Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

Thorax 1986;41:337-344

Editorial Immunoglobulins in the lung

The lungs are continually exposed to a vast range of meable membrane with respect to proteins in solu- antigenic material. The immunoglobulins of the lung tion. The transudation rate of a protein, and hence therefore have an important role in the neutralisation the secretion concentration, will depend on three fac- of antigens, which leads to cellular processing and tors: the "resistance" of the tissues to protein removal. The lung is undoubtedly a major "immu- diffusion, the plasma concentration of the protein, nological organ" since it contains a considerable and the effective size of the protein. We do not know amount of lymphoid tissue, with the ability to syn- whether the diffusion rate of proteins across the lung thesise immunoglobulins. The analysis of immuno- varies in different anatomical areas, although local globulins in the secretions of the lung has clarified our inflammation undoubtedly results in increased tran- understanding of their origins in this organ but the sudation.' Several proteins in lung secretions are details of the functions of immunoglobulins in the derived exclusively by diffusion from the blood. For lung are not well understood. Nevertheless, the secre- these proteins a correlation is observed between tions lining the airways are likely to be the site where the secretion:serum concentration ratios and their lung immunoglobulin function as a front line of effective sizes (expressed as Stokes' radius).1 Con- defence is most important. For this reason we can be sequently, if the secretion concentration of a protein fairly confident that studies of immunoglobulins in (relative to that in the serum) is higher than predicted lung secretions will be relevant to their significance for this size, simple transduration from the blood is in vivo. not the only source. The second possible source of a protein in the lung Origins of immunoglobulins in lung secretions secretions is local production by cells within lung tis- sues. Synthesis might be effected by cells actually So far our knowledge of the origins of the immuno- within the secretions. Alternatively, proteins may be http://thorax.bmj.com/ globulins in lung secretions has resulted almost synthesised by the epithelium or cells in the lamina entirely from the development of antisera that are propria and then either diffuse or become actively specific for the immunoglobulins and their discrete transported across the epithelium. subpopulations, such as the structural forms of IgA There is now evidence to suggest that, apart from and the immunoglobulin subclasses. These antisera IgD, for which no data are yet available, all the have been used for immunohistochemical studies, immunoglobulins enter the lung secretions by all of which have shown the presence of immunoglobulin these mechanisms. bearing plasma cells in the bronchial mucosa and in the secretions of the airways, thus confirming that IgA on September 26, 2021 by guest. Protected copyright. local synthesis of immunoglobulins does occur at IgA is the predominant immunoglobulin class in lung these sites. Furthermore, the development of specific secretions in contrast to blood plasma, where IgG is and sensitive immunological assays has allowed the found at higher concentrations. Most of the IgA in immunoglobulin composition of lung secretions to be the blood (about 90%) is monomer (mIgA),2 whereas studied in some detail. in lung secretions about halfthe IgA is dimeric (dIgA) In essence, there are two major sources of immu- and most of this is in the form of secretory IgA noglobulins in the lung secretions: (sIgA).3 Two subclasses of IgA, IgAl and IgA2, have Firstly, all the immunoglobulin classes are repre- been distinguished in serum and secretions by the use sented in the blood plasma. Thus a proportion of of specific . In the blood IgA2 comprises these proteins in lung secretions will be derived from 10-20% of the total IgA but in bronchoalveolar the vascular compartment by diffusion (transudation) lavage fluid samples it represents about 30%.4 These across the lung tissue, which behaves as a semiper- differences in the IgA composition of serum and lung secretions reflect the considerable local synthesis of IgA within the lung, although a proportion of the IgA Address for reprint requests: Dr D Burnett, Lung Immuno- in lung secretions is still derived from the blood by biochemical Research Laboratory, Clinical Teaching Block, General transudation. I Hospital, Birmingham B46NH. Dimeric IgA (dIgA), which is produced by plasma 337 Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

