Agents and Actions, vol. 31, 1/2 (1990)

0065-4299/90/020036-03 $1.50+ 0.20/0 9 1990 Birkh/iuser Verlag, Basel

Meeting report Asthma as an inflammatory disease: Are animal models relevant? Meeting of the British Inflammation Research Association (BIRAs) 29 March, 1989. Rh6ne-Poulenc, Dagenham M.T. Withnall and F.B. de Brit0 Biological Research,Rh6ne-Poulenc Ltd, Rainham Road South, Dagenham, EssexRMI0 7XS, UK Asthma is now recognised to be a chronic inflammatory disease, and our knowledge of its aetiology has advanced so that we can perhaps appreciate why some of the compounds of the 70's and early 80's failed in the clinic. But are our current models of asthma any more predictive than the discredited ones that they replaced? This was the subject of the BIRAs meeting held at the Dagenham research centre of Rh6ne-Poulenc Ltd on 29 March,/989. In his introduction, Dr. Lawrie Garland (Wellcome) summarized the evidence of inflammation in asthma and the types of models employed. He stressed that models had value in attempting to improve our understanding of the disease pathophysiology but may prove misleading if used as screens for "anti-asthma" drugs. The well-known example was the undue reliance on models such as rat PCA as screens for developing cromoglycatelike diaags. He emphasized that a more worthwhile approach was to develop pharmacologically welldefined agents (e.g. lipoxygenase inhibitors, PAF antagonists etc.) with which to probe the disease process itself. Dr. Penny Hutson (Southampton General Hospital) outlined the phases (immediate, 17 h, 72 h) of bronchoconstriction, and the accompanying cellular changes, in the allergen-induced guinea pig model. The ability of salbutamol to inhibit the increase in neutrophits at 17 h without affecting the late bronchoconstriction, coupled with the lack of effect of a neutrophil-depleting antiserum, argued against this cell being important in late phase

reactions. Eosinophil infiltration was marked and persistent in this model; whilst eosinophil depletion studies have yet to be carried out, the Southampton team hypothesize that this cell is involved in the development of bronchial hyper-reactivity. No further hyper-reactivity was seen after antigen challenge above that induced by sensitization (this is an important observation and was taken up later in the meeting). The effects of anti-asthma drugs were generally as would be expected from their clinical actions; it is interesting that the PAF antagonist WEB 2086 inhibited the 17 h and 72 h reactions when given at 6 h after antigen challenge, but not when given beforehand, suggesting that the release of PAF occurs several hours after the immediate bronchoconstriction. The fact that this compound reduced the number of eosinophils at 72 h but not at /7 h may bring into question the role of eosinophils in late onset reactions. Another observation that questioned this was the finding that guinea pigs that inhaled PAF did not show late onset responses although there were marked increases in bronchoalveolar lavage eosinophils and neutrophils at 17 h and 72 h. Some workers have argued that this model measures nasal responses rather than effects on the lower respiratory tract. Whilst accepting this possibility, Dr. Hutson cited the lack of effect on late onset reactions of the nasal decongestant ephedrine, and of anti-histamines, as arguments against it. Questions from the floor centred on evidence that cell infiltration can be distinguished

Agents and Actions,vol. 31, 1/2 (1990) from late phase reactions and late responses from hyper-reactivity, emphasising the complexity of asthma research. It was also suggested that the small changes in airway function measured in the model did not mimick clinical asthma, but this was contested by Dr. Tim Clay. In his own paper, Dr. Clay (Toxieol Labs, Ledbury) summarized the development work that led to the above model being utilized in its present form, and gave additional data on clinically used drugs. He confirmed that hyper-reactivity to methacholine developed during sensitization and lasted up to a month. Dr. Clive Page (King's College, London) wondered whether the measured late onset responses were, in fact, caused by a diurnal variation in airway conductance arising from the induced hyper-reactivity, rather than being a direct response to antigen. The ability of salbutamol to inhibit the late onset reaction when administered shortly before the expected time of the response led Dr. Nick Turner (Rh6ne-Poulenc, Dagenham) to suggest that the bronchoconstriction is a mechanical rather than an inflammatory effect. In the sensitized rabbit model reported by Dr. Tony Coyle (King's College, London), antigen challenge resulted in a progressive decline in dynamic compliance over a 6 h observation period. At 24 h after challenge, there was hyper-responsiveness to histamine. Platelet depletion did not abolish the immediate bronchoconstriction, but it prevented the subsequent progressive decline in compliance, and the 24 h hyper-responsiveness was lost. Eosinophil infiltration was also markedly reduced. PAF antagonists had a similar effect to anti-platelet serum and the view of the King's College group is that platelets are responsible for eosinophil recruitment in this model. In the discussion of this paper, it was brought out again that a greater degree of hyper-responsiveness developed during immunization than after antigen challenge. The apparent dissimilarity between the antigen load that the rabbits received and the repeated exposure to very low concentrations of allergen in clinical asthma was raised. Dr. Harry Smith (Beecham Pharmaceuticals) started with the premise that the eosinophil is important in asthma and developed a novel non-antigen-driven model that allowed the study of effects on airway function of a maintained eosinophilia. In this case, the eosinophilia was induced by i.v. injection of Sephadex beads, which led to lung

