfrom the earlier study of Toddet al. [14]. in which IgG3anti-mucin was observed to be specifically associated withsymptomatic pigeon fanciers. The reasons for the differencein our results are likely to relate partly to the techniquesemployed and partly to the way in which samples wereobtained. The earlier study used qualitative methods lesssensitive than the ELISAs used here, so that low levels of

IgG3 antibodies in some asymptomatic individuals couldhave been missed. Moreover, in the earlier study, thesamples from patients with pigeon fanciers’ lung weregenerally obtained at a time when the patient was beingseen by his doctor for active disease: in this study thesymptomatic individuals were not necessarily experiencingcurrent problems. Since IgG3 has a shorter half-life than theother subclasses [36] this could account for differences inour results.

Although precipitin positive pigeon fanciers also hadhigh titres of antibodies to pigeon serum, we could find nodifferences here between symptomatic and asymptomaticindividuals, either in terms of total IgG or IgG subclasses.This suggests that, although IgG1 appears to be a majorcomponent of this response, it is unlikely to account for thedevelopment of disease. One possibility is that the biochem-ical nature of the antigen itself also affects the properties ofthe immune complexes, and that pigeon mucin, being alarge, complex, highly glycosylated molecule (about 80%carbohydrate [37]) is intrinsically much more of a problemin terms of clearance of its immune complexes. In thiscontext it is interesting to note that other related diseases(e.g. Farmers’ lung, Summer type hypersensitivity pneumo-nitis) involve antigens which might include similarly com-plex, highly glycosylated molecules.

Although significant differences were seen in the IgG1anti-mucin titres of symptomatic and asymptomatic indivi-duals, these differences were small, and it is unclear whetherthey could result in differences in the immune complexessufficient to explain the clinical difference between groups.The composition of an immune complex will depend notonly on the concentration, but also on the functional affinityof each antibody isotype. It was possible that, if IgG1antibodies were of higher affinity, these might be over-represented in immune complexes. This idea was not,however, supported by our results which suggest that allsubclasses of anti-mucin are, on average, likely to be ofsimilar functional affinity. Another possibility which shouldbe considered is that the increased IgG1 anti-mucin levelsseen in the symptomatic individuals relates indirectly to thepathogenesis: IgG1 may, perhaps, correlate with a particulartype of cytokine present in the symptomatic group. Suchcytokines could be derived from T cells specific for theprotein component of mucin, or perhaps from other celltypes.

The individuals categorized as Group C in our studyrepresent an interesting population. These people describedsymptoms, apparently related to pigeon exposure, that wereindistinguishable from those in Group A subjects. How-ever, these individuals were precipitin-negative by ourcriteria, although a number (12/43) did show a reactionagainst faeces but not pigeon serum. Such a reaction wasalso observed in a similar proportion of individuals in group



D (24/81) and has occasionally been observed in unexposedcontrols. Because of the possibility that this may reflectreactions with bacterial components present in the pigeondroppings, these single reactions are normally disregardedfor diagnostic purposes. Nonetheless it is possible thatsome of these individuals should more correctly be classi-fied as group A or B. Taking this into account, there remain31 symptomatic individuals in this study who are precipitinnegative against both pigeon serum and droppings. Onepossibility is that these people have a form of disease inwhich type IV rather than type III reactions predominate. Itwould be important to verify that the symptoms describedby this group are related to pigeon antigens (e.g. byinhalation challenge), and if so to investigate their T-cellresponses to pigeon antigensin vitro.

This study confirms the importance of mucin as a majorantigen in pigeon fanciers’ lung. The IgG subclass profilesof the antibodies may provide a clue to the underlyingmechanisms of the disease and why only some individualsdevelop symptoms. IgG1 and IgG3 are usually associatedwith T-dependent antibody responses to proteins whilstIgG2 antibodies are normally associated with T-independentresponses to polysaccharides. The differences in immuneresponses between the symptomatic and asymptomaticindividuals may therefore relate to differences in T-cellresponses to pigeon intestinal mucin. We are currentlyinvestigating this possibility.