338 cells in the lamina propria, comprises two IgA mole- Since most of the blood IgA is monomeric, a large cules linked by another protein, , which is also proportion of mIgA in lung secretions should, the- synthesised by plasma cells. The dIgA is bound by a oretically, be derived from the plasma by transuda- receptor, secretory component (SC), which is a pro- tion. Nevertheless, lung secretion concentrations of tein inserted in the plasma membrane on the baso- mIgA are also higher than would be predicted if the lateral surfaces of some mucosal epithelial cells. The protein were derived only from the blood.3'1 This dIgA-SC complex on the epithelial plasma membrane suggests that most of the mIgA in bronchoalveolar is thought to be endocytosed, transported in vesicles lavage fluid and sputum is locally synthesised by across the cell, and released at the luminal surface as plasma cells within the lamina propria, although, in sIgA.' contrast to dIgA, monomer is not transported across This SC mediated transport of dIgA has been the epithelium by the SC mediated mechanism. defined mainly in studies of the gut mucosa. The high Immunoglobulin secreting cells are also located in the concentrations of sIgA in lung secretions support the mucosal secretions lining respiratory epithelium but hypothesis that this mechanism also operates in the one study of these cells, obtained by bronchoalveolar lung. Furthermore, immunohistochemical studies lavage, suggested that they do not contribute appre- have shown that SC and IgA are both present on the ciably to the IgA found in secretions.'2 epithelial surface, within vesicles ofbronchial glandu- Further evidence for local IgA synthesis in the lung lar epithelial cells and in the glandular lumen. In con- has been obtained from studies of the IgA subclasses. trast, cilated epithelium stains only faintly for SC and The proportion of IgA present as IgA2 is higher in is negative for IgA.6 This evidence supports the con- lung secretions than in blood plasma.4 This suggests cept of SC mediated transport of dIgA, as sIgA, that the proportion of IgA2 producing plasma cells of across the glandular epithelium of bronchi. Secretory the lung should be higher than that of non-mucosal component and IgA associated with J chain (sug- lymphoid tissue. Immunohistochemical studies using gesting that the IgA is dimeric) have also been located antisera specific for IgAI or IgA2 confirm this predic- on the plasma membrane and within pinocytic invag- tion. In bone marrow, tonsil, and peripheral lymph inations and vesicles of bronchiolar non-cilated epi- nodes 10-20% of the IgA plasma cells produce thelium and type II alveolar cells.7 These results IgA2.13 14 By comparison, 26-33% of the IgA showed that SC mediated transport of dIgA could plasma cells in bronchial mucosa produce IgA2.'3 also contribute appreciably to sIgA in the lower The proportion of IgA2 in blood or lung secretionshttp://thorax.bmj.com/ respiratory tract. therefore reflects the proportion of IgA2 producing IgA producing plasma cells are more abundant in plasma cells in non-mucosal lymphoid tissue and the glands and lamina propria of major bronchi than bronchial mucosa respectively. in the small bronchi, bronchioles, or alveolar sep- tae.78 These observations suggest that most sIgA IgG production should occur in the upper respiratory IgG is the predominant immunoglobulin class in tract. This is supported by a study in dogs that blood and some of the lung IgG is derived from the showed that sIgA concentrations were highest in plasma by transudation. Four subclasses of IgG have secretions from the upper respiratory tract.9 Another been described and each has been detected in bron- on September 26, 2021 by guest. Protected copyright. investigation,'0 which compared the IgA components choalveolar lavage fluid samples from normal sub- in sputum, bronchial washings, and bronchoalveolar jects.'5 In that study IgG subclass concentrations lavage fluid from patients with chronic bronchitis, were related to albumin measurements and the IgG highlighted the major problem with this kind of subclass:albumin ratios in the bronchoalveolar lavage study. The differential dilution of secretions, caused fluid and serum samples were compared. It was con- by sampling techniques, makes comparison ofprotein cluded that IgGI and IgG2 were derived wholly from concentrations in secretions from different levels of the blood because the IgG:albumin ratios were simi- the bronchial tree complicated. lar in serum and lavage samples. This result, however, Plasma cells in the lamina propria are not the exclu- would actually argue in favour of an appreciable sive source of sIgA in the upper or lower respiratory degree of local production of IgGI and IgG2 within tract. About 10% of plasma IgA is dimeric and Hai- the lung since it does not take into account the moto et al7 showed that IgA is present in endocytic different diffusion rates of albumin and IgG through vesicles of capillary endothelial cells, in the inter- biological tissues. Albumin, which has a Stokes radius cellular spaces adjoining endothelial cells, and on (Rs) of 3.5 nm, is smaller than IgG (Rs = 5.1 nm) and basement membrane. Some of this IgA was dimeric can therefore diffuse more easily from blood to secre- since J chain was also located in these areas. These tions. Indeed, it has been shown that if IgG were results suggested that dIgA in capillary blood might derived exclusively from the blood, the secre- be transported by SC into lung secretions. tion:serum concentration ratio relative to that for Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

