Eosinophil Accumulation and Airway Hyperreactivity

CORRESPONDENCE

Eosinophil accumulation and airway hyperreactivity

To the Editor: the lung during an allergic reaction was unaffected by SDZ PCO 400, and since the capacity of this drug to On the basis of correlation studies, UNDERWOOD et al. suppress allergic bronchospasm is not demonstrable fol- [1] have concluded that accumulation and activation of lowing bilateral vagal section, we incline to the opinion eosinophils determines the hyperreactivity that results that increased reactivity during allergic reactions in the from an allergic reaction in guinea-pig airways. In jus- guinea-pig is determined by modified behaviour of nerves tifying their policy of correlating responses at various (e.g. facilitated transmission across synapses in parasym- intervals following an allergic reaction, they suggested pathetic ganglia). that previous studies had used insufficient time points to Activated eosinophils secrete peptidoleukotrienes and permit definitive conclusion. In our case, this is a mis- cytotoxic proteins, materials that have been shown to representation since our analysis did not depend solely induce hyperreactivity in the guinea-pig. Hence, it must upon correlation data. We followed the precepts of DALE be acknowledged that accumulation and activation of [2], who was alert to the problem posed by physiological eosinophils may contribute to allergic hyperreactivity. processes for which there were many candidate media- However, the data presented by UNDERWOOD et al. [1] tors. He listed a series of criteria by which to judge can- do not exclude the possibility that increased reactivity didates. These included a requirement that the appearance which is manifested acutely might persist into the phase of a causative agent must coincide with, or antecede, the of eosinophil recruitment. It is suggested, therefore, that effect for which the agent is held responsible. additional experimental data will be needed to justify When studying accumulation of eosinophils during their assertion that eosinophil accumulation and activa- allergic reactions, we were surprised to observe near- tion determines hyperreactivity during allergic reactions. maximal accumulation of eosinophils in the airway lumen Use of multiple test spasmogens will provide an inci- following exposure to doses of antigen that were only sive test of the hypothesis. With this technique, it has weakly effective, or even ineffective, in causing increased been possible to distinguish hyperreactivity in which reactivity to intravenous injections of histamine [3]. Our eosinophil involvement is not suspected (e.g. in response finding of a lack of correlation between hyperreactivity to intravenous endotoxin) or unlikely (e.g. in response and accumulation of eosinophils in the airways ques- to subcutaneous infusion of racemic salbutamol) from tioned the dogma that allergic airway hyperreactivity was hyperreactivity that is known to be associated with eosino- determined by materials released from activated eosinophil accumulation within the airways (e.g. in response phils. To address this issue experimentally, we elected to intravenous infusion of platelet-activating factor (PAF)) to study the initial stages of an allergic reaction, which [6], and to differentiate these forms of hyperreactivity antecede accumulation and activation of eosinophils. from that which is manifest acutely following an aller- When sensitized guinea-pigs received an intravenous gic reaction [5]. It is already established that there are bolus of antigen at a low dose level, there was transient marked differences between patterns of increased airway bronchospasm. On resolution of this bronchospasm, reactivity to various spasmogens following exposure of increased reactivity to intravenous histamine was already sensitized animals to allergen, where histamine>acetyl- manifest and had comparable amplitude to hyperreac- choline>0, and that which follows intravenous infusion tivity that was evident several hours after exposure to of PAF, where histamine>0>acetylcholine [5, 6]. From antigen [4]. Similar results were obtained following infu- these observations, it seems unlikely that activation of sion of a low dose of antigen in passively sensitized eosinophils will prove pivotal in the genesis of allergic guinea-pigs, when airway reactivity was increased dis- hyperreactivity. Nevertheless, we endeavoured to demon- proportionately for different spasmogens (i.e. in rank strate hyperreactivity following intravenous infusion order: acetylcholine, serotonin, peptidoleukotriene E4, (when cells are entrapped within pulmonary capillaries) bradykinin, prostaglandin F2α, histamine and peptido- or intratracheal instillation of activated eosinophils, but leukotriene C4) [5]. By way of contrast, accumulation of without success (unpublished observations). Should eosinophils within the airway lumen was only detected others succeed, they might establish a rank order of spas- after an interval of some hours [3] and, as UNDERWOOD mogen reactivity and compare