Acknowledgements

We are thankful to Dr Gavin Boyd, Dr Philip Lynch and theBritish Pigeon Fanciers Medical Research Team, StobhillHospital, Glasgow, G21 3UW, UK, for their support andassistance in undertaking these studies. We thank Mr RodAdams for arranging our visits to Seaton Delaval and SouthBoldon marking stations and Ms Beverley Stevens forexpert technical assistance. Financial support for thisstudy was provided by the Wellcome Trust.

References

1 Barboriak JJ, Sosman AJ, Reed CE. Serological studies inpigeon breeders’ disease. J Lab Clin Med 1965; 65:600–4.

2 Fink JN, Tebo T, Barboriak JJ. Differences in immuneresponses of pigeon breeders to pigeon serum proteins. J LabClin Med 1969; 74:325–30.

3 Tebo HT, Fredricks WW, Roberts RC. The antigens of pigeonbreeders disease II. Isolation and characterisation of antigenPDE I. Int Arch Allergy Appl Immunol 1977; 54:553–9.

4 Banham SW, MacKenzie H, McSharry C, Lynch PP, Boyd G.Antibody against a pigeon bloom extract: a further antigen inpigeon fanciers’ lung. Clin Allergy 1982; 12:173–8.

5 Boyd G. Clinical and immunological studies in pulmonaryextrinsic allergic alveolitis. Scott Med J 1978; 23:267–76.

6 McSharry C, Banham SN, Lynch PP, Boyd G. Antibodymeasurement and extrinsic allergic alveolitis. Eur J RespirDis 1984; 65:259–65.

7 Hensley GT, Garancis JC, Cherayil GD, Fink JN. Lungbiopsies of pigeon breeders’ disease. Arch Path 1969;87:572–9.

8 Todd A, Coan RM, Allen A. Pigeon breeders’ lung: pigeonintestinal mucin, an antigen distinct from pigeon IgA. Clin ExpImmunol 1991; 85:453–8.

9 Sennekamp J, Vogel F, Steins R. Detection of IgG antibodiesagainst pigeon intestinal mucosa antigens in pigeon breederssera using the immunofluorescent technique. Int Arch AllergyAppl Immunol 1976; 50:674–83.

10 Eade OE, Hodges JR, Berrill WT et al. Immuno-fluorescentantibodies in patients with bird fanciers’ lung. Clin ExpImmunol 1978; 32:257–62.

11 Fredericks WW, Tebo TH. The antigens of pigeon breeders’disease. III. Immunologically related antigens of pigeon drop-ping extracts: PDE1,2,A and B. Int Arch Allergy Appl Immunol1980; 61:65–74.

12 Longbottom JL. Pigeon breeders’ disease: quantitative immu-noelectrophoretic studies of pigeon bloom antigen. Clin ExpAllergy 1989; 19:619–24.

13 Goudswaard J, Vaerman IP, Heremans IF. Three immunoglo-bulin classes in the pigeon (Columbia livia). Int Arch AllergyAppl Immunol 1977; 53:409–19.

14 Todd A, Coan RM, Allen A. Pigeon breeders’ lung: IgGsubclasses to pigeon intestinal mucin and IgA antigens. ClinExp Immunol 1993; 92:494–9.

15 Banham SW, McSharry CP, Lynch PP, Boyd G. Relationshipbetween avian exposure, humoral immune response and pigeonbreeders disease among Scottish pigeon fanciers. Thorax 1986;41:274–8.

16 Mackenzie DWR. In Basic Serodiagnostic Methods for Dis-eases Caused by Fungi and Actinomyces, Public HealthLaboratory Service Monograph Series, HMSO, London, U.K,1980.

17 Mantle M, Allen A. Isolation and characterisation of the nativeglycoprotein from pig small intestinal mucus. Biochem J 1981;195:267–75.

18 Hutton DA, Pearson JP, Allen A, Foster SNE. Mucolysis of thecolonic mucus barrier by faecal proteinases: inhibition byinteracting polyacrylate. Clin Sci 1990; 78:265–71.

19 Mantle M, Allen A. A colorimetric assay for glycoproteinsbased on the periodic acid/Schiff stain. Biochem Soc Trans1978; 6:607.