339 albumin would have a value of 0.5.1 Any value The concentrations of this immunoglobulin in greater than 0.5 therefore indicates a degree of local bronchoalveolar lavage fluid suggested appreciable production. The secretion:serum ratio of IgGI and local synthesis, probably by IgE producing plasma IgG2, relative to albumin, reported by Merrill etal' cells in the bronchial mucosa.16 approached 1.0, confirming significantly higher con- centrations in bronchoalveolar lavage fluid than could be accounted for by transudation from the Immunoglobulin functions in the lung blood. The results reported for IgG3 and IgG4 indi- cated that these subclasses are also locally produced The primary function of all immunoglobulins is the in the lung but to a relatively greater degree than recognition and binding of specific antigenic deter- IgGI and IgG2. minants, whether soluble (including toxins), particu- Plasma cells producing IgG have been located in late, or cellular, such as pathogens. The consequences bronchial mucosa8 16 and these are likely to be the of immunoglobulin binding depend on the nature of source of the locally produced IgG, since IgG pro- the antigen and, at its simplest, may be the physical ducing cells isolated from secretions obtained by prevention of antigen penetration through the epi- bronchoalveolar lavage probably do not normally thelium. The secondary effects may include com- contribute appreciably to the concentrations of this plement activation with target cell lysis, opsonisation immunoglobulin in secretions.'2 IgG synthesised by resulting in enhanced phagocytosis, or plasma cells in the bronchial mucosa will not, how- dependent cell mediated cytotoxicity by a variety of ever, be actively transported by SC and its movement effector cells. The ability to intiate these secondary across the epithelium probably still relies on passive effects, however, varies with the immunoglobulin diffusion. class and subclass concerned. Local production of IgG3 and IgG4 is relatively greater than that of IgGI and IgG2'5 and therefore Functions ofIgA plasma cells producing IgG3 and IgG4 are also likely The presence of high concentrations of IgA in lung to be relatively more frequent in the bronchial secretion suggests that this immunoglobulin class mucosa. Such a finding would support the idea of should have an important role in the neutralisation of selective production of these subclasses. No data are inhaled antigens or pathogens and their toxic prod- yet available, however, regarding IgG subclass pro- ucts. The complex structure of secretory IgA (sIgA) ducing plasma cells within the lung. Further studies may in itself confer advantages on this immuno- http://thorax.bmj.com/ will be necessary to test this possibility and clarify the globulin. Firstly, since sIgA has four antigen binding mechanism for IgG production in pulmonary tissues. sites, it may be an effective agglutinating antibody, preventing bacterial growth'7 and bacterial adher- IgM ence to the epithelium.'8 Secondly, secretory com- IgM is composed of five immunoglobulin subunits ponent has the ability to stabilise the genetic variant linked by a J chain and is consequently a large protein Am2 of the IgA2 subclass. The Am2 form is unstable (molecular weight 900 000). Diffusion of 1gM into since it lacks disulphide bridges between the heavy

lung secretions from the blood is greatly restricted by and light chains but human SC has been shown to on September 26, 2021 by guest. Protected copyright. its large size. Nevertheless, the concentrations of IgM bind dimeric and monomeric IgAm2 molecules, thus in bronchoalveolar lavage samples from normal preventing their dissociation.'9 It was shown that the subjects and sputum from bronchitic subjects are stabilisation of IgAm2 required an excess of SC in the higher than can be explained by simple diffusion from reaction mixture. Free SC is present in lung the blood.'"' As an illustration, in bronchoalveolar secretions20 and this mechanism might therefore lavage fluid the concentrations of IgM (relative to occur in vivo. albumin) are higher than those of the smaller protein The structural integrity of immunoglobulins in the a2 macroglobulin."IIgM producing plasma cells are lung secretions is an important requirement for their present in the bronchial mucosa, though less fre- antibody activity. Lung infection and inflammation quently than IgG or IgA producing cells.'6 IgM is are often associated with the presence of proteinases bound and transported across epithelial cells by SC derived from leucocytes and bacteria. It has been and this mechanism is therefore likely to contribute shown that the opsonic effect of IgG antibodies to appreciably to concentrations of this protein in lung Pseudomonas aeruginosa was reduced in broncho- secretions. alveolar lavage samples from patients with cystic fibrosis.2' This inactivity of the IgG antibodies was IgE attributed to the presence of proteinases that frag- IgE is present at low concentrations in serum and mented the immunoglobulin molecules. The concept bronchoalveolar lavage fluid from normal subjects.' s of proteinase mediated damage to immunoglobulins Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