this order with that already et al. [1] have demonstrated, there is an even greater established for an allergic reaction in this species [6]. delay before products of eosinophil activation are released Such evidence will either support eosinophil involve- in significant amounts. We contend, therefore, that it is ment unequivocally or provide conclusive grounds for highly unlikely that accumulation of eosinophils during rejecting this hypothesis. the initial phase of an allergic reaction determines the A complementary approach would be to ascertain changed behaviour of the airways. This inference was whether cyclosporin A suppressed allergic airway hyper- reinforced by demonstration that allergic hyperreactivity is reactivity. We were greatly impressed with the capacity of wholly suppressed by SDZ PCO 400, an opener of potas- nonimmunosuppressive doses of cyclosporin A to sup- sium channels [4]. Since accumulation of eosinophils in press eosinophil accumulation during active or passive 1332 CORRESPONDENCE allergic reactions in the guinea-pig. As well as provid- I.D. Chapman ing a lead for new chemical entities which might inhibit Dept of Pharmacology, Quintiles Scotland Ltd, Research activation of eosinophils selectively, this finding was for- Avenue South, Heriot Watt University Research Park, tuitous, for it allowed evaluation of allergic hyperreac- Riccarton, Edinburgh EH14 4AP, UK. tivity in the absence of eosinophil accumulation. The outcome of our studies was quite conclusive: doses of J. Morley cyclosporin A which suppressed eosinophil accumulation Dept of Applied Pharmacology, Cardiothoracic Institute, and activation did not diminish acute allergic broncho- Dovehouse Street, London SW3 6LY, UK. spasm and did not influence acute allergic hyperreactivity in the guinea-pig [7, 8] an observation reported in REPLY rats by others [9]. We did not extend our studies to From the authors: include later time-points, but we suggest that this would be more decisive than correlation studies. We are encouraged by the fact that our paper on air- Observations in laboratory animals are consistent with way inflammation and bronchial hyperresponsiveness clinical findings and indicate that, despite close associa- (BHR), as intended, stimulated discussion on this impor- tion in many circumstances, accumulation and activation tant, but controversial subject. In our study, we demon- of eosinophils within the airways and exacerbation of strated a relationship between eosinophils and their airway reactivity are effectively independent processes. Our cytotoxic products and BHR, which has now been con- assessment of the clinical and laboratory evidence has tested by Chapman and Morley (see above), who argue led us to conclude that pharmacologists should consider that eosinophil accumulation and BHR are two separate hyperreactivity of the airways and accumulation and acti- entities. We believe that this view does not embrace all vation of eosinophils as distinct and separate entities, available information and that there is now important with differing susceptibilities to inhibition by drugs [9]. evidence to support a role for activated eosinophils and The report by UNDERWOOD et al. [1] has not caused us their products (cytotoxic proteins, cytokines and other to modify our opinion. mediators) in the development of airway hyperresponsiveness. Chapman and Morley cite a number of their earlier References guinea-pig studies, in which they found either broncho- alveolar lavage (BAL) eosinophilia with no or little 1. Underwood S, Foster M, Raeburn D, Bottoms S, Karlsson increase in airway reactivity to bronchoconstrictors or, J-A. Time-course of antigen-induced airway inflamma- conversely, BHR before the accumulation of significant tion in the guinea-pig and its relationship to airway hyper- numbers of eosinophils. However, these observations fail responsiveness. Eur Respir J 1995; 8: 2104Ð2113. to take into account the generation of eosinophil medi- 2. Dale HH. Progress in autopharmacology: a survey of pre- ators from cells resident in the submucosa at the time of sent knowledge of the chemical regulation of certain challenge and occurrence of BHR. Even at baseline, functions by natural constituents of tissue. Johns Hopkins Med J 1930; 53: 297Ð347. guinea-pigs are known to have a small number of eosino- 3. Sanjar S, Aoki S, Kristersson A, Smith D, Morley J. phils residing in airway tissue and only by careful tis- Antigen challenge induces pulmonary airway eosino- sue studies (immunohistochemistry or similar techniques) phil accumulation and airway hyperreactivity in sensi- would it be possible to assess cell accumulation and medi- tized guinea-pigs: the effect of antiasthma drugs. Br J ator release. Unfortunately, the above studies did not Pharmacol 1990; 99: 679Ð686. include such measures. 