20 Devey ME, Bleasdale-Barr KM, Bird P, Amlot PL. Antibodiesof different human IgG subclasses show distinct patterns ofaffinity maturation after immunization with keyhole limpethaemocyanin. Immunol 1990; 70:168–74.

21 Pullen GR, Fitzgerald MG, Hosking CS. Antibody aviditydetermination by ELISA using thiocyanate elution. J ImmunolMeth 1986; 86:83–7.

22 Calvanico NJ, Ambegankor SP, Schlueter DP, Fink J. Immu-noglobulin levels in bronchoalveolar lavage fluid from pigeonbreeders. J Lab Clin Med 1980; 96:129–40.

23 Yoshizawa Y, Nomura A, Ohdama S et al. The significance of



complement activation in the pathogenesis of hypersensitivitypneumonitis. Sequential changes of complement componentsand chemotactic activities in bronchoalveolar lavage fluids. IntArch Allergy Appl Immunol 1988; 8:417–23.

24 Dall’Aglio PP, Perci A, Bertorelli G, Brianti E, Scarpa S. Studyof immune complexes in bronchoalveolar fluids. Respiration1988; 54 (Suppl. 1):36–41.

25 Yoshizawa Y, Miyashita Y, Inoue T, Sumi Y, Miyazaki Y,Sato T, Ohtsuka M. Sequential evaluation of clinical andimmunological findings in hypersensitivity pneumonitis:Serial subclass distribution of antibodies. Clin Immunol Immu-nopathol 1994; 73:330–7.

26 Drent M, van Valzen-Blad H, Diamant M et al. Bronchoalveo-lar lavage in extrinsic allergic alveolitis: effect of time elapsedsince antigen exposure. Eur Respir J 1993; 6:1276–81.

27 Sugiyama Y, Kasahara T, Mukaida N, Matsushima K, Kita-mura S. Chemokines in bronchoalveolar lavage fluid insummer-type hypersensitivity pneumonitis. Eur Respir J1995; 8:1084–90.

28 Tanoue M, Yoshizawa Y, Sato T, Yano H, Kimula Y, Miya-moto K. The role of complement-derived chemotactic factorsin lung injury induced by preformed immune complexes. IntArch Allergy Immunol 1993; 101:47–51.

29 DeRidder G, Berrens L. Precipitating and non-precipitatingcomplement consuming antibodies in pigeon breeders disease.Immunobiol 1979; 156:168–78.

30 Falconer AE, Carson R, Johnstone R, Bird P, Kehoe M, CalvertJE. Distinct IgG1 and IgG3 subclass responses to two strepto-coccal protein antigens in man: analysis of antibodies tostreptolysin O and M protein using standardised subclass-specific enzyme-linked immunosorbent assays. Immunol1993; 79:89–94.

31 Berrens L, Maesen FPV. An immunochemical study of pigeonbreeders disease I. The aselective antigens. Int Arch Allergy1972; 43:289–303.

32 Berrens L, Maesen FPV. An immunochemical study of pigeonbreeders disease II. The specific antigens. Int Arch Allergy1972; 43:327–36.

33 Fredricks WW, Tebo TH. The antigens of pigeon breedersdisease III. Immunologically related antigens of pigeon drop-ping extracts: PDE1,2,A and B. Int Arch Allergy Appl Immunol1980; 61:65–74.

34 Hammarstrom L, Smith CIE. IgG subclasses in bacterialinfections. Monogr Allergy 1986; 19:122–33.

35 Burton DR, Gregory L, Jefferis R. Aspects of the molecularstructure of IgG subclasses. Monogr Allergy 1986; 19:7–35.

36 Spiegelberg HL. Biological activities of immunoglobulins ofdifferent classes and subclasses. Adv Immunol 1974; 19:259–94.

37 Allen A, Pearson JP. Mucus glycoproteins of the normalgastrointestinal tract. Euro J Gastroenterol Hepatol 1993;5:193–9.



igg subclass responses to pigeon intestinal mucin are related to development of pigeon...

Documents