340 may have relevance to the presence of high concen- class specific response to stimulation by particular trations of IgA2 in the lung secretions. IgAl is sus- antigens. Polysaccharide antigens appear to evoke a ceptible to hydrolysis by proteinases produced by predominantly IgG2 response systemically,27 many pathogenic bacteria, cleavage of the IgAl whereas protein antigens favour the production of occurring at a Pro-Thr bond in the hinge region ofthe IgGI and IgG3 antibodies.28 heavy chain.22 IgA2 has a 13 deletion The biological effects of IgG antibodies in the lung including the susceptible Pro-Thr residues. IgA2 is are far from understood. Many factors, including the therefore resistant to proteolytic cleavage and its nature of the stimulating agents, antigen presenting function in lung secretions would be better main- cell populations, and cooperating T-cells within the tained in the presence of potentially damaging lung will influence the type of antibody response. In enzymes. turn, the effects of the antibodies produced will The ability of IgA to activate complement has been depend in part on the effector cell populations the subject of controversy. Several studies have present. Since many of these cells are now recognised shown that, under appropriate conditions, in vitro, as being phenotypically variable in different tissues, IgA may be effective in activating complement by the future studies may be more rewarding if the experi- classical or alternative pathways. The significance of ments investigate appropriate human lung material. these effects in vivo is, however, still unresolved and further studies are required. Similarly, there have Functions of IgM been conflicting results regarding the opsonising IgM is the most effective complement fixing immuno- ability of IgA antibodies. Secretory IgA has been globulin class and the 10 antigen binding sites on the shown to enhance phagocytosis by mouse alveolar pentameric structure ensure that it is a good aggluti- macrophages,23 but IgA antibodies to Haemophilus nating antibody. IgM is the first immunoglobulin to influenzae from bronchoalveolar lavage samples were be detected in the blood during the primary and also shown to block the opsonising activity of IgG secondary antibody responses and its main role is antibodies.24 Furthermore, although sIgA antibodies generally considered to be the neutralisation of alone may not induce antibody dependent cell medi- pathogens, especially viruses, in the vascular com- ated cytotoxicity by effector cells, they have been partment. The role of IgM in the lung secretions is shown to synergise with IgG in stimulating this mech- less well understood, although its ability to activate anism of target cell lysis.25 Clearly, the opsonising complement and opsonise pathogens is likely to be of ability of IgA may be dependent on experimental major importance. It is thought that IgM may behttp://thorax.bmj.com/ conditions in vitro and factors such as the presence necessary in replacing the functions of IgA in subjects of other immunoglobulins and effector cell with selective IgA deficiency. This conclusion is based subpopulations should be considered. Further studies on the observation that IgM producing plasma cells investigating the biological significance of IgA under and IgM concentrations in secretions were shown to conditions occurring in vivo may clarify our under- be raised in the intestine of patients with IgA standing of its function. deficiency.29 The relevance of these observations to the lung is not known.

Functions ofIgG on September 26, 2021 by guest. Protected copyright. The biological effects of IgG vary considerably with Functions of IgE the subclass concerned. Whereas IgGI, IgG2, and IgE is bound by Fc receptors on mast cells. Cross IgG3 can activate complement via the classical path- bridging of the cell bound IgE with specific antigen way, IgG4 is ineffective. The opsonising activity of an results in the release of several active substances from antibody will depend largely on the presence of rele- the mast cell, including histamine and slow reacting vant Fc receptors on effector cells. Naegel etal26 substance of anaphylaxis (leukotrienes C4, D4, and showed that about 25% of human alveolar macro- E4), which result in bronchiolar smooth muscle phages could bind IgG3 and 10% bound IgG4, contraction and increased blood vessel permeability. whereas no significant binding of IgGI or IgG2 was The mast cells also release potent enzymes, including observed. These results suggest therefore that IgG3 elastase, cathepsin G, and kininogenase.3031 Fur- and IgG4 may be the most important subclasses thermore, other cells are recruited to the