37 hyper-responsiveness demonstrable both in vivo and in vitro. Several agents having clinical value in asthma (e.g. cromoglycate) were not active in the model, although/~-agonists and corticosteroids reduced the number of blood eosinophils and lung hyper-responsiveness. There appeared to be a partial correlation between compounds' ability to inhibit SRS-A release and to decrease blood eosinophilia. In order to show activity, compounds had to be administered within 4 h of Sephadex administration and it was suggested that neutrophil-endothelial cell interaction at the site of embolization of the particle may trigger leukotriene synthesis which is required to induce the observed eosinophilia. In this model, there was evidence that eosinophils induced into the peritoneal cavity were activated. In discussion, it was apparent that there are still unanswered questions about the model; for example, in pulmonary vasculitis there is a marked lung eosinophilia without clinical symptoms of asthma. The inhalation route of administration is particularly important in asthma in offering the potential for high local concentrations of agents in the lungs and so minimizing the risk of unwanted systemic effects. Dr. Adrian Payne (Wellcome) gave a clear summary of the techniques available for administering materials, and the pitfalls to avoid. Moving on to the work at Welleome, Dr. Payne commented that they had not seen hyper-reactivity develop following sensitization to raise IgG antibodies, but hyper-reactivity had been seen in animals with viral infections.Turning to the subject of the meeting - t h e relevance of animal models to clinical asthma Dr. Payne expressed the view that it depended on the question being asked of the model. Appropriate models can be devised to demonstrate, for instance, that a compound inhibits 5-1ipoxygenase in vivo, and this activity would be expected to be demonstrable in man, but the models would not necessarily predict efficacy in clinical asthma. A similar point was made from the floor during discussion in order to gain benefit from any model, it is essential to understand the processes occurring in that model: all models have some factors in common with human asthma. Pharmaceutical companies want to know as early as possible in the development of a compound whether it is likely to be effective in man; Dr. Nell Barnes (London Chest Hospital) addressed the question of the value of animal models of asthma

38 by discussing the difficulty of developing human models of asthma. Dr. Barnes started by questioning the value of concentrating on antigen-induced asthma, which he considered to be one of a number of precipitants of clinical asthma. Having said that, Dr. Barnes went on to conclude that none of the models in humans are really predictive of clinical asthma. He considered immediate reactions, late reactions to antigen, exercise-induced asthma, cold air inhalation and bronchial hyper-responsiveness, but found flaws in all the models. In spite of corticosteroids being the "gold standard" for treating chronic asthma, it has been difficult to demonstrate that they reduce bronchial hyper-responsiveness. Dr. Barnes shared Dr. Payne's view that models are best used to probe mechanisms, and he suggested that animal models should be used, for example, to demonstrate 5-1ipoxygenase inhibition in vivo as a confirmation of bio-availability and then the compound should immediately be taken into full clinical trials. As an intermediate end point, 5-1ipoxygenase inhibition could be demonstrated in humans, but he did not think that other clinical models were predictive of activity in clinical asthma. Dr. Page questioned whether clinical

Agents and Actions, vol. 31, 1/2 (1990) studies were of sufficient duration to detect activity in a compound; Dr. Barnes' view was that the effects of steroids are clearly apparent within a month, and, since this was the type of compound that the pharmaceutical industry was trying to improve upon, this time span was a reasonable target for seeing clear effects. In his summing up, Dr. Garland reiterated that animal models should be used with caution, understood as much as possible, and designed to probe mechanisms in order to define new targets for drug action. Dr. Clay pointed out that we need to obtain as much information as possible from animal models for drug registration purposes, in addition to assessing the predictive value or relevance of the models. All in all, this was a very enjoyable and successful meeting. No great revelations were made - we shall still ponder the meaning of our models - but it was good to give the subject an airing in the questioning but fairly informal format of a BIRAs meeting.

Received 15 December1989, acceptedby M. J. Parnham, 4 January 1990