4. Chapman ID, Kristersson A, Mathelin G, et al. Effects Chapman and Morley, further argue that their experi- of a potassium channel opener (SDZ PCO 400) on guinea- ments with the potassium channel opener SDZ PCO 400 pig and human pulmonary airways. Br J Pharmacol 1992; conclusively demonstrates the separation between eosino- 106: 423Ð429. philia and airway reactivity. Reportedly, they found an in- 5. Hoshiko K, Morley J. Allergic bronchospasm and air- hibition of BHR despite no change in the number of cells way hyperreactivity in the guinea-pig. Jpn J Pharmacol in BAL fluid. However, the appropriate correlation would 1993; 63: 151Ð157. have been between eosinophil mediators and airway reac- 6. Morley J, Chapman ID, Hoshiko K, Mazzoni L. Acute tivity. Likewise, in a separate study, cyclosporin A reduced airway hyperreactivity in the guinea-pig. Eur Respir Rev the number of eosinophils in BAL fluid at a dose that 1995; 5 (29), 202Ð210. did not alter airway responsiveness. As already discussed, 7. Chapman ID, Hoshiko K, Mazzoni L, Pfannkuche H, the number of eosinophils in BAL does not necessarily Morley J. Cyclosporin A and the acute allergic reaction correlate with cells in the bronchial wall (the relevant in the guinea-pig. Br J Pharmacol 1993; 108: 209P. site of tissue damage), so these observations remain incon- 8. Chapman ID, Mazzoni L. Mechanisms of inhibition by clusive. cyclosporin A on pulmonary leukocyte accumulation. Trends Pharmacol Sci 1994; 15, 99Ð101. The third argument raised relates to changes in airway 9. Elwood W, Lotvall JO, Barnes PJ, Chung KF. Effect of reactivity by nonantigenic stimuli and the use of differ- dexamethasone and cyclosporin A on allergen-induced ent spasmogens. The authors used platelet-activating fac- airway hyperresponsiveness and inflammatory cell responses tor, isoprenaline and endotoxin to alter the sensitivity of in sensitized Brown Norway rats. Am Rev Respir Dis guinea-pig airways to various spasmogens. These dif- 1992; 145, 1289Ð1294. ferent approaches have little in common with antigen chal- 10. Chapman ID, Foster A, Morley J. The relationship between lenges, as also acknowledged, and are, therefore, unlikely inflammation and hyperreactivity of the airways in asth- to progress the debate about the role of the eosinophil ma. Clin Exp Immunol 1993; 23: 168Ð171. in asthma and BHR. CORRESPONDENCE 1333

Numerous studies in guinea-pigs and other species have References demonstrated a relationship between eosinophilia and airway reactivity. Particularly interesting are studies with 1. Oosterhout JM Van, Ladenius RC, Savelkoul HFJ, Ark anti-interleukin 5 (anti-IL-5) antibodies, which diminish I Van, Delsman KC, Nijkamp FP. Effect of anti-IL-5 and the antigen-induced eosinophilia and significantly reduce IL-5 on airway hyperreactivity and eosinophils in guinea- BHR [1, 2]. Interestingly, IL-5 knock-out mice are unable pigs. Am Rev Respir Dis 1993; 147: 548Ð552. to mount an eosinophilic response, and airway responsive- 2. Eum S-Y, Haile S, Lefort J, Huerre M, Vargaftig BB. ness is not altered after sensitization and antigen challenge Eosinophil recruitment into the respiratory epithelium [3]. Reconstitution with IL-5-producing virus, restored the following antigenic challenge in hyper-IgE mice is accom- mice's capability to generate eosinophils and caused a con- panied by interleukin 5-dependent bronchial hyperrespon- comitant increase in BHR [3]. siveness. Proc Natl Acad Sci USA 1995; 92: 12290Ð12294. The postulate by DALE [4] was originally developed 3. Foster PS, Hogan SP, Ramsay AJ, Matthaei KI, Young to define criteria for neurotransmitter candidates but can IG. Interleukin-5 deficiency abolishes eosinophilia, air- be applied to mediators of a wide range of autonomic ways hyperreactivity, and lung damage in a mouse asth- physiological processes. Our data are consistent with a ma model. J Exp Med 1996; 183: 195Ð201. role for eosinophil-derived cytotoxic proteins in enhanc- 4. Dale HH. Progress in autopharmacology: a survey of pre- ing allergic airway hyperresponsiveness, and are in agree- sent knowledge of the chemical regulation of certain Johns Hopkins ment with a previous study reporting that eosinophil functions by natural constituents of tissue. Med J 1930; 53: 297Ð347. activation, rather than accumulation, is required for devel- 5. Pretolani M, Ruffie C, Joseph D, et al. Role of eosinophil opment of BHR in the guinea-pig [5]. Obviously, as dis- activation in the bronchial reactivity of allergic guinea- cussed in our paper, other cells and proinflammatory pigs. Am J Respir Crit Care Med 1994; 149: 1167Ð1174. mediators and cytokines contribute to the complex series 6. Bousquet J, Chanez P, Lacoste JY, et al. Eosinophilic of events leading to the development of bronchial hyper- inflammation in asthma. N Eng J Med 1990; 323: responsiveness. A relationship between eosinophil media- 1033Ð1039. tors, tissue damage and BHR has also been reported in asthmatic subjects [6]. Inferentially, understanding the cellular and molecular mechanisms behind these and other S. Underwood, M. Foster, D. Raeburn, S. Bottoms, J- symptoms in asthma pathophysiology may translate into A. Karlsson improvements in current therapy and in the quality of Discovery Biology, Rhone-Poulenc Rorer, Dagenham, life of asthma patients: Essex, U.K.