inflam- concerned in the opsonisation of pathogens for matory process through the release of neutrophil and alveolar macrophages and may explain the relatively eosinophil chemotactic factors and platelet activating high degree of local synthesis of these proteins. factor. This IgE mediated inflammatory response is Further studies will be necessary to clarify the role of most evident in type 1 hypersensitivity reactions such IgG in inducing opsonisation by other effector cell as occur in allergic asthma, but it has been proposed populations in lung secretions. that this mechanism has an important role in the host Another factor to be considered is the class or sub- defence against metazoan parasites.32 In addition, Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

341 IgE may itself interact with other inflammatory cells. trations of IgG and IgA in lavage fluid.36 IgM was For instance, IgE has been shown to bind to receptors also increased in chronic hypersensitivity pneu- on alveolar macrophages causing the release of monitis but no significant increase in IgE concen- lysosomal enzymes.32 trations in lavage fluid were observed. This study also The factors responsible for high IgE antibody showed a significant reduction in IgG in lavage fluid concentrations in atopic individuals are not known, after corticosteroid treatment. although defects in T cell function have been The increased immunoglobulin concentrations in suggested.33 the lung secretions of patients with interstitial lung disease would appear to occur for two reasons. The first is an increase in the transudation of protein from Disease and lung immunoglobulins the blood, presumably because of inflammation within the lung. This conclusion is based on the obser- The immunoglobulin composition of the lung may be vation that albumin concentrations in lavage fluid are influenced by many factors. In some diseases, how- also frequently increased in sarcoidosis, "1 12 although ever, the aetiology or progression of the condition not in idiopathic pulmonary fibrosis.36 Secondly, appears to be directly related to abnormal immuno- local synthesis of immunoglobulins is likely to result globulin production or function. Thus autoantibodies from an increase in synthesis by plasma cells in the may be an important feature in Goodpastures's dis- lung, including cells within the secretions lining the ease and atopy is often associated with excessive IgE airways. Although no significant relationship was synthesis. The present discussion, however, will be observed between the number of immunoglobulin confined to two groups of lung diseases, the inter- secreting cells and immunoglobulin levels in stitial and obstructive lung conditions. bronchoalveolar lavage fluid from normal subjects, a correlation was reported for IgG (but not IgA or Interstitial lung diseases IgM) in lavage fluid from patients with sarcoidosis. 12 The interstitial lung diseases are characterised by Other studies have reported an increase in IgG secret- infiltration of the interstitium by inflammatory cells ing cells in lavage samples from patients with and the development of alveolitis, fibrotic changes, sarcoidosis34 37 and idiopathic pulmonary fibrosis.37 and granulomas. There appears to be a significant The results were not conclusive for other immuno- "immunological" component to the pathogenesis globulins since Hunninghake and Crystal34 observed of these diseases, indicated by the presence of an increase in IgM producing cells in lavage samples http://thorax.bmj.com/ activated T lymphocytes and a polyclonal stimulation from patients with sarcoidosis but Lawrence etat37 of immunoglobulin synthesis, including auto- reported no difference in comparison with controls. antibodies.34 Sarcoidosis is associated with a No increase in IgA secreting cells were found. Inter- significant increase in serum concentrations of IgA estingly, successful corticosteroid treatment reduced and a significant, though small, increase in serum the numbers of IgG producing cells in lavage samples IgG.12 from the patients with sarcoidosis.37 The increased The most striking changes in immunoglobulin immunoglobulin concentrations in lung secretions production in interstitial diseases are revealed by the from patients with interstitial lung disease clearly analysis of lung secretions from patients with these result mainly from local synthesis by plasma cells and, on September 26, 2021 by guest. Protected copyright. conditions. Several studies have shown greater IgG by contrast with the normal lung, a considerable pro- and IgA concentrations in bronchoalveolar lavage portion of this protein is derived from cells lining the fluid samples from patients with sarcoidosis than in airways. The source of these cells is not known. The samples from healthy controls." 12 35 The concen- predominant form (80%) of IgA in bronchoalveolar tration of IgM in lavage fluid in sarcoidosis was also lavage fluid from subjects with sarcoidosis, however, shown to be increased in two of these studies"1 35 but is monomer."1 Since this represents a higher propor- was reported to be similar to that in lavage fluid from tion ofmonomer than is present in the normal lung, it controls by Rankin etaL.'2 The increased immuno- was suggested that the plasma cells responsible for globulin concentrations in lavage fluid were related to IgA synthesis in the sarcoid lung were recruited from disease activity since they were observed only in the blood." Alternatively, subpopulations of B lym- patients with high intensity alveolitis, l " and clinical phocytes already within the lung could be stimulated improvement after corticosteroid treatment was asso- in these diseases. This explanation seems less likely in ciated with reduced immunoglobulin concentrations view ofthe polyclonal nature of the immunoglobulins in lavage fluid.35 produced.34 Idiopathic pulmonary fibrosis, fibrosing alveolitis, The reasons for the increased synthesis of immuno- and chronic hypersensitivity pneumonitis were also globulins in patients with interstitial lung diseases reported to be associated with increased concen- remains obscure. The phenomenon appears to be Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

342 localised specifically to the lung. This is supported by the microbial antigen challenge. the observed increase in immunoglobulin concentra- Several factors may explain why immunoglobulin tion and number of immunoglobulin producing cells deficiency has not been detected in many patients with in bronchoalveolar lavage fluid samples and by the chronic or recurrent infections. Firstly, the recognised increase in numbers ofT helper cells in the lung secre- incidence of immunoglobulin deficiency has increased tions obtained from patients with interstitial lung dis- since assays for IgG subclasses were introduced. eases. Furthermore, T cell numbers correlated with More specific assays for the measurement of IgA immunoglobulin concentrations and number of subclasses and antigen specific antibodies might immunoglobulin producing plasma cells and with reveal further causes of infection. Secondly, despite disease activity.3435 Perhaps most important was the normal immunoglobulin concentrations, deficiencies observation that these T cells could induce immuno- in effector cell function may be present in a propor- globulin synthesis in otherwise unstimulated B cells tion of patients. Thirdly, and relevant to all of these from the blood of normal individuals.34 The evidence possibilities, is the need to quantify the components so far therefore suggests that T cell mediated syn- of the immune system in the lungs of these patients thesis of immunoglobulins within the lung is a central rather than in the serum. Infections are often confined feature of interstitial lung diseases. The mechanisms to the lungs, and thus deficiencies in local immuno- leading on to lung damage are not yet clear. Further globulin synthesis alone may be a key aetiological fac- identification of the subpopulations of the immuno- tor. For instance, specific sIgA deficiency has been globulins produced in these conditions would be described in patients with normal serum IgA concen- interesting. This might help to establish whether there trations.20 The secretory IgA system may therefore be is a functional association between the immuno- independent of the systemic IgA system, and this globulins, effector cells, and pathological changes might also be the case for other immunoglobulins. At seen in the lungs. present little is known about how the immuno- globulin composition of the lung secretions of patients with recurrent or chronic respiratory infec- Infections and obstructive lung disease tions compares with that of normal individuals. Immunoglobulin deficiencies are often associated The IgA system has been studied in sputum sam- with chronic or recurrent pulmonary infections and ples obtained from patients with chronic bronchitis. several studies have investigated the immunoglobulin In those who were not apparently IgA deficient the

class or subclass deficiencies in the serum of patients presence of a clinical pulmonary infection was associ- http://thorax.bmj.com/ with these respiratory problems. The frequency of ated with significantly increased sputum mIgA and observed immunoglobulin deficiencies in these dIgA.3 An appreciable amount of the IgA (both patients differ considerably from study to study and monomer and dimer) was locally synthesised, pre- will reflect the groups of patient selected. For sumably in response to the infecting organisms. These instance, the frequency of selective serum IgA results suggest that the lungs of most patients with deficiency (<50mg/1) in caucasians is 1/300-1/2000 chronic bronchitis are capable of mounting an IgA but in the Japanese population it is much rarer, immune response to infecting organisms, although no having an incidence of only 1/18000.38 Furthermore, comparison with the healthy lung is yet possible. the specificity of the assays used will dictate whether Soutar8 reported that patients who had died from on September 26, 2021 by guest. Protected copyright. an immunoglobulin deficiency is detected. Serum chronic obstructive bronchitis were deficient in IgA deficiencies of IgG clearly are often overlooked unless producing plasma cells throughout the respiratory the IgG subclasses are quantified. For instance, tract, whereas patients with bronchitis who had died Umetsu et at39 studied 20 children with recurrent of other causes were normal, which suggests that fatal sinopulmonary infections who had normal total bronchitis was associated with a deficiency of lung serum IgG concentrations. All of these children were, IgA plasma cells. A more recent study reported however, deficient in one or more IgG subclass and increased IgA plasma cells in the bronchi of patients 15% were deficient in IgA. The most frequent IgG with chronic bronchitis.4" The plasma cells were of subclass deficiency was IgG2, and this was possibly both IgA subclasses but the increase was mainly in the the reason for the observed low titres of antibody to plasma cells producing IgAl. Clearly these different the capsular polysaccharide of Haemophilus results may be explained in part by the study of influenzae in these patients. different groups of patients. Further studies are Studies in other groups of patients have detected required, quantifying the IgA and IgG subclasses in lower frequencies of immunoglobulin deficiencies in bronchoalveolar lavage fluid from patients with the serum ofpatients with recurrent and chronic chest chronic obstructive lung disease and comparing the infections.40 Furthermore, in some patients immuno- concentrations with those of normal individuals. globulin concentrations were increased in response to Such comparisons will clarify whether immuno- Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

343 globulin deficiencies have a role in the aetiology of In: McGee JR, Mestecky J, eds. The secretory immune chronic obstructive lung disease. system. New York: Annals ofthe New York Academy of Science, 1983:803-5. D BURNETT 15 Merrill WW, Naegel GP, Olchowski JJ, Reynolds HY. University Department of Immunology subclass proteins in serum and and Lung Immunobiochemical Research lavage fluid of normal subjects: quantitation and com- Laboratory, General Hospital parison with immunoglobulins A and E. Am Rev Respir Birmingham B4 6NH Dis 1985;131:584-7. 16 Nijhuis-Heddes JMA, Lindeman J, Otto AJ, Snieders MW, Kievit-Tyson PA, Dijkman JH. Distribution of References immunoglobulin-containing cells in the bronchial mucosa of patients with chronic respiratory disease. Eur 1 Stockley RA, Mistry M, Bradwell AR, Burnett D. A J Respir Dis 1982;63:249-56. study of plasma proteins in the sol phase of sputum from 17 Reynolds HY, Thompson RE. Pulmonary host defences. patients with chronic bronchitis. Thorax 1979;34:777-82. I-Analysis of protein and lipids in bronchial secretions 2 Newkirk MM, Klein MH, Katz A, Fisher MM, Under- and antibody responses after vaccination with Pseudo- down BJ. Estimation of polymeric IgA in human serum: monas aeruginosa. J Immunol 1973;111:358-68. an assay based on binding of radiolabelled human secre- 18 Williams RC, Gibbons RJ. Inhibition of bacterial adher- tory component with applications in the study of IgA ence by secretory : a mechanism of nephropathy, monoclonal gammopathy and liver antigen disposal. Science 1972;177:697-9. disease. J Immunol 1983;130:1176-81. 19 Jerry LM, Kunkel HG, Adams L. Stabilization of dis- 3 Stockley RA, Afford SC, Burnett D. Assessment of 7S sociable IgA2 proteins by secretory component. J Immu- and llS immunoglobulin A in sputum. Am Rev Respir nol 1972;109:275-83. Dis 1980;122:959-64. 20 Stockley RA, Burnett D, Afford SC. The immunological 4 Delacroix DL, Dive C, Rambaud JC, Vaerman JP. IgA measurement of "free" secretory piece and its relation- subclasses in various secretions and in serum. Immu- ship to local IgA production. Clin Exp Immunol 1981; nology 1982;47:383-5. 45:124-30. 5 Mostov KE, Blobel G. A transmembrane precursor of 21 Fick RB, Naegel GP, Squier SU, Wood RE, Gee JBL, secretory component the receptor for transcellular trans- Reynolds HY. Proteins of the cystic fibrosis respiratory port of polymeric immunoglobulins. J Biol Chem tract. Fragmented immunoglobulin G opsonic antibody 1982;257:11816-21. causing defective opsonophagocytosis. J Clin Invest 6 Goodman MR, Link DW, Brown WR, Nakane PK. 1984;74:236-48. Ultrastructural evidence of transport of secretory IgA 22 Kilian M, Mestecky J, Kulhavy R, Tomana M, Butler across bronchial epithelium. Am Rev Respir Dis WT. IgA proteases from Haemophilus influenzae, Strep-

1981;123:115-9. tococcus pneumoniae, Neisseriae meningitis and Strep- http://thorax.bmj.com/ 7 Haimoto H, Nagura H, Imaizumi M, Watanabe K, tococcus sanguis: comparative immunochemical studies. lijima S. Immunoelectromicroscopic study on the trans- J Immunol 1980;124:2596-600. port of secretory IgA in the lower respiratory tract and 23 Richards CD, Gauldie J. IgA-mediated phagocytosis by alveoli. Virchows Archiv (Patho Anat) 1984;404:369-80. mouse alveolar macrophages. Am Rev Respir Dis 8 Soutar CA. Distribution of plasma cells and other cells 1985;132:82-5. containing immunoglobulin in the respiratory tract in 24 Musher DM, Goree A, Baughn RE, Birdsall H. Immu- chronic bronchitis. Thorax 1977;32:387-96. noglobulin A from bronchopulmonary secretions blocks 9 Kaltreider HB, Chan MKL. The class-specific immu- bactericidal and opsonising effects of antibody to non- noglobulin composition of fluids obtained from various typable Haemophilus influenzae. Infect Immun 1984; levels of the canine respiratory tract. J Immunol 45:36-40. 1976;116:423-9. 25 Shen L, Fanger MW. Secretory IgA antibodies synergise on September 26, 2021 by guest. Protected copyright. 10 Wiggins J, Hill SL, Stockley RA. The secretory IgA sys- with IgG in promoting ADCC by human poly- tem of lung secretions in chronic bronchitis: comparison mophonuclear cells, monocytes and lymphocytes. Cell of sputum with secretions obtained during fibreoptic Immunol 1981;59:75-81. bronchoscopy. Thorax 1984;39:517-23. 26 Naegel GP, Young KR, Reynolds HY. Receptors for 11 Delacroix DL, Marchandise FX, Francis C, Sibille Y. human IgG subclasses on human alveolar macrophages. Alpha-2-macroglobulin, monomeric and polymeric Am Rev Respir Dis 1984;129:413-8. immunoglobulin A, and in bron- 27 Siber GR, Schur PH, Aisenberg AC, Weitzman SA, choalveolar lavage. Am Rev Respir Dis 1985;132:829-35. Schiffman G. Correlation between serum IgG2 concen- 12 Rankin JA, Naegel GP, Schrader CE, Matthay RA, trations and the antibody response to bacterial poly- Reynolds HY. Airspace immunoglobulin production saccharide antigens. N Engi J Med 1980;303:178-82. and levels in bronchoalveolar lavage fluid of normal sub- 28 Stevens R, Dichek D, Keld B, Heiner D. IgGl is the jects and patients with sarcoidosis. Am Rev Respir Dis predominant subclass of in vivo- and in vitro-produced 1983;127:442-8. anti-tetanus toxoid antibodies and also serves as the 13 Andre C, Andre F, Fargier C. Distribution of IgAl and membane IgG molecule for delivering inhibitory signals IgA2 plasma cells in various normal human tissues and to anti-tetanus toxoid anti-producing B cells. J Clin in the jejunum of plasma IgA-deficient patients. Clin Exp Immunol 1983;3:65-9. Immunol 1978;33:327-31. 29 Brandtzaeg P, Fellanger I, Gjeruldsen ST. Immuno- 14 Crago SS, Kutteh WH, Prince SJ, Radl J, Haaijman JJ, globulin M: local synthesis and selective secretion in Mestecky J. Distribution of IgAl and IgA2 subclasses in patients with immunoglobulin A deficiency. Science human tissue: correlation with the presence of J-chain. 1968;160:789-91. Thorax: first published as 10.1136/thx.41.5.337 on 1 May 1986. Downloaded from

344 30 Meier HL, Heck LW, Schulman ES, MacGlashan DW. 1985;132: 1060-5. Purified human mast cells and basophils release human 36 Reynolds HY, Fulmar JD, Kazmierowski JA, Roberts elastase and cathepsin G by an IgE-mediated mech- WC, Frank MM, Crystal RG. Analysis of cellular and anism. Int Arch Allergy Appl Immunol 1985;77:179-83. protein content of bronchoalveolar lavage fluid from 31 Proud D, MacGlashan DW, Newball HH, Schulman patients with idiopathic pulmonary fibrosis and chronic ES, Lichtenstein LM. -mediated hypersensitivity pneumonitis. J Clin Invest 1977;59: release of kininogenase from purified human lung mast 165-75. cells. Am Rev Respir Dis 1985;132:405-8. 37 Lawrence EC, Martin RR, Blaese RM, et al. Increased 32 Kay AB, Moqbel R, Durham SR, et al. Leucocyte acti- bronchoalveolar IgG-secreting cells in interstitial lung vation initiated by IgE-dependent mechanisms in diseases. N Engi J Med 1980;302:1186-8. relation to helminthic parasitic disease and clinical mod- 38 Kanoh T, Mizumoto T, Yasuda N, et al. Selective IgA els of asthma. Int Arch Allergy App! Immunol deficiency in Japanese blood donors: frequency and 1985;77:69-72. statistical analysis. Vox Sang 1986;50:81-6. 33 de Weck AL, Stadler BM, Knutti-Muller J, Hofstetter 39 Umetsu DT, Ambrosino DM, Quinti I, Siber GR, Geha H, Luden C, Huesser C. Factors influencing human IgE RS. Recurrent sinopulmonary infection and impaired synthesis in vitro and in vivo. Int Arch Allergy Appl antibody response to bacterial capsular polysaccharide Immunol 1985;77:38-44. antigen in children with selective IgG-subclass 34 Hunninghake GW, Crystal RG. Mechanisms of hyper- deficiency. N Engi J Med 1985;313:1247-51. gammaglobulinaemia in pulmonary sarcoidosis. Site of 40 Stanley PJ, Corbo G, Cole PJ. Serum IgG subclasses in increased antibody production and role of T lympho- chronic and recurrent respiratory infections. Clin Exp cytes. J Clin Invest 1981;67:86-92. Immunol 1985;58:703-8. 35 Bauer W, Gorny MK, Baumann HR, Morell A. T- 41 Burnett D, Crocker J, Stockley RA. IgAl and IgA2 lymphocytes and immunoglobulin concentrations in plasma cells in the bronchial walls of subjects with and bronchoalveolar lavage of patients with sarcoidosis and without chronic bronchitis. Am Rev Respir Dis 1985;131, high and low intensity alveolotis. Am Rev Respir Dis part 2:22. http://thorax.bmj.com/ on September 26, 2021 by guest. Protected copyright.